You are on page 1of 10

ESTIMATION of PRIMARY OIL RESERVES

J . J. ARPS BRITISH-AMERICAN O I L PRODUCING C O .


MEMBER AlME DALLAS, TEX.

ABSTRACT recoverable oil reserves and a pro- chart is presented (Fig. 1) showing
jection of future rates are: [a.] the the three periods in the life of an
This paper reviews the methods exploitation and development of an imaginary oil property. Time is
currently in use for estimating pri- oil reservoir; [b.] the construction of shown on the horizontal axis, while
mary oil reserves and discusses the gasoline plants, pipelines and refiner- the cumulative production and esti-
principles on which these methods ies; [c.] the division of ownership in mated ultimate recovery are plotted
are based. Particular emphasis is unitized projects; [d.] the price to be vertically. No particular units are
placed on how these methods change paid in case of a sale or purchase of used and this schematical chart is
with the type of information avail- an oil property, and the magnitude not to scale.
able during the life cycle o f an oil of the loan which it will support; During the first period, before any
property. This paper contains various [e.] the proper depreciation rate for wells are drilled on a property, any
novel estimating methods and short- the investment in oil properties; and estimates will of necessity be of a
cuts heretofore unpublished. [f.] evaluation of the results of an very general nature, based on ex-
exploration program. perience from similar pools o r wells
INTRODUCTION This discussion will be confined to in the same area, and usually ex-
Estimating oil reserves is one of the various methods and tools which pressed in barrels per acre. This will
the most important phases of the are currently in use for estimating
work of a petroleum engineer since oil reserves to be obtained during the
the solutions to the problems he primary phase of an oil-producing
deals with usually depend on a com- reservoir and for a projection of the
parison of the estimated cost in future production rates. Reserves
terms of dollars, with the anticipated which may be obtained by secondary
result in terms of barrels of oil. His recovery methods or fluid injection
recommendations to management re- programs and gas and gas conden-
garding the best course of action are sate reserves will not be discussed
therefore normally based on the most in this paper.
favorable balance between these Unfortunately, reliable oil reserve
two. figures are most urgently needed dur-
ing the early stages when only a
Specific engineering problems minimum amount of information is
which require such a knowledge of available. Management's interest
in the oil recovery from a prop-
Original manuveript received in Petroleum
Branch office on Feb. 26, 1956. Revised manu- erty-aside from its use for account-
script received June 29. 1956. Paper presented ing purposes-usually declines when
at the Petroleum Conference-Economics and
Valuation, Dallas. Tex., March 29-30. 1956. the property approaches its economic
Discussion of this and all following techni-
cal papers is invited. Discussion in w r i t ~ n g( 3 limit, just at the time when the re-
copies) may be sent to the offices of the
Journal of Petroleum Technology. Any dis- liability of the estimates is at its best.
cussion offered after Dee. 31. lQ56, ghould hc
in the foim of a new paper. T o illustrate this general idea n

