With more than 36 years of petroleum industry experience in mud logging, core

analysis, and phase-behavior studies, Phillip L. Moses, reservoir fluid analysis

manager at Core Laboratories Inc. in Dallas, develops new equipment and
techniques for gas/liquid analyses and phase-relationship studies at elevated
pressures and temperatures. He has directed phase-behavior studies from
reservoirs in the U.S., Canada, South America, and Indonesia. A graduate of
Texas A&M U. with a BS degree in physics, he has completed basic and advanced
petroleum reservoir engineering courses at Texas A&M U. Moses has written
several technical papers and has given phase-behavior lectures to numerous
groups. He was a 1967-68 SPE Admissions Committee chairman.

Engineering Applications of Phase

Behavior of Crude Oil and
Condensate Systems
Phillip L. Moses, SPE, Core Laboratories, Inc.

Summary. Fluid samples must be taken early in the life of a reservoir to obtain samples truly representative
of the reservoir fluid. They should be taken only after a carefully planned well conditioning and testing
program. When the PVT data obtained from these samples are used, care should be taken to adjust FVF's and
gas/oil ratios (GOR's) for surface separator conditions.
Introduction
The proper development, engineering, and production
of an oil or gas reservoir requires a considerable
amount of planning. At the same time that plans are
formulated to develop the field, plans should also be
made for a data-gathering program to facilitate
reservoir engineering months and even years into the
future. This data-gathering plan should include a
sufficient number and variety of electric logs and
cores on key wells to describe the reservoir
adequately. Electric logs and core analyses evaluate
the reservoir rock. If reservoir engineering
calculations are to be made to optimize production
from a reservoir, including EOR, then the properties
of the reservoir fluids must also be known. The
properties of the reservoir water fall within narrow
ranges and are seldom studied at reservoir pressures
and temperatures. The properties of the reservoir
water determined are normally confined to chemical
analysis and possibly compatibility tests in cases of
injection projects. This paper is concerned primarily
with the study of the hydrocarbon fluids contained in
a reservoir.
Coring and logging programs should continue
throughout the development of a reservoir. The data
obtained from the last well drilled are as valuable as
the data obtained from the first well. This is usually
not the case for reservoir fluids. Samples
representative of the original reservoir can be
obtained only when the reservoir pressure is equal to
or higher than the original bubblepoint or dewpoint.
Copyright 1986 Society of Petroleum Engineers

Journal of Petroleum Technology, July 1986

Methods are available for extrapolating fluid data

obtained after some pressure decline. Remember,
however, that extrapolation is an educated man's
word for guessing. Plans for obtaining reservoir fluid
samples and analyses should be made early in the life
of a reservoir. Reservoir fluid samples should be
taken before significant reservoir pressure decline has
been experienced.

Oil Reservoirs
Oil reservoirs can be divided into two categories:
ordinary oil reservoirs and near-critical reservoirs.
The ordinary oil reservoirs are sometimes called
black-oil reservoirs. This misnomer does not reflect
the color of the reservoir fluids. It is meant only to
distinguish them from near-critical oil reservoir
fluids. The near-critical fluids will be discussed in
subsequent paragraphs.
Ordinary oils are characterized by GOR's up to
approximately 2,000 ft 3 /bbl [360 m 3 /m 3 ], oil
gravities up to 45API [0.8 f/cm 3 ], and FVF's of
less than 2 bbllbbl [2 m 3 1m ]. Remember that there
is no sharp dividing line between an ordinary and a
near-critical oil. Such factors as composition and
reservoir temperature greatly influence the behavior
of the reservoir fluid. It is often impossible to
determine whether a fluid should be studied as a
near-critical oil or as an ordinary oil until it is
actually in the laboratory being observed.
We have two methods for sampling ordinary or
noncritical oils: surface sampling and subsurface
715

sampling. These methods were discussed by

Reudelhuber l -3 and are not covered in detail in this
paper, but it should suffice to say that wells should
be carefully conditioned before sampling. If wells are
not conditioned properly and the samples are not
representative of the reservoir fluid, then the resulting
fluid study may yield invalid data. In subsurface
sampling, well conditioning usually consists of a
period of reduced flow followed by shut-in. In
separator sampling, it is imperative that the well be
stabilized, then tested for a sufficiently long period to
determine the GOR accurately.
The reservoir fluid study on a noncritical oil should
consist of five tests.
Pressure/Volume Relations. This is a constantcomposition expansion of the reservoir fluid at the
reservoir temperature during which the bubblepoint is
measured. Above the bubblepoint, the compressibility
of the single-phase fluid is measured. Below the
bubblepoint, the two-phase volume is measured as a
function of pressure.
Differential Vaporization. This test measures the
amount of gas in solution as a function of pressure
and the resultant shrinkage of the oil as this gas is
released from solution. Also measured are the
properties of the evolved gas, including the specific
gravity and deviation factor. The density of the oil
phase is also measured as a function of pressure.
Viscosity. Viscosity, which is resistance to flow,
should be measured as a function of pressure at
reservoir temperature.
These three tests are all conducted at reservoir
temperature, and the results describe the behavior of
the reservoir fluid as it exists in the reservoir.
Separator Tests. One or more separator tests should
be measured to determine the behavior of the
reservoir fluid as it passes up the tubing, through the
separator or separators, and finally into the stock
tank. The FVF, Bo, and gas in solution, R s , are
measured during these tests. It is usually
recommended that four of these tests be used to
determine the optimum separator pressure, which is
usually considered the separator pressure that results
in the minimum FVF. At the same pressure, the
stock-tank-oil gravity will be a maximum and the
total evolved gas-i.e., the separator gas and the
stock-tank gas-will be at a minimum. For most
midcontinent crudes, this optimum separator pressure
usually occurs in a range from about 90 to 120 psi
[621 to 827 kPa]. Obviously, some field producing
conditions do not allow the operation of the separator
at optimum pressure. If the gas-gathering line in the
field is at 1,000 psig [6895 kPa], the first-stage
separator must be operated at this pressure or higher.
Therefore, a second separator must be placed in the
flow stream to achieve a near-optimum FVF. The
optimum second-stage separator pressure may also be
determined by the PVT laboratory either
experimentally or through equilibrium ratio
calculations with the reservoir fluid composition and
computers.
As reservoir pressure is depleted and gas is evolved
from solution within the reservoir, the FVF of the
reservoir oil gradually becomes smaller. Ideally, the
716

FVF of the reservoir oil should be measured as a

function of reservoir pressure by placing a large
sample of oil in a PVT cell and pressure-depleting by
differential liberation at the reservoir temperature. At
each of several pressure levels during this differential
depletion, samples are removed and passed through a
separator or separators at surface conditions, and the
FVF and gas in solution are measured. Sufficient
pressure levels should be studied to obtain the data to
plot a curve of FVF and gas in solution as a function
of reservoir pressure. This method, described by
Dodson et al., 4 is an excellent way to study
noncritical oils and should be considered the preferred
method. Unfortunately, most reservoir fluid studies
contain only the separator data on the reservoir oil at
its original bubblepoint. The reservoir fluid report
does not contain a curve of FVF as a function of
reservoir pressure, but only the FVF's at the
bubblepoint. The FVF curve and gas-in-solution curve
must be constructed with a correlation first described
by Amyx et al. ,5 and later by Dake. 6 This
correlation, the adjustment of the differential data to
flash conditions, works reasonably well in most
instances and is far superior to making no correction
at all. Again, Dodson's method is superior.
It is my observation that 70 to 80 % of reservoir
engineers do not understand the conversion of
differential data to flash data; consequently, the
relative-oil-volume curve from the differential
liberation is used instead of the flash-formation-factor
curve. This can lead to errors of 10 to 20% or more
in calculation of oil in place (OIP) and recoverable
oil. An explanation of the conversion from
differential to flash is presented in the Appendix.
Composition of the Reservoir Fluid. Most of the
parameters measured in a reservoir fluid study can be
calculated with some degree of accuracy from the
composition. It is the most complete description of
reservoir fluid that can be made. In the past,
reservoir fluid compositions were usually measured to
include separation of the components methane through
hexane, with the heptanes and heavier components
grouped as a single component reported with the
average molecular weight and density. With the
development of more sophisticated equations of state
to calculate fluid properties, it was learned that a
more complete description of the heavy components
was necessary. It is now recommended that
compositional analyses of the reservoir fluid include a
separation of components through C 10 as a minimum.
The more sophisticated research laboratories now use
equations of state that require compositions through
C 30 or higher.

