Journal of Natural Gas Science and Engineering 10 (2013) 25e40

Contents lists available at SciVerse ScienceDirect

Journal of Natural Gas Science and Engineering

journal homepage: www.elsevier.com/locate/jngse

Wax formation assessment of condensate in South Pars gas processing plant sea
pipeline (a case study)
M.R. Rahimpour a, *, M. Davoudi b, c, S.M. Jokar a, I. Khoramdel a, c, A. Shariati a, M.R. Dehnavi d
a
Department of Chemical Engineering, School of Chemical and Petroleum Engineering, Shiraz University, Mollasadra Street, Shiraz 71345, Iran
b
Department of Chemical Engineering, Engineering Faculty, Persian Gulf University, Bushehr, Iran
c
Development & Engineering Management Department, South Pars Gas Complex Company, Assaluyeh, Iran
d
Department of Chemical Engineering Shahrood Branch, Islamic Azad University, Shahrood, Iran

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

Article history: The wax deposition from the gas condensate in South Pars gas processing plant causes a number of
Received 3 August 2012 severe problems. These problems include: (1) deposits form on the reboiler tubes of stabilizer column
Received in revised form and tend to reduce its duty (2) forcing periodic shut-down and removal of deposits (3) interrupting
12 October 2012
normal processing operations. An understanding of deposition, nature and propensity is necessary to
Accepted 22 October 2012
Available online 28 November 2012
mitigate the mentioned problems. In this work, the multi solid phase model is used to predict the wax
precipitation from gas condensate fluid. For five different reservoir fluids, several methods were inves-
tigated to split the heavy hydrocarbon fraction into pseudo fractions. The results show that the Al-
Keywords:
Wax
Meshari method is the most accurate one. Also, a set of consistent correlations were used to calculate
Multi solid phase model the critical points, fusion properties and the acentric factor of the single carbon number groups in the
Splitting method extended composition. Finally the best methods for predicting the wax formation are selected and used
Critical and fusion properties to predict the wax formation in the sea line of South Pars gas processing plant. The modeling shows that
South Pars gas processing plant wax precipitation starts at 293 K and 86 bar. At this pressure and temperature the pipeline is 94 km away
from the wellhead.
Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction well, transfer lines, treatment vessels, and storage tanks are all
negatively affected by the presence of paraffin waxes intermingled
Naturally occurring petroleum fluids are a mixture of hydro- with emulsions, solids, and corrosion by-products. Solids and
carbon compounds. The Refined petroleum products may be in waxes frequently combine to form inclusion aggregates, which
liquid, gas, or solid form. Some heavy products such as asphalts and complicate the treatment of waxes by the various methods. The
waxes are in solid form. But, under atmospheric condition most successful production and transfer of waxy oils to the refinery is
hydrocarbons are in liquid form. Certain properties such as critical essential for the profitability of the refinery. The broad volatility
constants, acentric factor, and molecular weight are specifications and melting point range of hydrocarbon components found in
of a compound regardless of being vapor or liquid. However, petroleum fluids cause formations of gas, liquid, and solid phases in
physical properties such as density, transport, or thermal properties response to changes in pressure, temperature, or composition.
depend on the state of the system and in many cases separate In order to proceed with the wax precipitation calculations,
methods are used to estimate these properties. Paraffin waxes are predictive thermodynamic models are needed. Currently, there are
among the many problems encountered in the production, trans- two types of solid models both based on flash type calculations.
port, and refining of gas condensate and crude oil. Some locations One is the solid solution model, and the other is the multi-solid
in the world have such severe problems with wax deposits that phase model.
wells are shut-in for pulling or scraping operations on a weekly or In the solid solution model the precipitation is considered as
monthly basis. The build-up of paraffins and asphaltenes represent a single solid solution and the activity model and equation of state
the organic equivalent of scale formation. Once the crude exits the are used for solideliquid and vaporeliquid equilibrium respec-
tively. Won et al. (1986, 1989), Pedersen et al. (1991a), Pedersen
(1995), Coutinho et al. (1995, 1996) and Coutinho (1998) used this
* Corresponding author. Tel.: þ98 711 2303071; fax: þ98 711 6287294. method by different activity models for describing non-idealities in
E-mail address: rahimpor@shirazu.ac.ir (M.R. Rahimpour). the solid and liquid phases. Chen et al. used the modified SRK EOS

1875-5100/$ e see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jngse.2012.10.003
26 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40

Table 1
Review of existing thermodynamic model for wax precipitation.

