Professional Documents
Culture Documents
net/publication/256712679
CITATIONS READS
8 223
6 authors, including:
S. M. Dessouky
Egyptian Petroleum Research Institute
52 PUBLICATIONS 129 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Hydrophobic polymers for gas and water shutt off control View project
All content following this page was uploaded by A. N. El-hoshoudy on 03 April 2018.
Fuel
journal homepage: www.elsevier.com/locate/fuel
h i g h l i g h t s
a r t i c l e i n f o a b s t r a c t
Article history: Density and viscosity are of the most important governing parameters of the fluid flow, either in the por-
Received 23 April 2012 ous media or in pipelines. Ideally, viscosity and density determined experimentally in the laboratory on
Received in revised form 9 May 2013 actual fluid samples taken from the field under study. However, in the absence of experimentally mea-
Accepted 14 May 2013
sured data, especially during the prospecting phase, or when only invalid samples are available, one
Available online 29 May 2013
can resort to empirically derived PVT correlations. Correlations are also needed for the calculation of mul-
tiphase flowing pressure gradients which occur in pipe. These calculations require the prediction of fluid
Keywords:
properties at various pressures and temperatures. Even though laboratory measurements of these prop-
Oil density
Oil viscosity
erties may be available as a function of pressure, they are usually measured under isothermal conditions.
PVT correlation The behavior of these properties as a function of temperature is usually predicted by using empirical cor-
Live oil relations. So it is of great importance to use accurate correlations to calculate the crude oil density and
Dead oil viscosity at various operating conditions. During the last decades, several correlations have been devel-
oped to estimate density and viscosity of oil at different reservoir conditions. However, these correlations
may be useful only in regional geological provinces and may not provide satisfactory results when
applied to crude oils from other regions since oil properties differ according to its source, origin and core
type .Also, crude oil composition is complex and often undefined. Therefore, based on Egyptian oil reser-
voirs data; new correlations have been developed for predicting density and viscosity of dead and live
crude oil. Validity and accuracy of these correlations have been confirmed by comparing the obtained
results of these correlations and other ones with experimental data for Egyptian oil samples. Checking
results of these correlations show that correlations developed by this study revealed more accurate
results than the literature correlations.
Ó 2013 Elsevier Ltd. All rights reserved.
0016-2361/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.fuel.2013.05.045
278 A.N. El-hoshoudy et al. / Fuel 112 (2013) 277–282
Nomenclature
API stock tank oil API gravity Pres reservoir pressure (psia)
Tres reservoir temperature (F) Pb bubble point pressure (psia)
qsc oil density, gm/cc at standard conditions cg gas specific gravity, air = 1
co oil specific gravity, water = 1 cgs separator gas gravity @100 psi, air = 1
Rs solution gas/oil ratio, scf/STB Rsb solution gas/oil ratio at bubble point (scf/STB)
lod dead oil viscosity (cp) lob gas-saturated oil viscosity (cp)
qod dead oil density (gm/cc) lo live oil viscosity (cp)
qo live oil density (gm/cc) qob gas-saturated oil density (gm/cc)
Er average percent relative error MWSTO molecular weight of stock-tank oil
Ea average absolute percent relative error S standard deviation
Emin. minimum absolute percent relative error Erms the root mean square error
Emax. maximum absolute percent relative error R correlation coefficient
density and viscosity above the bubble point due to compression of collected sample is non representative, (4) quality check lab anal-
the oil. However below the bubble point, increasing pressure cause ysis, (5) lab analyses are in error, (6) estimating the potential re-
an increase in solution gas, which in turn decrease the oil density serves to be found in an exploration prospects, and (7)
and viscosity. At atmospheric pressure and reservoir temperature, evaluating the original oil in place and reserve for a newly discov-
there is no dissolved gas in the oil [i.e., Rs = 0] and therefore the oil ered area before obtaining the laboratory analysis to justify a pri-
has its highest density and viscosity values. Thus, oil density corre- mary development plan [5]. All correlations were concerned with
lations all belong to three categories: dead oil, saturated oil, under- crudes from different locations and presumably of different charac-
saturated oil density correlations, also, oil viscosity correlations all teristics. Each study claimed that the resulting correlation would
belong to three categories: dead oil, saturated oil, undersaturated provide the best approximation of PVT properties for the local re-
oil viscosity correlations. Numerous correlations have been pro- gion compared to the other commonly used correlations. Studies
posed to calculate the oil density and viscosity [3]. These correla- performed by Macary and Batanony [6], Hanafy et al. [7], Glasso
tions are categorized into two types. The first type refers to black [3], Dokla and Osman [8], Marhoun [9], Labedi [1] all supported
oil type correlations which predict density and viscosity from this conclusion [5].