Near-Critical Oils
Near-critical oils have often been referred to as
volatile oils. Volatile oil is not an apt description
because virtually all reservoir fluids are volatile.
What is really meant is that the reservoir fluid
exhibits the properties of an oil existing in the
reservoir at a temperature near its critical
temperature. These properties include a high
shrinkage immediately below the bubblepoint. In
extreme cases, this shrinkage can be as much as 45 %
Journal of Petroleum Technology, July 1986

of the hydrocarbon pore space within 10 psi [69 kPa]

below the bubblepoint. GOR's are usually 2,000 to
3,000 ft3/bbl [360 to 540 m 3/m 3], and the oil
gravity is usually 40 API [0.83 g/cm 3] or higher.
Near-critical oils have FVF's of 2 or higher. The
compositions of near-critical oils are usually
characterized by 12.5 to 20 mol % heptanes plus,
35 % or more of methane through hexanes, and the
remainder ethane.
Near-critical oils were first discussed in the
literature by Reudelhuber and Hinds 7 and by Jacoby
and Berry. 8 Near-critical oils must be studied
differently in the laboratory and by the reservoir
engineer to arrive at an accurate prediction of
reservoir performance. To understand this, it is
necessary to consider that near-critical oils are
borderline to very rich gas condensates on a phase
diagram.
There is a fairly sharp dividing line between oils
and condensates from a compositional standpoint.
Reservoir fluids that contain heptanes and are heavier
in concentrations of more than 12.5 mol % are almost
always in the liquid phase in the reservoir. Those
with less than 12.5 mol % are almost always in the
gas phase in the reservoir. Oils have been observed
with heptanes and heavier concentrations as low as
10 % and condensates as high as 15.5 %. These cases
are rare, however, and usually have very high tank
liquid gravities.
As mentioned, a near-critical oil undergoes a very
high shrinkage as the pressure falls below the
bubblepoint. This high shrinkage creates a high gas
saturation in the pore space. Because of the gas/oil
relative-permeability characteristics of most reservoir
rocks, free gas achieves high mobility almost
immediately below the bubblepoint. It is fortunate
that this free gas is a rich gas condensate.
Conventional volumetric material-balance techniques
on ordinary oils make no provisions for treating this
mobile gas as a retrograde condensate. Instead, the
calculation procedures bring this free gas flowing in
the reservoir to the surface as free gas and add it to
the solution gas.
A properly performed reservoir fluid study on a
near-critical oil furnishes the data that will enable the
reservoir engineer to perform a compositional
material balance. In this manner, he can account for
production of retrograde condensate, as well as oil,
from the reservoir. Reudelhuber and Hinds 7 reported
that for the reservoir they studied, a compositional
material-balance calculation procedure would predict a
liquid recovery from the reservoir approximately four
times higher than conventional volumetric material
balance would. Jacoby and Berry 8 reported an
approximately 2.5-fold increase for the reservoir they
studied. Jacoby and Berry's study was done on a
reservoir in north Louisiana that was discovered in
late 1953. By conventional material-balance
techniques, Jacobr and Berry predicted that 880,000
bbl [140x 10 3 m ] oil would be produced from the
reservoir. By compositional material-balance
techniques, they predicted that 2.2 million bbl
[350x 10 3 m 3 ] would be produced. By 1965 the field
had been depleted, and Cordell and Ebert 9 presented
0

Journal of Petroleum Technology, July 1986

a case history. Actual recovery from the reservoir

was 2.4 million bbl [382XI0 3 m 3 ]. The excellent
agreement between the actual performance and the
predicted performance confirms the theory behind the
compositional material-balance approach.

Retrograde Gas-Condensate Reservoirs

A retrograde gas-condensate reservoir fluid is a
hydrocarbon system that is totally gas in the
reservoir. Upon pressure reduction, liquid condenses
from the gas to form a free liquid phase in the
reservoir. Retrograde condensate reservoirs are
characterized by gaslliquid ratios of approximately
3,000 to 150,000 ft 3 /bbl [540 to 27XI0 3 m 3 ].
Liquid gravities usually range from about 40 to
60 API [0.83 to 0.74 g/cm 3], although condensate
gravities as low as 29 API [0.88 g/cm 3] have been
reported. IO Color alone is not a good indicator of
whether a particular hydrocarbon liquid is condensate
or oil. The 29 API [0.88-g/cm 3] condensate was
black. High-gravity condensates and oils can be
water-white. We normally do not expect to see
retrograde behavior at reservoir pressures below
about 2,500 psi [17.2 MPa]. At these relatively low
pressures, the condensate is usually very light in
color and high in gravity. The lower gravities and
darker colors observed in condensates are indicators
of heavy hydrocarbons. High pressure is required to
vaporize heavy hydrocarbons; consequently, a
reservoir producing a dark condensate should be
expected to have a high dewpoint.
Gas-condensate reservoirs are almost always
sampled at the separator and recombined in the
producing gaslliquid ratio. Oil wells are conditioned
for subsurface sampling by a reduction in the flow
rate for a period of time and then shut-in until static
pressure has been achieved. If we were to attempt to
condition a gas-condensate well in the same manner,
we would find that the liquid condensate in the tubing
would coalesce and fall to the bottom of the tubing
when the well was shut in. A subsurface sample
would then retrieve a sample of this liquid. The
liquid would exhibit a bubblepoint rather than a
retrograde dewpoint. The composition of the liquid
would be totally different from the composition of the
reservoir fluid.
To obtain samples for reservoir fluid analysis from
a gas-condensate well, the well should ideally be
produced at a rate equal to or slightly above the
minimum stable rate. If, however, a well has been
producing at a stable rate for some time and the rate
is not excessive, then it is usually better to test at this
rate than to adjust the rate to the minimum stable
rate. The most important factor in a flow test is
stabilization. This includes stable wellhead pressure,
stable gas production, and stable liquid production.
For a well producing with a subsurface flowing
pressure below the dewpoint, the liquid saturations
and compositions in the drainage area must also be
stabilized. Once stabilization has been achieved, as a
barrel of liquid condenses from the reservoir fluid in
the vicinity of the well bore , then another barrel of
liquid must enter the wellbore. In this manner, the
saturations and compositions in the vicinity of the
717

TABLE 1-DEPLETION STUDY AT 256F.

