SPE 91505
On the Validation of PVT Compositional Laboratory Experiments
Samaniego-V. F.1, Bashbush B. J. L.2, Len G.A.3, Mazariegos U. C.1, Corona B. A. and Castillo P. P. F.1UNAM
2
Geoquest 3Pemex
Copyright 2004, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the SPE Annual Technical Conference and
Exhibition held in Houston, Texas, U.S.A., 2629 September 2004.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in a proposal 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 a proposal of not more than 300
words; illustrations may not be copied. The proposal 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
The validation of compositional PVT results can be
understood as a systematic correction of the inherent errors of
these experiments, in a way that the true characteristics of the
fluids are preserved.
When carrying out a compositional simulation study (flow in
the reservoir, pipeline flow, optimization of separation
conditions, etc.), five steps must be taken:
1. Collect representative samples;
2. Carry out careful laboratory experiments the basic
procedure for compositional systems gas condensate and
volatile oil, is the constant volume depletion (CVD);
3. Validation of the PVT results from step 2;
4. Perform a regression analysis on the validated results of
step 3, to fit the parameters of an equation of state (EOS);
5. Simulate the process.
This study presents original procedures for the validation
(checking of consistency or quality) of compositional PVT
results obtained from the CVD experiment. A review of the
literature indicates that programs for the validation of these
experiments have not been presented. This paper covers this
gap.
The present work discusses three programs developed by the
authors to validate PVT compositional CVD laboratory
results. The findings of this study clearly indicate that
performing step 4 of the above procedure, without a careful
work regarding step 3 can lead in many cases to very serious
errors in the regression analysis of step 4 and consequently, in
the simulation results of step 5.
In general the validation procedure of this study consists of
the two main steps as follows; 1. Correct the main PVT
inconsistencies, like negative equilibrium constants, crossing
of the equilibrium constants curves, graphed in terms of log Ki
vs. pressure, and kinks or jumps of these curves. The
corrections change the gas phase composition of the
components that show inconsistencies and then, the computer
program is run again to check whether the new Ki values
present inconsistencies; 2. The second step consists of
checking that the summation of the gas mole percentages of
all components results in a 100 value. The main graph to
adjust these summations is that of log Ki vs. the boiling
temperature Tb of each component.
Three different approaches for this process are discussed,
with advantages and disadvantages clearly stated. The first is
a conventional manual procedure for the two main steps of the
process. The second is a semi-automated procedure that
automatically performs the first validation step. It was
developed to study PVT reports where the detected error level
in the measurements was high, which rendered the first
procedure difficult to use, because of the many hours needed
to complete the analysis. The third procedure is an enhanced
program, where a first version for a fully automated validation
procedure is presented.
From the many hydrocarbon mixtures analyzed, an example
is presented for a gas-condensate, which clearly that serious
errors in the simulation results can occur if a proper validation
study of the PVT compositional results is not carried out.
Introduction
The basic PVT experiment for compositional fluids, gas
condensate and volatile oils, is the constant volume depletion
(CVD) experiment, designed to provide volumetric and
compositional data, for reservoirs producing by pressure
depletion 1-5.
As a result of drilling deeper and hotter formations, the
number of variable composition reservoirs has increased
considerably during the last few decades 6,7. These reservoirs
are often found at temperatures near the critical temperature of
the hydrocarbon mixture they contain. Fig. 1 shows a graph of
depth vs. temperature for a group of variable composition
reservoirs, including several critical point fluids. These fluids
are characterized by having considerable amounts of
intermediate hydrocarbons (C2-C6). Table 1 of reference 3
present the compositional analysis of 10 of these reservoir.
The temperature and composition of these fluids are the
reasons for obtaining large retrograde condensations from the