182 PETROLEUM TR i h S A L I I0\5, \ IM1


thcrcfore be called the "barrels per West Edmond, Spraberry and the will therefore be assumed that the
acre" period. The estimates can like are usually over-estimated dur- reservoir mechanism can be pre-
range all the way from AB (prop- ing the early stages. dicted and that the necessary data
erty non-productive) to CD, reprc- It should be emphasized that, as on the reservoir rock and reservoir
senting the most optimistic view. in all estimates, the accuracy of the fluids are o r can he made available.
The second period follows after results generally cannot be expected
one or more wells are drilled, and, to exceed the limitations imposed on
assuming that the property is pro- it by inaccuracies in the available Pools without an active water
ductive, it is now possible to pin the basic data. In other words, the bet- drive and which produce solely as
estimated ultimate recovery down ter and more complete the available the result of expansion of natural
within the much closer limits E F and data, the more reliable will be the gas liberated from solution in the
GH. During this time interval most result. Another factor which is often oil are said to produce under a de-
estimates are on a volumetric basis overlooked - particularly in cases pletion mechanism; also termed an
and are based on acre-feet of pay where property values are involved internal o r solution gas drive. When
and a unit recovery expressed in har- -is that additional investment in a free gas cap is prescnt this mec-
rels per acre-foot. The type of in- acquiring good basic data during the hanism may be supplemented by an
formation available during this per- early stages oftentimes pays off later. external o r gas cap drive. When the
iod consists of well logs, core anal- With good basic data available the reservoir permeability is sufficiently
ysis data, bottom-hole sample infor- engineer making the estimate nat- high, the oil viscosity low, and when
mation and a subsurface map. Inter- urally feels more sure of his results, the pay zone has sufficient dip or a
pretation of this data in the light of and will be less inclined to the cau- high vertical permeability, the deple-
early pressure behavior may lead to tious conservatism which sometimes tion mechanism may be followed o r
creeps in when many of the basic accompanied by gravity segregation.
conclusions regarding the type of
productive mechanism to be ex- parameters are based on guesswork When a depletion-type reservoir is
pected. Production decline curves or only. first opened to production its pores
other trends are not yet available contain :qterstitial water and oil with
during this second, or "barrels per VOLUMETRIC ESTIMATES gas in solution under pressure. N o
acre-foot", period. free gas is assumed to be present in
The third period follows after the oil zone. The interstitial water is
sufficient actual performance data on usually not produced, and its shrink-
the property have become available During the second, o r "barrels per age upon pressure reduction is negli-
to make a check of previous volume- acre-foot" period, reserve estimates gible, compared to some of the other
tric estimates against decline curve are based on a knowledge of the oil- factors governing the depletion-type
trends possible. Also, the pressure or gas-bearing reservoir volume and recovery.
hehavior may now make material the physical characteristics of the When this reservoir reaches the
balance work possible, thereby offer- reservoir rock and the reservoir fluids end of its primary producing life,
ing other valuable clues regarding contained therein. These data must and disregarding for the time bcing
the type of production mechanism be supplemented with an informed the possibility of gas cap drive o r
and active oil in place. This in- guess as to the reservoir mechanism gravity segregation, it contains the
creased amount of information which will govern the withdrawals same interstitial water as before, to-
causes the lines HJK and FLK, during the primary production per- gether with a certain amount of resi-
which delineate the range between iod. dual oil under low pressure. The
optimistic and pessimistic estimates, The petroleum engineering liter- void space vacated by the oil pro-
to converge gradually in point K. ature of the last 25 years has dealt duced and by the shrinkage of the
which represents the true ultimate extensively with the determination remaining oil is now filled with gas
rccovery or cumulative production of physical characteristics of the liberated from the oil. During the
at abandonment time. The dashed reservoir rock - such as porosity, depletion process this gas space has
curve BK represents the cumulative permeability, relative permeability to increased gradually to a maximum
recovery from the property. the different phases, wetting charac- value at abandonment time. The
It follows from the foregoing that teristics, capillary pressure, and in- amount of gas space thus created is
it is desirable for the engineer mak- terstitial water-as well as with the the key to the estimated ultimate re-
ing a reserve estimate to provide his significant physical characteristics of covery under a depletion mechanism.
client o r his management with the the reservoir fluids such as their vis- It is reached when the produced
possible sprcad in estimated ultimate cosity, compressibility, gas solubility, free gas-oil ratio in the reservoir,
recovery, both high and low, or, if shrinkage, and such other parameters which changes according to the rela-
he provides them with a single fig- as surface tension and interfacial tive permeability ratio relationship
ure, to point out the probable error tension. and the viscosities of oil and gas
in his estimate. Prediction of the reservoir mech- involved, causcs exhaustion of the
In reviewing the histories of many anism under which recovery will be available supply of gas in solution.
reserve estimates over an extended obtained during the primary phase
period of time in many different FINALFREEGASSPACE
of production also has found wide-
fields, it seems to be a common ex- FROM COREANALYSIS
spread treatment in the technical
perience that the really good fields literature since the middle thirties. If a sufficiently large number of
such as East Texas, Oklahoma City, particularly since the introduction of accurate determinations of the oil
Yates or Redwater, etc.. have been the material balance concept and a and water saturation on freshly-re-
generally under-estimated during the better understanding of the inter-re- covered core samples is available, an
early barrels per acre-foot period action between the oil reservoir and approximation of the total free gas
compared to their later perform- its adjacent aquifer, if such an aqui- space to be expected in a reservoir
ance, while the poorer ones such as fer is present. Tn this discussion it can be obtained from them. This
method is based o n the assumption the reservoir rock contains, in bar- tics of the reservoir rock and its
that the depletion process taking rels per acre-foot: fluid content, this table has been
place within t h e core upon the Interstitial water 7758.f.S~
found very helpful in estimating the
Free gas 7758.f.5,
rcduction in pressure by bringing it Reservoir oil 7758.f.11-Su--Sg) possible range of depletion recovery
to the surface is somewhat similar Stock tank oil 7758.f.ll-Sw-SI)
6.
factors. It may be noted that in
to the actual depletion process in in which S, stands for the final gas general the most important single
the reservoir. Possible loss of liquids space as a fraction of the total pore factor governing the recovery factor
from the core before analysis may space and B, is the single phase is the K , / K , relationship of thc res-
cause this total saturation to be too formation volume factor for oil at ervoir rock. Unconsolidated inter-
low. On the other hand. the smaller abandonment reservoir pressure. The granular material seems to be the
amount of gas in solution in the unit recovery factor, DR, for a de- most favorable, while increased ce-
residual oil has a tendency to reduce pletion-type reservoir is therefore: mentation or consolidation tends to
the final gas phase. Those using this affect recoveries unfavorably. Infor-
method hope that these two effccts mation obtained from various sources
somewhat compensate for each other. since publication of this paper:' seems
to indicate that the K,/K, curve No.
23 as shown in Fig. 2 for Wasson
Dolomite was very likely in error
In 1945 R. C. Craze and S. E. and the maximum case for lime-
Buckley', of the special API study Another approximatc approach to
the problcm of estimating recoveries stones and dolomites is therefore too
committee on Well Spacing and high. It is recommended that the
Allocation of Production, collected from depletion-type reservoirs is the
use of the correlations developed for maximum case for sands and sand-
a large amount of statistical data
on the performance of some 103 six different types of reservoir rocks,
oil reservoirs in the United States representing maximum, average and
for the purpose of a statistical minimum conditions for sands and
analysis of the well spacing prob- sandstones and for limestones, dolo-
lem. Twenty-seven of these fields mites and cherts" In this study these
appeared to be producing under a six types of reservoir rocks were as-
depletion-typc mechanism. They pro- sumed to be saturated with 12 syn-
duce from formations widely differ- thetic crude oil gas mixtures, repre-
ing in geologic age and characteris- senting gravities from 15" to 50"
tics and have been operated with dif- API, and gas solubilities from 60 to
ferent spacings and efficiencies. A 2,000 cu ft/bbl, and their production
study of the final gas saturation of performancc computed by means of
these fields shows considerable scat- the wcll-known depletion equation
tering of the data, as is to be ex- (see Eq. 2 on this page) in which the
pected. Most points seem to fall. symbols correspond to the letter sym-
however, between 30 and 40 per bols recently submitted by the Petro-
cent of the pore space, with the av- leum Branch Executive Committee of
erage at 30.4 per cent. the AIME', with the units defined as
The average crude oil viscosity
in Ref. 3.
under initial reservoir conditions of The results of this study, express-
these fields was 2.2 cp. whilc the ing recovery factors as a percentage
average solution gas-oil ratio was of stock tank oil in place, are repro-
378 cu ft/bbl. As shown by Muskat duced here as Table 1. In cases
and Taylor', higher solution gas-oil where no detailed data are available
ratio or lower oil viscosity in the concerning the physical characteris-
same reservoir rock tend to increase
the final gas saturation and vice
versa.
n .%, -
-- - -
B, B., AP - --
B, /I< k,, AP

