,.*

SPE 5800

@.

Enhanced Crude Oil Recovery Potential —An EstimateT. M. Doscher,* SPE-AIME, ShellOil Co.F. A. Wise,** SPE-A[ME, ShelI011Co.

Introduction and SummaryExamination of published API datal on domestic re-serves of crude oiI results in several interesting observa-tions concerning the present status and future directionfor the United States oil industry. Comparison of thelow discovery rate of new reserves that the industry hasexperienced in the past — notwithstanding the discov-ery of Pwdhoe Bay — with our present domestic de-mands for crude oil is striking. These discrepanciespromise to have a dramatic effect on our domestic’”oilindustry in years to come as the large fields discoveredin the i930’s and 1940’s are depleted. The foremostconclusion of this report, in common with that of manyothers, is that large, new domestic sources of crude oilneed to be discovered and developed soon. Or, alterna-tively, an economically viable, massive substitute fordomestic oil must be developed soon.

The second conclusion’is that, contrary to many opin-ions, the cumulative oil reservoir recovery efficiency inthe U.S. has been decreasing for a number of years,with the exception of a slight lift from the discovery ofPrudhoe Bay.”We can be certain that this is no waycaused by poor or mediocre engineering, or by any ar-bitrary or poorly conceived decisions by management.The changes that have occurred in recovery efficiencywith time are merely and exclusively caused by thesequence in which different types of oil fields we~edis-covered in various locations in the U.S.

Study of the recovery-efficiency data has permitted

‘Now a consultantin Houston,Texj**NowwithPetro-LawisCorp.,Houston,Tex.

..* .—

us to reach some conclusions on the potential of tertiaxyrecovery that we believe will be a shade better thanthose estimates that have been over-imbued with wish-ful thinking. The API data leave little doubt that there isand will be a very large amount of oil discovered thatwill not be produced with the technology we have,de-veloped to date. Thus, a third conclusion is that in addi-tion to the need for large, new domestic sources ‘ofpetroleum to be discovered, there is a potential for asizable addition to domestic reserves if recovery effi-ciency can be increased. It becomes apparent, how-ever, that the tertiary recovery potential caimot bequantified at this time, although targets can be spelledout. Were we faced with definitive success today in ex-p(rimental agd pilot operations, the time frame requiredfor execution of the projects wouid make 1 millionB/D by 1990 a difficult target.

Recovery Efficiency —Retrospective Analysis13g.1 relates the 1973estimates of ultimate recovery ofcrude oil reservoirs to their year of discovery; that is,reserves are attributed to the year in which the field wasdiscovered even though development and additions toreserves from the field were continued in subsequent‘years.1The data are not complete for the last few yearsbecause development drilling and evaluation is stillunder way, but the trend is incontrovertible. Except forPmdnoe Bay, our discovery rate since 1958 has aver-aged less than 1 billion bbl/yeav even with PrudhoeBay averaged out over the past 8 years, the discovery

-Examination of re&very ejlcien~y and distributkm of oil in place in the U.S. indicates-reservoir recovery ej?cieney has been quite static ftw many years. Superjlcial changes ineficiency have resulted from the sequence in which reservoirs have bee~ discovered@ the,methods used to book supplemental “reserves. Deveioprnent of tertiary recovery schemes isseen to have comparable importance to exploration in.froittier areas.

M.4Y, 1976. .. .

%,.575

rate is well under 2 billion bbl/year. The degree of suc-cess of exploration drilling also can be observed by ref-erence to Fig. 2,2 which portrays the relation betweenexploration success and wells drilled. It is a cornerstoneof our industry’s future that this trend has to be,.dramat-ically reversed when new frontiers are made avadable tothe driller.

Also shown in Fig. 1 is a curve indicating domesticconsumption of crude oil, both domestically producedand imported. In recent years it has been hovering inthe neighborhood of 5 to 6 billion bbl/year, Even morethan the commonly shown figures comparing totaldomestic production with total domestic demand, thesecurves demonstrate the alarming disparity between theadditions and withdrawals we have been makhig in ournational oil bank account. Unless the success of explo-ration returns — and quickly — to that experienced dur-ing the glorious 1930’s and 1940’s we are probably onthe road to closing out our nation’s oil account for lackof funds.

Now, let us look at the recovery efficiency with theultimate aim of determining if there is a significantamount of crude oil resources that we can view as a

s

o1920 1930- 1940 1s50 lwo tsm

YEARS

F@. 1 — Discovery rate and dcmestic demand.

lSO-JQelLLION eA

lm-

SO

a.60

ELS-.

ALASXAI

A

U.S.A.

● s , m) 1$0 200 .:

timcmnori tiusFig.2.— Cumulative recoverable oil discovered; total

industry, 1920 to 1969.

576 ,’.— -

target for enhanced recovery. Recovery efficiency fodiscovered resources can be calculated from 1920 to th(present using the API data, Pre-1920 discoveries angrouped together. Fig. 3 demonstrates that the recoveqefficiency of our more recently discovered fields is Iesthan the efficiency of that of earlier discoverie~. Admittedly, the data for the last few years are open to question, since not all the fields discovered had supplemental projects installed at this time; as a result, not all th(supplemental reserves were booked. On the other handthe Suppi8mentarytargets are recognized and projectam implemented more quickly than in the past. It is als(obvious that if only a small amount of primary ibooked,. then an even smaller amount of secondary ca]be booked. The trend is, again, incontrovertible.

In the face of our presently expanding technolog:concerning production and reservoir techniques, why ithis reduction occurring? We will attempt to answer thibv fwst breakimzdown the United States’ reserves on Is~ate-by-stateb&is. Borders between political units annot usually coincident with geological features, surfac(or subsurface, although there are marked exceptions.s 1is fortuitous, however, that in the U.S., state borders d(appear to circumscribe oil provinces. For discovenethus far, four states have contributed 71 percent of oudiscovered ultimate recovery:”Texas with 35 percentCalifornia with 14 percent, Louit.iana with 13 percentand Oklahoma with 9 percent. The ultimate recoverbroken down by states associated with discoveries i]10-year increments is shown graphically in Figs. 4i.and 4B.