gases and high shrinkage in liquid volumes from the oils.
The increased sophistication in the predictive techniques for
these compositional reservoirs, underlines the need for a
procedure that is able to determine the equilibrium ratios Ki
capable of reproducing the observed behavior of the fluid
samples in the laboratory. Several methods have been
2 SPE 91505
published to correlate Ki values 8-11. Two methods are
available to calculate these ratios based on conventional CVD
experiments where only the withdrawn vapor from the PVT is
analyzed using gas chromatography4 to determine
composition, yi. An alternate experimental set up has been
reported12, that allows also a direct liquid sampling at all
pressures considered in the study.
Inherent to all physical measurements and calculations are
two broad classes of errors, viz., determinate and
indeterminate errors13. Any error that is identified and
allowed for in magnitude and sign in the form of a correction
allowing for its effect is a determinate errors. It is common
to encounter errors that might theoretically be determined by
the adodption of a more refined method of measurement and
calculation. Were this economically justified. When error
estimates based upon past experience and best judgement are
inadequate for the purpose at hand, it becomes necessary to
devise practicable means for determining the accuracy. For
instance, in engineering work it oftern proves useful to employ
material balances as a basis for indicating probable limitations
on the accuracy of a result.
The main purpose of this paper is to present a discussion of
procedures for the validation (checking of consistency or
quality) of compositional PVT results obtained from CVD
experiments. Three programs developed by the authors to
validate PVT results are described. It will be demonstrated,
that this practice avoids the great deal of time and effort spent
in attempts to match inconsistent data, with a multi-variate
non-linear regression programa to calculate the parameters of
an equation of state (EOS).
A Compositional Simulation Procedure
The simulation of compositional flow problems (flow in the
reservoir, pipeline flow, optimization of separation conditions,
etc.) is usually a highly complex problem. When carring out
these types of studies, the following five steps procedure is
recommended:
1. Collect representative samples.
2. Conduct careful laboratory PVT experiments. As stated,
the basic experiment for compositional fluids is the
constant volume depletion (CVD).
3. Validation of the PVT results from step 2. This is the main
subject of the present paper.
4. Perform a regression analysis of the validated results of
step 3, to fit the parameters of an equation of state.
5. Simulate the process.
Consistency of the experimental data.
The five main checks for the consistency of the experimental
PVT data have been discussed in reference 3. Next a brief
review of these five items will be presented; an additional
sixth check will be included.
1. Plot the composition of the produced gas yi in a log scale
vs. pressure. Smooth curves should be the norm. Humps
in the graph will usually be an indication of PVT
experimental errors. Fig. 2 illustrates this problem for gas
condensate SBL-1.
2. A second check applicable to retrograde condensate fluids
is by means of the two phase deviation factor. The
calculated values should not differ in more than a unit in
the third decimal place in all instances. This check is not
definitive; therefore, satisfying it is not a conclusive proof
of consistency of the experimental data.
3. All compositional PVT reports include the analysis of the
liquid remaining in the cell at the end of the last depletion
stage. This composition should be compared against the
calculated liquid compositions. For most components with
initial mole fractions greater than 0.4 percent, the
difference between the calculated and the measured
compositions should be less than one percent.
4. Another useful plot can be that of the logarithm of the
liquid composition xi, versus pressure. As for check No. 1,
unexpected humps in the curves are indicative of
experimental errors.
5. The most important and sensitive check for the consistency
of the experimental data is a plot of the logarithm K values
obtained through the material balance procedure versus
pressure. The curves should plot in a parallel-like trend
with no humps inconsistent or crossings. Fig. 3 shows an
example of a PVT data set for gas condensate SBL-1. The
upper curve (higher K values) should correspond to
nitrogen, followed by the curves of methane and carbon
dioxide. Then, follows either the curve of ethane or the