Under initial conditions the res-


ervoir rock contains, in barrels per
TABLE I-PRIMARY RECOVERY I N PER CENT O F O I L I N PLACE FOR DEPLETION-TYPE RESERVOIRS
acre-foot :
Interstitial water 7758.f.S~- O i l Solution Oil
Reservoir o i l 7758.f.11-Su.1 GOR Grovity Sand or Sandstones Limestone, Dolomite or Chert
Stock tank oil 7758.f.11 -Swl ICu f t l b b l l ["APII Maximum Average Minimum Moximum Average Minimum
Bi
in which f stands for porosity as a
fraction and S, for interstitial hater
as a fraction of the pore space; B ,
is the single phase formation volume
factor for oil undcr initial condi-
tions.
U n d e r abandonment conditions
-
1Refercnces r i v e n at end of paper.
stones should be used for both types low permeability zone. In such a WATER DRIVE RESERVOIRS
of rock. case separate calculations should be
Natural water influx into oil reser-
Next in importance is crude oil made for each permeability bank
gravity with viscosity as its corollary. which is known to be continuous and voirs is usually from the edge inward
the results converted into rate-time parallel to the bedding planes (edge
Higher oil gravities and lower vis-
curves for each. The estimated ulti- water drive) or upward from below
cosities appear to improve the re-
covery. The effect of gas solubility mate should then be based on a su- (bottom water drive). Bottom water
on recovery factors is less pro- perposition of such curves for the drive is only possible when the reser-
nounced and shows no consistent different zones. In that case one may voir thickness exceeds the thickness
pattern. Apparently the beneficial ef- find that the more permeable banks of the oil column, so that the oil-
fects of lower viscosity and more are depleted and have yielded their water interface underlies the entire oil
effective gas sweep with higher gas- full unit recovery, while tighter zones reservoir. It is further only possible
oil ratios is sometimes offset by the are still partially saturated. The pro- when vertical permeabilities are high
higher formation volume factors. duction rate from these tighter zones and there is little or no horizontal
I n the previous refercnce3 the re- may, however, be so slow that the stratification with impervious shale
sults were expressed in Table 1 as combined rate for all zones has al- laminations.
recovery in barrels per acre-foot per ready reached the economic limit. In either case, water as the displac-
per cent porosity rather than as re- ing medium moves into the oil-bearing
covery in per cent of oil in place. section and replaces part of the oil
In general, these data seem to indi- originally prescnt. The key to a volu-
cate a recovery range from the poor- Additional recovery can some- metric estimate of the unit recovery
est combinations of 2 or 3 bbl/acre- times be realized when a high pres- by water drive is in the amount of oil
ft for each per cent porosity to the sure reservoir is undersaturated so which is not removed by the displac-
best combinations of 17 or 18 bbl/ that it may produce by virtue of the ing medium. This residual oil satura-
acre-ft/per cent. An over-all average expansion of its own reservoir oil, tion after water drive plays a role
seems to be around 10 bbl/acre-ft/ its interstitial water and of the res- similar to the final gas saturation in
per cent. Analysis of the Craze-Buck- ervoir rock itself, until the bubble the depletion-typc reservoir discussed
ley data indicates an over-all average point is reached. in the previous section.
of the order of 13 bbl/acre-ft for This additional recovery by ex- In order to determine the unit re-
each per cent porosity. pansion above the bubble point is covery, it is necessary to again com-
If the actual PVT data on the res- sometimes quite substantial. In the pare the amount of interstitial water
ervoir fluids are available, as well as case of the D-7 zone in the Ventura and oil with dissolved gas initially
all the necessary characteristics of Avenue field, described by E. V. present with the condition at aban-
the reservoir rock, the best way to Wattss, 40 per cent additional re- donment time, when the same intersti-
arrive at a volumetric estimate is to covery was indicated from this tial water is still present but only the
actually compute the pressure-satura- source. residual o r non-floodable oil is left.
tion relationship according to Eq. 2. The remainder of the original oil has
The accuracy of this type of calcula- at that time been removed by water
tion falls off rapidly if the pressure When conditions are favorable for displacement.
increments chosen are too large, par- gravity segregation or gravity drive,
ticularly during the final stages when such as low oil viscosity, large pores
the gas-oil ratio is increasing rapidly. and high vertical permeability or
In the study of Ref. 3 where compu- steep dips, the ultimate recovery can
tations were carried out with IBM be estimated from a knowledge of Calculation of the residual oil per-
equipment, pressure increments of 10 the residual oil saturation remaining centage from laboratory data on the
psi proved satisfactory. after such gravity segregation has relative permeability relationship for
BabsonI4 and Tarner'j have ad- taken place. Very little factual in- reservoir oil and water combined with
vanced other computation methods formation on residual saturation after the necessary viscosity data is possible
which require a much smaller num- gravity segregation is available in the by means of the Buckley and Leverett
ber of pressure increments and can literature. D. L. KatzGreports that in frontal drive method', which has been
therefore be handled by desk calcu- the Oklahoma City Wilcox reservoir modified and simplified by Pirsonbnd
lator. These methods are based on oil saturations were found in the gas Welge" The difficulty of obtaining a
a trial-and-error type of solution of zone varying between 1.0 and 25.9 reliable relative permeability relation-
material balance and gas-oil ratio per cent, while oil saturations of be- ship for the reservoir rock being stud-
equations. tween 52.7 and 92.8 per cent were ied has, however, somewhat restricted
found below the gas-oil contact. the practical use of this ingenious
These figures are in general agree- mcthod.
ment with actual recoveries com- RESIDUAL OIL SATURATION
The permeability distribution in puted by comparing the oil withdrawn FROM COREANALYSIS
most reservoirs is sufficiently non-
with the volume of sand depleted. The The method most commonly used
uniform in vertical and horizontal
directions so that the depletion cal- writer found by studying the receding is to consider the oil saturation as
culations as above on average ma- gas-oil contact, from detailed Okla- found by ordinary core analysis after
terial should be fairly representative. homa City Wilcox fluid level data, correction for shrinkage as the resid-
However, when distinct layers of and by comparing the oil produced ual oil saturation to be expected from
high and low permeability are known with the sand volume from which it flooding with water. This is based on
to be present the depletion process was drained, that recoveries as high the assumption that water from the
may advance much more rapidly in as 70 per cent of the pore space were drilling mud invades the pay section
a high permeability hank than in a not unusual. just ahead of the core bit in a manner
similar to the water displacement tor and the various significant parame- an oil-water contact will advance
process in the reservoir itself. ters. Their equation reads: more or less in the same manner as if
Recovery fraction = .I 1403 + the formation were entirely uniform.
2719 logK +
25569 S, - In such a case the unit recovery fac-
.I355 10gp<~- 1.5380 f - tor should be representative except
.00035 H . . . . . . (3) possibly for a correction to allow for
The analysis by R. C. Craze and in which K represents permeability in efficiency of the drainage pattern. In
S. E. Buckley' referred to in the previ- millidarcies, S, the interstitial water other reservoirs there may be distinct
ous section is also a valuable source saturation, p, the oil viscosity in centi- layers of higher and lower permeabili-
for statistical data on residual oil sat- poises, f the porosity fraction and H ties which appear to be more or less
uration after water drive. Some 70 of the pay thickness in feet. continuous across the reservoir. In
the 103 fields analyzed produced such a case water may advance much
wholly or partially under water drive Since Craze and Buckley's data was
arrived at by comparing indicated re- more rapidly through the high perme-
conditions. Like the depletion-type ability streaks than through the tighter
fields, they were widely distributed coveries per acre-foot with porosity
data from the reservoir rock, the rc- zones. Allowance must then be made
geographically and produced from for this permeability distribution. In
formations which differed in their sidual oil calculated by this method
