T.P.

3710

APPLICATION OF LABORATORY PVT DATA

TO RESERVOIR ENGINEERING PROBLEMS
C. R. DODSON, MEMBER AIME, UNIVERSITY OF SOUTHERN CALIFORNIA, lOS ANGElES, D. GOODWILL,
LA HABRA, CALIF., AND E. H. MAYER, TUPMAN, CALIF., JUNIOR MEMBERS AIME, STANDARD Oil CO. OF CALIF.

ABSTRACT originally in place in a pool was recovered from the ground

under uncurtaiIed production conditions caused some operators
The paper summarizes the historical background of the to re-examine and modify their production methods. However,
pressure-volume-temperature analyses of reservoir fluids, the though these works showed methods of production that have
errors involved in both the sampling and testing of reservoir subsequently resulted in greater oil recoveries, they in general
fluids, the type of information required of a PVT determina- went unnoticed, even though there was a fear that the nation's
tion, and the field conditions that limit the application of petroleum resources were being exhausted.
anyone analysis. Particular emphasis is placed on the neces-
sity for approximating as closely as possible the liberation The period 1924-1933 saw the industry take considerably
sequence occurring in the producing formation, flow string, more interest in reservoir behavior because of an unfounded
and surface separators. A combined differential and flash or fear of its inability to replace oil reserves, the possibility of
"composite" liberation is suggested as the best means of ap- government regulation, and the energy of one man, Henry L.
proximating this liberation sequence. Doherty. Doherty aroused heated discussions in the industry
concerning conservation. To prove or disprove his theories
the first experimental work on the reservoir behavior of pe-
troleum was undertaken. The papers of Dow and Calkin:
HISTORICAL BACKGROUND Beecher and Parkhurst,' and Mills and Heithecker9 are classics
Petroleum reservoir engineering commonly is considered to although their experimental procedures were crude. These
be one of the newest fields of petroleum science, yet much papers proved that the properties of petroleum in the reservoir
work of a theoretical and intuitive nature was done many are quite dissimilar to the properties measured at the well
years before the modern techniques of reservoir analyses were head. Gas dissolved in the oil phase was recognized as having
developed. considerable importance as had been predicted by the earlier
The period 1910-1924 saw considerable work in the field theorists.
of reservoir behavior done by the U. S. Bureau of Mines. This Slightly later Miller'· and his co-workers assembled the first
work,,-" although entirely theoretical, pointed out the impor- text correlating all the known data on the reservoir behavior
tance of gas in the recovery of oil from the reservoir. The of petroleum and used these data to show the economic value
statement by J. O. Lewis that 20 per cent or less of the oil of conservation. Quantitative application of pressure-volume-
temperature analyses of reservoir fluids was given in the paper
References given at end of paper. of Coleman, Wilde, and Moore"; this work showed that with
Manuscript received in the office of the Petroleum Branch, AIME, Sept. sufficient laboratory and production data prediction of reser-
18, 1962. Paper presented at the Petroleum Branch Fall Meeting in Los
Angeles, Calif., Oct. 23-24, 1952. vo;r behavior was possible. The refined sampling and labora-
Vol. 198, 1953 PETROlEUM TRANSACTIONS, AIME 287
T.P. 3710 APPLICATION OF lABORATORY PVT DATA TO RESERVOIR ENGINEERING PROBLEMS

tory techniques of Lindsly12~l. led to the present methods of ing conditions may change as well as the nature of the reservoir
reservoir fluid analysis. Such improved techniques make pos- fluid: thus, new PVT analyses will take into account these
sible the quantitative inve~tjgation of re~ervoir prohlem~ now factnr~ which were not anticipated in prior tests.
taking place.
Modern laboratory procedures still do not truly represent
reservoir behavior. This paper will present some proposals to
help correct these discrepancies. Bottom-H ole Samples
Many types of bottom-hole samplers have been devised and
de,cribed in the literature.",]·"·,24,25,,. The mechanical nature of
the devices introduces some error into the sampling operation;
however, these errors are probably of less importance than
those resulting from the other difficulties.
RESERVOIR TEMPERATURE AND PRESSURE
It is often questionable whether the fluids in the tubing at
SURVEYS, SAMPLING OF RESERVOIR FLUIDS
the point where a bottom-hole sample is taken are representa-
The applicability of the results of laboratory PVT deter- tive of the reservoir hydrocarbons. This is particularly true if
minations to problems in reservoir behavior depends upon there has been considerable drawdown of the well prior to
field information. To use the laboratory results in reservoir sampling, if several zones containing fluids of dissimilar prop-
studies it is necessary to have accurate field temperature, pres- erties are produced together, if the gas and liquid are not
sure, and production information, and to know the samples homogeneously mixed in the proportions in which they existed
tested are representative of the reservoir fluid. in the formation, and if both gas cap and oil zones of a reser·
voir are open simultaneously to the well bore.
Allowing the well to stand shut-in for a short time prior to
sampling usually assures a fairly representative reservoir
liquid sample, lacking only the gas formerly associated with
Temperature Surveys
it. Only in the case where the reservoir fluid was under-
Bi-metallis,17 maximum recording," and electrical resistance saturated initially and during production would the sample
thermometers"; thermocouples; and other methods" are avail· be truly representative.
able to obtain bottom-hole temperatures. It is essential that
the fluid within the well bore through which the measuring
device passes be in thermal equilibrium with the forma-
tion and that the recording instrument come to thermal equili. Recombined Samples
brium with the well fluids readily. For these reasons, static Often sampling of the fluids produced from a well is a
or shut-in temperature measurements are considered the most satisfactory and accurate method for obtaining a representa-
representative. tive reservoir fluid sample. Both the produced liquids and gas
are sampled, usually at the high pressure separator. These
two samples are then combined in the same proportion as the
two ph:tses are produced from the well. The liquid and gas
Pressure Surveys also may be combined in other ratios to give samples rep-
resentative of other reservoirs or of the originating reservoir
The merits and limitations of the various sub· surface pres-
as it existed at some previous time.
sure measuring devices are adequately discussed in liter-
ature.'·~23
There are also some inaccuracies possible in using this
method, and they must be corrected for if the results obtained
from the recombination are to be representative. Relative per-
meability conditions in the formation may allow unequal flow

Sampling of Reservoir Fluids

While the methods of obtaining bottom-hole temperatures
and pressures are well standardized and of known accuracy,
the accuracy of reservoir fluid sampling methods is open to
considerable question. The following discussion considers the 5

available procedures for obtaining a sample from a well.

Particular emphasis is placed on the limitations of these meth-
4
ods which may cause the sample to be non-representative of
the reservoir fluid.
/
~

The field situation usually determines the method of sam-

pling employed. The observed properties of the field sample V
and the sampling conditions should be recorded prior to the V/ ~

shipment of the sample to the laboratory for testing. These ~I

/
~v '-Vo/VTO
data are an integral part of the final PVT analyses. The ex-
pediency of obtaining several samples from a well is obvious.
The cost of such additional sampling is small after prepara-
'":l;
~>
'">
"II
/ .-- COMPOSITE LIBERATION-
SEPARATOR CONDITIONS: 100
TANK CONDITIONS
DIFFERENTIAL
. 0
SEPARATION-
~;~~~
tions have been made to take the first sample. Several samples
~ YMPERATURE 140°' I
provide a check on the results obtained and will indicate
the accuracy of the sampling technique.
.J
::i ,Ol---_
1 0 _.. ~ ~-- ----,;;
loao ""- 2000 3000 4000
PRESSURE (PSIG)

Re-sampling of the reservoir at various times during its life

will insure greater accuracy in the PVT studies. Field operat- FIG. 1 - PRESSURE-VOLUME RELATIONS OF LIQUID PHASE.