Hydrocarbon Analyses of Produced Well Stream, Mol%

Component
Carbon dioxide
Nitrogen
Methane
Ethane
Propane
iso-Butane
n-Butane
iso-Pentane
n-Pentane
Hexanes
Heptanes plus

6,010
0.01
0.11
68.93
8.63
5.34
1.15
2.33
0.93
0.85
1.73
9.99
100.00

5,000
0.01
0.12
70.69
8.67
5.26
1.10
2.21
0.86
0.76
1.48
8.84
100.00

4,000
0.01
0.12
73.60
8.72
5.20
1.05
2.09
0.78
0.70
1.25
6.48
--100.00

Reservoir Pressure
(psig)
3,000
2,100
0.01
0.01
0.13
0.13
76.60
77.77
8.82
8.96
5.16
5.16
1.01
1.01
1.99
1.98
0.73
0.72
0.65
0.63
1.08
1.01
3.82
2.62
100.00
100.00

1,200
0.01
0.12
77.04
9.37
5.44
1.10
2.15
0.77
0.68
1.07
2.25
--100.00

700
0.01
0.11
75.13
9.82
5.90
1.26
2.45
0.87
0.78
1.25
2.42
--100.00

Molecular weight of heptanes plus

Density of heptanes plus

158
0.827

146
0.817

134
0.805

123
0.794

115
0.784

110
0.779

109
0.778

Deviation factor
Equilibrium gas
Two-phase

1.140
1.140

1.015
1.016

0.897
0.921

0.853
0.851

0.865
0.799

0.902
0.722

0.938
0.612

Well stream produced

Cumulative percent of initial

0.000

6.624

17.478

32.927

49.901

68.146

77.902

700'
Trace
0.01
11.95
4.10
4.80
1.57
3.75
2.15
2.15
6.50
63.02
100.00
174
0.837

Composition of equilibrium liquid phase.

wellbore do not change. If the flowing rate is

changed, 3 months may be required to restabilize the
well. Once the well is stable, the gas and liquid
production rates should be measured for 48 hours or
more before sampling.
As is the case with oil reservoirs, gas-condensate
reservoirs should be sampled early in their life,
before significant pressure loss has occurred. Once
reservoir pressure has declined below the original
dewpoint, it is no longer possible to get samples that
represent the original reservoir fluid. When the
reservoir pressure falls below the retrograde
dewpoint, liquid condensate forms from the reservoir
fluid. Initially, there is no permeability to this liquid
phase, and only the remaining reservoir gas flows to
the wellbore. If we sample the well stream under
these conditions by taking samples of separator gas
and liquid and recombining them in the produced
gas/liquid ratio, the dewpoint of the mixture should
be expected to be the current reservoir pressure.
As reservoir pressure continues to fall, more and
more retrograde liquid condenses in the formation; at
some saturation point, this liquid will begin to flow
and enter the wellbore. If the well is tested and
sampled under these conditions, the resultant fluid
after recombination would yield a dewpoint higher
than the current reservoir pressure and could
conceivably be higher than the original reservoir
pressure. When the recombined reservoir fluid is
examined at the current reservoir pressure, some free
liquid will be found in the PVT cell. The amount of
gas in the cell relative to the amount of liquid is
usually interpreted as a measurement of the mobility
ratio in the reservoir at the drainage boundary.
A reservoir fluid study on a condensate reservoir
should include the composition of the separator gas,
718

separator liquid, and recombined reservoir fluid. In

the past, these compositions were carried only
through hexanes, with heptanes plus lumped together
as one fraction. I recommend that these compositions
be carried through decanes as a minimum, with the
undecanes and heavier lumped together as a single
fraction, to facilitate compositional modeling of gascondensate reservoirs. As indicated earlier, some of
the more sophisticated major producing companies
now request analyses to C 30 and higher.
The reservoir fluid study should include a
measurement of the retrograde dewpoint, the fluid
compressibility above the dewpoint, and the gas and
liquid volumes below the dewpoint during a constantcomposition expansion.
Finally, the fluid study should consist of a
simulated depletion. This depletion generally consists
of a series of expansions and constant-pressure
displacements of the reservoir fluid such that the
volume of the cell remains constant at the termination
of each displacement. This procedure is referred to as
a constant-volume depletion. The reservoir gas
produced during each constant-pressure displacement
is charged to analytical equipment, and the
composition and volume are determined. The
deviation factor of the gas produced, the two-phase
deviation factor of the hydrocarbons remaining in the
cell, and the volume of liquid remaining in the cell
should be measured at each of the depletion
pressures. The two-phase deviation factor is not
understood well by most reservoir engineers. The
most popular form of material balance on a
condensate reservoir is the PIZ-vs.-cumulativeproduction curve. The deviation factor used should be
the deviation factor of all of the hydrocarbons
remaining in the reservoir. This includes the
Journal of Petroleum Technology, July 1986

TABLE 2-CALCULATED RECOVERY DURING DEPLETION

Calculated Cumulative Recovery During Depletion
Reservoir Pressure
(psig)
Initial
in Place

6,010

66.24

3,000
329.27

2,100
499.01

700

4,000
174.78

1,200

1,000

681.46

779.02

181.74
777.15
38.52
38.45

0
0
0
0

10.08
53.18
2.26
2.29

21.83
145.16
5.17
5.38

31.89
283.78
8.03
8.73

39.76
440.02
10.51
11.85

47.36
608.25
13.21
15.51

51.91
696.75
14.99
18.05

1,841
835
368
179

0
0
0
0

126
58
26
12

344
163
73
35

674
331
155
73

1,050
526
256
122

1,474
749
374
177

1,709
873
441
206

Total "Plant Products" in

Second-Stage Gas, gal
Ethane
Propane
Butanes (total)
Pentanes plus

204
121
53
23

0
0
0
0

12
7
3
1

27
17
8
3

42
27
13
5

55
36
17
7

70
47
23
10

80
54
27
11

Total "Plant Products" in

Well Stream, gal
Ethane
Propane
Butanes (total)
Pentanes plus

2,295
1,461
1,104
7,352

0
0
0
0

153
95
70
408

404
250
178
890

767
468
325
1,322

1,171
707
486
1,680

1,626
979
674
2,037

1,880
1,137
789
2,249

Cumulative Recovery per

MMscf of Original Fluid
Well Stream, Mscf
Normal Temperature Separation *
StoCk-tank liquid, bbl
Primary-separator gas, Mscf
Second-stage gas, Mscf
Stock-tank gas, Mscf
Total "Plant Products" in
Primary Separator Gas, gal
Ethane
Propane
Butanes (total)
Pentanes plus

5,000

Calculated Instantaneous Recovery During Depletion

Reservoir Pressure
(psig)
6,010
Normal Temperature Separation *
Stock-tank liquid gravity, API at 60F
Separator gas/well-stream ratio, Mscf/MMscf
Primary-separator gas only
Primary- and second-stage separator gases
Separator-gas/stock-tank-liquid ratio, scf/STB
Primary-separator gas only
Primary- and second-stage separator gases
GPM from Smooth Well-Stream Compositions
Ethane plus
Propane plus
Butanes plus
Pentanes plus

5,000

4,000

3,000

2,100

1,200

700

49.3

51.7

55.4

60.4

64.6

67.5

68.6

777.15
815.67

802.85
837.04

847.45
874.26

897.28
915.77

920.44
935.04

922.04
936.84

907.14
925.38

4,276
4,488

5,277
5,502

7,828
8,076

13,774
14,058

19,863
20,178

22,121
22,476

19,475
19,867

12.212
9.917
8.456
7.352

10.953
8.648
7.209
6.158

9.175
6.856
5.434
4.437

7.509
5.164
3.752
2.800

6.851
4.469
3.057
2.108

6.970
4.479
2.990
1.959

7.574
4.963
3.349
2.171

'Primary separator at 450 psig and 75F; secondstage separator at 100 psig and 75F; stock tank at 75F.

