Petroleum 3 (2017) 340e347

Contents lists available at ScienceDirect

Petroleum
journal homepage: www.keaipublishing.com/en/journals/petlm

Robust correlation to predict dew point pressure of gas

condensate reservoirs
Mohammad Ali Ahmadi a, *, Adel Elsharkawy b
a
Department of Petroleum Engineering, Ahwaz Faculty of Petroleum Engineering, Petroleum University of Technology, Ahwaz, Iran
b
Department of Petroleum Engineering, College of Engineering & Petroleum, Kuwait University, Kuwait

a r t i c l e i n f o a b s t r a c t

Article history: When the bottom-hole flowing pressure in a gas condensate reservoir drops below the dew point
Received 31 December 2015 pressure, liquid starts to build up around the well bore resulting in gas productivity decline.
Received in revised form For this reason it is important to be able to accurately either measure or estimate the dew point
23 April 2016
pressure. The condensate formed in the reservoir will not flow until its saturation reaches the
Accepted 3 May 2016
critical saturation and in many cases it might not be entirely recovered. It order to maximize gas
production and condensate recovery, the reservoir pressure must be maintained close to the dew
Keywords:
point pressure. Several attempts have been made to predict the dew point pressure in case the gas
Dew point pressure
Gene Expression Programming
sample becomes unavailable or measured value is unreliable. Unfortunately, most of these attempts
Condensate gas have minor success rates and are based on limited data.
Modeling In this paper we present a robust, cheap, and easy model for predicting the dew point pressure
for gas condensate reservoirs. The new model is an intelligent based model called “Gene Expression
Programming” that is carried out to generate a precise and accurate correlation to estimate the dew
point pressure in condensate gas reservoirs. The new model has been trained and tested using a
large data bank collected for the literature. Precision of the suggested correlation has been
compared to published correlations. The validity of this model has also been compared to exper-
imental data and other published correlations.
Copyright © 2016, Southwest Petroleum University. Production and hosting by Elsevier B.V. on
behalf of KeAi Communications Co., Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction Because of their inverse thermo-dynamical behaviors, having

Huge amounts of time, energy, and money are routinely spent
and dedicated for the model to carefully, characterize effective
gas condensate reservoirs as the most valuable types of hydro-
carbon reservoirs. Gas deliverability losses as a result of
condensate blockage; represent great challenges to the project
economically. Therefore, a great degree of accuracy is definitely
required to obtain and estimate vital parameters such as dew
point pressure of gas condensate reservoirs [1e6].
* Corresponding author. Tel.: þ98 912 6364936.
E-mail address: ahmadi6776@yahoo.com (M.A. Ahmadi).
Peer review under responsibility of Southwest Petroleum University.
Production and Hosting by Elsevier on behalf of KeAi
a very detailed knowledge and observing physical trends of
reservoir temperature and pressure, as two most influencing
parameters, through proposing cheaper, faster and non-
laboratorial methods such as empirically derived equations,
equation of state (EOS) and gained from cutting-edge mathe-
matical solutions have mostly drawn related experts' attentions
[7e16]. For instance, researchers have made great attempts to
put forward techniques based on modern soft computing
methods to measure factors such as dew point pressure which is
defined as a pressure type border that passing it causes forming
condensate bank in vicinity of the well bore, although the
referred vital, critical and crucial attribute which behaves
extremely non-linear can be reported from running some con-
ventional, time-consuming and very much expensive experi-
ments such as constant composition expansion (CCE) and
constant volume depletion (CVD) which their generated results
might adversely be affected with some inherent technical and
monetary constraints such as restricted experimental budgets,

