Copyright 2001, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the 2001 SPE Annual Technical Conference and
Exhibition held in New Orleans, Louisiana, 30 September3 October 2001.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300
words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.
ABSTRACT
New empirical PVT correlations for Gulf of Mexico (GOM)
oils have been developed as a function of commonly available
field data. Correlations have been developed for:
bubblepoint pressure,
solution gas oil ratio at bubblepoint pressure,
oil formation volume factor at bubblepoint pressure,
undersaturated isothermal oil compressibility,
Oil Viscosity:
dead oil viscosity,
saturated oil viscosity, and
undersaturated oil viscosity.
For the development of correlations, we have covered a wide
range of data. More than one hundred pressure-volume-
temperature (PVT) reports from GOM have been used in the
development of correlations. Two of the published
correlations, Standing , and Petrosky and Farshad correlations
were tested using our GOM data set. Proposed correlations of
this study predicted the PVT properties of GOM oils better
than the correlations published in the literature, even when the
coefficients of the published correlations are tuned. Using the
correlations of this study, we have written a simple program
that can generate PVT datasets for use in reservoir simulation.
Laboratory measured differential liberation data for two
different oil samples were also compared against our
empirically generated differential liberation data.
INTRODUCTION
In reservoir studies, from material balance calculations to
simulation, fluid properties are always required to estimate the
in place volumes, surface volumes, and the transport
parameters that interact with the flow. The variations of PVT
properties during depletion phase are also needed to evaluate
the reservoir performance and to design surface and
subsurface facilities.
Ideally, PVT properties are experimentally measured in
laboratory. When such direct measurements are not available,
PVT correlations from the literature are often used.
Fundamentally, there are two different types of correlations in
the literature. The first group of correlations is developed
using randomly selected datasets. We would like to call such
correlations generic correlations. The second group of
correlations is developed using a certain geographical area or a
certain class/type of oil. Correlations using randomly selected
datasets may not be suitable for certain type of oils, or certain
geographical areas. Even though the authors of the generic
correlations want to cover a wide range of data, such
correlations still work better for certain types of oils.
Specialized correlations represent the properties of a certain
type of oil or geographical area (for which they are developed)
better than the general purpose correlations.
Crudes are complicated hydrocarbon mixtures, and their
detailed composition may not always be available. Even when
there is detailed compositional information available, it is
common practice to group the components into pseudo-
components (lumping). Black oil representation of oils is a
special case of lumping where the reservoir fluid is
represented by a separator gas component and a stock-tank oil
component. Black oil PVT data can often be correlated with
practically measurable quantities such as, oil gravity (API),
gas gravity, solution gas oil ratio (GOR), and reservoir
temperature. Note that representation of oils with only a few
parameters may not always be enough to characterize their
behavior with an acceptable tolerance. For example, a
naphthenic oil may have the same API, GOR, and gas gravity
as its parafinic counterpart. Although they may have totally
different viscosities, most correlations would predict the same
viscosity for those two oils.
We have tested the correlations published in the literature
against the available GOM data. Our work indicates that the
accuracy and the ranges of validity of those published