curve of hydrogen sulfide (depending on the fluid
composition and reservoir temperature). Underneath the
remaining component curves should plot in increasing
order of their molecular weights (or normal boiling points).
The K values of isobutane and iso-pentane should always
be higher than those corresponding to the normal
components. In summary, the following list present the
order that the K values curves should follow in a log Ki
versus pressure curve.
- Nitrogen
- Methane
- Carbon dioxide
- Ethane
- Hidrogen sulfide
- Propane
- Iso-butane
- Butane
- Iso-pentane
- Pentane
- Hexane
- Heptane+
6. A graph of log Ki versus the boiling temperature Ti of the
different components of the hydrocarbon mixture is also a
sensitive consistency data check9,14,15. For well performed
PVT experiments, this graph should closely follow a
straight line; as stated for previous checks, humps indicate
errors in the measurements. The order of the Ki ratios
should be the same discussed in check 5.
Description of the validation Programs.
Once one or more of the previous six consistency checks has
been applied to a specific CVD PVT data set for a gas
condensate or volatile oil and errors have been found, we can
proceed to correct them using one or more of the three
programs that we have developed. The validation process is
understood as a systematic correction of the inherent errors of
SPE 91505
these experiments, in a way that the true PVT characteristics
of the hydrocarbon mixture is preserved.
Program No. 1 KVOLCO. The program KVOLCO was
developed in 1977 and its methodology published four years
later3. The driver to work out the procedure stem from the
need to formulate an optimum development plan for a giant
high temperature retrograde condensate reservoir with the aid
of reservoir simulation. The incipient use of the Peng-
Robinson EOS had been selected to model the phase
behavior.16 However, numerous attempts to match the
laboratory data from a CVD had failed to render an acceptable
combination of consistent EOS parameters.
During subsequent attempts to validate the CVD data, the
methodology described in the original paper3 was formulated
and successfully applied to correct inconsistencies. Once
these were removed, the EOS matching proceeded smoothly
and the study continued its course.
At the conclusion of the study, a FORTRAN program was
written and the procedure was successfully extended to the
CVD laboratory data from a volatile oil reservoir. By the time
the SPE 10127 was presented, the program had undergone
several enhancements and had been applied to analyze the
compositional data from 28 retrograde condensate and volatile
oil reservoirs. To date, the authors have applied this process
to several hundreds sets of CVD PVT laboratory experiments
from variable composition reservoirs from many basins
throughout the world.
The method consists of a simple material balance procedure
to calculate the mole fractions of the liquid phase remaining in
a PVT cell during each one of the pressure stages in any
standard CVD experiment. From those calculations, the K-
values can be easily obtained and plotted. The analyses of this
semi-log graph along with a complementary graph of the
calculated liquid compositions serve as the basis for the
diagnostic of the quality of the lab results. Thus, the
application of a stepwise process to improve the internal
consistency of the data prior to attempting the match with a
multi-variate non-linear regression program to calculate the
parameters of any of the common EOS, becomes an essential
requirement.
As stated in the original paper, this practice avoids the great
deal of time and effort spent in attempts to match inconsistent
information.
Among the enhancements incorporated to the program are
an extrapolation procedure to calculate the fluid composition
and K-values of either the first drop of liquid at the dew point
or the first bubble of gas at the bubble point, according to the
case; and procedures to obtain the variation of the density and
the molecular weight of the plus fraction in the liquid.
This methodology was later incorporated into the Eclipse suite
of programs as a recommended preamble to any EOS
matching to validate the accuracy of the laboratory data. The
name used: COMB reflects the COmpositional Material
Balance process required to comb the data prior to
performing non-linear regression exercises which, if not
properly done, may skew the EOS characterization in the
wrong direction.
3
Program No. 2. VALCO-Compositional Validation.17 This
was the second developed program. It has two main options
that will be next described.
