60 Issue 2: 2015 collection

UDC 622.276.031:532.529.5 (470+571)

T.S. Yushchenko, A.I. Brusilovskiy

A new engineering method for creating and adapting

PVT model of natural gas condensate mixture

Keywords: At the present time, the design and monitoring of the development of virtually any
PVT model, hydrocarbon deposits are accompanied with hydrodynamical model building. One of the
constant volume important stages of hydrodynamical model building is the development of a PVT model of
depletion (CVD), the reservoir hydrocarbon mixture. In the engineering practice, the method most commonly
equation of state, used for building thermodynamic models of natural hydrocarbon systems is a method
gas condensate utilizing the Peng–Robinson three-parameter state equation and efficient algorithms of
mixture. modeling vapor-liquid equilibria of multicomponent systems.
The state equation has many adjustable parameters (critical pressure and temperature,
acentric factor of C6+ fractions, component shift parameter, pair interaction coefficients, etc.).
For this reason, building an adequate thermodynamic model of natural hydrocarbon
systems is not simple and requires the understanding and use of the fundamental premises
of physical chemistry.
Works [1–5] propose algorithms for building PVT models of natural hydrocarbon
systems (reservoir oils and gas condensate systems) and for their adjustment following
laboratory studies. For example, sources [3–5] pay the most attention to evaluating sample
representativeness, and adjustment following laboratory studies is performed by changing
the molecular mass of stabilized condensate (C5+ group). However, the potential contents of
the C5+ group in the reservoir gas are not maintained, and not all of the key parameters are
used for adjustment. Articles [1, 6, 7] emphasize the procedure for dividing the Cn+ group
into pseudocomponents using various methods and determining the properties of such
pseudocomponents (critical pressure and temperature, acentric factor) based on correlation,
but do not specify which state equation parameters are adjusted and in what way. It should be
noted that in many works the procedure of thermodynamic model adaptation to experimental
data is not detailed, and its use in practice requires experience and a creative approach.
An engineering phased approach to building an adequate thermodynamic model of
reservoir oil is described in work [8]. The authors developed an algorithm of successive
PVT model adaptation to the key reservoir oil properties: saturation pressure, volume factor
in the reservoir conditions, separated oil density, and dynamic viscosity of reservoir oil.
For gas condensate systems, the development of such a method is even more difficult
due to the retrograde condensation phenomenon and higher hydrocarbon sensitivity to
temperature and pressure changes. The most important parameters of reservoir gas that
should be considered during design are described in work [9].
This article deals with a new engineering method for building multicomponent PVT
models of natural gas condensate mixtures. The method is based on the reproduction of field
measurement results and basic laboratory thermodynamic studies of representative samples
during mathematical modeling using a state equation.

PVT model building

The process of building a PVT model of reservoir gas condensate mixture is based on
data about the component composition and the results of a number of laboratory studies of
thermodynamic, physical, and chemical properties of the reservoir system and stabilized
condensate, which are provided in technical reports [10].
In calculating stabilized condensate and dry gas reserves, data about the reservoir
gas Z factor and potential decrease in C5+ group contents in the reservoir gas are used.
Also important are the value of the gas/condensate ratio in field gas-condensate studies,