remaining gas phase and the retrograde liquid. The

two-phase deviation factor furnishes this information.
On lean gas-condensate reservoirs, use of the wrong
deviation factor will not result in serious error, but
use of the wrong deviation factor on a richcondensate reservoir will cause serious errors and
will generally lead to an understatement of reserves.
Table 1 illustrates data typically measured during a
depletion study.
The data measured during the depletion study are
then used for a recovery calculation for a unit volume
reservoir. The results of these calculations are
illustrated in Table 2. The unit volume chosen was 1
MMcf [28 x 10 3 m 3] original reservoir fluid at the
dewpoint pressure. Col. 1 in Table 2 illustrates the
amount of stock-tank liquid, primary-separator gas,
Journal of Petroleum Technology, July 1986

second-stage separator gas, stock-tank gas, etc., in

place in this unit volume reservoir. The amount of
stock-tank liquid in 1 MMcf [28x 10 3 m 3 ] reservoir
fluid depends on the temperature and pressure of the
separators at the surface. At this point, the reservoir
fluid study can be tailored to a specific field condition
by making the recovery calculations at the separator
conditions used in the field. Because the data reported
in Table 2 are the results of computer calculations, a
variety of separator conditions can be investigated
with a relatively small additional investment in
computer time. Note that in Col. 1 of the example,
181.74 bbl[29 m 3 ] ofstock-tank liquid were initially
in place in the unit volume reservoir. By the time the
reservoir pressure had been depleted to 700 psig [4.8
MPa] (Col. 8), only 51.91 bbl [8.3 m 3 ] had been
719

produced. The difference between the initial in place

and that produced at 700 psi [4.8 MPa] is 129.83 bbl
[20.6 m 3 ]. This amount remains in the reservoir at
this pressure as retrograde loss or unproduced at this
pressure. Similar figures are available for the
primary-separator gas, second-stage gas, etc. (For a
more detailed explanation of these recovery
calculations, refer to Ref. 11.) The recovery
calculations of the gas-condensate reservoir are made
with the assumption that the retrograde liquid does
not achieve mobility in the reservoir, which allows
for a finite solution of a recovery calculation as
opposed to the trial-and-error solution required for an
oil reservoir where two phases flow. This assumption
appears to be a good one for most gas-condensate
reservoirs. Only the very rich gas-condensate
reservoirs ever achieve sufficient liquid saturation to
achieve liquid mobility in significant amounts. In
cases where liquid mobility is significant, a
compositional material-balance approach is required to
predict reservoir performance.

Rsd

separator to stock-tank conditions,

bbllbbl [m 3 /m 3 ]
cubic feet of gas in solution at any
pressure less than the bubblepoint in 1
bbl [0.16 m 3 ] residual oil when measured by differential liberation, ft 3 /bbl
[m 3 /m 3 ]

R sdb

= cubic feet of gas in solution at the

RsJ

Rsjb

bubblepoint in 1 bbl [0.16 m 3 ] residual

oil when measured by differential
liberation at reservoir temperature,
ft 3 /bbl [m 3 /m 3 ]
cubic feet of separator and stock-tank gas
in solution at any pressure less than the
bubblepoint in 1 STB [0.16 m 3 ],
ft 3 /bbl [m 3 /m 3 ]
cubic feet of separator and stock-tank gas
in solution at the bubblepoint in 1 STB
[0.16 m 3 ], ft 3 /bbl [m 3 /m 3 ]

Conclusions

References

The two basic methods of collecting reservoir fluid

samples are subsurface and surface or separator
samples. In either case, the reservoir must be
sampled before a significant loss in pressure has been
experienced, and great care must be taken in
preparing a well for sampling. Both must be adhered
to if representative samples are to be obtained.
The studies performed in the laboratory must
recognize the character of the oil. For the laboratory
personnel or the reservoir engineer to treat a nearcritical oil as an ordinary oil would grossly understate
the producing potential of the field.
The reservoir engineer must make proper
adjustments in fluid data to account for the
differences in the flash and differential processes.

1. Reudelhuber, F.O.: "Separator Sampling of Gas-Condensate

Reservoirs," Oil & Gas J. (June 21, 1954) 138-40.
2. Reudelhuber, F.O.: "Better Separator Sampling of Crude Oil
Reservoirs," Oil & Gas J. (Nov. 8, 1954) 181-83.
3. Reudelhuber, F.O.: "Sampling Procedures for Oil Reservoir
Fluids," JPT (Dec. 1957) 15-18.
4. Dodson, C.R., Goodwill, D., and Mayer, E.H.: "Application of
Laboratory PVT Data to Reservoir Engineering Problems," JPT
(Dec. 1953) 287-298; Trans., AIME, 198.
5. Amyx, J.M., Bass, D.M., and Whiting, R.L.: Petroleum Reservoir
Engineering (1978) 67-70.
6. Dake, L.P.: Fundamentals o/Reservoir Engineering (1978) 67-70.
7. Reudelhuber, F.O. and Hinds, R.F.: "Compositional Material Balance Method for Prediction of Recovery from Volatile Oil Depletion
Drive Reservoirs," JPT (Jan. 1957) 19-26; Trans., AIME, 210.
8. Jacoby, R.H. and Berry, V.J. Jr.: "Method for Predicting Depletion
Performance of a Reservoir Producing Volatile Crude Oil," JPT
(Jan. 1957) 27-33; Trans., AIME, 210.
9. Cordell, J.C. and Ebert, C.K.: "A Case History-Comparison of
Predicted and Actual Performance of a Reservoir Producing Volatile
Crude Oil," JPT (Nov. 1965) 1291-93.
10. Kilgren, K.H.: "Phase Behavior of a High-Pressure Condensate
Reservoir Fluid," JPT(Aug. 1966) 1001-05; Trans., AIME, 237.
11. Moses, P.L.: "Consideration of Retrograde Loss in Determining
the Optimum Economic Operation of a Gas Condensate Reservoir, "
18th Southwestern Petroleum Short Course, Dept. of Petroleum
Engineering, Texas Tech U. (April 1971).

Nomenclature
B0

= barrels

Bod =

B odb =

B oj =

B ojb

720

of bubblepoint oil required to yield

1 STB [0.16 stock-tank m 3 ] oil at 60F
[16C], bbllbbl [m 3 /m 3 ]
barrels of oil at some reservoir pressure
other than the bubblepoint pressure required to yield 1 bbl [0.16 m 3 ] residual
oil at 60F [16C] when differentially
liberated to atmospheric pressure,
bbllbbl [m 3 /m 3 ]
barrels of bubblepoint oil required to yield
1 bbl [0.16 m 3 ] residual oil at 60F
[16C] when differentially liberated to
atmospheric pressure, bbl/bbl [m 3 /m 3 ]
barrels of oil at some reservoir pressure
other than the bubblepoint pressure required to yield 1 STB [0.16 m 3 ]
at 60F [16C] when flashed
through the separator to stock-tank
conditions, bbllbbl [m 3 /m 3 ]
barrels of bubblepoint oil required to yield
1 STB [0.16 m 3 ] at 60F [16C]
when flashed through the

Appendix-Toward a Better
Understanding of the Differential
Flash Process
Reservoir depletion and production consist of two
separate processes or a combination of them:
differential liberation of gas and flash liberation of
gas. The differential liberation is defined as a process
whereby gas is removed from oil as it is released
from solution. By contrast, in a flash liberation of
gas, all of the gas remains in contact with all of the
oil until equilibrium between the two phases is
attained.
Most people believe that the differential liberation
process more nearly represents the process that
occurs in an oil reservoir. Actually, the reservoir
process is a combination of differential and flash.
Immediately below the bubblepoint, while there is
little or no permeability to a gas phase, the process is
Journal of Petroleum Technology, July 1986

TABLE A-1-DIFFERENTIAL VAPORIZATION AT 220F

Pressure
(psi g)

Solution GOR, *

2,620
2,350
2,100
1,850
1,600
1,350
1,100
850
600
350
159

854
763
684
612
544
479
416
354
292
223
157

Rsd

Relative
Oil
Volume, Bod * *
1.600
1.554
1.515
1.479
1.445
1.412
1.382
1.351
1.320
1.283
1.244
1.075
1.000

1000.----...,.-...,.--,---.--,--,---,

900~---+----+---_1----~----~---+--~

;l

SOO~---+----+----+----~--~----~~--i

700~--~--~~--~--~----4-J~~4----4
/V

:c

~ 600~---+----+---~----~V~L-~---+--~

u:

o
L
~ 500 ~-+---+----j
~v----:F-+---+---+---i

i5 400~-~-~-~~~--}--~-~-~

'Cubic feet of gas at 14.65 psia and 60F per barrel of residual oil
at 60F.
.. Barrels of oil at indicated pressure and temperature per barrel of
residual oil at 60F.