Before 1920, California, with the discoveries of thegiant reservoirs in the Los Angeles Basin and the SanJoaquin Valley, was the state in which most of the ul-timate recovei’ywas found. Other states, particularly theAppalachian states and Oklahoma, were at the timeeven more important than Texas and Louisiana. Startingin the 1920’s and continuing through the 1950’s, Texasrapidly replaced California as the leading oil state; sub-sequently, Louisiana came on and vied with Texas asthe leading oil-finding state. At the end of the last dec-ade, Alaska replaced Louisiana.

When we examine recovery efficiency on a state-by-

Sor

soml YEARLY

10- . . .

.

-0 t 1 # I 1 1

EJRE19201920 1930 , 1940 1950 1960 1s70.

Ftg. 3- Recoveiy efficiency.-

JOURNAL OF Pi+ROWM TECHNOLOGY

state basis in light of 1974 technology, we find ratherstriking differences. Fig. 5 shows how recovery effi-ciency has varied in 10-year increments in the chiefproducing states. Something different happened duringthe 1930’s in Texas and Louisiana, but not in Oklahomaand California.

The beautiful reservoirs (highly permeable and mostwith a strong natural water drive) were being discoveredin East Texas and South Louisiana. The East Texasfield is perhaps the epitome of this discovery cycle.Compared with the shallow, more viscous oil reservoirsdiscovered in the 1920’s and earliei (many by iookingfor the subterranean pool feeding the surface seeps),these reservoirs of East Texas and South Louisiana werebonanzas in every sense of the word. Later, during the1940’s and 1950’s, the industry made its new’ dis-coveries by drilling deeper onshore or farther out intothe deltas of the Gulf of Mexicq the sands were finer,the carbonates were tighter, and permeability and poros-ity were lower. .4 water drive, if ,any, was weaker, Re-covery efficiency turned down again. The graduai,monotonous decline in recovery efficiency was inter-rupted in the late 1960’s with the discovery of thePrudhoe Bay accumulation — the largest in the U.S.,and tentatively reported to have a modest water drive.Fig. 5 shows recovery efficiency both with and withoutPrudhoe Bay. Not only were the reservoirs discoveredduring the 1930’s efficient, they were big. In fact, ofour 10 largest fields by remaining reserves, six werediscovered in the 1930 to 1940 period and seven aredeep in the heart of Texas,

On a cumulative basis, the trend is quite clear, asFig. 6 demonstrates. Cumulative recovery increasedsignificantly during the 1930’s, and has been going

BILLIONSOF SSLS.

4

S2 BILLION

so -.

8S ●IUEON

BO O. O.I.P.

75 SILLION70 72 SILUON

so - ~ TEXAS 6S BILLION

D CALIFORNIA -

~ LOUISIANA S1 BILLION

‘o w OKuHOmA ‘-----

1I

~ OTHERIII

40 - ULTIMATE RECOVERYII

SAV

mME IN7ERVAL

Fig. 4A — Recovery by states.

MAY, 1976.,

BILLIONBOF BBLS.100

so

80

70

60

60

40

30

20

10

,0 I

2 BILLION

%

66 SILLION

O.O.I.P.

76 BILLION

7S sN.LION

~ TEXAS ,

7

66 BILLION

u CALIFOR~IA

N LOUISIANA

~ OKLAHOMA. A

~ OTHER IA?\ULTIMATE REOOVERV

k I\

AVEPAGINGPRUOHOE SAYOVER 1966.16i

.

IE lS20 1620-2s lB20-36 1S40.46 16S0.66 1660.6970.76

. ALASKA

,6 BILLION

, LOWER 46

TIME IN7EIEVAL

Fig. 49 — Recovety by states.

701-

— TEXABHt.”., CALIFORM,A=~~ LoulsIANA

SOP ‘- OKLAHOMA

~~‘0 PRE 1920 1920.29 1930.3S 1S40-4S 1SS0-S9 1SS0.SS

TIME INTERVAL

Fig. 5 — Recovety efficiency by states.

50

5~ 40u~

“ [-

i “sO‘ ~~ij .20 .~.

10 -.5

I I I I J~RE 1S20 1S20

I1s30 1940 19s0 1960 lox

F@. 6 — Cumulative recovety efficiency.

. .

down in recent years with a bh of an inflection resultingfrom Prudhoe Bay.

A look at the microcosm of West Texas/SoutheastNew Mexico shows some of the detail of the nature ofthe historical oil-discovery cycle and the changes in re-covery efficiency characteristic of the U.S.

Fig. 7 shows that three fields (oldies and goodies, toborrow a phrase from the recording industry) accountfor about 20 percent of the area’s ultimate recovery.Yates field was discovered. in 1926, Wasson field wasdiscovered in 1936, and Kelly-Snyder field (ScurryCounty) was discovered in 1948.

Although there are some other sizable fields in tiearea, the industry has been.’ ‘grubbtn’ stumps” since1950, waiting for something better to come along.Perhaps even the biggest stumps have been alreadygrutibed. The few reservoirs that have been rooted outin West Texas in recent years are small and show fairlylow recovery efficiencies, The original oil in place inthis region is estimated at 89 billion bbl, or about 20percent of that found in the nation. The ultimate re-covery is estimated at 26 percent, compared with 32percent for the nation. Surprisingly, this significantlylower efficiency is not caused by the carbonates, whichwe will see give almost average values after waterflood-ing, but results from the poor sandstones in the region.

Past Estimates of Recovery Efficiency andThe Subsequent Effect of WaterfloodingThe preceding analysis is “look-back dat~’ as of Dec.31,, 1974. The recovery efficiency is that expected bythe appli~atifi-of 1974 technology to the reservoir. Butwhat if these reservoirs were evaluated with the tech-nology of, say, the 1930’s and 1940’s would, the es-timated reccnery. efficiencies be significantly lower?In other words, can we demolisirate a real increase inrecovery efficiency of a particular reservoir or class ofreservoirs.with time? If we could; we could then saythat developing technology did indeed increase recoveryefficiency, and we might thin be safe in projecting thisiqcrease.forward in.time. We are .no Alley Oop, so wecannot transport ourselves backward in time and make acomparable evaluation ourselves. We will have to relyon the best estimates made by competent engineers at

‘r .760

51- SCURRY co.