Manual correction: The inconsistencies observed in the PVT
results can be manually corrected, preserving as discussed the
true characteristics of the sampled fluids. This procedure can
be described by the two following steps.
1. Correction of the main inconsistencies. In this first step
the important inconsistencies like negative equilibrium
ratios, crossing of the log Ki versus pressure curves, and
humps, are corrected, through changes in the composition
of the specific component mole fraction in the gas phase.
Right after these changes the validation program is run to
check whether the inconsistencies have been take care of.
2. Summation of the gas mole percentages of all components
to 100. The summation of the gas mole percentages of all
components in a hydrocarbon mixture should be 100; this
is checked for each of the depletion pressures considered
in the experiment. With regard to the inaccuracies of the
PVT data, in this second step to comply with the 100
percent requirements, the secondary inconsistencies like
non-smooth log Ki versus pressure curves, need to be
corrected. For this purpose, it has been found that a graph
of log Ki versus the boiling temperature Tb is very useful,
where the second level inconsistencies show as deviations
with respec to a negative slope straight line. This
correction basically corresponds to the sixth check
previously discussed.
Some additional simple general rules for the proper use of this
method are as follows:
- Based on results of the present study, it is recommended to
avoid changes in methane, because of its important
sensitivity to alterations of composition.
- The necessary changes in the gas composition are for
usually found inconsistencies, on the order of tenths or
hundreds of a percent.
Semiautomated correction. Some PVT results present an
important number of inconsistencies, making difficult its
manual validation as previously discussed, among other
causes due to the long time needed.
Similarly to the manual correction, the following procedure
for this validation is recommended.
1. The first steps corrects the main inconsistencies already
mentioned in step 1) of the manual correction, of the PVT
data through and automated approach, based on a graph of
log Ki versus the boiling temperature Tb discussed in step 2
of the manual procedure. Fig. 4 shows an example for a
volatile oil SLB-VO-1. A minimum square straight line is
fitted to this data and the correction process starts
comparing the Ki ratios calculated for the original PVT
data in the descending component order discussed in
consistency check (5); in other words, the first K N 2 value
for nitrogen should be bigger than the second KC1 value for
methane, and this bigger than the third KCO2 value, etc.; in
case of an inconsistency it will show as a hump, and the
program will change the corresponding Ki value for that
given by the straight line at the X value of the boiling
4 SPE 91505
temperature for the particular troublesome component,
Kis1. In a similar way, Kis1 is compared with the original
K i +1 ratio; if Kisl> K i +1 , the program selects Kisl as the
equilibrium ratio for compoennet I; if this condition is not
met, K i +1 is assigned as K ( i +1) sl . As can be easily expected,
after each of these changes, the composition of the gas
phase yisl is calculated using the Kisl ratio. This process is
repeated for each of the pressure sconsidered in the
experiment.
Through the previously discussed procedure the program
automatically corrects all the main inconsistencies of the
PVT data.
2. Summation of the gas mole percentages of all components
to 100. As for the previous manual correction option, the
100 summation balance is checked manually for each of
the depletion pressures of the experiment.
Program No. 3 AUTOVAL18-AUTOmatic VALidation
This program presents a first version towards the perfect goal,
to have a fully automatic validation procedure. As for the
previous VALCO program, AUTOVAL considers two steps
for the correction of the PVT data inconsistencies.
1. Correction of the main inconsistencies. This first step is
the same as discussed for step 1 of the semi-automated
correction option of the VALCO program.
2. Summation of the gas mole percentages of all components
to 100. The Jones and Erbar19 computer oriented algorithm
for the determination of equilibrium ratios is used in this
step. This technique makes use of the fact that the natural
logarithm of the equilibrium constants versus the normal
boling temperature of the componenets will approximate a
straight line. This adjustment technique is applied pressure
point by pressure point to the Ki equilibrium values