L/

300~---+~~~+----+----~----~---+----i

200

100

1.S0

II

OL---L-~--L--~---L--~--~

1.70

400

SOO

G)

1.600

E 1.60
:s

(5

1600

2000

2400

2S00

Fig. A-2-Adjustment of gas-in-solution curve to separator

conditions.

>

i5

1200

Pressure: PSIG

1.50

CD

;>

tV

a;

1.40
1.0

a::

1.30

1.20

CD
CI

...

tV

1/

0.9 1-----1f----+V----:iiO"'I"''-----+--+--+-~

.c

(/)

1.10
1.00

''I(JV
0.7

L....._--'-_ _- ' -_ _ _ _ _ _- ' -_ _.l...-_-"-_-:-:'

400

r------r---,--,k--::---rV--;P------r----,

SOO

1200

1600

2000

2400

2S00

Pressure: PSIG

J----+----~--~--_1----~---+--~

0.6 L....._.....L.._ _l..-_-'-______

500

1000

1500

2000

...J....._-.J._~

2500

3000

3500

Pressure: PSIG

Fig. A-1-Relative oil volume vs. pressure.

Fig. A-3-0il-shrinkage curve.

primarily a flash process. As the reservoir gas

saturation reaches the critical saturation, gas begins to
flow and is removed from the reservoir oil. This is a
differential liberation of gas. Much of the gas,
however, remains in the reservoir with the oil as
pressure in the reservoir falls. This is a flash
liberation of gas. So the reservoir process begins as a
flash process and soon becomes a combination flash
and differential process. As pressure continues to
decline, more and more gas flows, bringing the
process closer to a differential process. Once oil and
gas enter the tubing, they flow together until they
reach the separator. In the separator they are brought
to equilibrium, and the gas and oil are separated.
This is a flash separation.
The reservoir process is simulated in the laboratory
by a differential liberation. The test is sometimes

referred to as a differential vaporization. The flash

liberation is simulated in the laboratory with separator
tests. It takes a marriage of the differential
vaporization and separator tests to prepare the
reservoir fluid data for engineering calculations.
In the laboratory, the differential liberation consists
of a series-usually 10 to I5-of flash liberations. An
infinite series of flash liberations is the equivalent of
a true differential liberation. At each pressure level,
gas is evolved and measured. The volume of oil
remaining is also measured at each depletion
pressure. This process is continued to atmospheric
pressure. The oil remaining at atmospheric pressure is
measured and converted to a volume at 60 P [16C].
This final volume is referred to as the residual oil.
The volume of oil at each of the higher pressures is
divided by the volume of residual oil at 60 P [16C].

Journal of Petroleum Technology, July 1986

721

TABLE A-2-SEPARATOR TESTS

Separator
Pressure
(psi g)

Temperature
(OF)

50
to 0

75
75

GOR, R Sfb *

StockTank Oil
Gravity
(OAPI at 60F)

FVF, B ofb * *

737
41

40.5

1.481

40.7

1.474

40.4

1.483

40.1

1.495

778
100
to 0

676
92

75
75

768
200
to 0

602
178

75
75

780
300
to 0

75
75

549
246
795

'GOR in cubic feet of gas at 14.65 psi a and 60F per barrel of stock-tank oil at 60F.
"FVF is barrels of saturated oil at 2.620 psig and 220F per barrel of stock-tank oil at 60F.

1.80

1000

IB-B BOlbl
0-

od

Bodb

1_70
1.600

1.60
CD

E
~

1.50

"0

1.445

>

(5 1.40

CD

;>
m 1.30

CD

a:

1.20

./

L, / . /V

./

/'

./

.-/

:c
In

-M

I
I

600

u.

_2

1445
(
.

1.474
1.600 )

BASIS: 100 PSIG

Separator@ 75F.
FVF = 1.474

I_

iii
a:

800

1200

1600

2000

2400

(5

enm

400

Cl

300

~{/

Fig_ A-4-Adjustment of oil-relative-volume curve to

separator conditions.

I
I

f
o

400

800

1.600

'l

jl

2800

Pressure: PSIG
100

/ AI76S -[(S54-763H.!.EiI]

500

200
400

:::1

7
v//~-I

700

1.474-

,,'"

800

/"

R.= R"b-(R'db-R'd)

)68

1.10 I

1_00

-I

900

1200

BASIS: 100 PSIG

Separator@75F.
Total GOR 768

1600

2000

2400

1_
2800

Pressure: PSIG

Table A-I and Fig. A-I illustrate these data. The

volumes of gas evolved are also divided by the
residual oil volume to calculate the solution GOR data
in Table A-I and Fig. A-2. These data are reported
in this form by long-standing, but unfortunate,
convention. The residual oil in the reservoir is never
at 60F [16C] but always at reservoir temperature.
Reporting these data relative to the residual oil at
60F [16C] gives the relative-oil-volume curve the
appearance of an FVF curve, leading to its misuse in
reservoir calculations. A better method of reporting
these data is in the form of a shrinkage curve. We
may convert the relative-oil-volume data in Fig_ A-I
and Table A-I to a shrinkage curve by dividing each
722

Fig. A-5-Adjustment of gas-in-solution curve to separator

conditions.

relative oil volume factor, Bod, by the relative oil

volume factor at the bubblepoint, B odb .
The shrinkage curve now has a value of I at the
bubblepoint and a value of less than I at subsequent
pressures below the bubblepoint, as in Fig. A-3_ As
pressure is reduced and gas is liberated, the oil
shrinks_ The shrinkage curve describes the volume of
this original barrel of oil in the reservoir as pressure
declines. It does not relate to a stock-tank barrel or
surface barrel.
Journal of Petroleum Technology, July 1986

We now know the behavior of the oil in the

reservoir as pressure declines. We must have a way
of bringing this oil to the surface through separators
and into a stock tank. This process is a flash process.
Most reservoir fluid studies include one or more
separator tests to simulate this flash process. Table
A-2 is a typical example of a set of separator tests.
During this test, the FVF is measured. The FVF is
the volume of oil and dissolved gas entering a
well bore at reservoir pressure and temperature
divided by the resulting stock-tank oil volume after it
passes through a separator.
The FVF is B 0; because separators result in a flash
separation, we should add a subscript, B oJ . In most
fluid studies, these separator tests are measured only
on the original oil at the bubblepoint. The FVF at the
bubblepoint is B ojb. To make solution-gas-drive or
other material-balance calculations, we need values of
BoJ at lower reservoir pressures. From a technical
standpoint, the ideal method for obtaining these data
is to place a large sample of reservoir oil in a cell,
heat it to reservoir temperature, and pressure-deplete
it with a differential process to simulate reservoir
depletion. At some pressure a few hundred psi below
the bubblepoint, a portion of the oil is removed from
the cell and pumped through a separator to obtain the
flash FVF, B oJ , at the lower reservoir pressure. This
should be repeated at several progressively lower
reservoir pressures until a complete curve of B oj vs.
reservoir pressure has been obtained. These data are
occasionally measured in this manner in the
laboratory; this method, which is the best for
obtaining data, is sometimes called the Dodson
method. 4 The process is time-consuming and
consequently adds to the cost of a study. Most studies
include only values of B ojb, the FVF at the
bubblepoint. The values of B oj at lower pressures
must be obtained by other means. A method has been
proposed for accomplishing this mathematically. 5,6 In
essence, the method calls for mUltiplying the flash
FVF at the bubblepoint, B ojb, by the shrinkage
factors at various reservoir pressures obtained earlier.
The shrinkage factor was calculated by dividing the
relative oil volume factors, Bod, by the relative oil
volume factor at the bubblepoint, B odb. If we
combine both calculations, we can start with the
differential-relative-volume curve and adjust it to
separator or flash conditions by
Bojb
Bo=Bod--' .......................... (A-l)
Bodb

This calculation is illustrated in Fig. A-4.