{

so

VA~ES WAS,SON‘1

+ t“

,1t.

I lr\ilAlr”A’El’”m 01 IJV t Vw, . w ~ ~’-~

1s20 1930 1940 1ss0 . 1960 1070Jo –,_

.Ftg. 7 — West Texas, Southeast N& Mexico discoverka.

work on the problem in those days.As recently as 1962, before the advent of the detaik

API statistical efforts, the Interstate Oil Compact Conmission (IOCC) estimated original oil in place for reervoirs discovered before 1949 by using a figure t“3.57 barrels of oil discovered for each barrel of cproduced plus proved primary reserves,”4 In othiwords, the best information available in 1962 indlcatian average primary recovery efficiency of 28 percenwhen almost 75 percent of the nation’s total ultimaaheady had been discovered. The amount of crude pnduced by,fluid injection by 1949 was negligible.

Let us compare the estimated primary recovery eflciency of 28 percent in 1949 with the API’s anticipate(proved and indicated) recovery efficiency of 32.9 pecent at the end of 1974. Indeed, there is an improvement. It is important to assess whether this increase ito be attributed to gradually improving technology or tlsome other factor. We believe it is primarily attributeto the iinplementation of a singIe technique - watelflooding. The difference in recovery with and witholwaterfloodingcould have been added in one fell,SWOOI‘The IOCC did just that for the period 1948 to 196(They had estimated a recovery efficiency varying fro]31.5 to 32.8 percent over this period. The API stttistics distribute this increase in recovery from WItefflooding over several decades by adopting-the prfcedure of booking waterflood and other supplementreserves only when the operation is actually imphmented. Thus, the answer to our question is that grafual improvements in technology did not result insteady improvement in recovery efficiency.

The contribution to our ultimate recovery from ttimplementation of waterfloods can be calculate fro]these considerations as amounting to (4.9/32.9) X-19[15 percent of the total ultimate; or, an increase of 21,billion bbl, 18 percent over primary.

Still another way to ccinii this estimate is to adtthe total revisions made in the API estimated reserve-between 19~0 and 1974, These total revisions add up .O29.4 billion bbl. However, in addition to oil booked aa result of implementation of fluid injection operationsrevisions include those resulting from other information(other than changes in proved acreage), such as change

TABLE 1— CHANGING ESTIMATES OFRECOVERY EFFICiENCw

Discoveries of 1955

Estimated Estimated RecoveryUitimate Oii in Piace Efficiency

Date of APi (thousands (thousands (thousandsEstimate of barreis) of barreis) of barreis)—.. — ——

19fi6lti68197019721974

(1974/1966)

19661968197019721974

~~

1,392 4,771 29.21,458 5,196 28.11,508 5,534 27.31,487 5,570 26.71,4!25 g 26.7

107 percent 117 percent ~wrcent

Dlscove”rles of 1960

663 2,537 26.9732 2,615 28.0 -:

.601 2,772 26.9.674 2,963923 3,082 %

=5 percent 121 percent lF&cerit

JOURNAL OF P&@JM &CHNCKQO}

. .

.

in estimates of original oil in place, the effectiveness ofthe drive mechanism, etc. These revisions were alwayspositive before 1950, so the 29.4 billion bbl probablyincludes more than just waterflood oil. (A couple ofbillion barrels certainly must be attributed to steam in-jection methods.) Thus, we are on quite firm ground inestimating waterflooding to have contributed an in-cremented oil volume of between 20 and 25 billion bbl.

We should stop here for a moment and give credit tothe @d-time reservoir and petroleum engineers whomade these early estimates, still valid today, with pen-cil, slide rule, and hand-cranked calculator. Their in-.sight more than compensated for the primitive, but ap-parently more than adequate, tools at their disposal. JJ(ealso must give credit to those operators in the Pennsyl-vania and New York oil fields who recognized the valueof waterflooding 80 or so years ago. It was their misfor-tune that the State of Pennsylvania outlawed the opera-tion of waterflooding, believing it would hurt the oilreservoirs. It was not until shortly after hard liquor wasoutlawed that waterflooding was legalized in 1921.

Extension of watefflooding to other regions was notrapid in the fo!lowing decade, primarily because of thelack of economic incentive. However, in the 1930’sfledgling efforts were reported in many other states. Amore thorough understanding of the process was de-veloped after World War H, and the modern technologywas subsequently developed. The greatest additions bywaterflooding were achieved during the last decade.

When one studies the API statistical data for the past10 years, it is also apparent that along with the steadyincrease in booked reserves owing to revisions and ex-tensions, there has been a concomitant increase in the

600

400

10(

:.

●,=9

I75% 0;O.LXI.P.BY 1960 /1

604 OFO.O.I.P. /

,,

LBY ;936

o0m

-ULTIMATE RECOVERY

,

. PUE

,Ftg. 8 — Cumulative original oil in place andultimate recovery.

i 1 n I I aB20 la20 Isao 1940 ISSO +360 1s70

MAY, 1976 ‘ -.... .

~ -a

estimate of original oil discovered. Table 1 shows hcenhanced technological insight and additional prodution performance over a period of 8 years led to changin estimated recovery efficiency for reservoirs discoered in pfior years. In some cases, as for 1955, susequent estimates of original oil in place have increasmore than estimates of ultimate recovery, and Iook-bacalculations of recovery efficiency decrease accoringly. This reveals an oversight in some prior attem[at demonstrating the effect of improved technology ~recovery efficiency. The increase in the estimated u]mate recovery was perceived, but the change in the ftimated original oil in place was not adequately assess— and sometimes was neglected. -‘

Technological developments have indeed permittoil to be produced from deeper, higher pressure, ahigher temperature reservoirs; in hostile, offshore, afrozen tundra environments and fzom Hades-like scgas environments. The accessibility to locationswhich oil has been discovered has been increased. Fnturing, acidization, and seiective completions hastimulated producing wells and have accelerated rec(cry. The cost of produced oil has been kept low and tproduction rates have been kept high enough to sati:our nation’s needs at low unit costs. It has been thtachievements that, in no uncefiain way, in times of wand of peace, have permitted our nation to reach amaintain its eminence. Although the continued i]plementation of supplementary recovery operations tplayed an important role, we have not yet succeededintroducing newer technology capable of furthercreasing recovery efficiency.