SPE 71633
PVT Properties and Viscosity Correlations for Gulf of Mexico Oils
Birol Dindoruk and Peter G. Christman, Shell Int. E & P. Inc.
2 DINDORUK AND P. G. CHRISTMAN SPE 71633
correlations are not suitable for our GOM resource base
(especially for the deepwater applications). Therefore, there is
a need for a suite of new PVT correlations for GOM oils.
INPUT DATA
All fluid samples were obtained from reservoirs in Gulf of
Mexico (GOM). We have used about 100 PVT laboratory
reports. Reservoir temperature (T), single stage flash data for
solution gas oil ratio (R
s
), gas gravity (
g
) and oil gravity
(API), and were used for the development of the correlations.
Statistical distributions of the input data are shown in Tables 1
and 2. As can be seen from Table 1, bubblepoint pressure of
the data ranged between 926 psia to 12230 psia.
Corresponding solution GORs ranged from 133 SCF/STB to
3050 SCF/STB. According to McCain
1
(1991), 3050
SCF/STB is close to the limit (3300 SCF/STB) at which the
reservoir fluid is retrograde gas at reservoir conditions.
Similar to P
bp
and GOR, oil formation volume factor (FVF) at
the bubblepoint pressure (B
obp
) varied between 1.08 to 2.9
RB/STB. We have also used separator oil formation volume
factors to develop a relationship between differential liberation
values (B
obpDL), and the separator adjusted values (B
obpSEP) for
separator temperature correction. Range of the separator oil
formation volume factors (1.08-2.74 RB/STB) is similar to the
range of differential liberation oil formation volume factors.
However, separator oil formation volume factors are either
equal to or lower than the differential liberation oil formation
volume factors. As per expected from the ranges of P
bp
and
B
obp
, bubblepoint oil compressibility range was also wide (5-
32 psi
-1
).
Compared to our work, Petrosky and Farshad
2
(1993) used data
covered smaller ranges of P
bp
(1574-6523 psia), R
s
(217-1406
SCF/STB), B
obp
(1.12-1.62 RB/STB), and c
o
(4-25 psi
-1
).
However, viscosity ranges considered for the viscosity
correlations were similar to ours. Some of our PVT reports did
not include viscosity measurements at desired conditions, and
some of our PVT reports exhibited too high dead oil
viscosities (outliers). After discarding the outliers, and the
incomplete datasets, more than 90 PVT reports (see Table 2
for details) were utilized for the development of the viscosity
correlations.
DEVELOPMENT OF THE PVT CORRELATIONS
Presented here are seven correlations developed using the
solver tool built in MS-Excel. A large number of functional
forms and their combinations were examined to obtain the
most appropriate correlations covering the full range of data.
During the development of the correlations, no attempt was
made to change the range of the input data in order to
minimize the average absolute relative error, as did Petrosky
and Farshad
3(
1995), who tried to obtain the best regression fit
while limiting the range of the input data. In their study,
optimum range of the input data is justified based on the
adopted functional representation of the subject PVT property,
and the deviation from the observed values (i.e., average
error). Our work also differs from the work of Vasquez and
Beggs
4
(1980). Vasquez and Beggs used more than one
equation to cover the full range of data where coefficients
change according to the API range of the oil. Contrary to
Vasquez and Beggs, we used a single equation to cover the
entire range of the data.
During each trial, the functional forms were either expanded
or truncated depending on the quality of the correlation, and
its associated error (relative deviation from the
measurements). For each prototype functional form, relative
error was plotted against the input parameters or group of
input parameters. If the error exhibited a trend with respect to
an input parameter, or group of input parameters, more terms
were added to the correlations.
Bubblepoint Pressure Correlation (P
bp
)
Bubblepoint pressure correlations in the literature are
functions of R
s
,
g
, API and T. The following equation is
proposed using our data:
where
Correlation coefficients in Eqs. 1a and 1b are presented in
Table 3. Eqs. 1a and 1b are similar to the correlations
developed by Petrosky and Farshad (1993), and
Standing
5
(1977). However, we have used three more
coefficients than Petrosky and Farshad. This should be
expected since our correlation covers a wider range of data. A
crossplot of measured versus the calculated values of the
bubblepoint pressure is shown in Fig. 1. The statistical quality
of the correlation is quantified in Table 4. The new
bubblepoint pressure correlation predicts the bubblepoint
pressures with an average relative error of 0.27%, and average
absolute relative error of 5. One of the reasons for the
improvement is that Petrosky and Farshad suggest the use of
the data from two stage separator tests. Nevertheless, tuning
their functional relationship does not lead to better results than
the results obtained from the proposed bubblepoint pressure
correlation. The statistical accuracy of the original and tuned
Petrosky and Farshad correlation is compared against other
correlations in Table 4.
We have also evaluated the accuracy of the Standing (1977)
correlation. The Standing correlation was selected, because it
is widely used in the petroleum industry for the quick
calculation of the PVT properties. The improvement in the
Standing correlation (due to tuning) is more than Petrosky and
Farshad correlation, because the Standing correlation is not
(1a) 10
11 8
10
9