http://dx.doi.org/10.1016/j.petlm.2016.05.001
2405-6561/Copyright © 2016, Southwest Petroleum University. Production and hosting by Elsevier B.V. on behalf of KeAi Communications Co., Ltd. This is an open
access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 M.A. Ahmadi, A. Elsharkawy / Petroleum 3 (2017) 340e347
having not abilities to acquire enough representative samples,
analyzing incompletely due to the limited number of samples
and natural errors of each test [17e26].
Calculations of dew point pressures (DPP) for retrograde gas
condensate reservoirs is possible using the equation of state
(EOS) which requires splitting the heptane plus fraction into
several pseudo-fractions, characterization of each of the sub-
fractions, and choice of the EOS. Even with proper character-
ization and splitting schemes, there is no guarantee that the EOS
calculations are accurate unless the EOS is properly tuned to
some experimental data such as constant volume depletion or
constant composition expansion. The DPP can also be calculated
using equilibrium ratios (K-values). However, the method in-
volves trial and error as the K-values are dependent on the
pressure and most of K-value approaches at high pressures are
not precise [29].
Kurata and Katz (1942) studied the behavior of gas conden-
sate reservoirs and volatile oils. They tried to correlate the DPP to
fluid properties using measurements of 49 gas condensate
samples [28]. Olds et al. (1945 and 1949) made the first attempts
to estimate the DPP for gas condensate reservoirs. Olds et al.
(1945) investigated the gas condensate fluid behavior of Paloma
filed [29,30]. They found that the dew point pressure is strongly
dependent of reservoir fluid composition. They observed that the
removal of the intermediate fraction from the mixture resulted
in a significant increase in the DPP. They also observed that the
temperature has little effect on the DPP compared to the effect of
the intermediate fraction. Olds et al. (1949) also studied the
behavior of San Joaquin Valley fields' gas condensate samples.
They presented a graphical correlation of DPP as a function of API
gravity, and gaseoil ratio (GOR) of the produced condensate. This
correlation has a limited application as it covers a very narrow
range of gas condensate compositions. Later, Reamer and Sage
(1950) made an attempt to correlate the DPP to gas condensate
fluid properties using five gas condensate samples [31]. They
presented various diagrams showing the effect of temperature
and GOR on DPP. Reamer and Sage could not establish a useful
correlation owing to the complication of the affection of
composition on the DPP.
Organick and Golding (1952) presented a graphical correla-
tion to estimate the saturation pressure for gas condensates and
volatile oils as a function of modified weight average equivalent
molecular weight and a molal average boiling point [32]. Their
graphical correlation is presented in the form of 14 working
charts, each including a group of plots that cannot be used for
computer calculations. Organick and Golding stated that simple
mixtures and pure components cannot be properly predicted
using their correlation.
Nemeth and Kennedy (1967) were the first to present a
mathematical correlation to predict the DPP for gas condensate
reservoirs [33]. They correlated the DPP to reservoir temperature
and gas condensate compositions. Later, Elsharkawy (2001 and
2002) presented another mathematical correlation to determine
the DPP for gas condensate reservoirs that is based on a large
data bank of experimental measurements and collected DPP
[12,27]. Elsharkawy also presented calculations of DPP using SRK
and PR EOS and two different characterization schemes of the
heptane plus fraction. His correlation showed that DPP calcula-
tion is dependent on EOS and characterization methods.
Morch et al. (2006) presented measurements and EOS
modeling of DPP of five synthetic natural gas condensate sam-
ples [34]. Their results showed a considerable deviation between
DPP calculated by EOS and measurements. Esmaeilzadah and
Samadi (2008) also made attempts to study gas condensate
behavior using EOS [35].
341
Loul et al. (2012) presented measurements and predictions of
DPP of five synthetic natural gas mixtures [36]. They used PR EOS
to model the DPP curve for five synthetic and two real gas
condensate samples. Loul et al. found that the binary interaction
number has a great effect on the DPP calculations using the EOS.
Hosseinkhni et al. (2014) used an expert system to calculate the
DPP for gas condensate reservoirs. However, they did not present
neither a mathematical model nor a graphical method to calcu-
late the DPP for gas condensate reservoir [37]. Finally, Galatro
and Cordero (2014) also measured the DPP of gas condensate
samples [38]. They presented a comparison of DPP calculation
using various schemes of characterization and binary interaction
number, and PR and SRK EOS. Marruffo el al. also presented a
model to estimate dew point pressure and heptane plus content
of gas condensate [39].
In addition, it has been noticed that EOS's are not acceptably
predicting the behavior of gas condensates [40e44]. Therefore,
many attempts based on specified advantages of soft computing
approaches such as dealing effectively with non-linearity, am-
biguity and uncertainty of the referenced issue have been made
to overcome hurdles of petroleum engineering issues such as
phase behavior determination and to present more handy, pre-
cise and appropriate solutions [45e55]. For example, a special
kind of an ANN to predict Pd through taking a set of thermo
dynamical and compositional parameters as input was imple-
mented by Jalali et al. [43]. Besides, Nowroozi et al. conducted an
ANFIS to foresee Pd by considering mostly compositional factors
[53].
The determination of the dew point pressure from constant
volume depletion and constant composition expansion is very
expensive and time consuming. Therefore, developing a new
high performance and precise correlation for estimating dew
point pressure for gas condensate reservoirs is the objective of
the current study. Throughout this work Gene Expression Pro-
gramming (GEP) is proposed to develop a new and user friendly
correlation with a high degree of efficiency for estimating the
dew point pressure (Pdew) in condensate gas reservoirs. This
study overcomes previous problems as it has successful been
applied to the predicting of many reservoir engineering param-
eters. Gas condensate data from literature, as well as from the
northern Persian Gulf oil fields located in the south of Iran are
used for the development of the proposed model, testing its
accuracy, and validity in comparison to the previously published
methods.
2. Gene Expression Programming (GEP) approach
2.1. Genetic programming
Genetic programming (GP) is a subset of evolutionary based
methods with a countless aptitude to spontaneously develop
programs of computer. GP inspired from principle of Darwinian
natural selection for the addressed evolutionary algorithm same
as genetic algorithm. The theory of genetic programming (GP)
was firstly proposed and evolved by Koza after such tests on
symbolic regression. The main distinction between genetic
programming and original genetic algorithm summarized to the
demonstration of the final solution. The referred target outcomes
of genetic programming (GP) are computer based softwares
which are demonstrated as tree topologies and formulated in a
language of function based programming [47] while original
genetic algorithm generates a string of numbers which illustrates
the final solution. Put it this way, through genetic programming
(GP), the developing individuals are parse trees than can
assorted in length all the time run in favor of fixed-length binary
342 M.A. Ahmadi, A. Elsharkawy / Petroleum 3 (2017) 340e347