Total Volumetric Target forTertiary RecoveryWe have already shown that the rate of discovery hinot kept pace with the rate at which we consume crucoil. We want to return to this point for a momentemphasize !hejustification for the ensuing di~cussionfthe need to increase recovery efficiency from knovreservoirs, as well as ~odiscover new resources in nefrontiers.

One-half the crude oil discovered in the U.S. wdiscovered by 1936, and 75 percent was discovered 11950, as shown in Fig. 8. We are living M these eaxdiscoveries. When discussion turns to the prospectsthe new frontiers, it is sobering to realize that t]erstwhile new frontier now peaking out, the GulfMexico, will contribute 6.3 billion bbl of estimated ttimate from.current fields. This is a sizable volumecrude oil, but in terms of national consumption itonly a trifle more than a single year’s supply. For coiparison, an increase in recovery efficiency of only tvpercentage points would add a reserve of slightly mcthan 8 billion bbl. Of course, we hope the new frontitwill exceed Prudhoe Bay and that recovery efflciencan be economically’increasedby more than a couulepoints but until the barrtk are-inthe pipeline, it is i]perative that we look on this goal of increasing recoveefficiency with comparable enthusiasm.

Let us now shift our attention from recovered oiithat left in the ground. That is our target for tertiflrecovery. By breaking down th-xe resources into Igional groupings, we can see in Table 2 that there a

. . .,...

,.

....-.,,.,,,!,

. .

.,

,

,.

. .

. . .

,.

,.

‘

TABLE 2 — ESTIMATED lJNRECOVERED OIL lN THE U.S.

Loukiana - southTexee - east end

Districte 2, 3, 5, and 6

Subtotal - Gulf Coast

Texas - south and westDistricts 1,4, 7S, 8S, 9, and 10CarbonatesSende

‘CaliforniaOklahoma

Subtotal - except Gulf Coast

Remaining U.S.

Total U.S.

Original Oil Estimatedin Place Ultimate Recovefy Estimated Residual

(billion barrels) (billion barrels) (percent) Billion Barrels Percent U.S.

30.96

35.17

66.13

67.9546.9783.2437,65

235.82

138.77

440.72

16.13 52.1 14.83

19,32 54.9 15.85

35.45 53.6 30.68 10.4

21.06 31.0 46.9011.16 . 23.8 35.8121.66 26.0 61.5812.44 33.0 25.21—.66.32 28.1 169.50 57.3

43.24 31.2 95.53 32.3

G 32.9 295.71 100

really two types of reservoirs, as already noted: goodones and middling ones. It is not just that there is anatural water drive prevalent along the Gulf Coast thequality of the sands, permeability, porosity, un~formity,and the low crude viscosity all aid and abet hi~ recov-ery efficiency. in at least one of these gian~fiejdm-with_a reported degree of uniformity better than somelaboratory sand packs (a Dykstra-Parson coefficient of0.2), the recovery is expected to reach some 80 percent.

Some Tertiary Processes andTertiary Targets ~.Assuming an initial saturation of 65 percent, residual oilsaturation in an average-~rforming’ Gulf Coast reser-voir would be less than 30 percent at 100-percentsweepefficiency. If, because of the geometry of the reservoirand the placement of wells, the sweep is onfi 70 per-cent, the residual in the swept zone would be of the“order cf only 10 percent, Such low values have beenconfirmed.8 Three-fourths of the residual oil in such areservoir is in the unswept poltions of the reservoir. It isdifficult to comeive how any tertiary process could getto this occluded oil, short of new drilling.

A fwst screening then would say that a fraction inexcess of one-half the Gulf Coast residualiis not at high

—,

TERTIARY 011 RECOVERED,

BBLS ACRE FT.‘-—

0 / II I I 1 I I I I

200 210 220 “230 240 2s0 200 2?0 2s0

enough satu~tion for tertiary recovery processes thatare cumently under consideration. On the other hand,those Gulf Coast reservoirs with less than average per-formance (higher than average residuals) will still havecharacteristics (greater uniformity, freer of clays, andhigher permeability) that are superior to those of mostother oil provinces for chemical flooding, Chemicalslugs introduced into these reservoirs will lead to moreefficient displacement because the slugs have a greateropportunity to maintain their integrity and can attainhigher velocities.7 On the other hand, Gogarty7 has in-dicated that in a good, clean Illinois Basin reservoir, apattern spacing of 2.5 to 5.0 acres is required for effi-cient displacement of oil by a surfactant system, In thecited reference, infill wells apparently can be drilled to1,000 ft for some $ 17,50d-.The resulting rate of retiirnis indicated to be only 8 to 10 percent at an anticipatedrecovery efficiency of 242 bbl per net acre-foot (ASO~= 0.031) and a sales value of $13.50/bbl (Fig, 9). Tak~ing into account the higher porosity for Gulf Coastsands (30 percent vs 19 percsnt) and an estimated lowerresidual (25 perdent~vs 40 percent), the recovexytargetwould be about the same as in Illinois. Sand thic~w.sswould not be suffkiiently greater in the Gulf Coast tojustify the obvious much higher cost of infill drilling inthe marshes and bays of South Louisiana, In lieu ofsurfactant systems that can be propagated over largerwell spacings or a much higher price than $13/bbl, atertiary process other than chemical flooding appears tobe needed as a gerxml tool for the Gulf Coast reservoirs.

Let us look at what appears from Table 2 to be one ofthe largest single targets for tertimy recovery — thecarbonates of West Texas. In textb~oks, carbonate res-ervoirs usually get short shrift. The rtiason for this isthe apparent complexity of the porous medium itself.Everyone has been to the beach as a child, and playedwith the sand and watched.water drain out of sand piles,The concept of a sandstone reservoir is regdily con+veyed; fluid flow inside the sandstone can be easily con-jured up in the mind’s eye. This, of course, is not thecue for carbonates.

However, whel~ one examines the performance ofcarbonate reservoirs, an unexpected consistency in their-.