,
_

+ a
R
a P
A
a
g
a
s
bp

(1b)
2
2
5
3 1
7
6
4 2

,
_

+
+

a
g
a
s
a a
R
a
API a T a
A

SPE 71633 PVT PROPERTIES AND VISCOSITY CORRELATIONS OF GULF OF MEXICO OIL 3
specifically designed for GOM oils like Petrosky and Farshad
correlation.
Solution Gas-Oil Ratio Correlation (R
sbp
)
Usually, there is no need to develop a new correlation for
solution gas oil ratio. Because, solution gas oil ratio can be
obtained by solving the bubblepoint pressure correlation for
solution gas oil ratio (equivalent of rewriting the P
bp
correlation in R
s
-explicit form). However, Eq. 1a cannot be
solved analytically for solution gas oil ratio. On the other
hand, numerical the solution of Eq. 1a for R
sbp
is trivial, if one
of the numerical root finding methods, like Newton-Raphson,
is employed. Due to the nonlinear dependency of the
variables, and the differences in the measurement errors for
P
bp
and R
s
, solution of Eq.1a for R
s
does not warrant the same
statistical accuracy for R
s
as P
bp
. Furthermore, the objective
function for P
bp
correlation is formed using the observed
values of P
bp
not R
s
. Note that when Eq. 1a is solved for R
s
,
P
bp
is used as an input parameter. A new correlation is
developed for bubblepoint solution gas oil ratio as a function
of P
bp
,
g
, API and T. Following equation is developed for
solution gas oil ratio:
where
Correlation coefficients in Eqs. 2a and 2b are presented in
Table 5. Contrary to Petrosky and Farshads conclusions, best
results were obtained by using the new correlation in Eqs. 2a
and 2b developed independent of the bubblepoint correlation
in Eq. 1a. The accuracy of both approaches for the calculation
of R
s
are shown in Table 6 and Figs. 2 and 3. Statistical
quality of the new solution GOR correlation (details shown in
Table 6) is about 1 percent better than the bubblepoint
pressure correlation. However, if the two outlying points are
excluded from the R
s
values obtained from the bubblepoint
pressure correlation (Fig. 3), R
s
s based on the P
bp
correlation
are more accurate for high P
bp
oils than the values obtained
from Eq. 2a. Eq. 2a predicts the solution GORs with an
average relative error of 1.28% and average absolute relative
error of 7.66%.
When Petrosky and Farshad correlation is used, deviation
from the unit slope line becomes more significant as R
s
(or
P
bp
) increases. The disagreement between the measured
values versus the predicted values gets significantly worse
beyond the range of the data used for the development of
Petrosky and Farshad correlation (maximum R
s
of 1406
SCF/STB). As expected, when the coefficients of Petrosky
and Farshad correlation are tuned, the overall quality of the
correlation improves. However the overall trend (for which the
tuned version of the correlation underestimates the solution
GOR for high solution GOR values) remains the same.
Similar results are obtained when original and tuned versions
of Standing correlation are used. Tuning the coefficients of
Standing correlation still results in the underestimation of R
s
for R
s
> 1500 SCF/STB. The apparent trend of
underestimation at high solution GORs suggests the existence
of a higher order functional relationship between R
s
and the
input variables. In this study, proposed functional form for R
s
remediates the problems of other correlations encountered at
high R
s
(or P
bp
) values.
Oil Formation Volume Factor (B
obp
)
Various functional groups of input variables (R
s
,
g
, API and
T) were formed to generate the best match to correlate the
bubblepoint oil formation volume factor (B
obp
) as a function of
R
s
,
g
, API and T. As shown in Eqs. 3a and 3b, the new B
o
correlation is a quadratic function of the generic variable A. In
Eq. 3a, the last term represents the additional temperature
correction for dead oils.
Where
Coefficients in Eqs. 3a and 3b are given in Table 7. Accuracy
of the new B
o
correlation is shown in Fig. 4. Only three points
exhibit relatively large deviation from the unit slope line.
Statistical parameters related to the accuracy of the correlation
are presented in Table 8. The proposed correlation predicts
the B
o
values with an average relative error of 0.11%, and
average absolute relative error of 2.0%.
When Petrosky and Farshad correlation is used, significant
scattering around the unit slope line is observed for B
o
values
of greater than 2.0 RB/STB. When the coefficients in the
Petrosky and Farshad correlation are tuned to our dataset,
some improvement of the results is observed. Both, the
original and tuned versions of Standing correlation also
exhibited wide scattering for B
o
> 2.0 RB/STB. Therefore, we
have adopted a more complex functional dependency of B
o
to
the input variables, R
s
,
g
, API and T, (Eqs. 3a and 3b).
Undersaturated Isothermal Oil Compressibility (c
obp
)
Most undersaturated isothermal oil compressibility
correlations in the literature are functions of R
s
,
o
,
g
and T.
(2a) 10
11
10
9
8
a
A a bp
sbp
g
a
a
P
R
1
1
]
1

,
_

+
(3a) ) 60 (
14
2
13 12 11
g
obp
API
T a A a A a a B

+ + +
(2b)
2
2
5
3 1
7
6
4 2

,
_

+
+

a
bp
a
a a
P
API
a
T a API a
A
(3b)
) 60 (
2
) 60 (
2
8
6 4
10
9
7
5
3
2 1

,
_

,
_

+ +

T
R
a
R a T a
R
A
a
g
a
s
a
s
a
a
o
a
g
a
s

4 DINDORUK AND P. G. CHRISTMAN SPE 71633

Using the same primary variables, the following correlation
was developed for the isothermal oil compressibility:
where
Coefficients of Eqs 4a and 4b are given in Table 9. A crossplot
of measured versus the calculated values of the oil
compressibility is shown in Fig. 7. Statistical accuracy of the
correlation is presented in Table 10. The proposed oil
compressibility correlation predicts the oil compressibility
values with an average relative error of 0.85% and average
absolute relative error of 6.21%. In the laboratory, oil
compressibility is calculated based on two volumetric
measurements at two distinct pressures. Oil compressibility is
assigned to the pressure interval (P) at which the
measurements are taken at the two bounds of that interval.
For the measurement of c
o
at P
bp
, the lower pressure is always
P
bp
and the upper pressure is P
bp
+ P. Due to the nature of the
undersaturated isothermal oil compressibility measurement,
the calculated c
o
value reflects the effects of averaging over
the pressure interval defined by P.
Using the PVT data set employed in the development of our
correlations, we have tested various c
o
correlations in the
literature. Among the correlations tested, the accuracy of
Petrosky and Farshad, and Vasquez and Beggs correlations are
presented.. Both versions of Petrosky and Farshad correlation
perform significantly poorer than the proposed c
o
correlation.
Similar results are also obtained when Vasquez and Beggs
correlation is used. Although tuning the coefficients of
Vasquez and Beggs correlation significantly improved the
results, the proposed correlation in Eq. 4a still yields better
results. This shows that further improvement of the results
requires a different functional form, as presented in Eqs. 4a
and 4b, for the oil compressibility correlation. Again,
comparison of the statistical accuracy of all the correlations
presented here is summarized in Table 10.
VISCOSITY CORRELATIONS
New viscosity correlations were developed for estimating dead
oil, saturated oil, and undersaturated oil viscosities. Similar to
other correlations in the literature, dead oil viscosity is used as
input for saturated oil viscosity correlation, and the saturated
oil viscosity is used as input for and undersaturated oil
viscosity correlation. Our results are compared against the
correlations developed by Petrosky and Farshad (1995), and
Standing (1977).
Dead Oil Viscosity Correlation (
oD
)
Dead oil viscosity correlations of the literature are functions of
API and T. The following equation is developed for
calculating dead oil viscosity:
where
Coefficients of Eqs. 5a and 5b are given in Table 11. As
shown in Eqs. 5a and 5b, the proposed dead oil viscosity
correlation is a function of two additional parameters linking