288 PETROlEUM TRANSACTIONS, AIME Vol. 198, 1953

 C. R. DODSON, D. GOODWILL, AND E. H. MAYER T.P. 3710

rates of liquid and its formerly associated gas to the well bore;

t~~~-~ -~
such a condition would obviously render a sample combined
on the basis of the primary trap gas-oil ratio inaccurate. The
necessity for accurate gauging and steady-state well conditions _ u un--+----
is immediately apparent. (If the well is producing both gas
from a gas cap and gas released from solution, the proper
combination of the fluid samples will depend on other infor-
;60oL--:--_______
a: I
I ______
I :
~
f L-- ~
mation besides the primary separator gas-oil ratio. General
correlations concerning the properties of reservoir fluids may
be useful in this regard"-30).
The economy resulting from the use of recombined well fluid
V t,
:3
a:
400

COMPOSITE UBERATION-
SEPARATOR
I
CONDITIONS 100 PStG, eOOF~
TANK CONDITIONS
.~ I
I 0 PSIG:, 80 0 r

samples instead of bottom-hole samples appears to be obvious;

s_~:~_ 200 __ --DIFFEREN1IAL SEPARATION- - - - - -
TEMPERATURE
0
140 F I I

however, the necessity for extremely accurate gauging and

___ ~__ _ ______ ~:~~A5U~E: AT 60° Fj & 14 7 PSI~
control of the well prior to and during the sampling operations III °0 - - 1000 2000 3000 4000
may partially obviate this apparent economy. PRESSURE (PS Ie)

FIG. 2 - SOlUTION GAS-OIL RATIO.

Flow Line Samples
For condensate fields, sampling of the fluids passing through tion conditions, if they exist, cannot be predicted by routine
the flow-line from a well or a number of wells may be em- laboratory PVT analyses. Unfortunately supersaturation can
ployed_ A common method uses a Pitot tube inserted in the have a considerable effect on reservoir engineering calcula-
flow-line through which the fluid is passing. Because of the tions.
unknown distribution of liquid and gas within the pipe, it is
difficult without proper calibration of equipment to obtain a
sample that will contain proper proportions of condensate
and gas. At Conditions Below or Near the Equilibrium
Gas Saturation
If there is such a low original gas saturation in the pool
that the formation has no permeability to gas, the vapor that
is released from solution due to pressure draw-down will re-
TYPES OF GAS LIBERATION AND main in the pores in contact with oil until an equilibrium gas
APPLICATION OF DATA OBTAINED FROM saturation is reached; then the gas will begin to flow. At a
THESE LIBERATIONS TO RESERVOIR distance from the well and under moderate drawdown condi-
PROBLEMS tions, both the oil and gas will be moving slowly after the
equilibrium gas saturation is reached. Therefore, flash libera-
A need exists for a comprehensive correlation between pres-
tion conditions are approximated over the entire formation
ently used laboratory gas liberations and the liberations occur-
before the equilibrium saturation is reached and at a distance
ring in the production of fluid from reservoir to stock tank.
from the well bore after the equilibrium saturation is reached.
Laboratory methods used in obtaining PVT relations should
In flash liberation the vapors formed are allowed to remain
endeavor to simulate the production liberation sequence.
in contact with the liquid phase until the desired equilibrium
conditions are reached. This type of liberation is characterized
at anyone pressure by a constant composition of the vapor
Reservoir Gas Liberation and liquid phases and a constant gas-oil ratio.
The only work" concerning true reservoir gas liberation Closer to the well bore (the distance will depend on the
that has an experimental rather than speculative basis is of magnitude of the draw-down) the oil will be moving more
a highly idealized nature. However, experimental data on reser- rapidly than the gas that evol'fes from it at this time. Under
voir constituents under normal production practices. Pirson" these conditions a differential liberation is approximated_ In
with the American Petroleum Institute Project 47. differential liberation the vapors formed are continuously re-
moved from contact with the liquid phase_ This type of libera-
tion is characterized by a varying composition of the liquid
and vapor phases and a continuously varying gas-oil ratio.
Possibility of Supersaturated Conditions The differential liberation condition probably applies to only
There is no definite information available that indicates ther- a small percentage of the drainage area of a well or reservoir
modynamic equilibrium is attained by the hydrocarbon reser- just prior to or just when equilibrium gas saturation has been
voir constituents under normal production practices_ Pirson" attained; therefore, laboratory flash liberation is used to simu-
and Sage" have both pointed out that supersaturation of the late field producing characteristics at this stage_
reservoir oil with gas can occur. Investigation to determine the
conditions necessary for this supersaturated condition is cur-
rently being conducteq under API Project 37 at the California
Institute of Technology. Results to date indicate that agitation At Conditions Above the Equilibrium Gas Saturation
is the most important factor causing the attainment of thermo- After the equilibrium gas saturation is reached, the differen-
dynamic equilibrium_ When he flow towards the well bore is tial separation process better represents the formation libera-
slow because of low formation permeability, slight pressure tion. The gas generally is more mobile than the oil and leaves
differential, or a low permeability zone immediately about the oil that originally contained it behind when both are flow-
the well bore, there is probably insufficient agitation or turbu- ing to the well bore_ The oil flowing to the well is constantly
lence to cause thermodynamic equilibrium_ These supersatura- liberating more gas due to the decreasing pressure. This is not

Vol. 198, 1953 PETROLEUM TRANSACTIONS, AIME 289

T.P. 3710 APPLICATION OF LABORATORY PVT DATA TO RESERVOIR ENGINEERING PROBLEMS

Flow-String Gas Liberation

The gas liberation occurring in the flow-string between the
! I producing formation and the well head is considered to be an
iII I iI adiabatic flash liberation. This belief is based on the fact that
~
"- .0400
\ i
I

I
i
all the fluid entering the flow-string must be removed from
it at the surface and there is sufficien~ agitation in the flow-

SEPA~A -I
::J
<.> string to induce equilibrium. There are several differences,
w
::; FLASH TlON however, between this flow·string liberation and a laboratory

-~
::J
..J
i TEMPERATURE 140·
flash liberation. The laboratory flash liberation allows only the
o I
> II gas liberated from the sample to stay in contact with the re-
<.>
i;: .0300 maining liquid until certain conditions are reached. This is

'\
<3
w usually done under isothermal condition". In the well bore,
Q. I
I
27TI gas from oil other than that in the well bore may be present,
en I