includes a recovery efficiency factor case the nature of the continuity of
geologic ages, physical properties, and different permeability streaks is un-
structural characteristics, contained which is not present when basing the
estimate on residual oil as found in known, the estimate which takes the
oils of different properties and pro- permeability distribution into account
duced under varying degrees of oper- the cores. This recovery factor re-
flects the effect of spacing, if any, and should be considered as a conserva-
ating efficiency. A study of the calcu- tive one, while the one based on a uni-
lated residual oil saturations for these of by-passing some of the oil in less
permeable strata by more rapid water form reservoir throughout should re-
fields shows a wide range; from 17.9 sult in an optimistic figure. W. E.
to 60.9 per cent of the pore space. influx through permeable zones. Also,
it includes the effect of having to Stiles l1 in 1949 showed how the re-
However, the data seem to relate covery from a waterflood property
themselves distinctly according to the abandon a property before the water-
may be computed, while taking into
oil viscosity and the reservoir permea- flooding action in all zones is com-
plete, because of the water-oil ratio account the permeability distribution.
bility. A simplified version of his method in
reaching the economic limit. When
The average correlation indicated using this data, therefore, one will not tabular form is shown on Table 2. The
by the authors between oil viscosity data on this tabulation are adopted
arrive at the unit recovery under 100
and residual saturation, both under from a Tensleep sand reservoir in
per cent efficient water flooding, as
reservoir conditions, is shown by the Wyoming where good statistical aver-
would be obtained from the core an-
following tabulation: ages on more than 3,000 core analy-
alysis data, but rather find an over-all
Reservoir Oil Viscosity Residual O i l Saturation ses were available. Part of these cores
average recovery per acre-foot.
(in CP) (per cent of pore space1 were taken with water-base mud,
0.2 30
CALCULATION OF TIIE UNIT which yielded the residual oil figures
RECOVERY UNDER A WATER on Line 6. Another portion was taken
DRIVEMECHANISM with oil-base mud and yielded the in-
Under initial conditions the reser- terstitial water figures of Line 8.
voir rock contains, in barrels per acre- The cores were divided into the five
The deviation of the individual data permeability groups shown on Line 1
from this average showed the follow- foot:
Interstitial water 7758.f.S~ and appeared to have a fractional dis-
ing trend against average formation Reservoir oil 7758.f.(l-SW)
tribution of the samples as shown on
Stock tank oil 7758.f.(l -Swl
permeability: Line 2. Line 3 carries the average per-
BO
Average Rerervoir
Deviation of Residual O i l
Saturation from Viscosity
Under abandonment conditions the meability for each group, while Line 4
Permeability Trend reservoir rock contains, in barrels per shows the capacity of each group in
(in md) (per cent of pare space)
50 112
acre-foot : terms of darcy feet. For this purpose
Interrtifial rmter 7758.1.5,
Residual oil
it was assumed that the calculation
(Stock tonk conditions) 7758.f.So applied to a representative 100-ft total
in which So stands for the residual oil section.
saturation under surface conditions as The relative water permeability K,
J fraction of the total pore space and
According t o these statistical behind the flood front is shown on
B,, is the single phase formation vol- Line 7, while the relative oil permea-
trends, the residual oil saturation un- ume factor for oil. The unit recovery
der reservoir conditions for a forma- bility K , ahead of the water in the
factor for a water drive reservoir clean section is listed on Line 9.
tion containing 1 cp oil and having an
W R is, therefore: These relative permeabilities were
average permeability of 500 md can
be estimated at 34.5 +
2, or 36.5 per based on laboratory measurements.
The unit recovery factor for each
cent of the pore space.
group was then computed by means
of Eq. 4. The formation volume fac-
tor B, used was 1.07.
I n another more recent statistical On Line 11 is shown the cumula-
study of Craze and Buckley's water Eq. 4 shows the unit recovery fac- tive +vet capacity in darcy feet under
drive field recovery data by Guthrie tor for a 100 per cent effective water the assumption that the five groups
and Greenberger'kultiple correla- drive. In many reservoirs the permea- are watering-out progrcssively in the
tion analysis methods were uscd to bility distribution is sufficiently non- order from high to low permeability.
determine the best-fitting correlation uniform in both vertical and horizon- The figures shown are the cumulatives
between the water drive recovery fac- tal directions so that a flood front or of those on Line 4. On Line 12 the
cumulative remaining or clearz oil ca- the produced stream vs cumulative oil ery may be obtained due to this so-
pacity is carried, which is equal to the recovery per acre-foot will be found called buoyancy effect.
rota1 capacity of 3.243 darcy ft minus on Fig. 2. Based on economic limit In addition to these buoyancy phe-
the wet capacity on Line 11. The considerations, a final oil percentage nomena there is also the effect of cap-
water-oil ratio of the produced stream of 2 per cent is indicated for this field illarity and preferential wetting of the
from such a linear water flood is and the corresponding recovery at reservoir rock by water. Imbibition of
shown on Line 13 based on: that time can be found by interpola- water from fractures and vugular ma-
tion as 292 hbl/acre-ft. terial into the low permeability matrix
As stated before, unless the conti- as the water advances may materially
nuity of the various permeability aid the above buoyancy mechanism,
zones across the structure is definitely but are much more difficult to eval-
or, with an oil viscosity of 5.5 cp and established, this recovery figure should uate quantitatively. To make a calcu-
a water viscosity of 0.46 cp: lation of the recovery under a buoy-
be treated as a conservative one.
(7) (11) ancy mechanism it is necessary to first
W O R = 12 X -- X --- . . ( 6 )
(0) (12) determine by statistical analysis of a
It will be noted that the WOR at In limestone pools producing un- large number of cores the average
the time all permeabilities down to der a bottom water drive, such as interval between high permeability
100 md are wet has reached 15.5. certain of the vugular D-3 Reef res- zones. A separate computation is then
When the watering-out process has ervoirs in Alberta, one finds an ex- made for each of the permeability
advanced down to 50 md the W O R is treme range in the permeabilities, ranges to determine what pcrcentage
36.0. When groups 3 and 4 succes- often running from microdarcies on of the matrix oil contained in a theo-
sively watcr out the W O R has in- up to the darcy range. Under those retical tube of such average length
creased to 76.5 and 307.9, respec- conditions t h e modified "Stiles" may be driven out during the produc-
tively. method heretofore described yields ing life of the reservoir under the
On Line 14 is shown the cumula- results which are decidedly too low. effect of this buoyancy phenomenon.
tive oil production from each group The reason is that in pools like the Surprisingly improved recoveries
at the time group 1 is watered-out Redwater D-3 there is a substantial are generally indicated by this method
(Min K,., = 100 md; W O R = density difference between the rising over what one would expect from a
15.5).In group 1 the recovery of 61.6 saltwater and the oil. While the water "Stiles" type calculation and the re-
listed is the product of Line 2 and rises and advances through the highly sults from recent studies of the rise in
Line 10 (full unit recovery). For each permeable vugular material, it may at water table seem to confirm the valid-
of the other groups the full recovery first bypass the low permeabilitv ity of this concept.
is reduced in the proportion of its matrix material leaving oil trapped
average permeability to 100 md, therein. However, the moment such ESTIMATES BASED ON
based on Stiles' assumption that the bypassing occurs, a buoyancy gradi- PERFORMANCE DATA
advance of a linear flood will be ap- ent is set up across this tight material
proximately proportional to its aver- which tends to drive the oil out ver- Estimates of ultimate recovery by
age permeability. The total recovery tically into the vugular material and extrapolation of a performance trend
for all groups corresponding to WOR fractures. In the case of Redwater fundamentally all follow the same
= 15.5 is shown as 175.4 bbl/
pattern. The two quantities one
D-3, where the density difference be- wishes to determine are usually either
acre-ft in the last column. In a similar tween salt water and oil is 0.26. while
-
remaining oil reserves or remaining
-
~