.,F@. 9 — Projected economics, 6,000-acm Maraflood

perfordiance is observed, Fig. 10 is a plot from a cur-procese development in Illinois; after-tax projections rent study being undertaken by the API Subcommittee

ehowingeffectof recovey efficiencys aoing and1crude price on rate of return. (From ogarty.~ “Eetimatod reektud In the halfof the-rekarvoireamehable to tertiary reooveqi

580 ““.,

JOURNALOF-PETROLEUM TECHNOLQOY.=-

.,

?

-.

.

on Recovery Efficiency.8 It represents a plot of initialand recovered oil in barrels per net acre-foot for a groupof randomly volunteered carbonate reservoirs from anumber of operators in Texas, (W,emight add that thedata from other solution gas driven carbonates willsuperimpose themselves on the Texas sample.) Suchstatistical consistency speaks for a rather consistent de-pletion mechanism being operative in carbonates. Weinterpret the apparent constancy of the recovery”effi-ciency to signify a statistically average occluded porevolume fraction in carbonates. Swept pores have littleoil left in them, and unswept pores have a lot (Fig. 11).If this is the correct picture, how do we get at the re-sidual? Chemical flooding “probably could reduce thelow residual in the swept pores to zero, but would nothave much chance of penetrating into the unsweptpores. Moreover, the calcium and magnesium in the in-terstitial brines and at mineral surfaces in the carbonatereservoir: would wreak havoc with the delicately bal-anced multicomponent chemical slugs containing ‘‘Sur-factant A + Surfactant B + cosurfactants+ oil + salt+water + oxidation inhibitor + bactericide. ” We do notdoubt for a moment that such a formulation could bedeveloped for carbonate% we do doubt its economicusefulness.

On the other hand, a miscible flood using carbondioxide does have promise. Under high pressure,super-critical carbon dioxide is partially to completelymiscible with many West Texas crude oils. Diffusionforces, aided by the low viscosity and density differ-ences between the miscible fluids, would promote thepenetration of the.carbon ‘dioxide into the formerly un-swept pores. But we must not overlook the histo-ryofLPG flooding.9 Starting in 1956, with the threat of theSuez crisis hanging over the Western world, LPG flood-

#fig was latched onto as a great hope for increasing bothrecove~ and production rates. Although crude oil wasless than 10@/galat the time, LPG was only pennies pergallon and could be viewed in this cpntext as a sacrificialreagent. Unfortunate y, it was found that permeabilityvariations within the reservoir and adverse mobilityratios, which could not be completely assessed in lab-oratory studies, mitigated economic recovery; the ratioof produced oil to injected LPG was too low even at

s.:Mllumls WmOwl.?W

mm . Wa

.—

-. UNU wmtmmm

mmw- suuwwmw.

iumw. wmmmo .

●*

9

On*sl*ak OIL w ●tact sa# .

Fig. 10 — Texaacarbonates, recovered oil vs original oilin place.

MAY, 1976 “’ “., ,

.

the then relatively low cost of the LPG. (Today, ofcourse, neither LPG nor natural gas can be thought ofas cheap, sacrificial reagents.) These adverse param-eters that obviously will appear in many carbonate res-ervoirs also will diminish the effectiveness of carbon‘&oxide.

Carbon dioxide, however, was introduced because itwas a far cheaper reagent than LPG and it was antici-pated that greater ratios of carbon dioxide to producedoil could be tolerated. Also, super-critical carbon di-oxide may have a density between the densities of oiland water and, thus, gross gravity instability would beavoided. The viscosity, however, is very low and thepotential for viscous fingering is still with us. In manycases,’0 attempts a~ made to restrict the fingering ofthe carbon dioxide by the alternate injection of water.The goals appear to be the achievement of some degreeof plugging of the more permeable paths and an im-provement in sweep. The plugging presumably resultsfrom Jamin effects — the high energy required to forcean alternating sequence of nonwetting and wett~ngphases through capillaries. Such effects, if they do occur,probably will be more pronounced in low-permeabilitypaths than in the higher, already-depleted paths throughthe reservoir, even though the latter accept‘more of theinjected fluid mixtures. But an even more significantdrawback to the use of water to counteract the fingeringof the carbon dioxide is the counter-effect it has on thebasic goal of the process.

A tertiary carbon dioxide flood is, in mechanistic,terms, a carbon dioxide flood of a water-filled reser-voir, and much of what the carbon dioxide does is todisplace the reservoir water. It would appe~r that it isbehind this front that the carbon dioxide will begin tocollect the distributed oil and, it is hoped, bank it upand get it moving. Alternating water with the carbondioxide promises to break up any oil bank forming inback of the initial water-carbon dioxide interface. Webelieve some other process schemes will have to be de-veloped to more efficiently reduce carbon dioxide fin-gering and circulation through the reservoir.

What can we say are the minimum requirements ofcarbon dioxide to recover a barrel of oi!, and what is thecost’?Merely to substitute for a barrel of stock-tank oil

CROSS-FLOW PORES

CONNECTED PORE SYSTEM /

/ /● ✎✎✎☛☛ ✎ ✎ ✎ ✎ ✎ ✎ ● ☛

● ☛✎✎☛✌ ● ✌☛☛ f“””

OCCLUDED PORE SYSTEM

● OIL SATURATION APPROACHES ORIGINAL

● OIL SATURATION APPROACHES ZERO

F@ 11 – Sketch for the concept of immiscible fiuid fiow incarbonates to account for obaewed statistical constancy of

recovery efficiency. Continuous poree swept to very iowrasiduais; occluded pores at almost originai saturation;’occludsd pore volume frfiction a “statistical constant.”

581

under anticipated subsurface conditions will require 2.5to 3.0 Mscf. Losses caused by solution in reservoir-retained oil and water, and by incomplete stripping ofcwbon dioxide from produced fluids (the recovered car-bon dioxide, at anticipated make-up costs, will be re-compressed and reinfected) could dottble this quantity.

Fig. 12 presents a location map of what we believeare the principal targets for carbon dioxide flooding inWest Texas,, and our arm’s-length estimate of the oilthat might be recovered by successful application of theprocess. The total of 4 billion bbl is indeed impressive.