To perform material-balance calculations, we must
also have the separator and stock-tank gas in solution

Journal of Petroleum Technology, July 1986

as a function of reservoir pressure. These values are

expressed as standard cubic feet per barrel and
usually are designated R sJ . The separator tests give us
this value at the bubblepoint, R sjb ' As pressure
declines in the reservoir, gas is evolved from
solution. The amount of gas remaining in solution in
the oil is then somewhat less. The differential
vaporization tells us how much gas was evolved from
the oil in the reservoir: (Rsdb -R sd ), where Rsdb is
the amount of gas in solution at the bubblepoint as
measured by differential vaporization at the reservoir
temperature and Rsd is the gas in solution at
subsequent pressures.
The units of Rsdb and Rsd are standard cubic feet
per barrel of residual oil. Because we must have the
gas in solution in terms of standard cubic feet per
barrel of stock-tank oil, this term must be converted
to a stock-tank basis. If we divide (Rsdb -Rsd) by
B odb, we have the gas evolved in terms of standard
cubic feet per barrel of bubblepoint oil. If we then
multiply by B ojb, we will have the gas evolved in
terms of standard cubic feet per barrel of stock-tank
oil. This expression now is (Rsdb -Rsd)(BojbIBodb)'
The gas remaining in solution then is R s = R sjb (R sdb - R sd )(B ojb IB odb) standard cubic feet per stocktank barrel. For every pressure studied during the
differential liberation, Rs may be calculated from this
equation. This calculation is illustrated in Fig. A-5.
It is a fairly common practice to use differential
vaporization data for material-balance calculations.
Values of Bod and Rsd are almost always higher than
the corresponding values from separator tests;
consequently, calculations of OIP and recoverable oil
will usually be lower than is correct. The differential
vaporization data should be converted to separator
flash conditions before use in calculations. The
methods presented in this paper are approximations.
For more accurate data, consider the method
proposed by Dodson et al. 4

51 Metric Conversion Factors

API 141.5/(131.5 + API)
bbl
ft3
ft3/bbl
OF
gal
psi

x 1.589 873
x 2.831 685
x 1.801 175
(OF-32)/1.8
x 3.785412
x 6.894 757

g/ cm 3

E-Ol
E-02
E-Ol

m3
m3

E-03
E+OO

m3
kPa

m 3 /m 3
C

JPT
This paper is SPE 15835. Distinguished Author Series articles are general, descriptive presentations that summarize the state of the art in an area of technology by describing recent developments for readers who are not specialists in the topiCS discussed.
Written by individuals recognized as experts in the areas. these articles provide key
references to more definitive work and present specific details only to illustrate the
technology. Purpose: To inform the general readership of recent advances in various
areas of petroleum engineering. A softbound anthology, SPE Distinguished Author Series: Dec. 19B1-Dec. 19B3, is available from SPE's Book Order Dept.

723

Discussion of Engineering Applications

of Phase Behavior of Crude Oil and
Condensate Systems
Fred H. Poettmann, SPE, Colorado School of Mines
Robert S. Thompson, SPE, Colorado School of Mines
We read with particular interest Phillip L. Moses' paper,
"Engineering Applications of Phase Behavior of Crude
Oil and Condensate Systems" (July 1986 JPT, Pages
715-23). We agree that many reservoir engineers do not
know how to interpret or to use a typical laboratory fluid
analysis report properly. In the course we teach on reservoir fluids, we cover the material discussed by the author.
To emphasize the importance of the proper use of the data,
the student is required to submit a term report in which
he optimizes two- or three-stage separator pressures to
maximize the production of stock-tank liquid. He does
this for a noncritical oil, critical oil, or gas condensate.
In the case of a noncritical oil, he is also asked to simulate a laboratory differential liberation test and Dodson's
experiment at the optimum separator conditions he previously determined. The input data the student is given are
the reservoir fluid composition and the reservoir temperature. The fluids are at or above the bubblepoint or upper
dewpoint pressure in the reservoir. To make these calculations, the student uses a Peng-Robinson equation-of-state
program.
The final product is a report for management that recommends separator conditions for -field separation and
PVT data for use in reservoir engineering calculations.
In the usual case where Boj and Rsj data are not available at pressures below the bubblepoint pressure, we emphasize the need for estimating Boj and Rsj from data in
the laboratory report on the reservoir fluid. We cover the
procedure described by the author. Depending on the
reservoir fluid, however, and the fact that the differential liberation carried out in the laboratory is not necessarily representative of what occurs in the reservoir, one
can obtain negative values of Rsj at the lower reservoir
pressures. In other words, the product of
Bojb
(Rsdb -Rsd)-Bodb

can be greater than R sjb at low pressure. Thus the amount

of gas liberated by the differential liberation process exceeds the amount of gas assumed to be available for liberation on the basis of flash liberation of bubblepoint oil.
A second procedure, not original with the authors, is
also described that does not require the use of the differentialliberation test. The procedure simply states that the
ratio of Boj and Rsj to B ojb and R sjb , respectively, as

Journal of Petroleum Technology, November 1986

calculated from Standing's correlation or similar correlations, is the same as the ratio for the actual fluid under
consideration. In other words,
Bot'
B.+=B+b - OJ
OJ
B
'
ojb

and

where Bot', B ojb ', Rst', and R sjb ' are obtained from
Standing's or a similar correlation. The values for B ojb
and R sjb are taken from a laboratory separator test and
are assumed to model field conditions accurately. In essence, this procedure assumes that the ratio of the properties of the fluid under study is statistically similar to
the ratio of the properties as obtained from the correlation. The problem of negative values for Rsj is also
avoided.
Another method for estimating Rsj is to use a procedure similar to what Moses described for obtaining B oj;
i.e.,

This, of course, requires the use of the laboratory differentialliberation data. The method, however, eliminates
the problem of calculating negative values for R sj .
In general, we find that the students have a difficult time
in assimilating the concepts of the phase rule and
hydrocarbon phase behavior, particularly when transposed
into petroleum engineering jargon, concepts, and units.
Understanding phase behavior is an important building
block in the study of petroleum engineering and is usually the students' first exposure to petroleum engineering.
We hope we can improve on Moses' personal" ... observation that 70 to 80 % of reservoir engineers do not
understand the conversion of differential data to flash
data ... " using the information available in a laboratory
report on a reservoir fluid.
(SPE 16046)

JPT

1263

Author's Reply to Discussion of Engineering

Applications of Phase Behavior of
Crude Oil and Condensate Systems
Phillip L. Moses, SPE, Core Laboratories Inc.
I commend F.H. Poettmann and R.S. Thompson for the
thoroughness with which they teach the conversion from
differential to flash. It is in the universities, through the
efforts of such professors, that petroleum engineering calculations can be upgraded.
Although I appreciate their comments, I'd like to reply
to those regarding their final method for calculating R sf .
It is my experience that the method presented in the paper
more accurately approximates laboratory-measured data
using the Dodson method. It is true that values of RSf
usually go negative at low pressures, usually below the
lOO-psi [689-kPa] range. The calculated curve must be
manually drawn to Rsf=O at atmospheric pressure. It is
my opinion that the greater accuracy of the data in the

