Correlations for Fluid
Physical Property Prediction
Introduction
The evaluation of physical properties of hydrocarbon
mixtures is an important step in the design of various
stages of oilfield operations. The fluid properties,
which change as pressure and temperature change,
must be evaluated for both reservoir engineering and
production design operations. For example, the
calculation of two-phase flowing pressure gradients
occurring in piping systems requires prediction of
fluid properties such as dissolved gas, formation
volume factor (FVF) , oil compressibility, and
viscosity at various locations in the pipe. Even
though laboratory measurements of these properties
may be available as a function of pressure, the
0149-213610006-6719$00.25
Copyright 1980 Society of Petroleum Engineers
968
measurements usually are made at reservoir tem-
perature only. The behavior of these properties with
temperature change usually is predicted empirically.
Increasing emphasis is being placed on the
necessity of having accurate estimates of fluid
physical properties in advance of laboratory PVT
analyses of the reservoir fluids. Many times these
properties are required at a time when the only in-
formation available consists of oil and gas gravities
and reservoir pressure and temperature. The purpose
of this study was to utilize a large base of laboratory-
measured PVT data to develop improved empirical
correlations to replace those commonly in use. Many
of the presently used correlations were developed
many years ago from limited data and are being used
beyond the range for which they were intended. The
JOURNAL OF PETROLEUM TECHNOLOGY
 present study was accomplished by first gathering
more than 600 laboratory PVT analyses from fields
all over the world. The data encompassed very wide
ranges of pressure, temperature, and oil properties
and included more than 6,000 measurements of Rs,
Bo and, IJ.0 at various pressures. Regression analysis
techniques then were used to correlate the laboratory
data. Empirical correlations were developed for the
following fluid properties as functions of pressure,
temperature, oil gravity, and gas gravity: (1) solution
gas/oil ratio (GOR), (2) oil FVF below Pb' (3) oil
FVF abovePb, and (4) oil viscosity abovePb'
To use the best approach for this study, a thorough
review of presently available methods was con-
ducted. A brief description of some of the previous
studies follows. A more detailed literature review
may be found in Ref. 1.
The best known method for calculating dissolved
gas and FVF is that of Standing. 2 In 1947 he
presented the results of a laboratory study of the
PVT behavior of 22 different crude-oil/natural-gas
mixtures from California oil fields. He developed
correlations for bubble-point pressures and oil FVF's
as empirical functions of solution GOR, gas gravity,
oil gravity, pressure, and temperature derived from
105 experimental data points. The correlation for
dissolved gas has been found to be inaccurate for
high-gravity crude oils, but it is still one of the most
widely used methods. His correlation for oil FVF,
although developed from limited data, is the method
most widely used in the petroleum industry today.
In 1958 Lasater, 3 presented a correlation of the
bubble-point pressure for black oil systems
developed from standard physical-chemical
equations of solutions. The correlation was based on
158 experimentally measured bubble-point pressures
of 137 independent systems. Lasater's correlation has
been found to be more accurate than Standing's
correlation for high-gravity crude oil systems, and
many engineers prefer using Standing's correlation
for crudes of API gravity less than 15 and the Lasater
correlation for lighter crudes.
Extensive research has been done on the deter-
mination of crude oil viscosities below the bubble-
point. In one of the early works found in the
literature, Beal 4 in 1945 presented correlations for
dead oil viscosity. He also developed a correlation
for viscosity of undersaturated crude oils.
The Beal correlation for gas-free or dead oil
viscosity is used widely throughout the oil industry
and is considered to be fairly accurate. One of the
limitations found in his presentation is the lack of
analytical expressions to evaluate viscosities of crude
oil systems below and above the bubble point.
In 1959 Chew and Connally5 developed a
correlation to cover a wider range of dissolved
GOR's and dead oil viscosities. The correlation was
presented in the form of an equation and in a
graphical form.