obtained in step 1 of this correction procedure. Reference
17 presents a detailed discussion of this technique.
One of the features of this program is its routine that takes
into account the associated uncertainty related to the
measurement of low liquid volumes, specially at depletion
pressures higher than the pressure of maximum retrograde
liquid volume. This routine uses a smoothing algorithm used
in well test analysis22, to obtain a curve that the user of the
program can evaluate jointly with the original (usually non-
smooth) curve, to conclude which better represents the PVT
behavior of the sample.
Comparison of Matching EOS Results to unvalidated and
Validate Measured PVT Data.
In the present section a discussion is presented with regard to a
comparison of results of matching and EOS to unvalidated and
validated experimentally measured PVT data. Previous
authors have clearly stated the need to check the experimental
data an fluido compositions for consistency3,4; as previously
stated, this practice avoids the great deal of time and effort
spent in attempts to match inconsistent information.
Before properly addressing the main topic of this section, it
is worth to make a brief comment regarding the most difficult
to analyze fluids, which are the called critical, physically
meaning that its temperature and pressure are very close to the
critical point values. While conducting a CVD experiment the
cell would require longer than the normal 0.5 hours
equilibration times at each pressure20, before measuring the
liquid volume; in addition, the removed gas from the cell at
each pressure is brought to atmospheric conditions, and the
amounts of surface gas and condensate are measured. Special
care must be taken to avoid liquid condensation in the
laboratory measurement system, which would result in errors
in the estimations through material balance3,4 of the densities
of the oil or the gas-condensate21. For cases where this
problem could not be solved, direct measurements of the
densities are recommended.
The example selected to show the problems that arise whn
matching an unvalidated PVT data set with an EOS, is a gas
condensate, SBL-1. Table 3 presents the basic information of
this fluid. Fig. 2 shows the CVD experimentally produced
(equilibrium) gas compositions, observing some problems
(crossings), specially for the C7+ heavy fraction.
A consistency check based on step 5 of the previously
presented procedure was performed, detecting inconsistencies
as shown in the results of Fig. 3, as crossing of the curves.
The results of the analysis of the PVT for this sample also
indicated important differences among values for the gas-
condensate density at dew-poing pressure, the experimentally
reported value, the calculated value based on the validated
PVT results, and the estimated value through static vertical
gradient measurement during a pressure buildup test; Fig. 5
and Table 2 present these results. This table includes results
for cases 1 and 2, which consider the gas condensate density at
dewpoint pressure estimated through the material balance
procedure and that experimentally estimated, respectively.
The calculated difference between the laboratory value,
0.2708 g/cm3, and the material balance value, 0.3606 g/cm3, is
25 percent.
A PVT compositional simulator was used to evaluate the
difference in results obtained through an EOS regression
process. Figs. 6,7 and 8 indicate that the value of the gas-
condensate at dewpoint influences the liquid dropout, the gas
deviation factor, z, and the equilibrium ratios, Ki. In this last
figure, it can be noticed that the differences in the Ki EOS
calculated values for cases 1, validated, and case 2,
unvalidated (original) is not important for the light
componenets, but clearly increased for the heavier C5 and C6
components and the heptane plus fraction.
Doing the previously described validation has, in many
occasions greatly simplified the matching process and resulted
in greater reliability in the application of consistent
compositional simulation practices.
Conclusions.
The purpose of this study has been to present a discussion on
the necessary process to be followed to validate CDV PVT
results, obtained for compositional fluids, gas-condensate and
volatile oils.
From the results of this works, the following conclusions
and/or remarks are pertinent.
(1). A simple procedure for a compositional study has been
described.
(2). A practical validations procedure for CVD PVT
compositional laboratory results consisting of two main steps
has been presented. It saves many hours of effort trying to
SPE 91505 5