~ 1
/BUBBLE TINT PSIG AND ,
0.02241 CU FT/LB
resulting in differences from the laboratory results in both
the gas and oil characteristics. This is particularly true in
.0200
i I 'I gas injection where part of the injected gas is by. passin•
1000 2000 3000 4000 5000 through partially depleted pores. Also separation of some gas
PRESSURE CPSIG)
from the liquid hydrocarbons at th~ tubing shoe may occur.
This gas may then be produced from the casing-tubing annulus
RIG. 3 - PRESSURE·VOLUME RELATIONS OF RESERVOIR FLUID. thus rendering more inaccurate the oil and gas characteristics
predicted from the laboratory flash liberation. And if a well
flows by heads, equilibrium between all produced fluids will
not be attained. Currently, however, the liberation process in
the tubing and casing is rei!arded as more nearly flash in
nature than differential.
a true laboratory differential liberation because the gas is con-
stantly in contact with oil on its trip to the well bore, and
the oil the gas is in contact with is not its originating liquid.
Whether there is sufficient agitation to allow the traveling gas Fluid Separator Gas Liberation
to come to equilibrium with the oil it by-passes is unknown.
It is probable that complete equilibrium between the gas and It is generally agreed that flash vaporization best represents
oil is never attained. the separator type liberation. When entering the separator the
It is also possible at this stage to have a combination of fluids are probably in equilibrium due to the agitation occur·
the differential and flash liberation in the reservoir due to the ring in the flow-string. During separation in the trap, a certain
pore size variance. When the larger pore spaces have attained amount of oil always remains inside the trap in contact with
gas equilibrium, there may be smaller pore spaces or capil· a constant volume of gas. Each unit volume of oil-gas mixture
laries where an equilibrium gas saturation has not been entering the separator is of approximately the same composi.
reached. These small capillaries probably contribute little to tion as the previous unit volume; thus homogeneous flow is
the total production of the field. and the dual liber'ttion effect essentially attained under conditions of ~teady well flow. Each
may therefore be neglected. subsequent separator results in another flash liberation with the
final flash liberation occurring in the stock tanks.
During gas injection it is known that very little of the gas
returned to the formation goes into solution in the remaining The variance between the laboratory and separator flash
reservoir oil due to lack of agitation. The gas goes either into liberation results from the laboratory flash being carried out at
the gas cap or by-passes through the larger, already depleted, a constant temperature for each trap, while the field separator
pores of the formation. The injected gas, mainly comprised of liberation may occur at greatly differing temperatures through-
methane and ethane due to prior recovery of the heavier con· out the year. However, this discrepancy between laboratory
stitutents, is of different composition than the gas that would and field results is slight in most cases.
be evolved by the oil remaining in the by-pass channels if the
reservoir pressure were allowed to decline. The volatile injected
gas will pick up some of the heavier constituents of the oil
and will mix with the liberated gases; the oil in turn will Composite Gas Liberation Method
dissolve some of the injected gas. This enrichment of rapidly In the previous discussion the reservoir liberation, after
moving gas will not produce complete equilibrium an:l is gas equilibrium was reached, was described as a declining-
not represented either by the flash or differential-type labora- pressure, relatively-constant-temperature, differential libera-
tory liberation. However, since the gas is moving rapidly past tion; the flowstring liberation was described as a declining-
the remaining oil, differential liberation is considered to afford pressure, declining-temperature, flash liberation; the separator
the best representation of the formation process. This non· and stock tank liberation was described as a reduced-pressure,
attainment of equilibrium causes computations involved in variable·temperature, flash liberation. To better approximate
gas injection operations to be in error if baeed on the assump- this reservoir and producing liberation cycle a composite type
tion that equilibrium is attained between the oil and injected liberation can be run in the laboratory. A differential libera-
gas. tion represents the reservoir vaporization; a series of flash lib-
Miller" believes the formation liberation to be an adiabatic erations represents the behavior of the oil passing up the flow-
rather than isothermal process as it is generally regarded. string, through the surface separators, and into the storage
Under certain conditions the use of isothermal laboratory data tanks. The limitations and errors involved in these approxima-
to approximate adiabatic reEervoir conditions might cause tions have been previously discussed.
errors of a measurable magnitude. However. Miller reports The laboratory procedure for this composite liberation fol-
that in most cases this error is slight. lows. A representative fluid sample is charged into the main test

290 PETROLEUM TRANSACTIONS, AIME Vol. 198, 1953

 c. R. DODSON, D. GOODWILL, AND E. H. MAYER T.P. 3710

so that the first Hash trapping will be eliminated from the

experimental procedure. This is in line with field operating
practice.
2.51---~-~--~----+-----,---+---I-----l
<Il In each case the oil produced from the trapping is corrected
""
<Il DIFFERENTIAL SEPARATlON- to a volume at standard conditions through use of appropriate
(5 'TEMPERATURE 140· F
~ 2.0 --t- ---~.---i---+-----+__ tables and its specific gravity noted. This final auxiliary cell
z . ! . volume corrected to standard conditions divided into the vol-
""
<.)
ume removed from the main cell allows evaluation of a forma-
tion volume factor based on tank oil or a close approximation
of it. Use of either the differential or flash liberation process
alone does not allow for measurement of tank oil properties
1.0 e--------'i'.----- - -j--.--+--~. ----+---+-----;-----1
, but only for a measurement of residual oil properties which
are not characteristic of the liquid produced from the well.
The volume of gas liberated from each trapping sequence
PRESSURE (PSIG)
(total volume at standard conditions) divided by the volume
of oil at standard conditions that evolved from this trapping
gives the solution gas-oil ratio for the pressure at which the
FIG. 4 - VISCOSITY OF LIQUID PHASE.
oil was removed from the main cell (this ratio is based on
tank oil and not residual oil).
Use of differential vaporization to represent the formation
liberation results in compositional changes of the final tank
oil and gas throughout the testing series. This phenomenon
cell at a pressure high enough to insure single phase behavior. occurs in the field also. Thus the composite liberation method
The temperature of the cell is then raised to that of the reser- makes possible prediction of the stock tank oil and gas gravity
voir. The sample is expanded with a reduction in pressure changes throughout the reservoir pressure decline.
(method similar either to the differential or flash liberation in Measurement of the liquid volumes in the main cell and
the undersaturated region) until the first bubble of gas j,; volumes of gas removed with adjustment for withdrawals of
evolved. At this point a small volume of oil from the main liquid for trapping separation allows the differential curves
test cell is removed at constant pressure to another, smaller. (for residual oil) of the sample to be determined simultan-
auxiliary test cell. The temperature of this cell is held to eously with the tank oil relations. The formation volume fac-
that existing in the first separator at the well head; the sample tor and gas in solution behavior of a sample as determined by
of oil in the auxiliary cell is flashed to the pressure of the these two methods (composite and differential liberations) are
first surface separator after noting the original volume of compared in Fig;:. 1 and 2. Note that there is a considerable
oil. The gas liberated from this first auxiliary cell flash is difference between the two curves, particularly at higher pres-
removed from the cell and its volume and specific gravity ,lIres where reservoir calculations based on PVT data are
measured. The remaining volume of oil in the auxiliary cell most commonly made.
is recorded. This oil is then flashed to the pressure of the The slope of the formation volume curve in the single phase
next separator while the cell is held at the temperature of or undersaturated region is the same for the differential, flash.
this separator. This flashing process is repeated as many times or composite liberation. The location of the volume curve in
as there are surface separators, including a final flash to the undersaturated region is considerably different for the
simulate the liberation to stock tank conditions. various processes becau~e of their variant nature after the
The laboratory making such a PVT analysis must select the bubble point has been reached.
trapping pressures when they are not specified. Usually a pri-
The numerical importance of the,e composite liberation
mary 450-500 psig trap and 50-60 psig secondary trap are
data for one field is shown in the material balance calculation
good approximations for most California conditions.
The maintenance of the auxiliary cell temperature at the
temperature in the surface traps results in more accurate
gas-oil ratio and shrinkage information than that given by the
usual flash and differential liberations run at higher reservoir
temperatures where more gas is liberated and more shrinkage
occurs.
The volume of oil remaining in the main cell is allowed .
:::-
0:
I
COMPOSITE LlBERATION- .I
SEPARATOR CONDITIONS: 100 PSIG, 80·F_
1.4
to expand through a pressure decrement while being held at ~ TANK CONDITIONS : 0 PSIG, 80· F
reservoir temperature. The gas evolved in this pressure decre- (J)
« DIFFERENTIAL SEPARATlON-
to TEMPERATURE 140· F
ment is then removed as in a differential liberation. Follow- IL -
0 1.2
ing the gas removal a small volume of oil from the main >-
test cell is charged at constant pressu, e to the auxiliary cell. >- VDIFFERENTIAL SEPARATION GAS
:>
The oil removed to the auxiliary is then mcce~sively flashed «
0: 1.0
to
at temperatures and pressures equal to those in the surface
separators and stock tanks with suitable volume readings and
gas measurements being taken after each step.
0
ii:
<>
"" 0.8
\ rCOMPOSITE LIBERATION GAS

~
The differential liberation of tile oil in the main test cell
Q.
Ul· ---....,
is continued with removal of portiulls of the differentially lib- 0 0
I---.,
0.6
erated oil at selected pressures. Each portion removed is Hashed 1000 2000 ~OOO 4000
PRESSURE (PSIG)
through what approximates the surface trapping arrangement.
Near the end of the composite liberation process the main
cell pressure may drop below the initial assumed trap pressure fIG. 5 - GAS SPECIFIC GRAVITY.