oD
to the P
bp
and R
sbp
of the original oil. This approach is
taken to capture some aspect of the oil type. Because, the same
amount of solution gas will cause different level of bubble
point pressures for paraffinic and aromatic oils. In this way we
capture some of the information about the oil type withoust
requiring additional data. Accuracy of the new dead oil
viscosity correlation is shown in Fig. 6. The graph on the
right is merely the modified scale version of the full scale one
on the left. Statistical parameters related to the accuracy of this
correlation are presented in Table 12. The proposed
correlation predicts the
oD
values with an average relative
error of -2.86% and average absolute relative error of 12.62%.
Considering the range of the data and the nature of dead oil
viscosity, the proposed correlation is better than any other
dead oil viscosity correlations tested. Dead oil viscosity is one
of the most unreliable properties to predict with correlations.
This results mostly from the large effect that oil type has on
viscosity. Two dead oil samples with identical API and T can
have orders of magnitude different viscosities. In this work,
we tried to limit this non-uniqueness to a certain degree by
incorporating the P
bp
and R
sbp
of the original oil.
Significant scattering around the unit slope line was observed
for
oD
values greater than 5 cp when Petrosky and Farshad
correlation is used. When the coefficients of Petrosky and
Farshad correlation are tuned to our dataset, some
improvement of the results is observed. Performance of
Standing correlation was significantly worse than the outcome
of the Petrosky and Farshad correlation. However, tuning the
coefficients of Standing correlation improved the calculated
viscosities even better than the values calculated from the
tuned version of Petrosky and Farshad correlation. The
statistical accuracy of the original and the tuned versions of
Petrosky and Farshad, and Standing correlations is presented
in Table 12.
Saturated Oil Viscosity Correlation (
obp
)
In the literature, saturated oil viscosity correlations require the
use of dead oil viscosity. The original approach by Chew and
Connally
6 (
1959) for correlating saturated oil viscosity in terms
of dead oil viscosity and solution gas oil ratio is used here.
( ) (4a) 10
6 2
13 12 11