English-language digest
Proceedings of Gas Industry Research and Development 61

stabilized condensate density, and dynamics of group fractions using the method described in
saturated condensate formation during laboratory work [11]. The critical temperature is calculated
studies of the constant volume depletion process. using the Lee–Kesler correlation, and the critical
It should be noted that the most important part of pressure and acentric factor using a system of
the saturated condensate loss curve is the interval equations consisting of Peng–Robinson state
between the initial condensation pressure and the equations for the standard conditions and the
abandonment pressure, which is usually at least Edmister correlation for the acentric factor.
30–40 bar for gas-condensate fields. The above 4. We perform adjustment for the saturated
parameters are the key parameters for building an condensate loss curve plotted following laboratory
adequate PVT model of a natural gas condensate studies of the constant volume depletion process.
system. We select the shift parameter and critical
temperature of the Cn+ group (if this group was
Thermodynamic model adaptation divided into a number of pseudocomponents,
For thermodynamic model adaptation to laboratory the adjustment can be done by changing the
study results, an engineering method is proposed, critical temperature and shift parameter of all the
consisting in successive adjustment of PVT model pseudocomponents), as well as the pair interaction
parameters in order to reproduce the key properties coefficients for interaction between methane and
of gas condensate mixture. Cn+ (or all the pseudocomponents into which the
We next describe the procedure for identifying Cn+ group is divided) as the parameters that will
PVT model parameters. be used for adjustment. The critical pressure and
1. We reproduce the potential level of the C5+ acentric factor are recalculated [11].
group in the reservoir gas ( L pot .ɋ5+ , g/m3 dry gas), by The most important interval of the curve for
ascertaining the molecular mass of the remaining which adjustment should be made is the interval
Cn+ ( M Cn). Since between the condensation beginning and the
n
abandonment pressure.
¦Z
i =0
C5+i M C5+i The adjustment is done as follows:
L pot .C5+ = , (1) a) we adjust stabilized condensate density
0.02404(1 − Z C5+ ) using the shift parameter of the Cn+ group (last
the molecular mass will be unambiguously fraction). If the calculated stabilized condensate
determined as: density is less than the experimental value, we
n −1
increase the shift parameter value;
0.02404(1 − Z C5+ ) Lpot .C5+ − ¦ Z C5+i M C5+i b) we adjust (a graphical representation should
i =0
M C n+ = . (2) preferably be used) the quantity of the condensate
Z Cn +
formed within the pressure range in which
Here, Z C5 and M C5i are the molar fraction and retrograde condensation takes place by changing
molecular mass of the ith fraction of the C5+ group the critical temperature of the Cn+ fraction. If the
in the reservoir gas, respectively; N is the number calculated curve is higher than the experimental
of components in the C5+ group. one, the critical temperature of the Cn+ group should
If the molecular mass of the last fraction (Cn+) be reduced. We then recalculate the critical pressure
is less than the molecular mass of the second to and acentric factor for the fraction according to the
the last fraction (Cn–1), the fractions Cn–1 and Cn+ new critical temperature value;
should be grouped and the molecular mass of the c) we adjust the quantity of the condensate
combined pseudofraction should be calculated formed near the initial condensation pressure using
using formula (2) for adjustment to the potential the pair interaction coefficient for the interaction
C5+ group contents in the reservoir gas. between methane and the Cn+ fraction (an increase
2. If the molar fraction of the remaining in the coefficient always results in increased initial
fractions (Cn+ group) is more than 2 %, we divide condensation pressure).
the remaining fractions (using laboratory report All the adjustment steps (a–c) should be
results or mathematically) so that the molar performed in sequence, in cycles, until we reach
fraction of the last component in the division does the desired accuracy of adjustment for the saturated
not exceed 1 %. condensate loss curve.
3. We calculate the critical temperature, The impact of the parameters on the saturated
critical pressure, and acentric factor of the C6+ condensate reservoir loss curve is shown in fig. 1.

Vesti Gazovoy Nauki scientific and technical collected book

62 Issue 2: 2015 collection

Saturated condensate loss volume, cm3/stand. m3 0.06

Shift parameters
0.05 of heaviest fraction

0.04

0.03

Pair interaction coefficient

0.02 Critical temperature for interaction between methane
of heaviest fraction and heaviest fraction
experiment
0.01
calculated value
after adjustment
0
2.5 7.5 12.5 17.5 22.5 27.5 32.5
Pressure, MPa

Fig. 1. Impact of state equation parameters on saturated condensate reservoir loss curve