Literature review Year Modeling method Solideliquid equilibrium theory Calculation method of fugacity or activity coefficient
Won et al. 1986 Solid Solution Ideal solid phase Solid activity coefficient ¼ 1
Non ideal liquid phase Modified regular solution
1989 Ideal solid phase Solid activity coefficient ¼ 1
Non ideal liquid phase Flory Huggins þ modified regular solution
Hansen et al. 1991 Solid Solution Ideal solid phase Solid activity coefficient ¼ 1
Non ideal liquid phase Polymer solution
Pedersen et al. 1991 Solid Solution Non ideal solid phase Modified regular solution
Non ideal liquid phase Modified regular solution
Coutinho et al. 1998 Solid Solution Non ideal solid phase UNIQUAC
Non ideal liquid phase UNIFAC þ Flory Free Volume
2005 Solid Solution Non ideal solid phase UNIQUAC
Non ideal liquid phase UNIFAC þ Flory Free Volume
Lira Galeana et al. 1996 Multi Solid Phase Ideal solid phase
Non ideal liquid phase Peng Robinson
Pan et al. 1997 Multi Solid Phase Ideal solid phase
Non ideal liquid phase Peng Robinson
Pauly et al. 2000 Solid Solution Non ideal solid phase UNIQUAC
Non ideal liquid phase UNIFAC
Nichita et al. 2001 Multi Solid Phase Ideal solid phase
Non ideal liquid phase SoveRedlichKuang (SRK)
Ji et al. 2003 Solid Solution Ideal solid phase Solid activity coefficient ¼ 1
Non ideal liquid phase UNIQUAC
Ghanae et al. 2007 Solid Solution Ideal liquid phase Liquid activity coefficient ¼ 1
Non ideal solid phase Wilson
Ghanae et al. 2008 Multi Solid Phase Ideal solid phase
Non ideal liquid phase Wilson

and the UNIQUAC equations to describe VLE and SLE, respectively. Critical properties, especially the critical temperature and
The correlations of volume parameter in the SRK EOS and the heat pressure, and the acentric factor are the important input parame-
of vaporization in UNIQUAC equations have been used in Chen’s ters for generalized correlations to estimate physical and thermo-
work (Chen et al., 2007). However, a shortcoming of this model is dynamic properties of fluids. Even small errors in prediction of
application of different thermodynamic approaches (EOS, activity these properties greatly affect calculated physical properties. For
coefficient) to describe the non-ideality of the liquid phase for VLE calculating these properties in industry some of the analytical
and SLE calculation. These types of calculations are thermody- correlations are used by Cavett (1962), Kesler and Lee (1976), Riazi
namically inconsistent and need to some adjustable parameters. and Daubert (1987), Twu (1984), and Riazi and Sahhaf (1996). Pan
The description of the solid and liquid phases non-ideality are the et al. proposed a correlation for calculating critical pressure of
main obstacle in the modeling of the solideliquid equilibrium of Paraffinic-Naphthenic-Aromatic (PNA) for heavy hydrocarbons
hydrocarbons in solid solution model and the accuracy depends to (Pan et al., 1996a). Most of these correlations use boiling point and
the use of activity coefficient model. Different activity coefficient specific gravity as the input parameters to estimate parameters
models have been suggested in the literature with and tuning such as molecular weight, critical constants, and acentric factor.
function and wax estimated value differ from each other Fusion properties used in stability analysis are related to the
considerably. solid precipitation or dissolution at different condition (This rela-
In the multi solid phase model, the precipitated hydrocarbon tionship will be shown in Eq. (5) later). Won in 1986 used the
which consists of pure components is assumed to be unstable and correlations between fusion temperature of pure component, heat
segregates from other precipitated species. The multi solid phase of fusion and molecular weight for normal alkanes (Won, 1986).
model is both very simple and accurate, requiring no adjustable Also, Chung improves Won’s heat of fusion correlation (Chung,
mixture parameters. Lira-Galeana et al. (1996), Pan et al. (1996a), 1992). In 1993 Erickson used correlations for calculating the
Nichita et al. (2001), employed this model to predict the wax
appearance temperature and wax weight percent. Vafaie-Sefti
et al., considered the vapor phase in addition to liquid and solid
phases in order to predict the pressure and composition effect on
the wax appearance point (Vafaie-Seftti and Mousavi-Dehghani,
2000). Escobar-Remolina developed a thermodynamic model
with a combination of ideal solution and multiple solid phase
formation. The fusion properties of the pure solid have been eval-
uated in Escobar’s work (Firoozababadi, 1999). Ghanaei et al., used
an activity model instead of equation of state in the liquid mixture
of multi solid phase model (Ghanaei et al., 2007). However the
consistency of this model depends on non-similarity in molecular
size. Esmaeilzadeh et al., surveyed different activity coefficients and
multi solid phase models. They proved that the predictive Wilson
method gives better result for non-ideality of solid phase
(Esmaeilzadeh et al., 2006). Existing thermodynamic models for
wax precipitation are summarized in Table 1. Fig. 1. Multi-solid phase model for wax precipitation.
 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40 27