manner the cumulative oil production the vertical permeabilities for matrix
for Min K,,.,., 50 md; WOR = productive life. Cumulative produc-
material are only a fraction of the tion and time are therefore normally
36.0, or 260.4 bbl/acre-ft. is com-
horizontal permeabilities, a simple selected as independent variables, and
puted on Line 15, and for Min K,c,.,
- 25, 10 and 0 md, respectively, calculation based on Darcy's law ap- plotted as abscissae. A varying char-
344.5, 418.3 and 55!),6 bb]/acre-ft, plied to a vertical tube with a cross- acteristic of the well performance
on Lines 16, 17 and 18. A graphical section of 1 Sq cm shows that during which can be easily measured and re-
presentation of the results of this com- the anticipated lifetime of the field corded is then selected as dependent
putation in terms of 011 percentage in very substantial additional oil recov- variable to produce a trend curve. For
extrapolation purposes this variable
has to meet two qualifications: (1)
TABLE 2-COMPUTATION OF THE WATER DRIVE RECOVERY FACTOR FOR A TENSLEEP SAND its value must be a continuous func-
RESERVOIR I N WYOMVNG B Y T H E "PERMEABILITY-BLOCK" OR MODIFIED STILES METHOD
tion of the independent variable and
GROUP I 2 3 4 5 Total
~ - - - change in a uniform manner; and (2)
1.
2.
Permeability Range ( m d )
Fraction of S a m p l e r
it must have a known end point.
3. Average Permeability (mdl By making a graph of the values
4. Capacity i n dorcy feet
5. Avg Porosity Fraction (I] of this continuously changing depend-
6. Avg Resid. Oil Fraction IS,,)
8. Avg 1nte:stitial Woter Froc. ISl,-) ent variable as ordinates against the
7. Relative Woter Perm. (KT,.) value of the independent variable
9. R e l a t i v e Oil Permeability ( K O )
10. E r t . U n i t Recovery Factor (bbl/acre-fl) (cumulative production or time) as
I I. C u m . "Wet" Cap. Z (41
12. C u m . "Clean Oil" Cap.=-3.243-111) abscissae and graphically extrapolat-
13. Water-Oil Ratio ing the apparent trend until the known
14. Cum. Rec. WOR - 15.5
Min Ka.,,r 100 end point is reached, an estimate of
-
:-
15. C u m . Rec. WOR 36.0
Min K w c t
-