PETROLEUM TRANSACTIONS, AIME 291

T.P. 3710 APPLICATION OF LABORATORY PVT DATA TO RESERVOIR ENGINEERING PROBLEMS

i~ determined. The viscosity of the reservoir oil preferably

COMPOSITE LlBERATlON-
should be determined in the lahoratory (\I!.;e of a rolling ball
, SEPARATOR CONDITIONS: 100 PSIG, 80°F
, TANK CONDiTIONS 0 PSIG,80°F viscosimeter is recommended'"). When lahoratory determina-
367 1 - - - - - - --; ---t!- ~----~---~~-

u.. I tion of the liquid visrositv i, not made. the rorrf'lation;.; of

0 Bear" can be used.
'"
I-
The deviation factor and den"ity of. th" differentially lib-
<t 36,6
Ii: erated gas ,should he determined during the funning of tlu,
'! difl'erentialliheration serie;.;. These experimental determination~
-~

>- 36,5
I- may then be compared to the deviation factors as determined
:>
<t by the Brown'" or some other similar method." A formation
a: ,
'" volullle factor for the gas (space in barrels (.ccupied in the
36~4 - ----- , -t---·--·~------+------- --- reservoir at a certain pressure bv one standard cubic foot of
gas that is not in solution') mav he computed using the devia-
tion information and gas laws. Use of the gas formation vol-
363
0 1000 2000 3000 lime factor. ,olution gas·oil ratio. and li(plid volume factor
PRESSURE (PSIG)
readilv gives the two.phase formation \'olume farto]'. To be of
"ignifican('(~ this two.phase factor must he romputed from the
FIG. 6 - GRAVITY OF COMPOSITE-LIBERATION TANK OIl. ('ompositt- liberation data and should not he obtained from
flash lilwration informatioll.
The thermal expansibilit\ and collipres"ihilit\ of the "at 11-
ratt"d nudt" sll<luld lw obtaint"d so that the obsened data may
he rt"vised to allY datllm. Slirface and interfacial tensiono.;
of Table 6. The material balance equation ~Iwwn in Table 6 ,hould ht" determilwd. The method of Hocutt" is typical.
was simplified before u~e by assuming: (1) No gas in solution :\11 data regarding field sampling methods. field sampling
in re'ervoir or encroached water. (2) The water compression conditions. and field tt'sls should he induded in the final report.
and thermal expansion are negligible or of compensating mag-
nitude, (3) No pore volume reduction due to sand compaction.
(4) TI'e water initially in place equals zero as it is inter-
The ""Minimum" PVT Requirements
,t:t;al and thus is assumed to be both immobile and to undergo
110 volumetric change.
Such a "miniIlItJlll" analysis would he one in which only the
1II0st important physical properties were measured. Other prop-
In pooh where the oil is highly volatile and the gas produced
('nies wOlild he approximated hy using the ha,.;ie test data.
contains considerable amounts of liquefiable constituents. con-
production information. and various corrt'lations.
o.;:deration mw't be given in material balance calculations to
Ca"-oil ratios and formation volume fa('tors a~ dett'rmined
the amounts of LPG and casinghead gasoline recoverable from
hy the composite and differential methods are important. The
both tl~e gas in solution in the oil produced and from the gas
gravity of the liberated gas can be obtained from this informa-
coming from the gas cap or residual reservoir oil.'" The solution
tion using the correlations of Standing" or Borden and Rzasa."
gas produced by the composite liberation can be analyzed for
From the test cell information and the liquid viscosity at one
Equid content. The method of Cook, Spencer, and Bobrowski'"
set of conditions the correlations of Beal'" will allow approxi-
;:lIows the liquids contained in the other produced gas to be
mation of reservoir liquid viscosities at other conditions. Esti-
computed. The formation volume factors for the liquids pro-
mated gas gravities will allow approximation of the viscosity of
duced will be obtained hy the composite liberation method.
the liberated gas through the correlations of Bicher and Katz.""
Estimated liberated gas deviation factors and gravities per-
mit computation of gas formation volume and two-phase forma-
tion volume factors as descrihed under "ideal"' PVT require-
ments.
REQUIREMENTS FOR PVT DETERMINATIONS Again the field sampling and field testing information are
an integral part of the report.
Evaluation of various producing methods or future produc·
tion possibilities may be gained through the combined use of
PVT information and operating data. The "ideal," "minimum."
and "no reservoir sample" PVT requirements are discussed. The '"No Reservoir Sample" PVT Requirements
With a knowledge of the gravity of the produced gas, grav-
ity of the tank oil, instantaneolls gas-oil ratio (assumed equal
to the solution gas-oil ratio or corrected to an estimated solu-
tioll ratio by other means), and bottom hole temperature and

The "Ideal" PVT Requirements

The formation volume and gas in solution relations should
he obtained hy flash, differential, and composite methods from Table 1 - Reservoir Data
the initial reservoir pressure, hubhle point pressure. or other HaLP: M:{ KID.
pressure as dictated by field conditions down to the pressure Cas·Oil ratio: No r"corel.
at the economic limit. The density of the oil and gas liberated Tank Oil Gravity: :n.5° A 1'1.
Initial R('servoir Pressure at mid-point of pprforations: 2,817 psig
by the differential and composite methods should be noted. at 4,611 ft IIPlow mean s('a lev('1.
Accurate measurement of the trapping conditions in the ('om- Initial Ht"st'l'voir Tt"mpt'ratllrt': No 1'1'('01'(1.
posite method is important. Initial Datlllll Pn'ssllrp: :!..76'2 psi~ al .+,450 ft Iwlo\\ lI1<'all "'il Ie,,'1.
Inilial Dallllll '!'Plllpnatllre: I.+OuF at '+,'+50 ft Iwl,)\, lIIeall "'n level.
Viscosity of the liberated gas can be determined frOll! the (:111'1'1'111 1)'1111111 I'n'o.;o.;lIl'l': 'Ii:!. I',i;.! at LISO fl Iwlll\\ IIwall't'a !1""'!
Bicher and Katz'" wrrelations when tIl(' gravity of thi~ phao.;(, a,.. or Ft,! •. I. It)S!.