The recent disposition of merchant carbon dioxidesupplies is shown in Fig. 13, and the cost of thesesupplies is shown in Fig. 14. The use is limited and thecosts are quite high, but considering t~e use the buyercan afford the tariff. However, if carbon dioxide is usedfor tertiary recovery in West Texas, and if our estimatesof consumption a~ in the ball park, then some 24 I~fof carbon dioxide will be required. If we assume it willbe used over a 20- to 25-year injection cycle, the de-mand, as shown in Fig. 15, jumps and eventually willreach more than 1 Tcf/year.

It is, of course, impossible for us to predict wheresuch a supply of carbon dioxide would come from, Ob-viously, exploratory searches are under way to find it inthe subsurface. (he thing certain is that we cannot af-ford to bum oil just to produce carbon dioxide burninga barrel of a typical crude will not produce much morethan 6 Mscf. But even if the carbon dioxide is availablefree of charge, perhaps from the stacks of a coal gasifi-cation plant, compressing and transporting it over a &-

tance of 500 miles, for example, puts a price tag on itof stout $ l/Mscf. Adding the mere operation of inject-ing t Mscf to recover a barrel of crude to the costs ofthe balance of the oilfield operatio~, recirculation ofcarbcm dioxide, and fighting the corrosiveness of thisacid gas, one could estimate an operating cost reaching$ 10/bbl. But if the carbon dioxide is not free (say it hasto be manufactured by specialty trianufacturing*), or ifit is to be recovered from stack gases, or shipped 1,000miles, or if direct consumption is greater than 6 Mscf,the direct operating costs to recover a barrel of oil couldrise by a factor of two or more.

How much of the potential may we expect to realize?It is obvious that although the technology of oil dis-”placement by carbon dioxide is still not ready to bewrapped up in a do-it-yourself treatise, there are otherfactors relating to reagent supply and costs, corrosionprevention, and pricing schedules that will have to begelled and matched with each other before a signific-antly meaningful estimate of the total potential of ter-tiary recovery with carbon dioxide can be made. Butestimating the total potential available is a relativelymeaningless exercise. It is the estimate of the rate ofsupply—the barrels per day we may expect to be putinto a pipeline some day in the future—that is impor-tant, We will ~eturnto this subject.

At this point, recall that another supplementary pro-cess, in addition to waterflooding, has been with us forsome time. Modem work on the use of steam to recover

●Such aa a modlflad PACE-S(3P(Shell Development Co., Ilceneor) facility.

i.- ! ~---- +- .--.+ .-. ;-. -------------

. . ----

Paw a...— ii–--r’““–-?-’2

E .LwAmomM’A

—-— i-

i,:

! “w’+’ =-p- EYA raoyoi- .:

Ls ,qg-y‘%, & .*.. .0 --

‘\. ,,; @ #*

.( .‘%s,,:“~h;,%:● .,.,,/ ‘..,

-r---- j-t

{+..— ..- ... .~. ,AT>!&I

I‘. ---.- . .

i ~

_i -. ~-.l-------i,l-- -------

HWOTHESIZEO ~ FLOOOFO$&NTIAL

FIELD

WAUOUSLAUWTSIILIWSLLAMEEMNOLEWLLEMUNKELLVWVOSRm~MeoLMMTnNo.~a ~u:3TERtWwAnu.e~wAnmssTasSANDtXLLSMCIELROVVATEE

vAeWM

TOTAM

ESTIMATEOC02REOUIRSWKUT- S84D~F

+ FR&TS,RE~TEO

●sHALLm RESERVOIRS- MECISILI?V00WTFULDUETOEXTREMELVLOWPRESSURE.

F@, 12 — Cwbon dioxide flood prospsets of West Texas and New Mexico.

582 — JOURNAL OF PETROL&JM TECHNO~Y

-f . .-.

..

petrole~m W* also initiated during the era of the Suezcrisis. The art and science of steam soaking andsteamflooding have been contributing an estimated200,000 B/12to California’sproduction total for rhost ofthe past decade. The added reserves are estimated to be~tween 1 and 2 billion bbl and may well be doubled inthe future as the result of the greater value realized byheavy cmde oil. The role that steam injection tech-niques has had in maintaining the health of at least oneWest Coast operation is shown in Fig. 16. Virtuallyall the oil produced by steam injection is from high-saturation, high-porosity, shallow reservoirs containinglow API-gravity crudes.

But steam will displace light oils, too, even moref effectively than heavy oils befause of the volatility of

the lighter oils and consequent distillation or strippingof the oil by the steam. Steam has a built-in, anti-bypassing capacity (built-in viscosity control, so tospeak) until relatively grossly depleted intervals be-tween wells are developed., Even then it continues tobe effective, since heat can be-transferredby conductionas well as convection.

The economic efficiency of steam drives are deter-mined principally by injection rate, spacing, thickness,porosity, and oil saturation (Table 3).

What is an acceptable economic steandoil ratio? InCalifornia operations, a value of 0.20 to 0.25 had beenacceptable in the past, and in more current operationswe would surmise that this ratio has been lowered to-ward a value of 0.15 because of the increased value ofheavy crudes. Local costs, the need for d@li~g wells,water treating, etc., of course, will affect the acceptableratio for any given operation. Considering the energybalance only, the ultimate steardoil ratio when usingyour own crude is 0.065, since 65 bbl of oil in a highlyefficient generator will convert about 1,000 bbl of waterto steam. These numbers permit one to estirn’atethat ittakes the equivalent of about % bbl of crude to coverthe costs of capital, operating expenses, and taxes forevery bamel of produced cmde at the 0.15 ratio.

This ratio could be conceivably reduced if. fuelscheaper than crude could be used for generating steam.We would like to pass on the following thought, Insome parts of the country; coal, lignite, and refinerybottoms, with rational controls on their utilization,could be significantly ‘cheaper boiler fuels than pro-duced crude or fuel oil. If, at such locations, thecheaper fuels were used, then the energy-basis steam/oil

. ratio could be lowered significantly below 0.065; addi-tional targets for steam recovery might then be re-vealed. This could ,be the cheapest coal liquefactionprocess available to us.