1264

higher pressure ranges justifies the inconvenience of the

problem at very low pressures. A similar problem occurs
when the differential relative oil volume factor is adjusted
to a flash FVF. Bod and B of must be the same value at
atmospheric pressure. The FVF at atmospheric pressure
is a thermal FVF and is the volume of oil at atmospheric
pressure and reservoir temperature divided by the volume
of oil at atmospheric pressure and 60F [16C]. The value
for Bod at atmospheric pressure may be used directly for
B of at atmospheric pressure. The calculated curve for B of
should be manually drawn to the value of Bod at atmospheric pressure.
(SPE 16177)

JPT

Journal of Petroleum Technology, November 1986

Discussion of Engineering Applications

of Phase Behavior of Crude Oil and
Condensate Systems
H.J. Smith, SPE, Consultant

Phillip L. Moses has produced an excellent paper, "Engineering Applications of Phase Behavior of Crude Oil
and Condensate Systems" (July 1986 1PT, Pages
715-23). Most petroleum engineers would benefit greatly from studying it, particularly the Near-Critical Oils and
Retrograde Gas-Condensate Reservoirs sections.
I have tested and sampled more than 400 wells over
the past 30 years in nearly all areas of the globe and have
done or participated in more than 100 reservoir studies
of all types of reservoirs.
I have never seen or heard of a condensate liquid that
shows color. To me, the distinguishing characteristic of
a condensate liquid is the complete absence of lubricating qualities; it feels like gasoline to the fingers. An oil,
on the other hand, has obvious lubricating qualities to the
touch. And an oil, in my experience, always shows color.
I have found oils as high as 55 API [0.76 g/cm 3] (Le-

hib field, Libya) and condensate liquids as low as 38 API

[0.83 g/cm 3] (Java Sea fields, Indonesia).
Also, in my experience, if a gas (cap or otherwise) is
not absolutely devoid of liquids-i.e., if it contains any
liquids at all-retrograde condensate behavior will always
occur, whether or not it can be observed and/or measured.
One can almost always say, I believe, that cap gas above
an oil leg will contain liquids. This does not mean, however, that a gas well test showing retrograde liquids can
always be said to indicate an oil leg downstructure, as
one of my clients recently insisted. A very dry gas zone
will never have an oil leg.
More study and a better understanding of the data Moses
presents so well will result in fewer clients' demanding
bottomhol~ fluid sampling of gas and gas-condensate
wells.
(SPE 16390)
JPT

Author's Reply to Discussion of Engineering

Applications of Phase Behavior of Crude Oil
and Condensate Systems
Phillip L. Moses, SPE, Core Laboratories Inc.
The classification of a hydrocarbon liquid as a condensate or an oil has nothing to do with its lubricating qualities. Oil is generally considered to be a hydrocarbon liquid
that is in the liquid phase in the reservoir, when removed
from the reservoir, and when obtained at the surface. Condensate, whether it be the product of retrograde condensation or normal condensation, is gaseous in the ground.
The liquid hydrocarbon obtained at the surface is the result of a temperature drop or a combination of pressure
and temperature drop as the gas is produced from the
reservoir to the surface. As the name implies, condensate is a hydrocarbon liquid condensed from a gaseous
phase.
Condensate produced from low-pressure reservoirs is
apt to be clear. Retrograde condensate, i.e., condensate
from high-pressure reservoirs, usually has some color.
The color ranges from light straw to black, as mentioned
in the paper.
Many gases contain condensable liquids but do not exhibit retrograde behavior. Condensable liquids usually
refer to propane and heavier hydrocarbons or ethane and
heavier hydrocarbons, which can be removed in a gasoline plant. Wet gases and dry gases may both contain substantial amounts of condensable liquids but do not exhibit
retrograde behavior. Dry gases produce no liquid at the
Journal of Petroleum Technology, February 1987

surface even though gas may contain sizable quantities

of propane and heavier hydrocarbons.
Smith is correct in stating that a gas cap above an oil
zone will contain "liquids." When a gas cap exists over
an oil zone, it is axiomatic that at the gilS/oil interface,
the gas and oil are at equilibrium. For every component
that exists in the oil, some amount of that same component will exist in the gas. The concentrations in each phase
are determined by the laws of physics or thermodynamics. Also, the bubblepoint of the oil at the gas/oil interface must be reservoir pressure at reservoir temperature.
Likewise, the dewpoint of the gas at the gas/oil interface
must be reservoir pressure at reservoir temperature. The
dewpoint will be retrograde in the case of high-pressure
reservoirs or normal in the case of low-pressure reservoirs. We may extend this to say that if an oil reservoir
is discovered with a bubblepoint equal to reservoir pressure, we should expect this reservoir to have a gas cap.
Also, if we discover a condensate reservoir with a dewpoint equal to reservoir pressure, we should expect this
reservoir to have an oil leg. No conclusions may be drawn
in regard to the size of the gas cap or the oil leg from
the PVT data.
(SPE 16440)

JPT
235

Discussion of Engineering Applications of Phase

Behavior of Crude Oil and Condensate Systems
K. Glowacki, SPE, New Mexico Inst. of Mining and Technology
In "Engineering Applications of Phase Behavior of Crude Oil and
Condensate Systems" (July 1986JPT, Pages 715-23), Moses said,
"Most people believe that the differential liberation process more
nearly represents the process that occurs in an oil reservoir. Actually, the reservoir process is a combination of differential and
flash. Imme<liately below the bubblepoint, while there is little or
no permeability to a gas phase, the process is primarily a flash
process. As the reservoir gas saturation reaches the critical saturation, gas begins to flow and is removed from the reservoir oil.
This is a differential liberation of gas. Much of the gas, however,
remains in the reservoir with the oil as pressure in the reservoir
falls. This is a flash liberation of gas. So the reservoir process begins
as a flash process and soon becomes a combination flash and differential process."
I believe that these statements are incorrect and produce misunderstanding and confusion. The following is meant to clarify some
of the statements.
1. In reservoir conditions, the differential process always occurs.
2. In production tubing and surface pipeline flow and in the separators, the flash process takes place (subject to some limiting assumption).
There is a significant difference between flash and differential
processes. Flash process refers to the conditions where the mass
of the considered system does not vary with changes in pressure
and temperature. The flash process in two-phase regions (vaporization or condensation) can be defined in terms of totl\l system composition. The total system composttion, Zi, can be measured at any
point outside the saturation line-e.g., Points A, B, and C in Fig.
D-l. As !l substitution, the following treatment can be used: the
total system composition in two-phase-region flash process remains
constant. The flash process may ensue for a composition Zi that
separates into two phases for the values of pressures and temperatures inside the saturation-curve area. After the temperature and
pressure are chosen, all the gas is in equilibrium with all the oil.
In other words, a change of pressure or temperature or both in a
flash process can change the equilibrium conditions according to
the Gibbs phase rule. This rule provides the number of independent
variables that, in tum, define intensive properties. Flash vaporization may be a batch or a continuous process. Treating two-phase
flow in tubing as a steady state and neglecting the gas storage effect
and gas slippage results in a flash process. In a horizontal flow and
in separators, a similar flash process occurs.
The same kind of eqUilibrium, but with fluid mass decreasing
differentially, is called a differential process (liberation or condensation). A scpematic representation of both processes, flash vaporization and differential vaporization (liberation), in laboratory
conditions can be found in basic handbooks (e.g., Refs. 1 and 2).
In reservoir conditions, the hydrocarbon PV (HCPV) remains
constant if the expansion of interstitial water and rock compressibility are neglected. For such constant HCPV, the differential