+ + A a A a a c
obp
(4b)
) 60 (
2
) 60 (
2
8
6 4
10
9
7
5
3
2 1

,
_

,
_

+ +

T
R
a
R a T a
R
A
a
g
a
s
a
s
a
a
o
a
g
a
s

( )
(5a)
log
8 6
4
7 5
3
a
sbp
a
bp
A a
oD
R a P a
API T a
+

(5b) log
2 1
a T a A +
SPE 71633 PVT PROPERTIES AND VISCOSITY CORRELATIONS OF GULF OF MEXICO OIL 5
The following equation is proposed using the data available to
us:
where
and
Coefficients in Eqs. 6a-6c are given in Table 13. Crossplot of
measured versus the calculated values of the saturated oil
viscosity is shown in Fig. 7. Statistical quality of the
correlation is presented in Table 14. As shown in Table 14, the
proposed correlation exhibits average relative error value of
3.05%, and average absolute relative error value of 13.2%.
Other correlations in the literature performed much worse than
the proposed correlation. Although the tuned versions of these
correlations looked much better than their original versions,
the predicted results were still not as good as expected. The
statistical accuracy of these correlations is compared in Table
14.
Undersaturated Oil Viscosity Correlation (
o
)
We have developed a new correlation for undersaturated oil
viscosity. The best regression analysis results were obtained
by using the following equation:
where
Coefficients of Eqs. 7a and 7b are given in Table 15. Eq. 7a
has the same functional form as in Petrosky and Farshad
correlation. However, the exponent A has a different
functional form. Accuracy of the proposed correlation is
shown in Fig. 80. Statistical accuracy of this correlation is
presented in Table 16. Using Eqs. 7a-7b, undersaturated oil
viscosities are predicted with an average relative error of
0.83% and average absolute relative error of 5.99%.
Accuracy of Petrosky and Farshad, and Standing correlations
was tested using the dataset employed in the development of
the new correlations.. Statistical parameters related to the
accuracy of the correlations presented are summarized in
Table 14. As shown in Table 14, tuned version of the
Standing correlation performs as good as the Petrosky and
Farshad correlation.
USE OF THE PROPOSED CORRELATIONS FOR THE
GENERATION OF PVT DECKS FOR RESERVOIR
SIMULATION
Differential liberation data were generated and compared
against the experimental results. Good agreement was found
between experimental values and calculated values. Here we
present two example cases.
In all the cases presented below, experimental data for the oil
viscosity, solution GOR, and oil formation volume factor are
compared against the proposed correlations. The bubblepoint
pressure is estimated using solution GOR.
Oil A
We generated differential liberation experiment for Oil A.
Input data used in the correlations are shown in Table 17.
Comparison against the experimental data is shown in Figs. 9-
11. Deviations from the experimental results were 0.2% for
bubblepoint pressure, -0.8% for bubblepoint oil formation
volume factor, 3.13% for bubblepoint oil viscosity, and 9.9%
bubblepoint oil compressibility.
For this oil, bubblepoint pressure is predicted almost exactly.
Therefore, the bubblepoint oil FVF is predicted better than the
bubblepoint oil FVF of oil in Oil B. Also, the rest of the oil
FVF curve is close to the experimental data (Fig. 9).
Oil B
We generated differential liberation experiment for Oil B.
Input data used in the correlations are shown in Table 18.
Comparison against the experimental data is shown in Figs.12-
14. Deviations from the experimental results were 5.8% for
bubblepoint pressure, -3.3% for Bubblepoint oil formation
volume factor, 0.47% for bubblepoint oil viscosity, and 1.2%
bubblepoint oil compressibility.
Underestimation of bubblepoint oil formation volume factor
(FVF) reflects the underestimation of bubblepoint pressure
(Fig. 12). However, the slope of the undersaturated oil FVF is
very close to the slope of the experimental data. Since the
bubblepoint solution GOR is used as an input parameter for
the correlations, the quality of the agreement against the
experimental data is better than the agreement obtained for oil
FVF.
CONCLUSIONS
New empirical PVT correlations for GOM oils have been
developed.
Proposed correlations are used to generate differential
liberation tables for reservoir simulation.
All of the proposed correlations have a wide range of
validity, and are superior to other published correlations
in the literature.
The proposed correlations can be tuned for other
basins/areas, or certain class of oils.
[ ] (6a)
B
oD obp
A
(6b)
) exp( ) exp(
5
3
2
1
4
s
a
s
s
R a
R a
R a
a
A +
(6c)
) exp( ) exp(
10
8
7
6
9
s
a
s
s
R a
R a
R a
a
B +
(7a) 10 ) (
6
A
bp obp o
P P a +
(7b) ) ( log log log
5 4 3 2 1 bp s obp s obp
P P a R a R a a a A + + + +
6 DINDORUK AND P. G. CHRISTMAN SPE 71633
NOMENCLATURE
API = API gravity at 60
o
F
B
o
= oil formation volume factor (RB/STB)
B
obp
= bubblepoint oil formation volume factor
(RB/STB)
c
o
= undersaturated oil compressibility (1/psi)
c
obp
= bubblepoint oil compressibility (1/psi)
P = pressure (psi)
P
bp
= bubblepoint pressure (psi)
P = pressure interval (psi)
R
s
= solution GOR (SCF/STB)
R
sbp
= bubblepoint solution GOR (SCF/STB)
T = reservoir temperature (
o
F)
T
sep
= separator temperature (
o
F)

g
= specific gravity of gas (
air
=1)

o
= specific gravity of oil (gm/cc,
water
=1 gm/cc)

o
= undersaturated oil viscosity (cp)

oD
= dead oil viscosity (cp)

obp
= bubblepoint oil viscosity (cp)
REFERENCES
1. McCain, W.D., Jr.: Reservoir-Fluid Property
Correlations State of the Art, SPE Reservoir
Engineering (May 1991) 266-272.
2. Petrosky, G.E., Jr., and Farshad, F.F.: Pressure-Volume-
Temperature Correlations for Gulf of Mexico Crude
Oils, paper SPE 26644 presented at the 68
th
Annual
Technical Conference and Exhibition of the SPE,
Houston, TX. (October 1993).
3. Petrosky, G.E., Jr., and Farshad, F.F.: Viscosity
Correlations for Gulf of Mexico Crude Oils, paper SPE
29468 presented at the Production Operations Symposium
of the SPE, Oklahoma City, OK. (April 1995).
4. Vasquez, M.E., and Beggs, H.D.: Correlations for Fluid
Physical Property Prediction, Journal of Petroleum
Technology (June 1980) 968-970.
5. Standing, M.B.: Volumetric and Phase Behavior of Oil
Filed Hydrocarbon Systems, SPE (1977).
6. Chew, J.N., and Connally, C.A.: A Viscosity
Correlation for Gas-Saturated Crude Oils, Trans., AIME
(1959) 216, 23-25.
SPE 71633 PVT PROPERTIES AND VISCOSITY CORRELATIONS OF GULF OF MEXICO OIL 7
Tables
Table 1: Data used for P
bp
, R
sbp
, B
obp
, and c
obp
correlations.
Quantity # of
Data
Points
Maximum Minimum Mean Standard
Deviation
T (
o
F) 104 276 117 174.2 32.7
P
bp
(psi) 104 12230 926 4616 2351
R
sbp
(SCF/STB) 104 3050 133 1023 653
o
API 104 40.00 14.70 30.95 4.87