50 the remaining reserves or remaining


16.

17.
C u m . Rec. WOR - 76.5
Min K n . , , t
C u m . Rec. WOR
25 -
--
307.9
life can be obtained. The basic as-
sumption in this procedure is that
Min K,,.,.t : 10 whatever causes governed the trend
18. Cum. Rec. WOR =. '/i
Min -I= 0 of a curve in the past will continue to
govern its trend in the future in a uni- timc is a constant fraction of the pro- types of decline curves the decline
form manner. duction rate. fraction is instantaneous and theoret-
This extrapolation procedure is With hyperbolic decline the drop ically applies only to a very short time
therefore strictly of an empirical na- in production per unit of time as a interval.
ture, and a mathematical expression fraction of the production rate is pro- An analysis of a large number of
of the curve based on physical con- portional to a fractional power of the actual production decline curves as-
siderations of the reservoir can only production rate, this power being be- sembled by W. W. Cutler, Jr.'3 indi-
he set up for a few simple cases. The tween 0 and 1. cates that most decline curves nor-
dependent variables most commonly With harmonic decline the drop in mally encountered are of the hyper-
selected, and their extrapolation, are production per unit of time as a frac- bolic type, with values for the ex-
as follows. tion of the production rate is directly ponent n between 0 and .7, while the
proportional to the production rate. majority fall between 0 and .4. The
Integration of the basic differential occurrence of harmonic decline ( n =
Rate of production is by far the equations leads to rate-time rclation- 1) is apparently rare.
most popular dependent variable used ships, as shown o n the classification As a matter of convenience the
when production is not restricted. In of production decline curves of Fig. 3. semi-log paper is most often used for
that case one commonly refers t o pro- After integrating these rate-time rate-time extrapolations, while regu-
duction decline curves. The two main equations a second time the rate- lar coordinate paper is favored for
types are rate-time and rate-cumula- cumulative relationships shown in ratecumulative extrapolations. Since
tive curves for each of the two inde- this same table are obtained. On the straight line extrapolation in these
pendent variables. Rate of production chart of Fig. 4 arc shown the trends cases assumes a constant percentage
has the advantage of always being of these three types of rate-time and decline, it will be obvious that such
readily available and accurately re- rate-cumulative curves on regular co- extrapolations therefore generally
corded. When no major changes in ordinate paper, semi-log paper, and provide results which are too conserv-
operating procedure are made, and no log-log paper. ative. Experienced engineers usually
stimulation treatments are applied, Inspection of this chart shows that allow for this by graphically flattening
the curves normally show a fairly in the case of constant percentage de- the declinc slope in the later stages.
smooth declining trcnd over extended cline thc rate-time curve becomes a
periods. This trend usually lends itself straight line on semi-log paper, while
well to extrapolation. T h e second re- the rate-cumulative curve straightens Another popular variable which is
quirement is also easily met, since out on regular coordinate paper. I n oftentimes substituted for the pro-
known o r estimated operating costs either case the tangent of the angle of duction rate in water drive fields is the
make it possible to determine the eco- slope is equal t o the decline fraction. oil percentage of the total fluid pro-
nomic limit rate, o r the end point of I n the case of hyperbolic-type dc- duced. Since projections of this oil
the curve. This economic limit rate is cline curves the rate-time relation- percentage vs time are not often
the production rate which will just ship as well as the rate-cumulative re- required, one usually finds this oil
meet the direct operating expenses of lationship can be straightened out percentage variable o n l y plotted
a well. In determining this economic after shifting to become straight lines against cumulative. An example of
limit it is often advisable not t o use on log-log paper. T h e shifted rate- this o n semi-log paper is shown for a
the operating expenditures charged t o cumulative curve in this case assumes T a r Springs reservoir in the Calvin
a well but to analyze these costs a reverse slope. Besides the extra field of Illinois in Fig. 5. The end
closely and determine how much work involved in shifting, this type of point in this case is the lowest oil
would actually be saved if the well paper also has the disadvantage that percentage which, combined with the
were abandoned. This saving yields a the horizontal scale on which the un- fluid-producing capacity of the lease,
more reliablc yardstick of the true known variable is plotted usually will just cover operating expenses.
economic limit of production, since becomes rather crowded at the point CUMULATIVE GAS VERSUS
certain expenses may have to be con- where the answer is desired. F o r this CUMULATIVE OIL
tinued if other wells on the lease are reason, special graph paper for hyper-
kept in operation. T h e following is a It is a characteristic of most oil
bolic decline was designed", making
sample calculation of the economic reservoirs that only a fraction of the
it possible t o plot either time or cumu-
limit for a well: oil in place is recoverable by primary
lative on a linear scale and still obtain
Crude Price per bbl $2.80 production methods. Gas, on the
Gar Revenue per bbl $0.20 the advantage of straight line extra-
other hand, moves much more freely
Total $3.00 polation. through the reservoir and it can gen-
Royalty (12.5 per cent) $0.375
Local Taxer (per h ' bbl) $0.125 In the case of harmonic decline it
erally be assumed that at abandon-
Leaves N e t Income per Gross Barrel $2.50 may be noted that the rate-time rela-
Ertimatsd direct operating cost ment time only the solution gas in the
ot Economic Limit $250 per month tionship can also be straightened out remaining oil at the then-prevailing
Estimated Economic o n log-log paper after shifting, and
Limit Rote 100 gross bbl/month pressure plus the free gas at that same
T h e mathematical background of assumcs a slope of 45". It may be of
pressure are left in the reservoir. In
these rate-time and rate-cumulative interest that in this specific case a plot other words, even though it is not
curves in connection with the so- of the inverse of the production rate known exactly how much oil may be
called loss ratio method of extrapola- versus time on a linear scale should recovered, a much firmcr idea is gen-
tion has been discussed in detail in a also yield a straight line. T h e rate- erally available of the amount of gas
previous publication". cumulative relationship for harmonic
decline becomes a straight line on that will be produced during the pri-
Three types of decline curves are mary production period. This pro-
commonly recognized: constant per- semi-log paper. T h e decline fraction
in this case is equal to the rate times vides us with the possibility of an end
centage decline, hyperbolic decline
and harmonic decline. the tangent of the slope angle. point to ;I performance curve. The
With constant percentage decline It should be stressed that in this cumulative gas-cumulative oil method
the drop in production per unit of mathematical treatment of different is illustrated with the chart on Fig. 6.
II HYPERBOLIC m. HARMONIC
DECLINE DECLINE DECLINE
I -