Little and Kennedy6 in 1966 presented a
correlation pertaining to the prediction of viscosity
of complex hydrocarbon systems from pressure
temperature, and composition. Although the
correlation was found reliable for both liquid and
JUNE 1980
gases, it requires compositional analysis of mixtures.
Beggs and Robinson 7 in 1975 presented a
correlation to estimate the viscosity of crude oil
systems based on PVT analyses of reservoir fluid
samples. They developed a correlation for the dead
oil viscosity as a function of temperature and oil
specific gravity. An expression also was given to
calculate saturated oil viscosity.
The literature review revealed that the most widely
used correlation for dead oil viscosity was developed
by Beal and for live oil viscosity by Chew and
Connally. Correlations developed by Standing and
by Lasater are the most commonly used for
predicting dissolved gas and oil FVF's.
Limiting factors commonly found throughout the
research are the lack of mathematical expressions for
ease of handling of computations and the use of data
from specific geographical regions in the develop-
ment of correlations. Hence, a bias factor is induced
in results obtained when extrapolating the correlation
for use on crudes of different characteristics.
Correlation Development
It was found that all of the desired fluid properties
could be correlated as functions of pressure, tem-
perature, oil gravity, and gas gravity. The gas gravity
is a strong correlating parameter and, unfortunately,
usually is one of the variables of most questionable
accuracy, because it depends on conditions at which
the gas/oil separation is made. For this reason, it was
decided to use a value of gas gravity obtained from
particular conditions of separator pressure in all of
the correlations that require gas gravity. The value of
100 psig was chosen as a reference pressure because it
was found to be the pressure resulting in minimum
oil shrinkage for the separator tests available. Using
a pressure of 0 psig also was considered, but it was
felt that the value of 100 psig more closely represents
field conditions and will result in less error if the
actual separator pressure from which a gas sample is
obtained is unknown.
Therefore, to use all of the other correlations
developed in this study, the gas gravity first must be
corrected to the value that would have resulted from
a separation at 100 psig. Regression analysis resulted
in the following equation for correcting gas gravity
for separator conditions.
'Ygs = 'Y gp [l. + 5.912 X 10 - 5("( 0)
. (T)log(P/114.7), ............... (1)
where
'Y gs gas gravity (air = 1) that would result from
separator conditions of 100 psig,
'Y gp = gas gravity obtained at separator conditions
ofpand T,
P = actual separator pressure, psia,
T = actual separator temperature, OF, and
'Yo = oil gravity, API.
Dissolved Gas
An improved correlation was obtained by dividing
the measured data into two groups, based on oil
%9
gravity. This was done both for the dissolved gas and where
oil FVF correlations. The division was made at a
value of oil gravity of 30 API. The following m=C I P C 2 exp(C 3 +C4p), ............... (7)
equation was obtained by regression analysis using
5,008 measured data points. and C I = 2.6,
C 2 = 1.187,
C
Rs = C I 'YgsP 2 exp [C 3 (Yo/(T + 460) J...... (2) C 3 = -11.513,and
C4 = - 8.98 x 10 - 5 .
Values for the coefficients are as follows.
Coefficient 'Yo 5,30 'Yo >30 Conclusions
CI 0.0362 0.0178 Improved prediction methods have been developed
C2 1.0937 l.l870 for some of the most commonly required crude oil
C3 25.7240 23.9310 PVT properties. Use of these prediction methods
requires no graphs or charts, and the methods are
Oil FVF (P5,Pb) convenient to use with computers or calculators. A
The oil formation volume factor for saturated oil was much larger data base was used in this study than was
correlated as a function of dissolved gas, tem- used in previous studies; therefore, the results should
perature, oil gravity, and gas gravity. The following be applicable to a wider range of oil properties.
equation was found to be the best form to reproduce Nomenclature
the measured data.