5. We adjust stabilized condensate density • high condensate – 100 to 500 g/m3;

using the shift parameter of the Cn+ group (last • unique condensate – over 500 g/m3.
fraction) again. Increasing the shift parameter For low condensate deposits, gas condensate
value also increases stabilized condensate density. studies are not conducted, and the condensate
6. We adjust the calculated gas/condensate ratio recovery ratio is evaluated using correlations or
using the shift parameter for the gas components graphical relationships [10].
C2–C4 (ethane, propane, butanes). Decreasing As the potential C5+ content in the reservoir
their values increases the gas/condensate ratio. gas condensate mixture increases, the percentage
The stabilized condensate density will be slightly of heavy fractions, the properties of which are
changed as a result, and it should be adjusted using difficult to determine, also increases. For this
the shift parameter of the Cn+ fraction. reason, building an adequate PVT model becomes
7. We adjust the reservoir gas Z factor in more and more difficult as the condensate content
the reservoir conditions by changing the shift in the reservoir system increases.
parameter for methane. Increasing the shift Let’s consider the utilization of the PVT model
parameter decreases the Z factor, and vice versa. adjustment method developed using the example
Thus, the method allows adjustment for of high condensate mixture with a potential
the following key parameters of gas condensate C5+ content in the reservoir gas of 231 g/m3 dry
studies: gas. The main properties of the gas condensate
• initial condensate content in reservoir gas; mixture are: reservoir temperature 86 °C; reservoir
• reservoir gas Z factor; pressure 27.36 MPa; initial condensation pressure
• gas/condensate ratio adopted for gas 26.6 MPa; Z factor in reservoir conditions 0.89;
condensate studies; gas/condensate ratio 465 cm3/m3; stabilized
• stabilized condensate density; condensate density 755.8 kg/m3; molar mass of
• initial retrograde condensation pressure; stabilized condensate 130 g/mol.
• results of constant volume depletion The technical report provides the following
studies. reservoir gas composition,% mol.: N2 – 0.734;
CO2 – 0.413; C1 – 79.745; C2 – 7.135; C3 – 4.885;
Example of building PVT model iC4 – 1.187; nC4 – 1.165; C5 – 0.988; C6 – 0.499;
of gas condensate mixture C7 – 1.152; C8 – 0.577; C9 – 0.391; C10 – 0.276;
Reservoir gas condensate mixtures are divided into C11+ – 0.853. The molar fraction of C11+ is close
the following groups by the stabilized condensate to 1 %, and a zone can be identified near the
(C5+) content [12]: initial condensation pressure on the saturated
• low condensate – max. 25 g/m3; condensate reservoir loss curve (fig. 2), in which
• medium condensate – 25 to 100 g/m3; condensate is formed very gradually. We could

English-language digest
Proceedings of Gas Industry Research and Development 63

Saturated condensate loss volume, cm3/stand. m3 0.05

0.04

0.03

0.02

experiment
0.01 calculated value before adjustment
calculated value after adjustment
0
2.5 7.5 12.5 17.5 22.5 27.5
Pressure, MPa

Fig. 2. Saturated condensate loss curve following modeling of constant volume depletion process