available field-measured PVT data, such as reservoir temperature, Multiple linear/nonlinear least-squares regression analysis will
oil API gravity, solution gas–oil ratio, saturation pressure, reservoir be used to develop the new correlations. In addition, accuracy of
pressure. The second type refers to oil composition models which developed correlations determined by comparing the obtained re-
derive mostly from the principle of corresponding states and its sults with the published ones through statistical error means (Er,
extensions. In these correlations beside previous properties, other Ea, Emax, Emin, S, Erms, and r) and graphical error means (cross plot
properties such as reservoir fluid composition, molar mass, critical analysis). Furthermore, the new correlations will be validated
temperature and acentric factor of components are used [4]. The using other experimental data sets not used in the correlations
reservoir fluid data have many applications in different areas of development [10].
the Exploration and Production process. While reservoir engineers
generally have the greatest claim on such data, reservoir fluid anal-
2. PVT correlations
yses are also quite valuable to geologists and production special-
ists. The process of collecting fluid samples may be repeated
The frequently used empirical correlations for the prediction of
during different phases of a field since discovery till its mature
dead, saturated and undersaturated oil densities and viscosities are
phase. A geologist may use correlations along with an oil or gas
reviewed in the following sections. For Egyptian oils, Hanafy et al.
gravity measurement from a near-by well for help in obtaining
[7] reported that Ahmed [11] correlation is the most accurate for
an estimate of the potential reserves to be found in an exploration
determining undersaturated oil density while Katz [12] and Stand-
prospect. After the exploration well is drilled and successful, a well
ing [13] correlations are the best to estimate dead and gas satu-
test may allow those same correlations to be used with the known
rated oil densities respectively, also, correlation after Beggs and
gravity, gas–oil ratio, and pressure data from the discovery well. In
Robinson [4] is the best to estimate dead oil viscosity while, the
an ideal situation, a fluid sample may be recovered from the dis-
gas saturated and undersaturated oil viscosities are best deter-
covery well for analysis. This more precise information on the
mined with Khan’s [14] correlations.
properties of the hydrocarbon accumulation may be used by geol-
ogists and engineers to justify further development drilling. One or
several of the development wells may then be completed and res- 2.1. Dead oil density correlations
ervoir fluid samples retrieved. The laboratory analysis of such sam-
ples provides the more accurate information needed to help plan Standing expressed the mathematical form of Katz’s chart for
the development of the field, design production facilities, deter- dead oil density [12]. Ahmed published two correlations; one cor-
mine the size and cost of equipment, and thereby make economic relation uses the stock tank oil molecular weight while the second
decisions. After production has been established, further sampling correlation approximates the estimation of dead oil density if the
and analysis may by requested by the engineer to evaluate poten- stock-tank oil molecular weight is not available [11].
tial improved recovery projects, consequently, empirical correla-
tions are of great importance to determine these physical 2.2. Gas saturated oil density correlations
properties with high precision at each stage. One can resort to
empirical PVT correlations to estimate the reservoir fluid data in Standing [13] by using 105 data points [22 fields], published his
the following cases: (1) inability to obtain a representative sample, correlation, In this correlation, a weighted average separator and
(2) sample volume is insufficient to obtain a complete analysis, (3) stock tank gas specific gravities should be used for cg, however
A.N. El-hoshoudy et al. / Fuel 112 (2013) 277–282 279
the use of separator gravity gives adequate results. Hanafy et al. [5] Table 1
correlated the bubble point oil density to the bubble point oil vol- Data ranges of Egyptian crude oils.