292 PETROLEUM TRANSACTIONS, AIME Vol. 198, 1953

 C. R. DODSON, D. GOODWill, AND E. H. MAYER T.P. 3710

Table 2 -_. Well Sampling Data ,ite methud were 100 psig and /:lO°F for the oeparator, and
Total Dppth: S,:JHO fl. I'lu,.:,.:t'd to: S.2:-lh 1'1.
o psig and 80°F for the tank.
Compj P tiol1: Date: ])PI·. I. 11)42. From tlH~ dat a (Tahles I through 4 J curves were constructed
( :a~i ng ~iZf~: hllt If!. (I-'i/-!:,. I Ihrollgh 61. It is to he noted that only three poinls
ll .. pth: S,:'7:1 fl. ',('n' ddt'rllli" .. d hy tlw composite method. The general form
I't~rlo,.aliofl: Clln pt"rforatl·(1.
Intpr\,al: S.II7·S.IS1 1'1.
of till' ,'.omposil(~ liheration curves was dictated hy the three
Tuhin,.: Size: 21f2 in. poinls ohtaine!1 and the shape of the differential curves. The
ll"pt It of tubi 11": ;;IJOe: :-;.1 ()O I!. shape of Ihe composite curves in the under-saturated area
Packer: S,074 ft. wa" determined hy the linear relationship of the differential
Commenb: Pack!'r "PI to i"olatt' IIl'l'er 1\ att'r ,allik
Samplin,.: Date: ,\1ay I, 19S1. ('urve; an offset was made to allow for the different final
I'rodlll'tion Data: (Durin,.: "amplin,.: I properties, The symhols, abhreviations, and suhscripts used are
Tuhin~ Prf's~tlrt': 480 I'sig Ihose given in the appendix.
(:asin,.: Pressure: Not recorded
~"parator Data
\faterial halance techniques were used in conjunction with
Pn"':-.:--lIn-· T'-llIperature the PVT data to calculate the tank oil uriginally in place in
Fi rst Sta~~: "l01lf' \j,,",- Ihe reservoir. The equation employed, production clata used,
Second Sta,.: .. : 'iOIlt' \jOlW
and rt";ults obtained are shown in Table 6. Note that at a
Tank: ~flt n-'(·ordt·d 'iot rPI'ord"d
Oil and (;a, Halt' pre.;surt' of 2,070 ]>~ig the tank oil initially in place for the
Fir,t Sta,.:": \Ollt' composite data was 292.9 x 10'; bbl against 231.8 x 10" bbl for
St'('ond Sta,.:!': \on!' Ihe difft'rential dala. at a pressure of 1.560 p~ig a similar differ-
Tank: \jot rf'I'orded
ellce was noted. The magnitude of the difference obtained
Ca,,·Oil Hatin: S.Mh I'll ft lI!'r bhl
Tank Oil Cravity: ::I4.:·\" .-\1'1 between the two types of calculations shows the necessity of
\Vat .. r: Tra('!' using liberation data that approximates conditions cluring the
(:ommf'nts: Produ(,tion rates durin:! ~ampling pf'riod: production sequence.
Gro.'>' Fluids - 74.4 HI D
Total Gas - 420 Mcf!D

PHYSICAL ANOMALIES IN THE PETROLEUM

RESERVOIR AND THEIR EFFECT ON
pres:;ure taken at the iime the above properties were mea:;- USE OF PVT DATA
ured, an estimate can be made of thc reservoir fluid properties.
The formation volume factors and solution gas-uil ratios can It is recognized that pressures and temperatures vary with
he determined for any condition using the correlations of ,Iruetural position throughout a pool. Even when these prop-
Standing." Borden and Rzasa," Katz,'" or Standing and Katz."· erties of the reservoir are corrected to a fixed datum, variation
i.< noted. These anomalies immediately pose problems with
The gas and liquid viscosities, gravity of the gas liberated,
regard to the use of PVT data, It is common to obtain an
gas deviation factors, ga~ formation volume factors, and two
average pool temperature and pressure by volumetrically or
phase furmation volume factors may be obtained a~ outlined
areally weighting these properties corrected to a datum; these
in "minimum" PVT requirements.
average values are then used with the laboratory data 111
Craphical representation of the data obtained frolll these reservoir calculations.
correlations can he extrapolated to cover futurc conditions.
A number of authors have noted that the properties of
fluids in one reservoir are not constant through its entire
extent. This is partic'ularly tflle of pools with considerable
closure."·H For example, in the fifth zone of the Newhall-
Potrero field in California" a variation of tank oil gravities
SAMPLE RESERVOIR PVT ANALYSIS 0
from 25 API downstructure to 39° API upstructure within a
closure of over 2,000 ft has been noted, Variation in properties
The following PVT report is an example of a practical of samples from various structural positions in the Weber sand-
analysis. The results are ha!'ed on all of the previously stated stone reservoir, Rangely field, Colorado, has been noted.'" In
"minimum" PVT requirements with some of the elements of the Weber re~ervoir the bubble point pressures were found
the "ideal" PVT requirements.
to vary from 2,560 psi a to 1,960 psia over 840 ft of relief;
The recombined sample was obtained frum a reservoir hav- similar variations were noted with respect to solution gas-oil
ing a primary gas-cap. Initial reservoir pressure and tempera- ratio, formation volume factor, and properties of the liberated
ture conditions were 2.762 psig and 140°F, respectively, gas phase.
Separator gas was added to the separator fluid charged into Such findings as these were predicted by the work of Sage
the test cell in measured amounts until the recombined sample and Lacey." Their theoretical thermo-dynamic calculations
showed a bubble point of 2,730 psig at 140°F. The formation when converted to an adiabatic basis would most probably
volume factor at these conditions was 1.2522 bbl reservoir check clo~ely the results obtained in the Rangely field. Their
oil per barrel of tank oil. The solution gas-oil ratio was 603 findings indicate that such a gradation of properties represents
qandard cu ft of gas per barrel of tank oil. more a condition of equilibrium than a situation of non-equili-
A portion of the recombined sample was introduced into a brium, Sage and Lacey also predicted that a stratification of
rolling ball viscosimeter where the viscosity of the liquid densities and properties with structural position would occur
phase as a function of pressure was measured. in the gas cap of a reservoir; while such a condition has not
The system ill the cell was then separated under flash con- yet oet'll proVt-tt to (-xis!, lack of gas-cal' sampling may he
ditiuns at the reservoir temperature of 140°F. Following these the reason.
tests, the differential and composite liberations were simul· Such clIlIdiliulh empha,;ize the fact that analyses uf samples
taneuusly performed. Trapping conditions used in the compu- from a number uf wells at various structural po~itions should

Vol. 198, 1953 PETROLEUM TRANSACTIONS, AIME 293

T.P. 3710 APPLICATION OF LABORATORY PVT DATA TO RESERVOIR ENGINEERING PROBLEMS

Table 3 - Pressure-Volume-Temperature and Table 4 - Reservoir Oil- Composite Liberation