BASIC
CHARACTERISTIC
I DECLINE I S CONSTANT
a-o
D E C L I N E I S PROPORTIONAL
TO A F R A C T I O N A L POWER (R 1
OF T H E P R O D U C T I O N R A T E
D E C L I N E IS PROPORTIONAL.
TO PRODUCTION R A T E
a = 8
I I o < n

D:,.qOz - dkr
1 FOR INITILL CONOITIONS ( FOR WITIK CONOITIONS:

S u b s t ~ t u t e From Rote-itme Equotron Subslrtuie From Rote - ttme Equolkon:


I Substttute From Rote - tune Equotbon:
I
To Find

I 1 a, = L- I
RELATIONSHIP D I
1
I D - D e c l ~ n e as o f r o c t l o n o f productLon r o t e ql = Productron rote o t tbme t

( D, = I n t t ~ o l decline 0, = Cumulotcve o t l productron ol t i m e I

, q , = l n l t l o l praduction role K = Constont

1 t = Time n = Exponent

FIG.3-CLASSIFICATION
OF PRODUCTION
DECLINE
CURVES.
- ---
R A T E - TIME CURVES
COORDINATE

TIME I l l LOG TIME (11

R A T E - CUMULATIVE CURVES
COORMNkTE SEMI-LOG

100
I '
.,
100.000 o 04000 100 tpoo IOWO 100.000
CUMULLTIVE 101 CUYULLTIVE (01 LOG CUMULATNE 101

I- CONSTANT PERCENTAGE DECLINE n = 0 D= 03


--- HYPERBOLIC DECLINE n = % 0,:lO
110 ---HYPERBOLIC DECLINE (SHIFTED ON LOG - LOG I

ID -------- HARMONIC DECLINE n = I D,:30


ma - - - - - - - H A R M O N I C DECLINE (SHIFTED ON LOG - LOG)