Table 2
Summary for the reservoir fluids used in the splitting study (Pedersen et al., 1991b).

Oil No. 1 10 11 12 15

Comp. Mol % MW Mol % MW Mol % MW Mol % MW Mol % MW

C1 1.139 0.016 0.021
C2 0.507 0.145 0.100 0.173 0.254
C3 0.481 1.392 0.118 1.605 1.236
i-C4 0.563 1.180 0.106 1.148 0.588
C4 0.634 3.088 0.099 3.596 2.512
i-C5 1.113 2.980 0.162 3.086 1.955
C5 0.515 3.802 0.038 4.171 3.485
C6 2.003 7.207 0.458 7.841 6.842
C7 5.478 90.9 11.333 94.1 2.194 90.8 11.11 94.1 12.8.5 92.2
C8 8.756 105 12.465 107 2.847 106.5 13.43 105.4 13.98 105.4
C9 7.222 117.7 7.784 122 1.932 122.3 9.419 119 9.190 119
C10 5.414 132 5.314 136 5.750 135 5.583 135 6.435 134
C11 5.323 148 5.033 147 4.874 149 4.890 148 5.118 148
C12 4.571 159 3.989 161 5.660 162 3.864 162 4.111 161
C13 5.289 172 3.869 175 6.607 176 4.298 175 4.231 175
C14 4.72 185 3.627 189 6.149 189 3.272 188 3.682 188
C15 4.445 197 3.165 203 5.551 202 2.274 203 3.044 202
C16 3.559 209 2.311 214 5.321 213 2.791 216 2.255 214
C17 3.642 227 2.472 230 5.022 230 2.311 232 2.405 230
C18 3.104 243 2.815 244 4.016 244 1.960 246 2.006 245
C19 2.717 254 2.110 258 4.176 256 1.821 256 1.766 257
C20 2.597 262 14.4þ 418 38.8þ 473 11.3þ 388 12.0þ 399
C21 1.936 281
C22 2.039 293
C23 1.661 307
C24 1.616 320
C25 1.421 333
C26 1.233 346
C27 1.426 361
C28 1.343 374
C29 1.3 381
Cþ30 13.23 624
SpGr. þ 0.953 0.880 0.963 0.872 0.887