process QCcurs always as differential vaporization or differential

condensation. Differential vaporization takes place when the
reservoir temperature is less than the critical temperature of solution (TR < Tc) and during retrograde-gas reservoir depletion, but
only in the region's pressure and temperature where the retrograde
liquid is vaporized.
In differential condensation, the oil reservoir pressure is maintained constant or almost constant-for example, by gas injection.
Differential condensation can also occur just below the dewpoint
in a gas-condensate reservoir.
Above the bubblepoint and the dewpoint curves, the apparent
value of vaporization andlor condensation is zero, but because the
mass of the fluid in a depleted reservoir is changing as a result of
decreasing pressure, the process could be assumed to be differential.
One important statement has to be added: there is no qualitative
difference between the reservoir fluid in either the differential or
the flash process, if pressure and temperature fall into the area
outside the saturation curve (Fig. D-l). A schematic of differential
vaporization of oil in reservoir conditions is given in Fig. D-2 and
described in Table D-l. As Fig. D-2 indicates, six hypothetical cases
are distinguished. Study Fig. D-2 and Table D-l simultaneously
with Fig. D-l and Fig. D-3.
Consider a first sample, where there is a fixed mass of oil at given
temperature, pressure, and HCPV. When the pressure P I drops to
P2, the volume of oil increases but the HCPV does not change,
so the difference in volumes equals the total oil production when
the pressure changes from PA to P2 (Sample 2).
Single phase
gas
B

Single phose
liquid

.
.... ---$---........
...

Cricondenbor

Cr\'fICOI
- :2potnt ,-.........

Afl

+
Bubble
poinl

.tl II I

:3 /'///

Ao$!tr,

70 /.0

/7&'0,
/1,

"'

'....

...

",

/,//

...... .......

.....

1$0

)40

/
I
1/

\30

,10

,-

\20

ricondenlherm

I
J

II

/1/
/ /
I
/
/ I /1
/ I
I
Two phose
/
I
I
Ireg,!,n
I
II
/
/
/
/
I ///1
/
I
/ 1/ / /
Reservoir temperature --

Fig. D.1-Pressure/temperature phase diagram for hydrocar.

bon reservolrfluld mixture.

TABLE D1-SAMPLE DESCRIPTION FOR FIG. 02

Reservoir
Parameters
HCPV, V
Hydrocarbon
mass in PV, m
Pressure, P
Temperature, T
Total reservoir
composition,

V, =V

2
V 2 =V

3
V3=V

m,=m
P,"'PA

m2<m,
P2<P,

zit =ZiA

Yi

Xi

Sample
4

V4 =V

5
Vs=V

6
V6=V

m3<m2
Pb<P2

m4<m3
P4<Pb

mS<m4
PS<P4

m6<mS
P6<PS

zi2 =zit

Zib =zit

Z/4*ZI1

zl5 *ZI1 *Z14

z/6 *ZiI *Z/4 *ZI5

Yit =Y/A =0

Yi2 =YiA =0

Y/3 =Yib *0

YI4 *Yit *Y/3

YI5 *YI1 *Y13 *Yi4

Y/6*YI1 *Y/3*YI4*YI5

Xit

xi2 =X/1 =Zi'

X/3 =Xib *X/1

XI4 *X I1 *X/3

XIS

Zi

Assumed constant reservoir temperature

Gas composition,
Oil composition,

=ZIA

=Zit

Journal of Petroleum Technology, June 1989

*XI1 *X/3 *X14

X/6 *X it

*X/3XI4

*XI5
633

Fig. D-2-Schematlc of differential vaporization In reservoir conditions P, >P2 >Pb >P4 >Pwt

Pi

(producing GOR)

time _____

Fig. D-3-Schematlc of reservoir pressure and GOR vs. time

In reservoir differential vaporization process.

The third sample is considered at the bubblepoint; the oil volume

change betweenp2 andpb resembles that between PI andp2' Beginning at Pb, the first gas bubbles are released. Pressure P4 corresponds to the lowest GOR value (Fig. 0-3) and coincides with
the highest pressure in a two-phase region, in which only one phase
(oil) still flows. Pressure P4 could be called a gas-flow saturation

pressure. Between Pb and P4, compositions xi' Yi' and Zi are

changed. Hydrocarbon mass is decreasing, so it is the differential
process, which is contrary to .Moses' belief that this is a flash
process.
At Point 5, HCPV remains constant as in Steps 1 through 4; the
oil volume has changed and the system is in a two-phase region.
An amount of released gas exceeds the gas-flow saturation pressure
P4; gas begins to run and is partially removed from the HCPV.
This is how a two-phase flow is generated. Sample 6 characterizes
the same process very close to the bottomhole area. The reservoirfluid mass difference in Steps 1 and 6 equals the total production
from an HCPV.
In conclusion, the flash process occurs whenever we are dealing
with a closed system or a steady-state flow-e.g., a two-phase flow
in vertical tubing, in horizontal pipe flow, and in separators. For
any open system, like a reservoir formation, or for an unsteadystate flow, the differential process properly describes the quasiequilibrium conditions.

Reference.
I. Amyx, J.M., Bass, D.M., and Whiting, R.L.: Petroleum Reservoir Engineering, McGraw-Hili Book Co., New York City (1960).
2. Craft, B.C. and Hawkins, M.P.: Applied Petroleum Reservoir Engineering, Prentice-Hall, Englewood Cliffs, NJ (1959).

JPT

(SPE 19214)

Author's Reply to Discussion of

Engineering Applications of Phase Behavior
of Crude Oil and Condensate Systems
P.L. Moses, SPE, Cawley, Gillespie & Assocs.
Glowacki's definition of a differential liberation is "The same kind
of equilibrium, but with its fluid mass decreasing differentially, is
called a differential process (liberation or condensation)." He does
not give a reference for his definition. It is not one that is in common
usage in petroleum engineering. Glowacki states that a reservoir
fluid always undergoes a differential process even above the bubblepoint where no phase change occurs. Differential liberation implies the liberation of gas. Craft and Hawkins 1 state, "In the flash
liberation process all of the gas evolved during a reduction in
pressure remains in contact, and presumably in equilibrium with
the liquid phase from which it is liberated. In the differential process,
on the other hand, the gas evolved during a pressure reduction is
removed from contact with the liquid phase as rapidly as it is
liberated." McCain 2 states, "Differential vaporization is a process
in which gas is removed from contact with the liquid as fast as it
is formed." Katz et al. 3 state, "Differential vaporization differs
from equilibrium flash vaporization in that the vapor is removed
from contact with the liquid as fast as it is formed." The Petroleum
634

Engineering Handbook 4 and Amyx et al. 5 make similar comments

regarding differential separation of gases and liquids.

Glowacki's statements are consistent with his definitions. Unfortunately, his definition is not one in common usage in petroleum
engineering.

Reference.
I. Craft, B.C. and Hawkins, M.P.: Applied Petroleum Reservoir Engineering, Prentice-Hall, Englewood Cliffs, NJ (1959).
2. McCain, W.O. Jr.: The Properties o/Petroleum Fluids, Petroleum Publishing Co., Tulsa (1973).
3. Katz, D.L. et al.: Handbook o/Natural Gas Engineering, McGraw-Hili
Book Co. Inc., New York City (1959).
4. Petroleum Engineering Handbook, H.B. Bradley (ed.), SPE, Richardson,
TX (1987).
5. Amyx, J. W. et al.: Petroleum Reservoir Engineering Physical Properties,
McGraw-Hili Book Co. Inc., New York City (1960).
(SPE 19893)

JPT
Journal of Petroleum Technology, June 1989