Viscosity Data (Gas and Oil Separator Tests)
Reservoir Data Gas-Oil Ratio Gravity Format-
Type of Sample: Recombined Specific tion
Tempel'- Separator Tank Gravity Volume
Room Temperature: 80°F Barometer: 14.4 psi a Pressure ature Oil Oil Tank Oil Gas Factor
Specific
Specific Gravity eu ft per eu ft per Bo
Gas-Oil Ratio
Differential Separation at 14QOF Volume
140 0 F (Air =
Gas
1)
psig of bbl bbl 0 API Air = 1 Yo/V']"!J
Pressure: Relative Viscosity Flash Differen-
Solution Liberated Volume Liquid Separation tiai
Scpal'ation
Cconti- 2730* 140 566.3 36.7 36.4 0.6744 1.2552
psig ell ft per IJbI poises ('u ft per Ib
1986 140 403.3 40.6 36.6 0.7259 1.1752
4964 0.02188
4460 0.6475 Sl1 140 136.8 58.5 36.6 0.7547 1.0776
4002 0.02209
4000 0.6178
0.5955 At pressures above bubble point pressure
3525
3414 1.3341 0.02224 3414 140 1.2428
3183 0.5751
2926 1.3416 0.02237 2926 140 1.2518
2917 0.5589
2748 1.3437 0.02240 *Bubble Point Pressure
0
2741 0.5513 NOTE: Separator Conditions 100 psig and 80 F
2730 687.24 o 1.3442 0.02241 Tank Temperature - RO°F
2661 0.02258 Tank Pressure - 0 psig
2620 0.5904
2428 0.02319
2390 0.6376
2236 563.35 123.89 1.2973 0.6291
2010 0.7012
2000 0.02489 solution in the reservoir oil at different pool pOSItIOns varies.
1986 506.38 180.86 1.2733 0.6305 When reservoir gas correlations are used to obtain other gas
1735 448.38 238.86 1.2483 0.6356 properties, it is important that a representative average gas
1700 0.02685 gravity be used. The methods for obtaining an average value
1600 0.8102
1485 391.03 296.21 1.2258 0.6266 are similar to those described with respect to the reservoir
1470 0.02908 liquid phase.
1314 0.03114 While no factual evidence supports the premise that a
1255 0.9450
1235 340.37 346.87 1.2077 0.6271 density stratification with structural position occurs in a gas
1194 0.03326 cap or gas reservoir, there are no conclusive data to refute this
1066 0.03590 assumption. Such variations, if they occur, will be important
1025 1.0430
985 282.77 404.46 1.1858 0.6449
because the size and behavior of the gas cap are important in
958 0.03884 the recovery of oil from the reservoir."' eo
907 0.04062 Little study has been made of the changes in properties of
842 0.04290
773 1.1617 a gas cap during the depletion of a pool. It is known that the
735 230.12 457.12 1.1645 0.6692 specific gravity of the gas differentially liberated from the
565 1.2876 reservoir fluid increases as the pressure decreases. Some of
435 170.77 516.46 1.1421 0.6911 this solution gas migrates to the gas cap or forms a secondary
275 1.4981
235 111.01 576.23 1.1132 0.7430 gas cap. It then is easy to see that the gas cap in any reservoir
138 1.6186 would vary with respect to properties during depletion. Yet,
50 49.63 637.61 1.0797 0.9815 as far as the authors know, such a variation has not been
o 0 687.24 1.0425 2.1916 1.3315 considered in any reservoir calculations.
Gravity of Residual Oil at 60 F: 35.7 0 API 0
Similarly the variation of gas cap properties has not been
Gas-Oil ratio based on volume of gas released and final residual oil considered, to the author's knowledge, in fields where gas
corrected to 60°F and 14.7 psia. injection is being employed. It is generally agreed that the
VO/V""d at 14.7 psia and 60 0 = 1.00 drier and lighter gas injected into the pool partly migrates
to the gas cap. The composition of this injected dry gas, which
is diluting the existing gas cap, is not similar to that of the
be considered before making any reservoir study. Where suffi- gas liberated from solution in the reservoir oil. A gain the
cient test data are not available, the correlations of Standing" importance of testing samples from the reservoir gas cap is
or Borden and Rzasa" may be helpful. apparent. Data on the variation of the properties of the gas
Correction of PVT data for such structural anomalies has cap with pressure decline would be of considerable use in
been suggested. One method·' involves weighting the PVT data calculations of future productivity.
from wells at various positions on the basis of the tank oil
originally in place. Selection of one PVT analysis considered
representative of an entire pool is another way of correcting
for structural variation of the properties of the reservoir fluids.
Such methods immediately imply that more than one well
has been sampled or that production data and fluid property SUGGESTED FUTURE INVESTIGATION
correlations have been employed. The errors involved in present methods of PVT allaly,;is
Since the properties of the reservoir fluid vary with structural 1 ave been discussed. One of the most important errors results
position in the pool, it is reasonable to assume, and it has been from lack of knowledge as to the nature of the gas liberation
observed, that the specific gravity of the ~as liberated from occurring in the formation. Correlation of relative permeability

294 PETROLEUM TRANSACTIONS, AIME Vol. 198, 1953

 C. R. DODSON, D. GOODWILL, AND E. H. MAYER T.P. 3710

Table 5 - Compressibility and Specific Velume Data regard to gas-cap properties may be quite inaccurate. partic-
Type of Sample: Recombined ularly where gas injection operations are occurring.
Specific Volume No determination of the physical behavior of a reservoir fluid
At Bubble Point Conditions: 2,730 poig and 140°F as far as the authors know has considered the effect of gas
Measured: 0.02241 eu ftjlb injection on the results of the analysis. An answer to the ques-
Calculated by Composition: 0.02190 cu ftjlh
Calculated by Flash Vaporization: Not Run tion of whether the present methods of approximating reservoir
Calculated by Stage Separations gas liberations are applicable when gas of non-equilibrium
at 100 psig and 0 psig 0.02119 cu ftjlb composition is coming in contact with the oil (i.e., by-passing)
Unsaturated Oil Above Bubble Point Pressure would be extremely helpful in laboratory work and in the
Specific Volume at 4964 psig and 140°F: 0.02188 cu ftjlb
Thermal Expansion: Not run application of the results of laboratory tests.
Compressibility at 140°F:
from 2,730 psig b) 3,414 psig. 10.937 x lO- G hhl
per bbl per psi
from 3,414 psig to 4,002 psig. 11.880 x 10-· bbl
per bbl per psi
from 4,002 psig to 4,964 psig. 9.815 x 10- 6 bbl
per bbl per psi
NOMENCLATURE
Symbols
Quantity Units Symbol
Table 6 - Material Balance Computations Using
Composite and Differential Liberation Data Cumulative production Barrels Op
of stock tank oil
Equations used (all symbols are defined in the appendix) :
Cumulative production Barrels
of water
Deviation factor for gas None z
llV. d
(l) Formation volume factor
(2)
1. Single phase Barrels reservoir space B
UiW =B iw • • • • (3)
per stock tank barrel
Vo = Bo + (5'0-50) B•. (4)
Vw =
Bw + (Sw' -Sw) B. (5) 2. Two phase Barrels reservoir space U
B-~~_Z_ (6) per stock tank barrel
• - P R • 520 • 5.62
Gas-oil ratio
Pressures 1. Cumulative produced Standard cubic feet per
QUANTITY* 2070 psig 1560 psig UNITS
gas-oil ratio barrel stock tank oil
acre feet.
III 0.1770 0.1770
acre feeto 2. Solution gas-oil Standard cubic feet per
0,. 11.20 x 10' 18.50 X 106 bbho ratio barrel stock tank oil
WF. 0 0 bbl
34.26 x 10" 67.08 X 10' std cu ft Pressure - absolute Pounds per square inch p
G"
Sf' 3059 3626 cu ftjbbho Ratio of original reservoir gas Barrels of gas at reser- m
W p 0.7950 x lOG 1.249 x 106 bbl cap volume to original vol- voir conditions per
Tank Oil 0 1** ume of reservoir oil barrel of oil at reser-
( calculated) 292.9 x lOG 275.6 X lOG bbho
voir conditions
Differential **
0, (calculated) 231.8 x 106 229.2 x 10' bbl' •• 'ri Change in reservoir pore vol- Barrels II Vsd*
*All PVT data used in computations are obtained from Tables 1 ,to ume due to rock compaction
5 and Figs. 1 to 5.
**For this material balance V,w =
1, Sw =
0, Vw 1, = Solution ratio of Standard cubic feet per S ..
Sw' = 0, llV,d = 0, W, = 0 gas in water volume of water at
standard conditions
Space occupied in the reser- Barrels reservoir space
voir by 1 volume of gas at per standard cubic foot
and phase behavior studies would give much valuable informa- standard conditions
tion in this regard. These studies may show that reservoir
evaluation can be made on the basis that thermodynamic Stock tank oil volume in place Barrels o
equilibrium attains in the reservoir; or conversely they may at any time
permit the magnitude of the error involved in this assump- Temperature - absolute Degrees Rankine T
tion to be approximated if equilibrium does not occur.
That the properties ef reservoir oil vary with structural posi- Volume Barrels, standard cubic V
tion has been pointed out. Further studies of this condition feet
would be valuable to the industry. Whether this variation Water encroachment - gross Barrels
represents a condition of equilibrium or non-equilibrium also
should be investigated. Water at standard conditions Barrels If'
Properties of the vapor phase existing in the gas caps of in place at any time
various reservoirs should be studied. Present assumptions in • A negative numerical value in most cases as a reduction is shown.