The calculated”valuesof steam/oil ratio in Fig. 17 &‘ for a reservoir having a porosity of 20, percent and a

recoverable oil saturation of 30 porosity percent using

TABLE 3 — FACTORS DETERMINING STEAM/OIL RATiO

( ).‘-’i -ht3,6&~~ , ~ ;Oil Produced = ~

Steam Injected A’

where

h,= formation thickneee, grose hJh, = net/groes8S. = decrease in oil saturation i= Injedlon rate

d=porosfW A’= arsa/injector

UOUIDS k GASES

. ao~””80110 (DRY ICE)

o~.1ss0 14s6 1s70 197s

Fig. 1L3— Carbon dioxide marketed industrial usage.

(FOB PLANT]6.00 ?/MCF

4.00-‘\

--.*-*, SOLID (ORY ICEI

S.oo

I,loulos a d’

.

‘\2.00

-.“*--

-***.*+.

SOURCE: US. DEPARTMENT OF COMMERCE

.OLIIIIIIII~J‘19s0 191==, 1s70 1s7s

Fig. 14— Carbon dioxide marketed industrial usage value,FOB plant.

~PM~Tl%fTEnTIARY +RECOVERY OF CRUOE 011

’200 -—,.. .—a~

:’.

*OO —-+-----

:

/200

.

o~l::f+fi\:,,,,,,,,l -

1ss0 1s65 1970 1875 1s8s 1ss0

Fig. 15 — Carbon dioxide industrial exploration andproduction usage.

ANNUAL AVG.DAILY OIL PROD.

M Sslwo

‘r TD7AL Oi& PSODUCTIOS

.- ?.. ,-...

0-1’-..i 1 t I t. t B 1 4

s46s0607$~. ss ~~”~=~

wAR

“~. 16— Sheil Oii 00. Weet Coaet Div. sourcesof production.

MAY, 1976 583

.

.

a recentiy published analysis.11Calculations (Table 4)were made for 5- and 10-acre spacing and injectionrates of 500 and 1,000 BWPD at pressures of 300 and1,000 lb, and assume a total injection of 1 PV of water.,Thesecalculations do not include any beneficial effectsof distillation.

The ratios are not economic if fuel oil is used togenerate steam. If cheaper fuels could be used underthe economic conditions noted, then economic recoveryby steam injection from low-porosity, depleted, lightoil reservoirs becomes a possibility—at least in someoil provinces. The process should be enhanced by dis-ti11ation12 or by still-to-be-developed synergisticmechanisms (possibly concurrent gas or air injection, oruse of only a fraction of a pore volume of steam as aprecursor, because it appears that nothing else will bankup residual, disconnected oil as steam does), and thencompleting the displacement with water, polymers, andsurfactantso

Estimated Barrels per Day ofTertiary Oil — Circa 1990How do”we go about estimating the barrels of tertiaryoil that will be added to pipeline runs in the comingdecades? Let us go back and look at the West Texascarbonate potential. We will assume a 20-year period isrequired to inject the primary amount of carbon dioxide;this is probably just enough time to pay out investmentin facilities and inject the reagent. We will assume weget started in 1980, that there is a 5-year delay to pro-duction response, and that the oil production ensuesover a 30-year. period. Our 4-billion bbl target thenwould’be ~roduced between 1985 and 2015 at the rateof 365,000 B/D. This goal seems too optimistic since itwould require the availability of 1 Tcf of carbon dioxideper year by 1980. We will discount it by 50 percent andset our goal on an average daily production of some

.175,000 B/D during the last 15 or so years of thiscentury.

The one chemical flood that has been announced isthe Marathon operation,’ referred to earlier. Its infenedperformance will average about 6,000 B7D over aperiod of some 15 years if the entire 6,000 acres in theRobhson field are developed. Were the Marathon ex-perience to be duplicated 10 times each year during thenext decade or so, a production rate from chemicalflooding would reach 600,000 B/D by about 1990, and750,000 B/D a few years later if still more opportunitiescan be found for development to be continued at thatpace. However, the estimate is overly optimistic, sincesufllcient unrecovered low-viscosity oil to fulfill the es-timate probably does not exist in reservoirs within about1,000 ft. Significant price increases, as already indi-cated, would be required to extend the usefulness of the

TABLE 4 — CALCULATEDSTEAM/OIL RATtO1.0 PV water injected se 70 percent eteam

S= O.3O-10=O.2OGross thickneee = 20 or 40 ft

Pieesure Injection Steam/Oii Ratio

(psi) Rate (band) %Acre Spacing 10-Acre Spacing

I,ooo 0.08/0.10 0.08/0.081,% 0.10/0.13 0.08/0,10

300 0.13/0.16 Q.1OIO.131,x 0.16/0.19 (1,13/0.18

584.

,.

process to deeper reservoirs, or a significantly superiorprocess would have to be developed. For the decade ofthe 1980’s, then, we would again discount the calcula-tion by 50 percent, and look for the production of some300,000 B/D by 1990. This would require the rede-velopment of some 300,000 acres by chemical floodingoperations.

Steamflooding will be extended to at least ~dditionalheavy oil targets in “California, and we have already in-dicated this could result in an increase in reserves of asmuch as 2. billion bbl. Extended over a period of 25”years, an average enhanced production of some 200,000B/D could be achieved. We~will not discount this esti-mate.:.Adding the three together results in a total of675,000 B/D by 1990. Allowing or hoping for furtherimprovements in the technology of these processes, theextension of steamflooding to some light oil projects,new technology to be invented, and, eventually, the’realization that energy will be relatively more costly toour nation than it has been historically, it is foreseeablethat tertiary oil might get to 1 million B/D by 1990.

This will be a very welcome addition to our nation’ssupplies: Pursuing these projects will be profitable ven-tures 10 those operators who have the amenable reser-voirs, and who have had the necessary technology de-veloped for them. This supply rate obviously will notalter the over-allenergy deficiency that the count~ willbe facing. By all accounts, the zenith of the age of pe-troleum in North America has been passed, and is cur-rently being reached in the rest of the world outsidecommunist areas, as shown by the supply projectionsdelineated in Figs. 17 and 18.2It is high time for us allto realize what the expectation for long-term petroleumsupplies actually is. We must have a national energypolicy that explores every possible way of continuing tomake an economically useful supply of energy availableto all of us. -

We should heed the conservationists who have lob-

?mnumm**~

[

!