g
(air=1) 104 1.0270 0.6017 0.7552 0.0838
c
obp
(psi
-1
) 99 31.91 5.02 12.03 4.62
B
obpDL (RB/STB) 99 2.8984 1.0844 1.5400 0.3898
B
obpSEP (RB/STB) 99 2.7381 1.0845 1.5219 0.3246
Table 2: Data used for
oD,

obp,

o
correlations.
Quantity # of
Data
Points
Maximum Minimum Mean Standard
Deviation
T (
o
F) 95 276 121 171.2 30.21
P
bp
(psi) 95 12230 926 4711 2406
R
sbp
(SCF/STB) 95 3050 133 1054 670
o
API 95 40.0 17.4 30.78 4.64

g
(air=1) 95 1.027 0.6017 0.7536 0.0815
c
obp
(psi
-1
) 95 31.91 5.02 12.06 4.69
P-P
bp
(psi) 93 10140 202 3080 2801

oD
(cp) 95 62.63 0.896 5.2546 7.6444

obp
(cp) 95 8.7 0.1610 1.1979 1.5585

o
(cp) 93 10.600 0.2110 1.4869 1.8410
Table 3: Coefficients for the proposed P
bp
correlation.
Coefficient (Pbp
correlation)
Value
a
1
1.42828E-10
a
2
2.844591797
a
3
-6.74896E-04
a
4
1.225226436
a
5
0.033383304
a
6
-0.272945957
a
7
-0.084226069
a
8
1.869979257
a
9
1.221486524
a
10
1.370508349
a
11
0.011688308
8 PVT PROPERTIES AND VISCOSITY CORRELATIONS OF GULF OF MEXICO OIL SPE 71633
Table 4: Statistical accuracy of bubblepoint pressure correlations.
Quantity This Study Petrosky
& Farshad
Tuned
Petrosky
& Farshad
Standing Tuned
Standing
Average Relative
Error (%)
-0.27 -1.10 -0.95 -10.36 -1.05
Standard Deviation
(%)
7.51 13.45 9.91 14.08 10.24
Average Absolute
Relative Error (%)
5.70 10.30 8.10 15.23 8.41
Standard Deviation
(%)
4.86 8.63 5.73 8.52 5.89
Table 5: Coefficients for the proposed Rsbp correlation.
Coefficient (Rsbp
correlation)
Value
a1 4.86996E-06
a2 5.730982539
a3 9.92510E-03
a4 1.776179364
a5 44.25002680
a6 2.702889206
a7 0.744335673
a8 3.359754970
a9 28.10133245
a10 1.579050160
a11 0.928131344
Table 6: Statistical accuracy of R
sbp
correlations.
Quantity This
Study
This Study (P
bp
correlation)
Petrosky
& Farshad
Tuned
Petrosky
& Farshad
Standing Tuned
Standing
Average Relative
Error (%)
-1.28 1.40 2.72 -1.75 17.39 -1.77
Standard Deviation
(%)
9.89 11.49 16.61 13.14 19.12 13.25
Average Absolute
Relative Error (%)
7.66 8.20 13.07 10.55 22.05 10.66
Standard Deviation
(%)
6.33 8.14 10.53 7.96 13.41 8.00
Table 7: Coefficients for the proposed B
obp
correlation.
Coefficient (B
obp
correlation)
Value
a
1
2.510755E+00
a
2
-4.852538E+00
a
3
1.183500E+01
a
4
1.365428E+05
a
5
2.252880E+00
a
6
1.007190E+01
a
7
4.450849E-01
a
8
5.352624E+00
a
9
-6.309052E-01
a
10
9.000749E-01
a
11
9.871766E-01
a
12
7.865146E-04
a
13
2.689173E-06
a
14
1.100001E-05
SPE 71633 PVT PROPERTIES AND VISCOSITY CORRELATIONS OF GULF OF MEXICO OIL 9
Table 8: Statistical accuracy of B
obp
correlations.
Quantity This Study Petrosky
& Farshad
Tuned
Petrosky
& Farshad
Standing Tuned
Standing
Average Relative
Error (%)
-0.11 -0.60 -0.18 2.96 -0.17
Standard Deviation
(%)
3.17 5.19 3.93 4.64 4.11
Average Absolute
Relative Error (%)
2.00 3.14 2.42 4.18 2.32
Standard Deviation
(%)
2.44 4.17 3.09 3.57 3.39
Table 9: Coefficients of the proposed c
obp
correlation.
Coefficient (c
obp
correlation)
Value
a
1
0.980922372
a
2
0.021003077
a
3
0.338486128
a
4
20.00006358
a
5
0.300001059
a
6
-0.876813622
a
7
1.759732076
a
8
2.749114986
a
9
-1.713572145
a
10
9.999932841
a
11
4.487462368
a
12
0.005197040
a
13
0.000012580
Table 10: Statistical accuracy of the undersaturated isothermal oil compressibility correlations.
Quantity This Study Petrosky &
Farshad
Tuned
Petrosky &
Farshad
Vasquez
& Beggs
Tuned Vasquez &
Beggs
Average Relative
Error (%)
-0.85 9.81 -2.57 16.17 -1.93
Standard Deviation
(%)
8.95 14.19 12.59 28.62 13.84
Average Absolute
Relative Error (%)
6.21 13.83 9.41 23.58 10.56
Standard Deviation
(%)
6.47 10.27 8.71 22.84 9.09
Table 11: Coefficients for the proposed dead oil viscosity correlation.
Coefficient (
oD
correlation)
Value
a
1
14.505357625
a
2
-44.868655416
a
3
9.36579E+09
a
4
-4.194017808
a
5
-3.1461171E-9
a
6
1.517652716
a
7
0.010433654
a
8
-0.000776880
10 PVT PROPERTIES AND VISCOSITY CORRELATIONS OF GULF OF MEXICO OIL SPE 71633
Table 12: Statistical accuracy of dead oil viscosity correlations.
Quantity This Study Petrosky
& Farshad
Tuned
Petrosky &
Farshad
Standing Tuned
Standing
Average Relative
Error (%)
-2.86 -1.80 -3.73 -2.60 -3.51
Standard Deviation
(%)
16.74 20.14 19.06 25.53 18.50
Average Absolute
Relative Error (%)
12.62 15.49 15.14 20.96 14.88
Standard Deviation
(%)
11.30 12.90 12.07 14.65 11.43
Table 13: Coefficients of the proposed saturated oil viscosity correlation
Coefficient (
obp
correlation)
Value
a
1
1.000000E+00
a
2
4.740729E-04
a
3
-1.023451E-02
a
4
6.600358E-01
a
5
1.075080E-03
a
6
1.000000E+00
a
7
-2.191172E-05
a
8
-1.660981E-02
a
9
4.233179E-01
a
10
-2.273945E-04
Table 14: Statistical accuracy of saturated oil viscosity correlations.
Quantity This Study Petrosky &
Farshad
Tuned
Petrosky &
Farshad
Standing Tuned
Standing
Average Relative
Error (%)
-3.05 0.26 -4.36 15.10 -6.63
Standard Deviation
(%)
17.29 25.97 20.52 52.28 31.82
Average Absolute
Relative Error (%)
13.20 17.60 15.90 28.20 20.40
Standard Deviation
(%)
11.51 19.05 13.57 46.48 25.21
Table 15: Coefficients for the proposed undersaturated oil viscosity correlation.
Coefficient (
o
correlation)
Value
a
1
0.776644115
a
2
0.987658646
a
3
-0.190564677
a
4
0.009147711
a
5
-0.000019111
a
6
0.000063340
SPE 71633 PVT PROPERTIES AND VISCOSITY CORRELATIONS OF GULF OF MEXICO OIL 11
Table 16: Statistical accuracy of the undersaturated oil viscosity correlations.
Quantity This Study Petrosky &
Farshad
Tuned
Petrosky &
Farshad
Standing Tuned
Standing
Average Relative
Error (%)
-0.83 3.51 -1.97 -4.26 -2.00
Standard Deviation
(%)
8.42 12.89 8.58 9.43 8.65
Average Absolute
Relative Error (%)
5.99 8.90 6.00 6.88 6.01
Standard Deviation
(%)
6.08 10.06 6.53 7.72 6.51
Table 17: Input data used for generating differential liberation
table for Oil A.
T (
o
F) 160
T
sep
(
o
F) 68
R
sbp
(SCF/STB) 813
API 27.4