temperature and heat of fusion as a function of number of carbon 1996b; Firoozababadi, 1999). They analyzed various methods for
atoms for non-normal components (Erickson et al., 1993). For calculating the critical properties of heavy petroleum fractions.
aromatic, naphthene and iso-paraffin the different correlations Their evaluation was based on the calculation of the cloud points
have been used by Lira-Galeana et al. in 1996. Pan proposed for different oils. In this work in order to predict the wax
a correlation to estimate the enthalpy of fusion for aromatic species appearance temperature (WAT) at which wax precipitate, a ther-
(Pan and Firoozabadi, 1997). Also, Mei et al., considered the heat modynamic model based on multi-solid phase model is used. For
capacity difference between the liquid and the solid phase on the this purpose the methods for splitting the plus fraction and critical
solid precipitation in his correlations (Mei et al., 1999). Ji et al. and fusion properties are compared and the best methods are
presented a set of correlations for estimation of fusion properties selected. The SRK EOS is used for describing gas and liquid
on the basis of odd or even number of carbon atoms. They showed properties. Finally, the combinations of best methods are used to
that the wax disappearance temperature represents a true solide predict the WAT of the sea pipeline which feeds the South Pars gas
liquid equilibrium point in contrast to wax appearance tempera- processing plant of Iran.
ture (Ji et al., 2004). This proves that the solids are pure rather than
mixed. In 2006 Escobar used the correlations for calculating fusion
properties and heat capacity with the purpose of developing multi- 2. Thermodynamic model
conceptual properties (Escobar-Remolina, 2006). Jeirani et al., have
suggested a set of correlations for estimation of the critical prop- In order to determine the mole fraction of component “i” in solid
erties and acentric factors of long-chain n-paraffins (Jeirani et al., phase, the k-value for solideliquid equilibrium can be determined
2007). Most of the correlations for both critical and fusion prop- as Eq. (1):
erties are detailed and compared in this paper in Appendix A.
An accurate characterization of reservoir fluid or crude oil xSi fLi fiPure
L
KiSL ¼ ¼ (1)
requires a complete analysis of the mixture with known mole xLi S S
fi fiPure
fraction and carbon number. For the mixtures where the compo-
sition of heavy hydrocarbons is represented by a single hydro- In solid solution models, it is assumed that the precipitation
carbon plus fraction, it is important to know the distribution of forms a solid solution and k-value can be calculated as follows:
carbon numbers to describe the mixture properly.
One of the main problems for calculating thermodynamic fis ðP; T; xs Þ ¼ xsi gsi ðP; T; xs ÞfiPure
s
ðP; TÞ (2)
properties of heavy petroleum fractions is the lack of sufficient
methods to predict basic characteristics of heavy hydrocarbons. xSi gLi fiPure
L

Choosing the characterization method strongly influences the KiSL ¼ ¼ (3)

xLi S S
gi fiPure
wax modeling results. Firoozabadi et al. have studied extensively
the wax and asphaltene precipitation in reservoir fluids (Pan et al., The determination of gsi ðP; T; xs Þ is required in Eq. (3).
28 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40

Table 3
Summary for the reservoir fluids used in the splitting study (Pedersen et al., 1991b).

Oil no. 1 10 11 12 15

Comp Mol % Mol Wt. Mol % Mol Wt. Mol % Mol Wt. Mol % Mol Wt. Mol % Mol Wt.
C1 1.139 16 0.016 16 0.021 16
C2 0.507 30 0.145 30 0.1 30 0.173 30 0.254 30
C3 0.481 44 1.392 44 0.118 44 1.605 44.09 1.236 44
i-C4 0.563 58 1.18 58 0.106 58 1.148 58 0.588 58
C4 0.634 58 3.088 58 0.099 58 3.596 58 2.512 58
i-C5 1.113 72 2.98 72 0.162 72 3.086 72 1.955 72
C5 0.515 72 3.802 72 0.038 72 4.171 72 3.485 72
C6 2.003 86 7.207 86 0.458 86 7.841 86 6.842 86
Cþ7 94.04 251 80.29 195 98.919 322 78.32 178 83.1 178
Sp.Gr.þ 0.877 0.834 0.896 0.825 0.829

The relationship between fugacity of component “i” in the solid Phase stability analysis is based on calculations in the multi-
and liquid phase can be derived by calculating the Gibbs free energy solid phase model, such that each component “i” could be consid-
change of pure component “i” from solid to liquid phase. The ered as a solid phase only if it satisfies Eq. (5). This could help us to
following expression allows the fugacity of a pure solid component determine the number of precipitating components.
S
(fpure;i L
) to be related to the sub cooled liquid fugacity (fpure;i )
according to Eq. (4) (Chueh and Prauznits, 1967)
s
f f fi ðP; T; zÞ fpure;i ðP; TÞ  0 i ¼ l; .; c (5)
S
! ZTi ZTi DC LS
fpure;i ðP; TÞ DHif Tif 1 LS 1 pi
ln L
¼ 1 þ DCpi dT dT
fpure;i ðP; TÞ RTif T RT R T
T T
fi(P,T,z) is the fugacity of component “i” in feed with composition “z”
Pif
1
Z at pressure “p” and temperature “T”.
þ DVi dP Each component which satisfies Eq. (5) would precipitate as
RT
P a solid wax. Furthermore, Eq. (6) should be solved to calculate the
(4) condition of wax precipitation.