Fig. 17 — &ude/LNG — historical develo ment andprojection of poeeible availabiii y$vorl$outside

communist areas and Nort America.

mwuclwl~s~,

-. I!

F .18 — Crude/LNG — historical development andYpro ectlons of possible availability, North America only.

JOURNAL OF PETRLXEUM TECHNOLOGY

bied (rightly so, in many instances) for the establish-ment of wild life and nature preserves. We must beginto lobby for the establishment of “people preserves”— American People — to be designated around everyshale mine, coal strip, and oil derrick on our shores.

AcknowledgmentsAppreciation is expressed to A. T. Guernsey for longago conversations that gave rise to this analysiw toC. H. Siebenhausen for important suggestions on thepresentation; to P. A. Payne for putting together someof the data on carbon dioxide; and to Shell Oil Co. fortime to think and permission to publish this paper.

References1. Reserves (# Crade Oil, Nataral Gas Liqaids, aad Namral Gas

in the United Stwes and Canada, API, published annually onDec. 31.

2. “Availability of OIL” SheIl Intemationele Petroleum MLJB,V.,handout lectures (Jsn. 1975)Appendices 9 and 10.

3. Lobeck, A. K.: Geomorpho/ogy, McGraw-HUlBook Co., Inc.,New York (1939) 459.

4. Ball., Interstate 011 Compact Commission Committee (Dec.1966)VIII, No. 2.

5. Bauer, G. G.: “Watefflooding in the Bradford Field,” Prod.Monthly (Nov. 1946) XI, No. 1, 22.

6. Richardson,J. E.’,Wyman, R. E., Jordan, J. P., and Mhchell, F.R.: “Methods for DeterminingResidual Oil With Pulsed NeutronCapture Logs,” J. Per. Tech. (May 1973) 593-606; Trans.,AIME, 255.

7. Gogsrty, W. B.: “ Status of Surfactant or Miceller Methods,”’J.Pet. Tech. (Jsn. 1976)93-102.

8. Subcommittee on Recovery Efficiency, API, Dallas. Personalcommunicationfrom C. 0: LHes, Asst. Dkector.

9. Craig, F. F., Jr.: “ A Current Appraisal of Field Miscible Slugprojects;’ J. Pet. Tech. (May 1970) 529-536.

10. improved Oil-Recave~ Field Repat@, Society of Petroleum En-gineers of AIME, Dallas (June 1975) 1, No. 1, 71, 103.

11. Stegemeier, G. L. attd Myhill, N. A.: “Stesm-lXve Correlationand Pre,@ction,*’paper SPE 5572 presented at the SPE.AIME50th AMUd Fall Techtiicsl Conference and Exhibition, Dallas,

- Sept. 28-Od. 1, 1975;12. I-Iagoort,J., Leijnee, A., and van Poelgeest, F.: “Steem-S&lp

Drive: A Potential Tertiary RecovecyProcess,” paper SPE 5570presented at the SPE-AIME 50th Annual Fall Technical Confer-ence and Exhibition, Dallas, Sept. 28-Ott. 1, 1975.

APPENDIXThe possibility that discovered original oil in place hasbeen incorrectly estimated will be examined briefly.Seventy-five percent of this oil was discovered before1950, and some improvements in making such estimateshave undoubtedly occurred. The annual API statisticalefforts include a re-estimation in terms of the current

.. .

MAY, 1976 ,

technology at the time of reassessment, as well as both~riginal oil in place and ultimate production.

One possible source of error in estimating the originaloil in place (OOIP) may not be fully accounted foq thatis the effect of pressure on porosity, Porosity before1950 would have been estimated from laboratory mea-surements It ambient conditions, However, it is wellknown that under subsurface conditions.the porosity isless than that at 1 atrn. The extent of the differencedepends very much on the nature of the reservoir rockand its burial history. Most ?xtreme differences will befound in unconsolidated sandstones.

Suppose the porosity was originally estimated fromco-nventionalcore analysis to be 36 percent. Using logs,the water saturation was calculated to be 25 percent.From these values, oil properties, and maps that wereconstructed of the reservoir, the initial oil in place wasdetermined by vohunetric analysis. This value for origi-nal oil in place would be used in association with pres-sure analysis and decline curves to determine the recov-ery efficiency quoted today.

Modem technology — logs and stressed core analysis— has demonstrated that many of the early porosityvalues determined from core analysis are optimistic, Ifporosity values are later redetermined and found to be,say, 28 percent, using Axchie’s equation with somesimplifying assumptions results in a water saturationvalue of 32 percent. Thus, our volumetric analysiswould become

OOIP,Ctu,l= 0.71 OOIPearlY~,ti~ate.

These numbers are only estimates of what could hap-pen. How much of this kind of error is in the estimateof original oil in place and, therefore, also of recoveryefficiency? Corrections of this type would be mainlyapplicable to sandstone reservoirs discovered before the1960’s. Even if the correction calculated above appliedto one-half the oil discovered before 1950, the over-alleffect on original oil in place would be a reduction of10 percent, and estimated recovery efficiency wouldbe boosted a few percentage points. The correction isviewed as an outside limit, and, even at that, it wouldnot result in any significant change in the conclusionspresented in this report. J3?T

Original manuscriptrecokd h society of Petroleum Engineers office Jan. 30,1976. PaPar accepted for publication Feb. 13,197$ Re~ised manuscript reoelvedMarch 9, 1976. Paper (SPE,5S00) was first presented et the SPE-AIME Improved011Recovery Sympoelum, held in Tulsa, Okla., March 22-24, 1976. @ Copyright1976 American Inetitute of Mining, Metallurgical, and Petroleum Engineere, Inc.

.-. ... .

. . 585—----,,

Is

$“ 10

0-l

5

00 40 80 120

CUMULATIVE DAYS OF STEAM INJECTION

!!

O FIELD RESULTS

.

CORRELATION(EQ.9)

RUBINSHTEIN

Fig. 7 - Field measured ad theoreticalcumulative heat loss from formati on.