strings. To develop the superior values for a submitted set of the
model variables, conventional optimization approaches like as
the genetic algorithm (GA) are mostly implemented in parameter
optimization. Genetic programming (GP) gives the basic topol-
ogy of the rough approach together with the values of its pa-
rameters. Based on the fitness value determined via fitness
function proposed in Genetic Programming, GP optimizes a
population of aforementioned computer program [55e62].
The addressed original genetic programming (GP) model is
alluded to as tree-based genetic programming (GP). Each
member in referred tree-based genetic programming (GP) is a
ranked topology tree containing functions and required termi-
nals. The implemented functions and terminals through
addressed tree are collected from an assortment of proposed/
introduced function groups and a group of terminals. To clarify
this point, the considered function through addressed tree may
comprise the Boolean logic functions, basic math operations or
any other math functions such as þ, , , /, AND, OR, NOT.
Moreover, the implemented terminal assortment T consisting of
the arguments for the functions and can contain logical con-
stants, variables, numerical constants, etc. To generate computer
approach in a tree-like topology with an origin point with
branches expanding from each function and closing in a termi-
nal, the addressed functions and terminals are selected randomly
and constructed together.
It should be noted that, Gene Expression Programming (GEP)
is a linear branch of addressed GP. The linear branch of GP as-
sembles an explicit difference between the phenotype and the
genotype of an individual. Therefore, the individuals are expli-
cated as linear series [55e62].
2.2. Gene Expression Programming (GEP)
Gene Expression Programming (GEP) developed firstly by
Ferreira [59] which is a natural development of genetic pro-
gramming (GP) [55]. Most of the implemented GEP operators are
the same as genetic algorithm (GA) with minor alterations. Main
elements of Gene Expression Programming (GEP) are as follow:
terminal set, control parameters, function set, fitness function,
and stop criteria [59]. GEP implements a constant length of
character series to demonstrate routs to the targets that are
thence presented as parse trees of various shapes and sizes. It
worth mentioning that, these trees are called Gene Expression
Programming (GEP) expression trees (ETs). It worth bearing in
mind that, one of the major benefits of GEP approach is summa-
rized to the generation of genetic diversity is extremely facilitated,
as genetic operators act at the chromosome level [46]. Unique
multi-genic nature is robust advantage of GEP which permits the
evolution of more complex programs comprised of various sub-
programs [53]. Any GEP gene comprises a list of symbols with a
fixed length which can be any element from a function set such as
{þ,,,/,√} and a terminal set such as {X,Y,8} [55e62].
The terminal set and the function set have to have the fold
characteristic: every function has to be capable to carry any value
of data type which can be restored by a function or presumed by
a terminal. To clarify this statement following set represents a
generic GEP gene with the terminal sets and given function:
The conversion tackles to the first position in the K-expres-
sion which relevant to the solution of the expression tree (ET)
and pores in the string one by one [53]. The above GEP statement
can also be represented in a mathematical formulation form as
below:
 X Y
X
(2)
Y
As illustrated previously, GEP genes contain consistent length
which is figured out at first for a given issue. Hence, what assorts
in GEP is not the length of the genes whereas the size of the
relevant ETs [46e54]. It is worth bearing in mind that there exist
a specific number of suspend components which are not
appropriate for genome mapping. Therefore, the length of GEP
gene may be same or greater than the valid length of a K-
expression. GEP utilizes a headetail approach to assure the
truthfully of a randomly collected genome. Each GEP gene is
comprised of a head and a tail while the head may consist both
function and terminal symbols, although the tail may have only
terminal symbols.
3. Data gathering
Before explaining the proposed correlation for the determi-
nation of dew point pressure in retrograde gas reservoirs, it
would be better to study the relationship between the addressed
target and input variables. Figs. 1e3 explain the dependency of
the dew point pressure in gas condensate reservoirs on the
reservoir temperature, Specific gravity of C7þ and molecular
weight of C7þ, correspondingly. As illustrated in Fig. 1, the low
correlation coefficient between dew point pressure and the
reservoir temperature demonstrates that reservoir temperature
has low effect on the mentioned output. Furthermore, Fig. 2 in-
dicates that increasing the amount of C7þ molecular weight has a
significant influence and increases the dew point pressure. This
agrees with previous studies that the molecular weight of C7þ
has the greatest impact on the dew point pressure in condensate
gas reservoirs. Finally, Fig. 3 shows that increasing C7þ specific
gravity results in an increase in the dew point pressure. The
correlation coefficient of the fitted line between C7þ specific
gravity and dew point pressure indicates that the C7þ specific
gravity has also an important impact on the value of dew point
pressure in retrograded gas reservoirs.
The details of the gas condensate samples used in this study
are reported in Table 1. This table shows that the data used in
the model development is covering a pressure range of

^:=:^::X:Y:X:Y (1)

where X and Y are variables; this point should be noted that “.” is
component of separating to simplify reading. The addressed
presentation is called a K-expression [55e62]. A K-expression
can be depicted as a layout which known an expression tree (ET)
in GEP. Fig. 1. Dependency of the DPP on the reservoir temperature.
 M.A. Ahmadi, A. Elsharkawy / Petroleum 3 (2017) 340e347 343

input variables (temperature, gas composition, MWC7þ, SGC7þ).