ume factor, to account for the compression of the oil above the Parameter Minimum Maximum Average
bubble-point pressure; the density of the crude is first calculated [Pres.] psig 793.00 10393.00 4120.00
at the bubble-point pressure and reservoir temperature. The calcu- [Tres.] F 119.00 301.00 227.00
lated density is then adjusted by using the isothermal compress- [Pb] psig 46.00 4804.00 2415.00
ibility coefficient. Rsb [SCF/STB] 10.00 7890.00 1390.00
APIo[degree] 21.00 52.00 38.00
Rs [SCF/STB] 9.00 7834.00 1361.00
2.3. Undersaturated oil density correlations cg [air = 1] 0.71 1.46 0.99
co 0.77 0.94 0.83
Vasquez and Beggs [15] by using 4036 data points published his MWSTO 113.00 200.00 145.00
correlation which derived from the isothermal oil compressibility cgs 0.62 1.55 0.93
qsc [lb/ft3] 48.00 58.00 52.00
correlations and is similar to the undersaturated oil formation vol- qod [g/cc] 0.70 0.87 0.77
ume factor correlation. Ahmed [11] published his correlation qob [lb/ft3] 10.00 54.00 39.00
which derived from the isothermal oil compressibility correlations. lod [CP] 0.54 8.26 1.88
lob [CP] 0.02 5.69 0.69
Beal [16] reported a correlation using 753 data points for his
analysis. He correlated oil gravity and temperature covering a
range of 100–220 F. Beggs and Robinson [4] presented their corre-
lation using 460 dead oil observations. Petrosky and Farashad [17]
developed a correlation with an average absolute error of 14.47%.
Glasso [3] also developed a correlation using temperature range
of 50–300 F for 26 crude oil samples. Kartoatmodjo and Schmidt
[2] developed the dead oil viscosity correlation based on data bank
consisting of 661 samples. The correlation parameters were the oil
API gravity and reservoir temperature. The data used in developing
this correlation was ranging from 0.506 to 682 cp.
Chew and Connally [18] presented their work for predicting oil Fig. 1. New model for dead oil density.
Fig. 3. New model for live oil density. Fig. 7. New model for live oil viscosity.
Table 2
Statistical accuracy of dead oil density correlations for the data used in model
Fig. 4. Cross plot of live oil density Correlations. validation.
Table 3
Statistical accuracy of live oil density correlations for the data used in model
validation.
Fig. 5. New model for dead oil viscosity. 1. Standing 28.1876 28.1876 64.3313 2.0757 14.8663 31.7519 0.7804
and Katz
2. Standing 0.6791 5.1230 18.2144 0.0084 7.1197 7.0329 0.9513
3. Hanafy et al. 0.4012 3.9851 9.5632 0.1254 4.5231 4.213 0.9231
4. This study 80.2764 3.4480 8.5539 0.0525 4.1358 4.0757 0.9636
Table 4
Statistical accuracy of dead oil viscosity correlations for the data used in model validation.
Table 5
Statistical accuracy of live oil viscosity correlations for the data used in model validation.
The accuracy and reliability of each mentioned correlation was 1. A new empirical correlations employed a multiple linear regres-
checked by using both statistical and graphical error means sion analyses were developed to estimate the physical proper-
analyses. ties of crude oils such as oil density and oil viscosity with an
excellent accuracy using field production and experimental
5.1.1. Statistical error analysis data for the Egyptian crude oil systems.
The accuracy of fluid property correlations relative to the exper- 2. The development approach of these new correlations is based
imental data of Egyptian oils is determined in this study using se- on some readily available properties and parameters. Most of
ven statistical means (Er, Ea, Emax, Emin, S, Erms, and r), where lower these parameters were not considered in the existing
values of Er and Ea implies a better correlation. The correlation correlations.
coefficient (r-value) lies between zero and one; a value close to 3. Comparative evaluation of the developed correlations and the
one indicate a good correlation among the given dependent and existing correlations from the literature was made using statis-
independent variables [5]. tical and graphical error analyses. It indicates that:
282 A.N. El-hoshoudy et al. / Fuel 112 (2013) 277–282
(a) The new correlations are much simpler and outperform the symposium on oil field chemistry held in Houston, Texas, 18–21 Febreuary,
1997.
existing correlations.