Vol. 198, 1953 PETROLEUM TRANSACTIONS, AIME 295

T.P. 3710 APPLICATION OF LABORATORY PVT DATA TO RESERVOIR ENGINEERING PROBLEMS

Subscripts 13. Lindsly, B. E.: "Solubility and Liberation of Gas from

Natural Oil-Gas Solutions," Bull. 554, USBM, (1933).
Quantity Units Symbol
14. Lindsly, B. E.: "The Effect of Gas Withdrawal Upon Res-
Encroached - gross value E ervoir Fluids," Trans. AIME, (1934) 107, 94.
IS. Lindsly, B. E.: "A Study of 'Bottom Hole' Samples of
Cas g East Texas Crude Oil," R.I. 3212, USBM, (1934).
Initial conditions 16. Lindsly, B. E., "A Bureau of Mines Study of a 'Bottom
Hole' Sample from the Crescent Pool, Okla.," Pet. Eng.,
Oil o (Feb.-March, 1936), 7, (5) 34.
17. Dale, C. R.: "Thermal Logging of Producing Oil Wells,"
TIe;,ervoir conditions R
Oil and Gas Jour., (1942) 40, (45), 49.
Ee3idual oil Product of differential resid 18. Tapper, W.: "Caliper and Temperature Logging," Sub-
or flash liberation surface Geologic Methods, 2nd ed., Colorado School of
Mines, Golden, Colo. (1950).
Sand sd 19. Briggs, F.: "Temperature Bomb," Oil Weekly, (1934)
Tank oil Product of composite TO 110, (7), 15.
liberation 20. Sclater, K. c., and Stephenson, B. R.: "Method of Obtain-
ing Bottom Hole Data," Oil and Gas Jour. (October 25,
Time after original No subscript shown ex- 1928).
cept where more than 21. Hawthorn, D. G.: "Subsurface Pressures in Oil Wells and
one subsequent time Their Field Application," Trans. AIME, (1933) 103, 148.
IS considered, in 22. Parks, E. K., and Gibbs, C. W.: "Instrument and Equip-
which case use num- ment for Recording Subsurface Pressures," Trans. AIME.
bers I denoting origi- (1934) 107, 42.
nal conditions, 2 first 23. Exline, P. G.: "A Precision Gage for Sub-surface Pressure
time after original, Measurements," Drill. and Prod. Prac. jar 1936, API,
etc. (1937), 116.
Water 24. Schilthius, R. J.: "Technique of Securing and Examining
w
Subsurface Samples of Oil and Gas," Drill. and Prod.
Prac. jar 1935, API (1936), 120.
25. Exline, P. G.: "New Apparatus for Securing and Exam-
ining Subsurface Samples of Oil," Drill. and Prod. Prac.
jar 1936, API, (1937), 126.
26. Grandone, P., and Cook, A. B.: "Collecting and Examin-
REFERENCES ing Subsurface Samples of Petroleum," Tech. Paper 629,
1. Huntley, L. G.: "Possible Causes of Decline of Oil Wells," USBM, (194
U. S. Bureau of Mines, Tech. Paper 51, (1913). 27. Standing, M '.' A Pressure-Volume-Temperature Cor-
2. McMurray, W. F., and Lewis, J. 0.: "Underground Wastes relation for Ih'A s of California Oils and Gases," Drill.
in Oil and Gas Fields and Methods of Preventivn," U. S. and Prod. Prac. /0/ 1947, API, (1948), 275.
Bureau of Mines, Tech. Paper 130, (1916). 28. Borden, Guy, an: ,tza~a, M. J.: "Correlation of Bottom
3. Blatchley, R. S.: "Waste of Oil and Gas in Mid-Continent Hole Sample Dai,d ': Trans. AIME, (1950), 189, 345.
Fields," U. S. Bureau of Mines, Tech. Paper 45, (1914). 29. Katz, D. L.: "Prediction of Shrinkage of Crude Oil,"
4. Lewis, J. 0., and McMurray, W. F.: "The Use of Mud- Drill. and Prod. Prac. jot' 1942, API, (1943),137.
Laden Fluid in Oil and Gas Wells," U. S. Bureau of Mines, 30. Standing, M. B., and Katz, D. L.: "Density of Crude Oils
Bull. 134, (1916). Saturated with Natural Gas," Trans. AIME, (1942) 146,
5. Lewis, J. 0.: "Methods for Increasing the Recovery from 140.
Oil Sands," U. S. Bureau of Mines, Bull. 148, (1917). 31. Miller, F. C.: "Steady Flow of Two-Phase Single-Com-
6. Lewis, J. 0., and Beal, c.: "Some New Methods of Esti- ponent Fluids Through Porous Media," Trans. AIME,
mating Future Production of Oil Wells," Bull. AIME, (1951) 192, 205.
(1918), 134. 32. Pirson, S. J.: Elements of Oil Reservoir Engineering, Mc-
7. Dow, D. B., and Calkin, L. P.: "Solubility and Effects of Graw-Hill Book Co., Inc., New York, (1950).
Natural Gas and Air in Crude Oils," R. I. 2732, USBM, 33. Sage, B. H.: "Unequilibrium Conditions," Talk presenter!
(1926) . to Pacific Petroleum Chapter, AIME, April 23, 1951.
8. Beecher, C. E., and Parkhurst, I. P.: "Effect of Dissolved 34. Cook, A. B., Spencer, G. B., and Bobrowski, F. P.: "Spe-
Gas Upon the Viscosity and Surface Tension of Crude cial Considerations in Predicting Reservoir Performance
Oil," Trans. AIME, (1926) 51. of Highly Volatile Oil Reservoirs," Trans. AIME, (1951),
9. Mills, R. Van A., and Heithecker, R. E.: "Volumetric 192, 37.
and API Gravity Changes Due to Solution of Gas in Crude 35. Bicher, L., and Katz, D. L.: "Viscosity of Natural Gases,"
Oils," USBM, R.1. 2893, (1928). Trans. AIME, (1944) 155, 246.
10. Miller, H. c.: Function of Natural Gas in the Production 36. Hocott, C. R., and Buckley, S. E.: "Measurements of Vis-
of Oil. API, (1929), New York. cosities of Oils under Reservoir Conditions," Trans, AIME,
II. Coleman, Stewart, Wilde, H. D., Jr., and Moore, T. V.: (1941) 142, 131.
"Quantitative Effect of Gas-Oil Ratios on Decline of Aver- 37. Beal, C.: "The Viscosity of Air, Water, Natural Gas,
age Rock Pressure," Trans. AIME, (1931) 86, 174. Crude Oil, and Its Associated Gases at Oil Field Tem-
12. Lindsly, B. E.: "Preliminary Report on an Investigation peratures and Pressures," Trans. AIME, (1946) 165, 94.
of the Bureau of Mines Regarding Solubility of Natural 38. Brown, G. G.: "Natural Gas Under Pressure," Proc.,
Gas in Crude Oil," Trans. AIME, (1932), 92, 252. Natural Gasoline Association of America, (May, 1940).