The finalized outcome dew point pressure correlation as
following as:

Pd ¼ 888:278  T  C1  3:60639 þ ðTÞ2  0:00785623

þ C1  1467:87 þ A  0:989073
(3)

where the coefficients A, B and C can be calculated from the

following equations:
   
A ¼ 140:909  N2 ð1=3Þ  B  0:484983 þ N2 ð1=3Þ  C
 
Fig. 2. Dependency of the DPP on the molecular weight of C7þ.
 0:576219 þ N2 ð2=3Þ  1746:2 þ B  0:290811
 
þ B2  3:33869  1005 þ C  0:484502
(4)
   
B ¼ 6991:4 þ MWC7þ 3  0:00165952  MWC7þ 3
   
 C7þ ð1=3Þ  0:00122815  MWC7þ 6  5:7182
     
 1011  C4 ð1=3Þ  C7þ ð1=3Þ  79241 þ C4 ð2=3Þ
 
 31517:4 þ C7þ ð1=3Þ  50917:8  C7þ ð2=3Þ  31614:2
(5)
 
C ¼ 29014  C1  52127:9 þ C1  SGC7þ 3  79848:3 þ C1
     
Fig. 3. Dependency of the DPP on the molecular weight of SG (C7þ).  C7þ ð1=3Þ  12633:6 þ C1 2  11116:5  SGC7þ 3
   
 58526:6 þ SGC7þ 3  C7þ ð1=3Þ  58263:6
Table 1  
Parameters of statistics of the DPP data employed in this study.  C7þ ð2=3Þ  43792:7
Variables Min Max Average (6)
Dew point pressure, psia 1400 10,790 4855.778
In the above correlation, T denotes the reservoir temperature
Temperature,  F 76 320 216.84
Hydrogen sulfide, mole fraction 0 0.2986 0.006547
and in  F, C1 to C7þ and N2 represent the gas composition in
Carbon dioxide, mole fraction 0 0.9192 0.018655 mole percent, MWC7þ and SGC7þ are the molecular weight and
Nitrogen, mole fraction 0 0.4322 0.011896 specific gravity of the heptane plus fraction, and the calculated
Methane, mole fraction 0.0349 0.9668 0.793606 dew point pressure is in psia. It worth mentioning that, by
Ethane, mole fraction 0.0037 0.1513 0.061711
performing sensitivity analysis between the gained output re-
Propane, mole fraction 0.0011 0.109 0.031852
Butanes, mole fraction 0.0022 0.203 0.020132 sults of the above correlation and equivalent actual DPP data
Pentanes, mole fraction 0.0006 0.055 0.010608 samples, the accurate numbers of the digits of the coefficients
Hexanes, mole fraction 0.0005 0.051 0.008054 and constants were determined. The exploited functions and
Heptanes-plus, mole fraction 0.0019 0.1356 0.038114
terminals through GEP method for deriving the above dew
Molecular weight C7þ 109 225 151.2775
Specific gravity C7þ 0.737 0.868 0.792548
point pressure correlation are summarized in Table 2. Table 2
explicates the used algorithm parameters and variables
comprising the numbers of handled chromosomes, genes, and
the utilized gene operators such mutation, inversion and
recombination.
1400e10,790 psia and a temperature from 76 to 320  F. The
data bank contains some sour gas condensate that has as high
as 30% H2S and others having 92% CO2. It also covers a very 5. Results and discussion
wide range of gas condensate compositions; some samples are
very lean and have methane as high as 91% and low heptane 5.1. Testing the model
plus, other are very rich and have a heptane content as high as
13%. Fig. 4 shows a comparison of the outputs gained from the
proposed correlation and the actual DPP data relative to real data
index. As shown in this figure, the gained outcomes of the newly
4. Proposed model developed correlation totally match the relevant measured real
dew point pressure data. In order to prove the effectiveness and
The GEP methods indicate the necessary variables, which lead precision of the above proposed correlation in dew point pres-
to the most robust integration of accurate correlations from the sure estimation, the cross-plot of the predicted data by the
344 M.A. Ahmadi, A. Elsharkawy / Petroleum 3 (2017) 340e347