S
Where fpure;i L
and fpure;i are pure component fugacity in the solid
and liquid phase, respectively. In the right side of Eq. (4), the term 1  
s
shows the effect of solideliquid transition, the terms 2 and 3 show fil P; T; xli ¼ fpure ðP; TÞ i ¼ ðc cs þ 1Þ; :::; c (6)
the effect of specific heat capacity change between solid and liquid
phase and the last term presents the effect of the pressure change cs is the number of solid phases and is determined from Eq. (5).
from fusion pressure (1 atm) to system pressure known as Poynting The mass balances for the non-precipitating and precipitating
correction term. components are shown in Eqs. (7) and (8) respectively. As it is
mentioned in these equations the solid phase is considered pure
and immiscible.
2.1. Multi-solid phase model

A number of studies have shown that when binary normal

alkane mixtures are cooled, the precipitation is unstable and
segregates into two solid phases (Dorset and Snyder, 1996). Based
on these observations, Lira-Galeana et al. (1996) developed a ther-
modynamic multi-solid wax model in which each solid phase is
described as a pure and immiscible phase. In this model we assume
that the feed enters with composition “z” at pressure “p” and
temperature “T” and separates into L mol liquid with composition xli
P s
and nj moles pure multi solid phase. The use of this model for
wax precipitation is shown in Fig. 1.

Table 4
Comparison results, absolute average relative error (AARE) for splitting hydrocarbon
plus fraction.

Oil # Absolute average relative error, j((Cal-Exp).Exp)j

Cþ
n Al-Meshari Ahmed Katz Pedersen
1 Cþ
7 0.11 0.18 0.74 3.67
10 Cþ
7 0.16 0.31 0.35 0.17
11 Cþ
7 0.64 0.63 1.95 0.51
12 Cþ
7 0.36 0.59 0.42 0.18
Fig. 2. Comparing mole fractions versus molecular weights of the extended analysis of
15 Cþ
7 0.16 0.28 0.33 0.14
theCþ7 fraction of the “Oil 1” system reported by Pedersen et al.
 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40 29

0.2

Experimental Katz Method

0.16
Ahmed Method Pedersen

Molar Distribution
Al-Meshari
0.12

(oil 12)

0.08

0.04

0
50 100 150 200 250 300

Molecular Weight
Fig. 3. Comparing mole fractions versus molecular weights of the extended analysis of
theCþ7 fraction of the “Oil 10” system reported by Pedersen et al. Fig. 5. Comparing mole fractions versus molecular weights of the extended analysis of
theCþ7 fraction of the “Oil 12” system reported by Pedersen et al.

2 3
c nsj
xli 41
X
zi 5 ¼ 0 i ¼ 1; :::; ðc cs Þ (7) c
F xli ¼ 1
X
j ¼ c cs þ1 (9)
i¼1

The (c þ cs) equations are used to calculate the (c þ cs) unknown

variables. The unknown variables are xli andnsj .
2 3
c nsj nsi
xli 41
X
zi 5 ¼ 0 i ¼ ðc cs þ 1Þ; :::; c 1; cs > 1
j ¼ c cs þ1
F F
2.2. Splitting the heavy fraction
(8)
As mentioned in the introduction, the key focus for calculating
“nsj ” and “F” are the moles of solid phase “j” and feed respectively. thermodynamic properties of wax precipitation in petroleum
The boundary condition for component “i” in liquid phase is as mixtures is estimating the thermodynamic properties of heavy
follow: hydrocarbon components. Table 2 shows composition, molecular
weight, and heavy fraction specific gravities of five oil systems
given by Pedersen et al. The types of crude oils shown in Table 2
cover a variety of oil mixtures.
0.2