not accurately adjust to the initial condensation properties of the reservoir hydrocarbon system
pressure when adjusting the PVT model for the both for hydrodynamical modeling of deposit
saturated condensate loss curve with the above exploitation utilizing a multicomponent filtration
component composition. The calculated value of model and when using pseudobinary models such
the initial condensation pressure was lower than as black oil with regard to condensate solubility
the experimental value. in the gas phase (with a live oil + wet gas option).
In such cases, it is necessary that the molar Consider specifics of data preparation for their
fraction of the last fraction in the PVT model further use in hydrodynamical modeling.
does not exceed 0.1 % to enable simultaneous
adjustment to the initial condensation pressure 1. Multicomponent filtration model
and the saturated condensate loss curve. For When a PVT model is used in the compositional
this reason, the C11+ fraction was divided into modeling of a field exploitation process, the
five pseudofractions mathematically using the number of components in the reservoir mixture
Whitson gamma distribution method [13, 14] so model will be very important, since the speed of
that the molar fraction of the last fraction is less field exploitation process calculations depends on
than 0.1 %. After the division, the pseudofractions it. For this reason, before using a PVT model built
had the following molar fractions, % mol: by the proposed method, it should be modified by
1C11+ – 0.514; 2C11+ – 0.170; 3C11+ – 0.104; reducing the number of components in the mixture
4C11+ – 0.046; 5C11+ – 0.020. Pseudofraction through grouping. The component grouping
properties were calculated using the method methods are described in works [2, 11].
described in work [11]. Some recommendations concerning compo-
Thus, using the algorithm proposed, it was nent grouping rules are:
possible to adapt the PVT model to the main • gas (CO2, N2, H2S, C1–C4) and oil (C5 and
properties of gas condensate mixtures with high heavier) components cannot be grouped with each
accuracy (the maximum error was less than 1 %), other, since it may affect the redistribution of the
including saturated condensate loss curves during vapor-liquid equilibria in the system;
studying the constant volume depletion process • it is recommended to group gas compo-
(see fig. 2). nents as follows: N2–C1, CO2–H2S–C2, C3–C4;
• if gas injection is planned (for example,
Building relationships between PVT properties CO2), this component should not be grouped with
of gas condensate mixture and pressure the others;
for field exploitation modeling • it is recommended to group heavy
PVT models built using the method proposed in components so that the law of molar fraction
the article provide reliable information about the distribution vs. the molecular mass of components

Vesti Gazovoy Nauki scientific and technical collected book

64 Issue 2: 2015 collection

is maintained. For more accurate modeling of For building the required relationships,
retrograde processes near the initial condensation data from the constant volume depletion (CVD)
pressure, several last fractions should not be process at the reservoir temperature and the infield
grouped together. separation configuration, in which virtually all
After grouping components into pseudo- the C5+ group fraction will be transformed from
components, it is recommended to check how the gas phase to the liquid phase (condensate),
accurate the modified PVT model reproduces the are used. For that purpose, the temperature at one
main properties of the gas condensate mixture. of the separation stages should be –25°C or less.
If necessary, the model should be adjusted using Such an approach allows to correctly reproduce
the method described above. stabilized condensate production.
We now provide possible grouping of mixture Obtaining the required relationships vs. pres-
components for the above example. The grouping sure is implemented in existing computer simu-
resulted in the following component composition lators (for example, Eclipse, VIP, MORE,
of the mixture, % mol.: (N2+C1) – 80.479; PVTsim). The gas and liquid phase property values
(CO2+C2) – 7.548; (C3+C4) – 7.237; (C5+C6) – 1.487; are determined by modeling stage separation at
(C7+C8+C9+C10) – 2.396; (C11+C12+C13+C14) – each step of the CVD process for the reservoir gas
– 0.6122; (C15+C16+C17) – 0.1206; C18+ – 0.1202 condensate mixture.
(a total of 8 components). After component Implementing this method requires the use
grouping, mixture properties changed insig- of an adequate multicomponent thermodynamic
nificantly and were adapted using the method model of the reservoir gas condensate system of
proposed in this article. Fig. 3 shows the saturated the exploitation target under study. The building
condensate loss curve before and after component algorithm for such models is described above.
grouping.
***
2. Black oil type model (pseudobinary model In conclusion, we can make the following
considering condensate solubility in gas phase) observations.
An adequate multicomponent PVT model of a An engineering method is proposed for
gas condensate system can be efficiently used for building a multicomponent model of reservoir
building relationships between PVT properties of gas condensate mixture that does not require
a natural hydrocarbon system and pressure during the use of the regression analysis procedure
the design of the field exploitation process using a to adequately describe PVT properties when
black oil model with the live oil + wet gas option. using state equations. The method is based on

0.06
Saturated condensate loss volume, cm3/stand. m3

0.05

0.04

0.03

0.02
experiment
0.01 calculated values before grouping
calculated values after grouping

0
2.5 7.5 12.5 17.5 22.5 27.5
Pressure, MPa

Fig. 3. Saturated condensate loss curve calculated using model

before and after component grouping

English-language digest
Proceedings of Gas Industry Research and Development 65

successive identification of composite model for reserves calculations and field exploitation
parameters in order to accurately reproduce design purposes.
the main results of experimental studies of The method can be efficiently used for
representative samples. designing field exploitation using both composite
The method developed allows to correlate models and black oil type models with the live
reservoir gas condensate system properties used oil + wet gas option.