Table 2 robustness and integrity of the studied approach in the deter-

The Gene Expression Programming (GEP) parameters executed in computational mination of required target [35e43]. In addition, the relative
steps.
deviations of the predicted values from the proposed model to
GEP algorithm parameters Value the actual dew point pressure data are shown in Fig. 6. As shown
Number of chromosomes 40 this figure, the maximum relative error for our developed cor-
Head size 8 relation is 30% which accounts to 2400 psi pressure however
Number of genes 8 some other points have the same deviations. Finally, to have a
Linking function Plus (þ)
fair comparison between our proposed correlation and others
Generations without change 2000
Fitness function Mean square error (MSE) which use similar input parameters. Figs. 7 and 8 depict the
Mutation 0.044 gained statistical parameters of the all used models including
Inversion 0.1 average absolute relative deviation percent (AARD%) and relative
IS transposition 0.1
mean square error (RMSE), respectively. Furthermore, the values
RIS transposition 0.1
One-point recombination 0.3
of the aforementioned parameters (RMSE and AARD) are re-
Two-point recombination 0.3 ported in Table 3. This table shows that the proposed model has
Gene recombination 0.1 the highest correlation coefficient (97%) and the lowest AARD of
Gene transposition 0.1 8% followed by Nemeth and Kennedy correlation (1967) that has
Constant per gene 2
a correlation coefficient of 64% and an AARD of 35%. The other
Operators used þ
e methods reported in Table 3 have low coefficients of correlations
/ and high AARD.

Power
5.2. Validity of the model

To validate the correlation proposed in this paper, measure-

ments of dew point pressure versus temperature of condensate
gas samples from China reported by Liu et al. (2013) [63] are
used. Fig. 9 depicts the variation of the dew point pressure versus
corresponding temperature for one of the previously mentioned
gas samples. The performances of the proposed model, as well as

Fig. 4. Contrast between proposed correlation output and experimental DPP versus
relevant data index.

Fig. 6. Relative deviations of the developed correlation outcomes versus corre-

sponding dew point pressure data.

Fig. 5. Scatter plot of the developed correlation outputs versus corresponding DPP.

model versus the actual measurements is shown through Fig. 5.

As demonstrated in Fig. 5, the predicted outcomes from the
suggested correlation entirely follow the fitted straight line with
high value of correlation coefficient (96.7%). Most of the data Fig. 7. Comparison between RMSE of the outputs gained via conventional and
points fall closely to the 45 line indicating a high degree of proposed models.
 M.A. Ahmadi, A. Elsharkawy / Petroleum 3 (2017) 340e347 345

Fig. 8. Comparison between AARD percent of the outputs gained via conventional
and proposed models. Fig. 10. Relative importance of each input variable on the dew point pressure in the
retrograded gas condensate reservoirs.

Table 3
Parameters of statistics of the developed correlation against routine correlations.

Approach RMSE R2 AARD% ANOVA analysis agree with Elsharkawy (2003) sensitivity studies
Nemeth and Kennedy (1967) 770.657 0.64 35.55
of the effect of temperature, heptane plus molecular weight, and
Elsharkawy (2001) 903.487 0.32 50.13 specific gravity on the predicted dew point pressure by his model
Humoud and Al-Marhoun (2001) 768.589 0.4761 38.75 and both SRK EOS and PR EOS.
Marruffo (2001) 823.753 0.3844 40.76
ACEa model (2011) 889.875 0.3364 45.68
Our proposed correlation 508.49 0.9688 8.1003 6. Conclusions
a
Alternating conditional expectations (ACE).
The accurate prediction of dew point pressure for gas
condensate reservoirs is very important to maintain the
reservoir pressure close to the dew point ensuring maximum
gas production and minimum condensate blockage. For this
reason, enormous attempts have been made in this work to
develop a novel and accurate model to predict dew point
pressure for retrograded gas condensate reservoirs. The new
model is developed from a large data bank that is comprised of
a larger range of pressures, temperatures, and various gas
condensate compositions. The new model has the following
features:

(1) Validity of the model has been studied using measured

Fig. 9. Variation of dew point pressure versus corresponding temperature for gas
condensate samples from China gas field reported by Liu et al. (2013) [63].
the previously published correlations are compared to the
measured dew point pressure in Fig. 9.
This figure indicates that the newly developed model
perfectly matches the measured data.
5.3. Analysis of Variance (ANOVA)
It is important in any model development to study the rela-
tive importance of the input variables. To depict the relative
significance of each input parameters on the DPP, the effective
statistical approach which is normally used through the “Anal-
ysis of Variance (ANOVA)” was carried out. Fig. 10 shows the
ANOVA analysis for the dew point pressure in condensate gas
reservoirs. It is clear, in Fig. 10, that the molecular weight of C7þ
fraction has the maximum effect on the dew point pressure
followed by the specific gravity of the plus fraction, methane
content, and reservoir temperature. Other important input var-
iables such as mole fraction of C2eC6 have little effect. The
data. The new model shows acceptable agreement be-
tween forecasted dew point pressure values and the
measured ones.
(2) Precision of the correlation is compared to previously
published correlations. The newly proposed performed
better than all published correlations. It has the highest
correlation coefficient and the lowest errors.
(3) ANOVA analysis has been studied for the data bank and
the proposed model. The significance of the input vari-
ables agrees with equation of state calculations and pre-
viously published literature.
(4) The proposed dew point pressure model that is based on
GEP is user friendly, robust and cheap and can be imple-
mented in the full-field reservoir simulators for conden-
sate gas production schemes.
(5) The new model can be used for planning CCE tests and to
predict the dew point pressure for gas condensate reser-
voirs when the measured one is unavailable.
References
[1] F.S. Alavi, D. Mowla, F. Esmaeilzadeh, Production performance analysis of
Sarkhoon gas condensate reservoir, J. Pet. Sci. Eng. 75 (1e2) (2010) 44e53.
[2] M. Jamiolahmady, H. Mahdiyar, P. Ghahri, M. Sohrabi, A new method for
productivity calculation of perforated wells in gas condensate reservoirs,
J. Pet. Sci. Eng. 77 (3e4) (2011) 263e273.
346 M.A. Ahmadi, A. Elsharkawy / Petroleum 3 (2017) 340e347
[3] M.R. Maleki, F. Rashidi, H. Mahani, E. Khamehchi, A simulation study of the
enhancement of condensate recovery from one of the Iranian naturally
fractured condensate reservoirs, J. Pet. Sci. Eng. 92e93 (0) (2012) 158e166.
[4] J. Mazloom, F. Rashidi, R. Kelly, A novel approach to the calculation of the
distance of a sealing fault from a well in a gas condensate reservoir, J. Pet.
Sci. Eng. 45 (3e4) (2004) 149e158.
[5] J.R. Ursin, Fluid flow in gas condensate reservoirs: the interplay of forces
and their relative strengths, J. Pet. Sci. Eng. 41 (4) (2004) 253e267.
[6] L. Vega, M.A. Barrufet, Analysis of a non-volumetric gas-condensate
reservoir using a generalized material balance equation with fluid prop-
erties from an equation of state, J. Pet. Sci. Eng. 48 (1e2) (2005) 105e119.
[7] L.F. Ayala, J.P. Kouassi, The similarity theory applied to the analysis of
multiphase flow in gas-condensate reservoirs, Energy Fuels v (2007)
1592e1600.
[8] A.M. Elsharkawy, S.G. Foda, EOS simulation and GRNN modeling of the
constant volume depletion behavior of gas condensate reservoirs, Energy
Fuels 12 (2) (1998) 353e364.
[9] F. Gozalpour, A. Danesh, A.C. Todd, B. Tohidi, Viscosity and density data of
two North Sea gas condensate samples at temperatures to 423 K and
pressures to 140 MPa, J. Chem. Eng. Data 46 (5) (2001) 1305e1308.
[10] S. Zendehboudi, M.A. Ahmadi, L. James, I. Chatzis, Prediction of condensate-
to-gas ratio for retrograde gas condensate reservoirs using artificial neural
network with particle swarm optimization, Energy Fuels 26 (6) (2012)
3432e3447.
[11] M. El Aily, M.H.M. Khalil, S.M. Desouky, M.H. Batanoni, M.R.M. Mahmoud,
Experimental studies on constant massevolume depletion of gas-
condensate systems, Egypt. J. Pet. 22 (1) (2013) 129e136.
[12] A.M. Elsharkawy, Predicting the dew point pressure for gas condensate
reservoirs: empirical models and equations of state, Fluid Phase Equilibria
193 (1e2) (2002) 147e165.
[13] D.W. Waples, The kinetics of in-reservoir oil destruction and gas forma-
tion: constraints from experimental and empirical data, and from ther-
modynamics, Org. Geochem. 31 (6) (2000) 553e575.
[14] K.L. Yan, H. Liu, C.Y. Sun, Q.L. Ma, G.J. Chen, D.J. Shen, X.J. Xiao, H.Y. Wang,