perform an EOS regression analysis based on inconsistent 16. Peng, DY. And Robinson, D.B.: A New-two Constant EOS,
data. Ind. & Eng. Chem. Fund. (1976) 15, No. 1., 59.
(3). The development of three different programs for the 17. Mazariegos U.C.E.: Desarrollo de un paquete de Cmputo para
validationf of PVT results was discussed. Determinar la Consistencia de un Anlisis Composicional de
Fluidos, M. Sc. Report, School of Engineering, National
(4). It was demonstrated through and example of the many
University of Mxico (1990).
analyzed by the authors, that serious errors in the simulation of 18. Corona B. A. A.: Validacin Automatizada de Anlisis PVT
a compositional process results can occur if a proper Composicionales, M. Sc. report, Shool of Engineering,
validation study of the pvt results is not carried out. National University of Mxico (1993).
19. Jones, D. M. and Erbar, J.: Computer Determination of Data-
Matched Equilibrium Ratios, J. Pet. Tech. (Aug. 1970) 1007-
References 1011.
1. Standing, M.B.: Volumetric and Phase Behavior of Oil 20. Eyton, D.G.P.: Practical Limitations in Obtaining PVT Data
Field Hydrocarbon Systems, eight edition, SPE, for gas Condensate systems, paper SPE 15765 presented at the
Fifth Middle East Show, Manama, Bahrain, March 7-10, 1987.
Richardson, Texas (1977).
21. Monter O.J.: personal Communication (No. 1997), Mexican
2. Reudelhuber, F.O. and Hinds, R.F.: Compositional Petroleum Institute, Mxico, D.F.
Material Balance Method for Prediction of Recovery 22. Standard Practice for conducting an Inter.-Laboratory Test
from Volatile-Oil Depletion-Drive Reservoirs, JPT Program to Determine the Precision of the Test Methods,
(January 1957); Trans., AIME, 210. ASTM E691-79.
3. Bashbush, J.L.: A Method to Determine K-Values from 23. Drohm, J.K., Goldthorpe, W.H. and Trangove, R.: Enhancing
Laboratory Data, paper SPE 10127 presented at the 56th the Evaluation of PVT Data, paper OSEA 88174 presented at
Annual Technical Conference and Exhibition, held in the 7th Offshore South East Asia, Singapore, 2-5 February
San Antonio, Texas, October 5-7, 1981. 1988.
24. Drohm, J.K., Trangove, R.D. and Goldthorpe, W.H.: On the
4. Whitson, C.H. and Torp, S.B.: Evaluating Constant-
Quality of Data from Standard Gas-Condensate PVT
Volume Depletion Data, J. Pet. Tech. (March, 1983) Experiments, paper SPE 17758 presented at SPE Gas
610-620 Technology Symposium, Dallas, Tx., June 13-15, 1988.
5. Whitson, C.H. and Brule, M.R.: Phase Behavior, SPE 25. Kool, H., Azari, M., Soliman, M.Y., Proett, M.A., Irani, C.A.
monograph Vol. 20, Richardson, Tx. (2000). and Dybdahl, B.: Testing of Gas Condensate Reservoirs-
6. Craft, B.C. and Hawkins, M.F.: Applied Petroleum Sampling, Test Design and Analysuis, paper SPE 68688
Rservoir Engineering, Prentice-Hall, Englewoods Cliffs, presented at SPE Asia Pacific Oil and Gas Conference, Jakarta,
N.J.(1959). Indonesia, 17-19 April 2001.
7. Len G.A. and Samaniego V.F.: Influencia de los 26. McCain, W.D., Jr.: The Properties of Petroleum Fluids, Second
Edition, Pennwell, Tulsa, OK. (1990), Chapter 5.
Resultados PVT en la Caracterizacin de Fluidos de
27. Samaniego V.F. and Cinco Ley, H.: Well Test Analysis in
Yacimientos Mediante la Ecuacin de Estado, Ingeniera Carbonate Reservoirs, in G.V. Chilingarian, S.I. Mazullo and
Petrolera (April 2002) 40-48. H.H. Rieke, III (Editors), Carbonate Reservoir
8. Dykstra, H. and Mueller, T.D.: Calculations of Phase Characterization: A Geologic-Engineering Analysis, Part II,
Compositions and Properties for Lean or Enriched Gas Elsevier
Drive; SPEJ (Sept. 1965)
9. Hoffmann, A.E., Crump, J.S. and Hocott, C.R.:
Equilibrium Constants for a Gas-Condensate System,
Trans., AIME (1953)198, 1-10
10. Scientific Software Corporation, N-Comp Users
Manual, Denver, Colorado(1981).
11. Abel, W. Jackson, R.R. and Wattenbarger, R.A.:
Simulation of a Partial Pressure Maintenance Gas
Cycling Project with A Compositional Model, Carson
Creek Field, Alberta, JPT (Jan. 1970).
12. Khan, M.S. and Hatamian, H.: Improved Method of
Compostional Analysis of Liquid at High-Pressure
Condition in PVT Study of Gas Condensate, paper SPE
21429 presented at SPE Middle East Show, Bahrain, 16-
19 November 1991.
13. Mickley, H.S., Sherwood, T.K. and Reed, C.E.: Applied
Mathematics in Chemical Engineering, McGraw-Hill,
New York (1957).
14. Moses, P.L. and Donohoe, C.W.: Gas Condensate
Reservoirs, in H.B. Bradley (Editor): Petroleum
Engineering Handbook, SPE, Richardson, Texas (1987).
15. Brinkam, F.H. and Sicking, J:N.: Equilibrium Ratios for
Reservoir Studies, Trans., AIME (1960) 219, 240-246.
 6 SPE 91505