[8] Dokla ME, Osman ME. Correlations of PVT properties for the UAE crudes. SPE
(b) The new correlations can be used to estimate oil density and Formation Eval 1992;7:41–6.
oil viscosity for a wide range of crude oils ranging from [9] Al-Marhoun MA. PVT correlations for Middle East crude oils. J Pet Technol
heavy to light oils. 1988;40(5):650–66.
[10] Younger MS. Handbook for linear regression. USA: Wadsworth, Inc., Belmont;
4. The new correlations are considered more relevant and applica- 1999.
ble to the Egyptian crude oils properties and conditions than the [11] Ahmed T. Compositional modeling of tyler and mission canyon formation oils
existing correlations, since the new models have been devel- with CO2 and lean gases. Final report submitted to Montana’s On a New Track
for Science [MONTS], Montana National Science Foundation Grant, Program,
oped based on a large number of data for the Egyptian fluid 1988.
samples. [12] Katz DL, Standing MB. Density of natural gases. Trans AIME
1942;146(140):149.
[13] Standing MB. Volumetric and phase behavior of oi1 field hydrocarbon systems.
9th ed. Dallas: Society of Petroleum Engineers; 1981.
References [14] Khan SA. Viscosity correlations for Saudi Arabian crude oils, SPE Paper
15720. Dallas: Society of Petroleum Engineers; 1987.
[1] Labedi R. Improved correlations for predicting the viscosity of light crudes. J [15] Vasquez ME, Beggs. HD. Correlation for fluid physical property prediction. J Pet
Pet Sci Eng 1992;8:221–34. Technol 1980;32(6):968–70.
[2] Kartoatmodjo F, Schmidt Z. Large data bank improves crude physical property [16] Beal C. Viscosity of air, water, natural gas, crude oil and its associated gases at
correlation. Oil Gas J 1994;4:51–5. oil field temperature and pressures. Trans he AIME 1946;165(114):127.
[3] Glaso O. Generalized pressure–volume–temperature correlation for crude oil [17] Petrosky GE, Farshad FF. Production Operations Symposium. SPE 29468,
system. JPT 1980;2:785–95. Oklahoma City, Oklahoma; 1995
[4] Beggs HD, Robinson JR. Estimating the viscosity of crude oil systems. JPT [18] Chew J, Connally CA. Viscosity correlation for gas saturated crude oil. Trans
1975;9:1140–1. AIME 1959;216:23–5.
[5] Hanafy HH, Macary SM, ElNady YM, Bayomi AA, El-Batanoney MH. A new [19] Abu-Khamsim A, Al-Marhoun M. Development of a new correlation for bubble
approach for predicting the crude oil properties. SPE 37439, presented at the point viscosity. Arab J Sci Eng 1991;16(2A):99.
SPE production operation symposium held in Oklahoma City, Oklahoma, 9–11 [20] Almehaideb RA. Improved PVT correlations for UAE crude oils. Paper SPE
March 1997. 37691. Presented at the SPE Middle East Oil Conference and Exhibition,
[6] Macary, SM, El-Batanoney MH. Derivation of PVT correlations for the gulf of Bahrain; 1997.
suez crude oils. In: EGPC 11th Petroleum Exploration and Production [21] De Ghetto G, Paone F, Villa M. Reliability analysis on PVT correlations. Paper
Conference, 1992 SPE 28904. Presented at the European Petroleum Conference London; 1994.
[7] Hanafy, HH, Macary SM, ElNady YM, Bayomi AA, El-Batanoney MH. Empirical [22] Elsharkawy AM, Alikhan AA. Models for predicting the viscosity of Middle East
PVT Correlations Applied to Egyptian Crude Oils Exemplify Significance of crude oils. Fuel 1999;78:891–903.
Using Regional Correlations. SPE 37295, presented at the SPE international