Experimental Katz Method

0.2

0.16
Ahmed Method Pedersen Experimental Katz Method

0.16
Al-Meshari Ahmed Method Pedersen
0.12
Mole Fraction

(oil 11)
Mole Fraction

0.12 Al-Meshari

(oil 15)
0.08
0.08

0.04
0.04

0 0
50 100 150 200 250 300 50 100 150 200 250 300

Molecular Weight Molecular Weight

Fig. 4. Comparing mole fractions versus molecular weights of the extended analysis of Fig. 6. Comparing mole fractions versus molecular weights of the extended analysis of
theCþ7 fraction of the “Oil 11” system reported by Pedersen et al. theCþ7 fraction of the “Oil 15” system reported by Pedersen et al.
30 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40

The experimental data in Table 2 shows that the composi- used for comparing these methods. The methods are explained in
tions of components heavier than C7 exhibit exponential molar Appendix A.
distribution. Therefore in the first step the components heavier Experimental and calculated mole fractions have been
than C7 have been lumped as a single component, Cþ þ
7 . For C7 the compared together and an absolute average relative error is re-
only measured data is Molecular weight and specific gravity. ported in Table 4. Also, the methods are compared in Figs. 2e6. As it
Table 3 shows compositions, component molecular weights, and can be seen in all cases the Al-Meshari method is the most accurate
heavy fraction specific gravities of the oil systems after method to split the hydrocarbon plus fractions.
grouping.
Splitting the heavy fraction in to the pseudo components make 2.3. Critical and fusion properties
it easier to describe the heavy fraction. Without having analytical
data this would lead to severe errors in calculating the phase The Multi-solid model is based on precipitation of heavy
behavior of mixtures. If Cþ 7 splits to a number of pseudo compo- components with average properties such as fusion temperature,
nents with a single carbon number are made, the errors will reduce fusion enthalpy, acentric factor and critical temperature and pres-
significantly. Each pseudo component should have a prescribed set sure. These properties are predicted with characterization
of physical properties. In this work, the MATLAB program is used for methods. Therefore, the choice of characterization method strongly
applying the four common splitting methods that were cited in influences the modeling results. In this section, attempts have been
literature. These four methods are known as Katz (1983), Ahmed made to compare the most common characterization methods.
et al. (1985), Pedersen et al. and Al-Meshari (2004) methods. The Recently Escobar-Remolina (2006) proposed the fusion tempera-
mentioned five data sets proposed by Pedersen et al. (1991b) were ture correlation (Eq. B-2-23).

Fig. 7. Comparing the (a) fusion enthalpy (b) fusion temperature (c) critical pressure (d) critical temperature correlations.
 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40 31

By using experimental data the Escobar correlation was modi- the mentioned correlations were compared. Also, we simplify the
fied and the new following correlations have been suggested in this averaged critical pressure using the Pan correlation (considering P,
work for the fusion temperature: N and A in one group) for heavy hydrocarbons (M  280) and Riazi
and Al-Sahhaf correlation for light hydrocarbons (M < 280). For
Tif ¼ 3  10 6
Mi3 4:1  10 3
Mi2 þ 1:9226Mi Mi < 600 most adjustments, we suggest the following correlations for the
critical properties:
(10)
Pc;i ¼ 53:975 expð 0:0056Mi Þ Mi < 280 (12)
15760:33
Tif ¼ 370:12 þ 0:002403Mi 139:67 Mi  600
Mi Pc;i ¼ 27:894 expð 0:004Mi Þ Mi  280 (13)
(11)
For heat of fusion and heat capacity, the correlations suggested Tc;i ¼ 220:41  lnðMi Þ 450:97 (14)
by Escobar, have been used in this study (Eqs. B-2-24 and B-2-25)
(Escobar-Remolina, 2006). Where the units of Tc and Pc are Kelvin and bar respectively.
Critical properties are needed to perform liquidesolid equilibria In this research, the acentric factor for each SCN group extended
using the cubic EOS. It is almost impossible to measured directly from the heavy fraction was calculated using the Edminster
critical properties of a long-chain n-paraffin, because of the thermal correlation (Edminster, 1958). The methodology for finding an
decomposition at high temperatures. The critical temperature (Tc) accurate set of correlations for fusion and critical properties of SCN
and pressure (Pc) can be adjusted to make the wax weight percent Groups (Eqs. 10, 11 and 12e14) are indicated in Appendix C. Fig. 7
curve pass through the experimental data. a through d shows the comparison of the correlations.
There are many correlations for fusion temperature, fusion
enthalpy, critical pressure and critical temperature in literature. The 3. Results and discussions
correlations are shown in Appendix B. In this work, the average
properties method was applied for undefined hydrocarbons to all Programming with MATLAB software was used to predict the
five oil samples that are proposed by Pedersen et al. (1991b) and wax weight percent for the oil systems given by Pedersen et al