g
(
air
= 1) 0.7310
Table 18: Input data used for generating differential liberation
table for Oil B.
T (
o
F) 230
T
sep
(
o
F) 60
R
sbp
(SCF/STB) 1049
API 33.3

g
(
air
= 1) 0.7072
FIGURES
Figure 1: Crossplot of measured bubblepoint pressures versus Figure 2: Crossplot of measured Rsbp versus calculated Rsbp from Eq. 2 .
calculated bubblepoint pressures (this study).
0
2000
4000
6000
8000
10000
12000
14000
0 2000 4000 6000 8000 10000 12000 14000
Measured Bubblepoint Pressure (psia)
C
a
l
c
u
l
a
t
e
d

B
u
b
b
l
e
p
o
i
n
t

P
r
e
s
s
u
r
e

(
P
s
i
a
)
Unit Slope Line
0
500
1000
1500
2000
2500
3000
3500
4000
0 500 1000 1500 2000 2500 3000 3500 4000
Measured Solution Gas Oil Ratio (SCF/STB)
C
a
l
c
u
l
a
t
e
d

S
o
l
u
t
i
o
n

G
a
s

O
i
l

R
a
t
i
o

(
S
C
F
/
S
T
B
)
Unit Slope Line
12 B. DINDORUK AND P. G. CHRISTMAN SPE 71633
Figure 3: Crossplot of measured Rsbp versus calculated Rsbp from Eq. 1 (this study).
Figure 4: Crossplot of measured Bobp versus calculated Bobp (this study).
Figure 5: Crossplot of measured cobp versus calculated cobp (this study).
.
0
500
1000
1500
2000
2500
3000
3500
4000
0 500 1000 1500 2000 2500 3000 3500 4000
Measured Solution Gas Oil Ratio (SCF/STB)
C
a
l
c
u
l
a
t
e
d