296 PETROlEUM TRANSACTIONS, AIME VoL 198, 1953

 C. R. DODSON, D. GOODWILL, AND E. H. MAYER T.P. 3710

39. Standing, M. B., and Katz, D. L.: "Density of Natural liberation method during approximately the first third of
Gases," Trans. AIME, (1942) 146, 140. pressure depletion below the bubble point, breaking up toward
40. Hocott, C. R.: "Interfacial Tension Between Water and the differential liberation curve during the second third of
Oil Under Reservoir Conditions," Trans. AIME, (1939) pressure depletion where gas is being removed approximately
132, 184. as fast as it is liberated in the reservoir, and during the
41. Espach and Fry, J.: "Variable Characteristics of the Oil last third, the formation-volume probably would remain well
in the Tensleep Sandstone Reservoir, Elk Basin Field, above the composite liberation curve and probably closer to
Wyoming and Montana," Trans. AIME, (1951) 192, 75. the differential liberation curve since large quantities of com-
42. Graybeal, O. A.: "California's Prolific Potrero Oil Field," paratively dry (largely methane) reservoir gas would be pass-
Oil and Gas Jour., (1951) 50, (10), 80. ing through the surface separators in intimate contact with
43. Cupps, c. Q., Lipstate, P. H., Jr., and Fry, J.: "Variance the oil, causing an abnormal amount of shrinkage of the
in Characteristics of the Oil in the Weber Sandstone Res· separator oil. There is no discussion in the paper of this
ervoir, Rangely Field, Colo." R.I. 4761, USBM, (1951). last effect (large amounts of excess gas production above
44. Sage, B. H., and Lacey, W. N.: "Gravitational Concentra- the solution GOR) on the formation-volume factor of the
tion Gradients in Static Columns of Hydrocarbon Fluid," oil during the last stages of depletion; in fact, it is neglected
Trans. AIME, (1939) 132, 120. in the laboratory method.
45. Burtchaell, E. P.: "Reservoir Performance of a High Since the formation volume factor is by definition a ratio
Relief Pool," Trans. AIME, (1949) 186, 171. of an end product (either tank oil or residual oil) to a
46. Muskat, M., and Taylor, M. 0.: "Effect of Reservoir reservoir volume of oil, it is difficult even to theorize on
Fluid and Rock Characteristics on Production Histories of the shape of the curve throughout the true liberation sequence
Gas-Drive Reservoirs," Trans. AIME, (1946) 165, 78. as the absolute values depend upon an exact duplication of
47. Babson, E. C.: "Prediction of Reservoir Behavior from the reservoir production process. At least the composite libera-
Laboratory Data," Trans. AIME, (1944) 155, 120. tion method proposed by Dodson, et ai, goes a step further
48. Patton, E. c., Jr.: "Evaluation of Pressure Maintenance along this line in that it considers the tank oil as a variable
by Internal Gas Injection in Volumetrically Controlled instead of a constant which is the manner in which the
Reservoirs," Trans. AIME (1947) 170, 112. residual oil is used in the differential liberation method.
49. Tarner, J.: "How Different Size Gas Caps and Pressure It should be well to remember that neither of the laboratory
Maintenance Programs Affect the Amount of Recoverable methods can exactly duplicate the actual behavior of the
Oil," Oil Weekly, (June 12, 1944) 114, (2),32. solution-gas drive reservoir and the limitations of both of
50. Muskat, M.: Physical Principles of Oil Production, Mc- these methods should be kept in mind.
Graw-Hill Book Co., Inc., New York, (1949).
With reference to Fig. 2 of the paper (solution GOR
curve), the curve obtained by the composite liberation gives
the solution GOR at surface conditions, however, only 20 to
DISCUSSION 25 per cent of this oil ever experiences these conditions. The
By William L. Boyd, Continental Oil Co., Ponca City, Okla. remaining 75 to 80 per cent of the oil remains in the reservoir
It is generally acknowledged that little is known concerning and is best represented by the differential curve, i.e., gas
the exact nature of gas liberation from a solution-gas drive in solution at reservoir pressure and temperatures instead of
oil reservoir. The composite liberation as suggested by Dodson, separating conditions which are at much lower pressures
et al, and the standard differential liberation are laboratory and temperatures.
methods to approximate the true liberation sequence as it It should be mentioned also that there can be no com-
actually occurs in the production process. In applying these parison between the two curves illustrated in Fig. 5. Naturally,
data to reservoir engineering calculations, one would prefer the specific gravity of the composite liberation gas is higher
to use the laboratory data which more nearly approximates than the differentially liberated gas because the pressure-
the true liberation sequence. temperature conditions are entirely unrelated. Without close
Differential liberation as performed in the PVT labora- inspection, the graph would lead one to believed that the
tory is approached by the true liberation sequence only in composite liberation gas specific gravity was 0.73 in the
the intermediate stage of pressure depletion after the relative reservoir at 200 Ib, whereas, it is really the specific gravity
permeability to the reservoir gas is such to allow the flow of of the solution gas liberated at 100 Ib to 0 Ib at 80°F
gas along with the oil so that the rate of gas production equals from the oil which was removed from the reservoir at 2000 lb.
rate of gas liberation in the reservoir. Even this is not entirely As such, these data could not be used in the conventional
correct, as the flowing gas in the reservoir is the gas which material balance calculation.
was in equilibrium with its originating oil and not necessarily The foregoing comments are not meant to detract from
the oil flowing toward the well bore. the data as presented in this paper, but to point out some
On the other hand, the composite liberation makes the of the difficulties in applying correctly these, or any other
assumption of differential liberation in the reservoir plus a set, of laboratory PVT data to reservoir engineering problems.
flash liberation of the oil through the flow string and sep-
arating equipment. This method assumes that the flowing gas
from the primary or secondary gas cap (the differentially
liberated gas) never comes in contact with the produced oil
and its solution gas.· This type of reservoir behavior is AUTHORS' REPLY TO MR. BOYD
obtained in the primary stages of depletion of a solution- Boyd's discussion is appreciated by the authors.
gas drive reservoir where gas saturations are not large enough The section of the paper titled "Types of Gas Liberation
to permit the flow of gas. and Application of Data Obtained from These Liberations to
One could, therefore, reason that in actual reservoir Reservoir Problems" discusses the formation liberations occur-
behavior, the formation volume factor curve for the true ring at various conditions of gas saturation. The differential
liberation sequence would follow the laboratory composite vaporization has been considered most representative of the

Vol. 198, 1953 PETROLEUM TRANSACTIONS, AIME 297

T.P. 3710 APPLICATION OF LABORATORY PVT DATA TO RESERVOIR ENGINEERING PROBLEMS

formation liberation. However, before equilibrium gas satura· A flashing of the produced oil will occur during the flow up
tion is reached, where the oil is of a highly volatile nature, the tubing and through the separators and tanks. If large
or where the producing gas-oil ratio is in excess of the amounts of gas are produced with the oil in the field, make-up
solution gas-oil ratio, the differential process is not completely gas can be injected into the auxiliary cell in the laboratory
representative of the formation liberation. along with the differentially liberated oil to make the flash
At conditions below the equilibruim gas saturation, a flash trapping stage of the composite liberation more representative.
liberation best represents the formation process. Special tech· The liberation to which any produced oil is subjected i~ a
niques must be used to simulate the formation liberation composite of the vaporization in the formation and that
when the re'3ervoir contains oil of high volatility; the paper occurring from well bore to stock tank. This dual process
of Cook, Spencer and Bobrowski (ref. 34 above) outlines a of vaporization occurring during production must be con-
method that will approximate the formation vaporization under sidered in any laboratory technique seeking to similate actual
,;nch circumstances. production.
Conservation of reservoir energy usually dictates that no It is recognized that the oil residual in the reservoir does
well or pool will be produced at gas·oil ratios greatly exceed· not undergo the composite vaporization. However, material
ing the solution gas·oil ratio. However, during production of' balance techniques commonly involve the tank oil in place
depletion type pools and when gas is being injected into ~t any time despite the fact that all this oil in place actually
the reservoir, the gas-oil ratio sometimes may be high. Under will not be produced to the ~tock tanks. Such balance cal-
these circumstances the differential vaporization is not en- culations, therefore. must use volume data that is hased
tirely repre~entative of the formation liberation because a on tank oil.
unit volume of oil produced or remaining in the reservoir Boyd is correct in his statements about Fig. 5_ It also
comes in contact with amounts of gas that are in excess of is to be pointed out that the gas formation volume factoL
those which the volume of oil would evolve during pressure B., used in the material balance is calculated using differential
decline. The gas under such circumstances generally is moving gas gravity data only. B. enters the "Schilthius" type material
much more rapidly than the oil and there is probably in- balance equation only in conversion of the volume of the
sufficient flow agitation to allow attainment of equilibrium. gas cap to surface conditions. Since the gas cap is composed
While the differential liberation is not completely repre· of differentially liberated gas and the original free gas
sentative of the reservoir process in such instances, it rep· (except under circumstances of gas injection), it is felt the
resents the formation vaporization more closely than any volumetric behavior of the gas cap is best represented by
other type. differential data.

298 PETROLEUM TRANSACTIONS, AIME Vol. 198, 1953