Table 1. Analyses of several hydrocarbon mixtures,

ranging from gas condensate to black oils.
COMP. SBL-GC-1 SBL-GC-2 SBL-VO-1 SBL-BO-1

GAS & GAS & VOLATILE BACK

COND. COND. OIL OIL
N2 2.43 0.4 0.634 0.4
CO2 3.369 1.94 0.938 2.96
H2S 1.227 0.67 0.000 1.13
C1 67.161 61.55 65.734 28.2
C2 6.528 9.04 6.212 8.38
C3 3.658 4.61 3.017 5.68
IC4 0.869 1.08 0.554 1.64
NC4 2.153 1.88 1.664 3.27
IC5 1.107 1.13 0.928 1.03 Fig. 1. Variation of temperature versus depth for gas
NC5 1.287 1.33 1.195 1.71 condensate and volatile oil in Basin, SBL-1.
C6 1.988 2.24 1.857 3.78
C7+ 8.223 14.13 17.267 41.82
100.000 100.000 100.000 100.000
Pr or Pb, 271 372 356 150
2
kg/cm
TEMP (C) 101 153 175 101
GOR 1172 759 72
(M3/M3)
lsc 0.7496 0.7919 0.9159
C2-C6 17.59 21.31 15.427 25.49

Table 2. Comparison of results for the condensate density

and the gas-condensate ratio (GCR), experimental,
calculated (EOS) and field measurements.
PARAMETER EXPERIMENTAL CASE CASE FIELD
1 2 MEASUREMENT
(EOS) (EOS) Fig. 2. Composition of the produced gas versus pressure
GAS 0.2724 0.3606 0.2724 for gas condensate SBL-1.
CONDENSATE
DENSITY, AT
pr(g/cm3)
GCR 1170 1055 845 1171
CONDENSATE 0.7494 0.7325 0.5907 0.7496
DENSITY AT
Sc(g/cm3)
OBJECTIVE 0.5313 2.16
FUNCTION

Table 3. Data for SBL-1 well gas-condensate sample.

Initial reservoir pressure, p i = 297 kg/cm2
Dewpoint pressure, p d = 271 kg/cm2
Density of the gas-condensate at p d =0.2724 g/cm3
Reservoir temperature, T =101 C
Separator-gas/condensate ratio, R sp =170 m3/m3
Separator-condensate density, csp =0.7494 g/cm3
Maximum condensate dropout = 26 percent
Fig. 3. Calculated equilibrium constants for the original
PVT data, using the method of Bashbush3, for gas
condensate SBL-1.
SPE 91505 7

Fig. 7. Comparison of the results for the gas desviation

factor z, experimental and calculated (EOS).
Fig. 4. Schematic PVT inconsistency correction graph of
log Ki versus normal boiling point Tb.

Fig. 8. Comparison of EOS calculated equilibrium ratios

for cases 1 and 2.
Fig. 5. Comparison of the gas densities calculated through
several methods, for gas con densate SBL-1.

Fig. 6. Comparison of the results for the CVD liquid

dropout, experimental and calculated (EOS).