Measurement and calculation of gas compressibility factor for condensate
gas and natural gas under pressure up to 116 MPa, J. Chem. Thermodyn. 63
(0) (2013) 38e43.
[15] M. Bonyadi, F. Esmaeilzadeh, Prediction of gas condensate properties by
EsmaeilzadeheRoshanfekr equation of state, Fluid Phase Equilibria 260 (2)
(2007) 326e334.
[16] C. Li, Y. Peng, J. Dong, L. Chen, Prediction of the dew point pressure for gas
condensate using a modified PengeRobinson equation of state and a four-
coefficient molar distribution function, J. Nat. Gas Sci. Eng. 27 (Part 2)
(November 2015) 967e978.
[17] S. Takeya, M. Kida, H. Minami, H. Sakagami, A. Hachikubo, N. Takahashi,
H. Shoji, V. Soloviev, K. Wallmann, N. Biebow, A. Obzhirov, A. Salomatin,
J. Poort, Structure and thermal expansion of natural gas clathrate hydrates,
Chem. Eng. Sci. 61 (8) (2006) 2670e2674.
[18] W. Yan, L.K. Wang, L.Y. Yang, T.M. Guo, A systematic experimental study on
the phase behavior of complex fluid mixtures up to near-critical region,
Fluid Phase Equilibria 190 (1e2) (2001) 159e178.
[19] G.D. Henderson, A. Danesh, D.H. Tehrani, J.M. Peden, The effect of velocity
and interfacial tension on relative permeability of gas condensate fluids in
the wellbore region, J. Pet. Sci. Eng. 17 (3e4) (1997) 265e273.
[20] R.S. Metcalfe, W.J. Raby, Phase equilibria for a rich gas condensate-nitrogen
system, Fluid Phase Equilibria 29 (0) (1986) 563e573.
[21] E. Moaseri, O. Mostaghisi, A. Shahsavand, B. Bazubandi, M. Karimi,
J. Ahmadi, Experimental study and techno-economical evaluation of
Khangiran sour natural gas condensate desulfurization process, J. Nat. Gas
Sci. Eng. 12 (0) (2013) 34e42.
[22] G.W. van Graas, A. Elin Gilje, T.P. Isom, L. Aase Tau, The effects of phase
fractionation on the composition of oils, condensates and gases, Org.
Geochem. 31 (12) (2000) 1419e1439.
[23] X. Wang, K.K. Mohanty, Critical condensate saturation in porous media,
J. Colloid Interface Sci. 214 (2) (1999) 416e426.
[24] G.D. Henderson, A. Danesh, D.H. Tehrani, S. Al-Shaidi, J.M. Peden, Mea-
surement and correlation of gas condensate relative permeability by the
steady-state method, SPE Reserv. Eval. Eng. 1 (2) (1998) 134e140.
[25] J.M. Lombard, D.G. Longeron, F.J.M. Kalaydjian, Influence of connate water
and condensate saturation on inertial effects in gas/condensate reservoirs,
SPE J. 5 (3) (2000) 301e308.
[26] P. Sigmund, A.M. Cameron, Recovery of retrograde condensed liquids by
revaporization during dry gas injection, J. Can. Pet. Technol. 16 (1) (1977)
1e15.
[27] A.M. Elsharkawy, Characterization of the plus fraction and prediction of the
dew point pressure for gas condensate reservoirs, in: SPE Western
Regional Meeting, 26e30 March, Bakersfield, California, 2001. SPE paper
68666.
[28] K. Karata, D.L. Katz, Critical properties of volatile hydrocarbons mixtures,
AIChE J. 38 (1942) 99e1021.
[29] R.H. Olds, B.H. Sage, W.N. Lacy, Phase equilibrium in hydrocarbon system,
composition of dew point, AIME 160 (1945) 77e99.
[30] R.H. Olds, B.H. Sage, W.N. Lacy, Volumetric phase behavior of oil and gas
from several Sna Joaquin Valy field, AIME 179 (1949) 287e302.
[31] H.H. Reamer, B.H. Sage, Phase behavior of hydrocarbon system, AIME 189
(1950) 261e268.
[32] E.I. Organic, B.H. Golding, Prediction of saturation pressure of gas
condensate and volatile oil mixtures, AIME 195 (1952) 135e148.
[33] L.K. Nemeth, H.T. Kennedy, A correlation of dewpoint pressure with fluid
composition and temperature, SPE J. 7 (2) (1967) 99e104.
[34] O. Morch, K.H. Nasrifar, O. Bollland, E. Solbraa, A.O. Fredheim, L.H. Gjersten,
Measurements and modeling of hydrocarbon dew points for five synthetic
natural gas mixtures, Fuild Phase Equilibria 239 (2006) 138e145.
[35] F. Esmaeilzadah, F. Samadi, Modification of Roshanfekr equation of state to
improve volumetric prediction of gas condensate reservoir, Fluid Phase
Equilibria 267 (2008) 113e118.
[36] V. Lou, G. Pappa, C. Boukouvalas, S. Skouras, E. Solrbaa, K. Christine,
E. Voutas, Measurement and prediction of dew point curves of natural gas
mixtures, Fluid Phase Equilibria 334 (2012) 1e9.
[37] H.R. Hosseinkani, F. Esmaeilzadeh, D. Mowla, Application of expert system
fro accurate determination of dew-point pressure of gas condensate res-
ervoirs, J. Nat. Gas Sci. Eng. 18 (2014) 296e303.
[38] D. Glatro, F. Coredro, Consideration for the dew point calculations in rich
natural gas, J. Nat. Gas Sci. Eng. 18 (2014) 112e119.
[39] I. Marruffo, J. Maita, J. Him, G. Rojas, Statistical forecast models to deter-
mine retrograde dew pressure and C7þ percentage of gas condensates on