S
o
l
u
t
i
o
n

G
a
s

O
i
l

R
a
t
i
o

(
S
C
F
/
S
T
B
)
Unit Slope Line
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3
Measured Oil Formation Volume Factor (BBL/STB)
C
a
l
c
u
l
a
t
e
d

O
i
l

F
o
r
m
a
t
i
o
n

V
o
l
u
m
e

F
a
c
t
o
r

(
B
B
L
/
S
T
B
)
Unit Slope Line
0
5
10
15
20
25
30
0 5 10 15 20 25 30
Measured Undersaturated Oil Compressibility (1/psi x 1E6)
C
a
l
c
u
l
a
t
e
d

U
n
d
e
r
s
a
t
u
r
a
t
e
d

O
i
l

C
o
m
p
r
e
s
s
i
b
i
l
i
t
y

(
1
/
p
s
i

x

1
E
6
)
Unit Slope Line
SPE 71633 PVT PROPERTIES AND VISCOSITY CORRELATIONS FOR GULF OF MEXICO OILS 13
Figure 6: Crossplot of oD versus calculated oD (this study).
Figure 7: Crossplot of measured saturated oil viscosity versus calculated saturated oil viscosity (this study).
Figure 8: Crossplot of measured undersaturated oil viscosity versus calculated undersaturated oil viscosity (this study).
0
5
10
15
20
25
30
35
0 5 10 15 20 25 30 35
Measured Dead Oil Viscosity (cp)
C
a
l
c
u
l
a
t
e
d

D
e
a
d

O
i
l

V
i
s
c
o
s
i
t
y

(
c
p
)
Unit Slope Line
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Measured Dead Oil Viscosity (cp)
C
a
l
c
u
l
a
t
e
d

D
e
a
d

O
i
l

V
i
s
c
o
s
i
t
y

(
c
p
)
Unit Slope Line
0
1
2
3
0 1 2 3
Measured Oil Viscosity at Bubblepoint Pressure (cp)
C
a
lc
u
la
t
e
d
O
il V
is
c
o
s
it
y
a
t
B
u
b
b
le
p
o
in
t
P
r
e
s
s
u
r
e
(
c
p
)Unit Slope Line
0
2
4
6
8
10
12
14
16
18
20
0 2 4 6 8 10 12 14 16 18 20
Measured Undersaturated Oil Viscosity (cp)
C
a
lc
u
la
t
e
d
U
n
d
e
r
s
a
t
u
r
a
t
e
d
O
il V
is
c
o
s
it
y
(
c
p
)
Unit Slope Line
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0 0.5 1 1.5 2 2.5 3
Measured Undersaturated Oil Viscosity (cp)
C
a
lc
u
la
t
e
d
U
n
d
e
r
s
a
t
u
r
a
t
e
d
O
il V
is
c
o
s
it
y
(
c
p
)
Unit Slope Line
0
2
4
6
8
10
12
14
16
18
20
0 2 4 6 8 10 12 14 16 18 20
Measured Oil Viscos ity at Bubblepoint Press ure( cp)
C
a
lc
u
la
t
e
d

O
il
V
is
c
o
s
i
t
y
a
t
B
u
b
b
le
p
o
in
t
P
r
e
s
s
u
r
e
(
c
p
)
Unit Slope L ine
14 B. DINDORUK AND P. G. CHRISTMAN SPE 71633
Figure 9: Comparison of oil formation volume Figure 10: Comparison of solution GOR data
Factor against the proposed empirical correlation (Oil A). against the proposed empirical correlation (Oil A)
Figure 11: Comparison of oil viscosity data Figure 12: Comparison of oil formation volume factor data
against the proposed empirical correlation (Oil A) against the proposed empirical correlation (Oil B).
Figure 13: Comparison of solution GOR data Figure 14: Comparison of oil viscosity data
against the proposed empirical correlation (Oil B). against the proposed empirical correlation (Oil B)
0
100
200
300
400
500
600
700
800
900
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Pressure ( ps i)
R
s
(S
C
F
/S
T
B
Calculated (Cor relatio n)
Data
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Pressure ( ps i)
O
il
V
is
c
o
s
i
t
y
(
c
p
)
Calculated (Cor relatio n)
Data
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
0 1000 2000 3000 4000 5000 6000 7000 8000
Pressure ( ps i)
O
il
V
i
s
c
o
s
i
t
y
(
c
p
)
Calculated (Cor relatio n)
Data
0
200
400
600
800
1000
1200
0 1000 2000 3000 4000 5000 6000 7000 8000
Pressure ( ps i)
R
s
(
S
C
F
/
S
T
B
)
Calculated (Cor relatio n)
Data
1.00
1.05
1.10
1.15
1.20
1.25
1.30
1.35
1.40
1.45
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Pressure (psi)
B
o

(
B
B
L
/
S
T
B
)
Calculated
Data
1.0
1.1
1.2
1.3
1.4
1.5
1.6
0 1000 2000 3000 4000 5000 6000 7000 8000
Pressure (psi)
B
o

(
B
B
L
/
S
T
B
)
Calculated
Data