Bo = 1 +CIR s + C 2(T -60) <'Yoi'Y gs )
Bo = oil FVF, bbl/STB (m 3 Istock-tank m 3 )
Bob oil FVF atPb, bbl/STB (m 3 Istock-tank m 3 )
+C3Rs(T-60)('Yoi'Ygs)' ............. (3) Co oil compressibility, vol/vol-psi (vol/vol kPa)
The values for the coefficients depend on oil gravity P pressure, psia (kPa)
and are given by the following. Pb bubble-point pressure, psia (kPa)
Rs dissolved GOR, scflSTB (std m 3 Istock
Coefficient "105,30 "10 >30 tank m 3 )
CI 4.677 x 10- 4 4.670 x 10- 4 T = temperature, OF CC)
C2 1.751 x 10- 5 1.100 x 10- 5 "I g gas gravity (air = 1)
C3 -1.811xlO- 8 1.337 x 10- 9 "10 oil gravity, API (gl cm 3 )
~o oil viscosity, cp (Pa s)
Oil FVF (P?:Pb) ~ob oil viscosity at P b, cp (Pa s)
The change in volume of an undersaturated liquid
depends on the isothermal compressibility of the
References
liquid. The FVF change as pressure is increased I. Vasquez A., M.E.: "Correlations for Fluid Physical Property
Prediction," MS thesis, U. of Tulsa, Tulsa, OK (1976).
above bubble-point pressure is calculated from 2. Standing, M.B.: "A Pressure-Volume-Temperature
Correlation for Mixtures of California Oil and Gases," Drill.
Bo =Bob exp!co(p-Pb)j ................ (4)
and Prod. Prac., API (1947) 275-286.
A correlation for oil compressibility, co' was 3. Lasater, J.A.: "Bubble Point Pressure Correlation," Trans.,
Al ME (1958) 213,379-81.
developed as a function of R s ' T, "10' "18' and p. A 4. Beal, c.: "The Viscosity of Air, Natural Gas, Crude Oil and Its
total of 4,036 data points was used In a linear Associated Gases at Oil Field Temperature and Pressures,
regression model to obtain Trans., AIME (1946) 165, 94-112.
5. Chew, J. and Connally, C.A. Jr.: "A Viscosity Correlation for
Co =(al +a2 R s +a3 T+a4"1gs +a5'Yo)/a6P, .. (5) Gas Saturated Crude Oils," Trans., AIME (1959) 216,23-25.
6. Little, J.E. and Kennedy, H.T.: "A Correlation of the
where Viscosity of Hydrocarbon Systems with Pressure, Tem-
perature, and Composition," Soc. Pel. Eng. 1. (June 1968)
157-162; Trans., AI ME, 243.
al -1433.0, 7. Beggs, H.D. and Robinson, J.F.: "Estimating the Viscosity of
a2 5.0, Crude Oil Systems," 1. Pel. Tech. (Sept. 1975) 1140-1141.
a3 17.2,
a4 - 1180.0, SI Metric Conversion Factors
a5 12.61, and degree API 141.5/(131.5 + API) g/cm 3
a6 10 5 . degree F CF - 32)11.8 C
Undersaturated Oil Viscosity (P?:Pb)
psi x 6.894 757 E + 00 = kPa
As pressure is increased on an undersaturated oil, the
Milton Vazquez, SPE
viscosity of the oil increases. The viscosity at
INTEVEP S.A.
pressures above the bubble-point pressure is ex-
H. Dale Beggs, SPE
pressed as a function of viscosity at bubble-point
U.ofTulsa
pressure ~ob' pressure p, and P b' The viscosity of the 'Now with Shell Oil Co.
oil at P b may be obtained from the Beggs and Original manuscript received in SOCiety of Petroleum Engineers office July 5,
Robinson 7 correlation. The equation is 1977. Paper accepted for publication April 21, 1978. Revised manuscript
received March 17, 1980. Paper (SPE 6719) first presented at the SPE 52nd
Annual Fall Technical Conference and Exhibition, held in Denver, Oct. 912,
~o =~ob(P/Pb)m, ...................... (6) 1977.

970 JOURNAL OF PETROLEUM TECHNOLOGY