References
1. Tokarev D.K. Adapting scheme for creation of 8. Brusilovskiy A.I. Rational approach to forming
adequate models for gas-condensate systems models of stratum oils for hydrodynamic
(by an example of Achimov deposits at the calculations at designing and monitoring of
Urengoy oil-and-gas-condensate field) / development / A.I. Brusilovskiy, A.N. Nugayeva,
D.K. Tokarev, D.G. Fateev, A.G. Kozubovskiy I.Ye. Khvatova // Vestnik TsKR Rosnedra. –
et al. – Territoriya Neftegaz. – 2012. – № 12. – 2009. – № 4. – P. 48–56.
P. 46–53.
9. Whitson C.H. Characterizing hydrocarbon plus
2. Brusilovskiy A.I. Procedure for application fraction / C.H. Whitson // SPE Journal. – 1983. –
cubic equations of state in modelling natural P. 683–694.
gas-condensate mixtures / A.I. Brusilovskiy. –
10. R Gazprom 2-3.3-303-2009. Guidance for
Gazovaya promyshlennost’. – 2004. – № 4. –
studding natural gas-condensate systems aimed
P. 16–19.
at calculation of balance and recoverable
3. Shchebetov A.V. Quality evaluation and modelling reserves of natural gas components, designing,
of gas-condensate studies in terms of initial data analysis and control of OAO “Gazprom” fields’
uncertainty / A.V. Shchebetov, M.V. Galkin. – development. – Moscow: Gazprom OJSC:
Gazovaya promyshlennost’. – 2009. – № 9. – Gazprom VNIIGAZ LLC, 2009. – 39 p.
P. 40–44.
11. Brusilovskiy A.I. Phase transformations
4. Zurita A.R. An efficient tuning strategy to calibrate at development of oil and gas fields /
cubic EOS for compositional simulation / Rafael A.I. Brusilovskiy. – Moscow: Graal. – 2002. –
Aguilar Zurita, W.D. McCain // SPE 77382. – 575 p.
2002.
12. Poroskun V.I. Principles of oil and fuel
5. Schebetov A. Quality check of gas-condensate gases reserves classifying and calculating /
PVT studies and EOS modeling under input data V.I. Poroskun, G.A. Gabrielyants, Yu.A. Podturkin
uncertainty / A. Schebetov, A. Rimoldi, M. Piana // et al. // Informational and analytical bulletin:
SPE 133258. – 2010. supplement to “Nedropolzovaniye XXI vek”
magazine. – Moscow: National Association for
6. Hosein R. Extended analysis for gas condensate
Subsoil Use Auditing. – 2007. – 40 p.
systems / Raffie Hosein, W.D. McCain // SPE
Reservoir Evaluation and Engineering. – 2009. – 13. Whitson C.H. Gas condensate PVT – What’s really
P. 163–166. important and Why? / С.H. Whitson, O. Fevang,
T. Yang // Proc. of Optimization of gas condensate
7. Joergensen M. Optimization of pseudo-
fields conference. – London: IBC, 1999.
component selection for compositional studies
of reservoir fluids / M. Joergensen, E.H. Stenby // 14. Whitson C.H. Phase behavior / C.H. Whitson,
SPE 30789. – 1995. M.R. Brule // SPE monograph volume. – Texas,
2000. – (Henry L. Doherty series).

Vesti Gazovoy Nauki scientific and technical collected book