Fig. 8. Calculation work-flow for modeling.

32 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40

a 8 b 8

Exp. (Pedersen et al.,1991) Exp. (Pedersen et al., 1991)

Pedersen et al., (1991) Pedersen et al., (1991)

6 Lira Gleana et al., (1996) 6

Lira-Galeana et al., (1996)

This Work This work

Wax Weight Percent

Wax Weight Percent

4 4
(oil 1)
(oil 10)

2 2

0 0
230 240 250 260 270 280 290 300 310 320 230 240 250 260 270 280 290 300 310 320 330

Temperature (K) Temperature (K)

c 16 d 8

Exp. (Pedersen et al., 1991) Exp. (Pedersen et al., 1991)

14
Pedersen et al., (1991)
Pedersen et al., (1991)

12 6
Lira-Galeana et al., (1996)
Lira-Galeana et al., (1996)
Wax Weight Percent

Wax Weight Percent

This work
10 This work

8 4
(oil 11) (oil 12)

4 2

0 0
250 260 270 280 290 300 310 320 230 240 250 260 270 280 290 300 310 320

Temperature (K) Temperature (K)

e 7

Exp. (Pedersen et al., 1991)

6
Pedersen et al., (1991)

5
Lira-Galeana et al., (1996)
Wax Weight Percent

4 This work

3
(oil 15)

0
230 240 250 260 270 280 290 300 310 320

Temperature (K)

Fig. 9. Comparing different models for predicting the wax weight percent for different oils (a) No. 1, (b) No. 10, (c) No. 11, (d) No. 12, (e) No. 15.
 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40 33

Fig. 10. Gas condensate stabilization unit.

(1991b). The trend of modeling is shown in Fig. 8. The SRK EOS was to separate liquid and slugs from gas. The liquids are sent to the two
used throughout this modeling. The result of each model is Condensate Stabilization Units (one per phase). High vapor pres-
compared with the experimental data and the other models sure values of condensate are due to the pressure of dissolved
(Fig. 9aee) for each of Pedersen’s oil systems. volatile hydrocarbons called “light ends” such as methane, ethane,
In Fig. 9aee the following calculation options were used. propane, butanes. To safely store condensate in tanks at atmo-
spheric conditions or to transport condensate safely in pipelines at
1 Calculating critical pressure and temperature using proposed a defined operating pressure, the condensate vapor pressure value
correlations (Eqs. (12)e(14)) and acentric factor using the must be controlled or adjusted to meet pipeline, storage, or tanker
Edminster correlation. Reid vapor pressure (RVP) specifications. The function of the
2 Calculating fusion temperature using proposed correlation stabilization unit (Fig. 10) is to remove the lightest components
(Eqs. (10) and (11)),heat of fusion and heat capacity using from the raw feed and to produce a liquid product. The stabilization
Escobar correlations (Eqs. B-2-24 and B-2-25) unit will give a stabilized condensate having a Reid Vapor Pressure
3 Using the Al-Meshari method to split the heavy hydrocarbon (RVP) of 10 psia in summer and 12 psia in winter. As it is shown in
fractions. Fig. 10 the raw condensate from the slug catchers is preheated
(E101 A/B), then flashed (D-101) before going through a desalter (D-
As is shown in figure aee, the results of this study are in good 105). The condensate light components are then removed in the
agreement with the experimental values. condensate stabilizer (C-101). The off-gas is compressed by (K-101)