FIG.&-THREE TYPESOF PRODUCTION


DECLINE
CURVESON COORDINATE,
SEMI-LOGAND LOG-LOG
GRAPHPAPER.
Cun~~llallvc oil production is plottcd tour progresses to the top of the sand estimates has often been overlooked.
on the horizontal scalc while the cu- the lease is ready for abandonment. Reference is made to the comparison
n~ulativc gas production is plotted An example of this is shown on the between the recovery analysis based
on the vertical scale. As is normal in chart of Fig. 7. on vol~lmetricdata with the one ob-
depletion-type fields, the trend of the By plotting this type of chart for tained from a projection of individual
curve appears to steepen with in- many leases in the East Texas field it well performance.
creasing gas-oil ratios. has been noted that the rise in the Oftentimes it is rlifiicult to fit the
From a volumetric calculation an water table appears to be more o r less projected performance to the volu-
estimate is made of the total gas to be proportional to the cumulative pro- metric estimate. If both types of esti-
released from this reservoir down to duction. mates are based on good, reliable
an assumed abandonment pressure. information but cannot be reconcilcd,
This figurc, which in this case was KELATIONSHIP
BLTWEEN some important conclusions may be
R'SERvES AND
1.42 billion cu ft, is marked on the drawn from this discrepancy. If the
chart as a horizontal line and repre- From the rate-cumulativc equation performance indicates a substantially
sents the ceiling of the cumulative gas- for constant Percentage decline shown lower ultimate recovery than the vol-
cumulativc oil curve. on Fig. 3 it may he noted that the umetric calculation would indicate, it
B~ extrapolating the current trend rem"ning reserves are equal to the might mean that there is something
until it intersects the estimate for the difference between the present pro- fundamentally wrong in the produc-
total gas available, an estimate can duction rate and the production rate tion practices used. Possi'ily morc
be obtained for the total oil rccovery at the economic limit, divided by the drainage points are nccded or the
during the primary period. decline as a fraction. The same time wells need stimulation treatments o r
units should be used for dctermining cleanout jobs.
OIL-WATERCONTACT OR both decline and production rates. On the other hand, if the well per-
ABANDOXMENT CON.I'OUR This leads to thc following short- formance projection indicates an ulti-
vs SANDTOP cut: When the decline is 1 per cent mate recovery well in excess of the
Another method which is somc- per month the remaining reserves are volumetric estimate, it could mean
times practiced in the larger water 100 times the difference in monthly that the subsurface interpretation used
drive fields, such as East Texas, is to production rates; for 2 per cent de- may be in error and that there may be
choose the depth of the oil-water con- cline per month this ratio equals 50; a larger oil reservoir on hand than
tact or abandonment contour as the for 3 per cent it is 33 1/3; for 4 per current subsurface interpretation in-
dependent variable to be plotted cent it is 25, etc. dicates. In that case it might be highly
against the cumulative oil recovery as When production rates are on a desirable to look for a possible exten-
the independent variable. The end daily or annual basis the same for- sion to such an oil reservoir.
point of this type of performance mula holds, provided the decline is
curve is the average depth of the top calculated on the same time basis. REFERENCES
of the sand for a given lease. The
method of extrapolation in this case I. Craze, R. C., and Buckley, S. E.:
CONCLUSION "A Factual Analysis of the Ef-
is based on the simple assumption
that whenever the abandonment con- A very important use of rcscrvc fect of Well Spacing on Oil Re-
covery," Drill.and Prod. Prac.
API (1945) 144.
2. Muskat, M., and Taylor, M. 0.:
"Effect of Reservoir Fluid and
Rock Characteristics on Produc-
tion Histories of Gas-Drive Res-
ervoirs," Tmns. AIME (1946)
165. 78.

CUMULATIVE PRODUCTION I N THOUSANDS OF BARRELS

FIG. 5-OIL PERCENTAGE-CUMULATIVE


RELATIONSHIP
ON SEMI-LOG
PAPER-TAR SPRINGSSANDPRODUCTION,
CALVINFIELD.ILL.
PETROLEUM TRANSACTIONS. A I M E
C W L A T I V E OIL PROWCTION PER WELL IN THOUSANDS Of BARRELS
FIG.7-ABANDONMENTLINE (SUBSEA)VS CUMULATIVE
OIL RECOVERY-
WOODBINESANGEAST TEXASFIELD.

3. Arps, J. J., and Roberts, T. C;.: C.: "Mechanism of Fluid Dis- Flood Calculations," Trans.
"The Effect of the Relative Per- placement in Sands," T r a n s . AIME ( 1949) 186, 9.
meability Ratio, the Oil Gravity, AIME (1942) 146, 107. 12. Arps, J. J.: "Analysis of Decline
and the Solution Gas-Oil Ratio 8. Pirson, S. J.: Elements of Oil Curves," Trans. AIME (1945)
on the Primary Recovery from a Reservoir Engineering, McGraw- 160, 228.
Depletion Type Reservoir," Hi11 (1950) 285. 13. Cutlcr, W. W., Jr.: "Estimation
Trans. AlME (1955) 204, 120. 9. Welge, H. J.: "A Simplified of Underground Oil Reserves by
4. "Letter Symbols for Reservoir Method for Computing Oil Re- Well Production Curves," USBM
Engineering," Jour. of Pet. Tech. covery by Gas or Water Drive," Bull. 228 (1924).
(Jan., 1955) 38. Trans. AIME (1952) 195, 91. 14. Babson, E. C.: "Prediction of
5. Watts, E. V. : "Some Aspects of 10. Guthrie, R. K., and Greenberg- Reservoir Behavior from Lab-
High Pressures in the D-7 Zone er, M. H.: " I h e Use of Multiple oratory Data," Trans. AIME
of the Ventura Avenue Field," Correlation Analyses for Inter- ( 1944) 155, 120.
Trans. AIME ( 1948) 174, 19 1. preting Petroleum Engineering 15. Tarner, 5 . : "How Different Size
6. Katz, D. L.: "Possibilities of Data," API Paper 901-31-G, Gas Caps and Pressurc Mainte-
Secondary Recovery for the New Orleans, La. ( M a r c h , nance Programs Affect Amount
Oklahoma City Wilcox Sand," 1955). of Recoverable Oil," Oil Weekly
Trans. AIME (1942) 146,28. 11. Stiles, W. E.: "Use of Permea- (June 12, 1944) 32.
7. Buckley, S. E., and Leverett, M. bility Distribution in Water ***