basis of production test data of eastern Venezuelan reservoirs, in: SPE Latin
American and Caribbean Petroleum Engineering Conference, 2001. Paper
No. 69393, Buenos Aires, Argentina.
[40] K.T. Potsch, L. Braeuer, A novel graphical method for determining dewpoint
pressures of gas condensates, in: European Petroleum Conference, 1996.
Paper No. 36919, Milan, Italy.
[41] M. Al-Dhamen, M. Al-Marhoun, New correlations for dew-point pressure
for gas condensate, in: SPE Saudi Arabia Section Young Professionals
Technical Symposium, 2011. Dhahran, Saudi Arabia.
[42] L.F. Ayala, T. Ertekin, M. Adewumi, Optimized exploitation of gas-
condensate reservoirs using neuro-simulation, in: SPE Asia Pacific Oil
and Gas Conference and Exhibition, 2004. Perth, Australia.
[43] F. Jalali, Y. Abdy, M.K. Akbari, Dewpoint pressure estimation of gas
condensate reservoirs, using artificial neural network (ANN), in: EUROPEC/
EAGE Conference and Exhibition, 2007. London, U.K.
[44] M.A. Ahmadi, Neural network based unified particle swarm optimization
for prediction of asphaltene precipitation, Fluid Phase Equilibria 314 (0)
(2012) 46e51.
[45] M.A. Ahmadi, M. Golshadi, Neural network based swarm concept for
prediction asphaltene precipitation due to natural depletion, J. Pet. Sci. Eng.
v (2012) 40e49.
[46] M.A. Ahmadi, M. Ebadi, A. Shokrollahi, S.M.J. Majidi, Evolving artificial
neural network and imperialist competitive algorithm for prediction oil
flow rate of the reservoir, Appl. Soft Comput. 13 (2) (2013) 1085e1098.
[47] M.A. Ahmadi, M. Pournik, S.R. Shadizadeh, Toward connectionist model for
predicting bubble point pressure of crude oils: application of artificial in-
telligence, Petroleum 1 (04) (2015) 307e317.
[48] M.A. Ahmadi, S. Zendehboudi, M. Dusseault, I. Chatzis, Evolving simple-to-
use method to determine watereoil relative permeability in petroleum
reservoirs, Petroleum 2 (01) (2016) 67e78.
[49] M.A. Ahmadi, S.R. Shadizadeh, New approach for prediction of asphaltene
precipitation due to natural depletion by using evolutionary algorithm
concept, Fuel 102 (2012) 716e723.
[50] M.A. Ahmadi, A. Bahadori, Determination of oil well production perfor-
mance using artificial neural network (ANN) linked to the particle swarm
optimization (PSO) tool, Petroleum 1 (02) (2015) 118e132.
[51] M.A. Ahmadi, Prediction of asphaltene precipitation using artificial neural
network optimized by imperialist competitive algorithm, J. Pet. Explor.
Prod. Technol. 1 (2011) 99e106.
[52] M.A. Ahmadi, A. Bahadori, A LSSVM approach for determining well
placement and conning phenomena in horizontal wells, Fuel 153 (1 August
2015) 276e283.
[53] S. Nowroozi, M. Ranjbar, H. Hashemipour, M. Schaffie, Development of a
neural fuzzy system for advanced prediction of dew point pressure in gas
condensate reservoirs, Fuel Process. Technol. 90 (3) (2009) 452e457.
[54] M.A. Al-Dhamen, New Models for Estimating the Dew-point Pressure of
Gas Condensate Reservoir, SPE paper 155410, 2010.
[55] M.A. Ahmadi, S. Zendehboudi, L. James, A. Elkamel, M. Dusseault, I. Chatsiz,
A. Lohi, New Tools to determine bubble point pressure of crude oils:
experimental and modeling study, J. Pet. Sci. Eng. 123C (2014) 205e214.
[56] A.H. Alavi, M. Ameri, A.H. Gandomi, M.R. Mirzahosseini, Formulation of
flow number of asphalt mixes using a hybrid computational method,
Constr. Build. Mater. 25 (2011) 1338e1355.
~o, M.A. Gonçalves, J.P. Papa, B. Zhang, W. Fan, et al.,
[57] R.S. Torres, A.X. Falca
A genetic programming framework for content-based image retrieval,
Pattern Recogn. 42 (2009) 283e292.
[58] M. Oltean, C.A. Grosan, Comparison of several linear genetic programming
techniques, Complex Syst. 14 (4) (2003) 1e29.
[59] C. Ferreira, Gene expression programming: a new adaptive algorithm for
solving problems, Complex Syst. 13 (2) (2001) 87e129.
[60] A.H. Gandomi, A.H. Alavi, M.R. Mirzahosseini, F. Moghadas Nejad,
Nonlinear genetic-based models for prediction of flow number of asphalt
mixtures, J. Mater. Civ. Eng. ASCE 23 (3) (2011) 1e18.
 M.A. Ahmadi, A. Elsharkawy / Petroleum 3 (2017) 340e347 347

[61] C. Ferreira, Gene Expression Programming: Mathematical Modeling by An
Artificial Intelligence, second ed., Springer-Verlag, Germany, 2006.
[62] M.A. Ahmadi, R. Haghbakhsh, R. Soleimani, M. Bazrgar Bajestani, Estima-
tion of H2S solubility in ionic liquids using a rigorous method, J. Supercrit.
Fluids 92 (2014) 60e69.
[63] H. Liu, C. Sun, K. Yan, Q. Ma, J. Wang, G. Chen, X. Xiao, H. Wang, X. Zheng,
Phase behavior and compressibility factor of two China gas condensate
samples at pressure up to 95 Mpa, Fluid Phase Equilibria 337 (2013)
363e369.