4. Case study of South Pars gas complex

The South Pars gas refinery is the biggest gas processing plant in
southern part of Iran and has been designed for gas from the South
Pars gas field. The South Pars gas field is geologically an extension of
Qatar’s 10.2  1012 standard m3 North Field. South Pars was initially
identified in 1988 and estimated to contain at least 3.4  1012
standard m3 of natural gas (Javanmardi et al., 2006; Rahimpour
et al., 2012). The Phases 4 & 5 onshore complex is located on the
Iranian coast of the Persian Gulf in the area of Asalooyeh village. The
total capacity of the Phases 4 & 5 onshore facilities is 2000 MMSCFD
of reservoir fluid. They will include all processing units to produce
sales gas, an ethane cut of petrochemical feedstock quality,
commercial grade propane and butane for exporting and stabilized
condensate. The reservoir fluid is delivered to the onshore plant via
two 32" sea pipelines feeds to the complex. These two sea pipelines
originate from the wellhead platforms (SPD5 and SPD6) and route
multi-phase flow (gas, condensate and glycolated water) to
reception facilities onshore. Raw feed received from the two plat-
forms, flows to finger-type slug catchers. The slug catchers are used Fig. 11. The debris build-up on the reboilers tubes.
34 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40

Table 5
The hydrocarbon composition of the sea pipeline.

Component Mole fraction Component Mole fraction

H2O 0.015579 C12CUT 0.001077
N2 0.033062 C13CUT 0.000880
CO2 0.017411 C14CUT 0.000588
H2S 0.006572 C15CUT 0.000391
C1 0.809659 C16CUT 0.000293
C2 0.051747 C17CUT 0.000196
C3 0.019173 C18CUT 0.000196
iC4 0.004109 C19CUT 0.000098
NC4 0.006946 Cþ
20 0.000293
IC5 0.002838 COS 0.000003
NC5 0.002838 CH4S 0.000029
C6 CUT 0.004109 ETSH 0.000245
C7CUT 0.004597 PR1THIOL 0.000138
C8CUT 0.004988 BU1THOL 0.000042
C9CUT 0.003423 HX1THIOL 0.000082
C10CUT 0.002446 MEG 0.004391
C11CUT 0.001565 Total 1.000000

Fig. 12. Collecting the debris for sampling (The black and gritty debris is surrounded
by residual clear liquid condensate).
wax we could reduce debris formation significantly. The prediction
of wax precipitation from the beginning in the sea pipelines could
and returned to the HP separators and the stabilized condensate is help us to prevent the wax formation before it becomes severe and
cooled and sent to storage ready for export (South-Pars Gas forms emulsions.
Company, 2012). As can be seen from the sample analyses, the wax molecules
Unfortunately the build-up of debris on the reboiler (E101 A/B) can combine with the other chemical species present in gas
tubes (Fig. 11) tends to reduce the heat transfer duty and forces condensate. The complex mixture of chemical species present in
periodic shut-down for cleaning, interrupting normal processing gas condensate makes it necessary to consider several factors
operations. when designing a treatment program for field application. In
As it is shown in Fig. 12 the debris samples were collected from these cases it is helpful to consider a likely arrangement of this
the reboilers tubes. complex aggregate and the nature of the solid. Also it helps to look
The analysis of the debris with Environmental Scanning Electron at both options of whether the forming wax crystal serves as
Microscope (ESEM) method shows that the debris is composed of: a collection point for the solid, or the solid serves as a nucleation
point for the onset of wax crystallization. Considering the fact that
 Iron (Fe) the emulsion is not resolved with heat and a de-emulsifier alone,
 Sulfur (S)
 Potassium (K), sodium (Na) and calcium (Ca)
 Chlorine (Cl)
Table 6
 Carbon (C) and oxygen (O) The variation of temperature with pressure along the pipeline.

Pipeline length (m) Pipeline pressure (bar) Pipeline temperature ( C)

Results for 3 spot analyses are presented in Fig. 13.
The material is dominated by iron sulphides/oxides and salts 25 101 62
1025 101 60
(e.g. KCl, NaCl). The high carbon content observed for one sample
2025 100 58
likely represents crystallized heavy hydrocarbons from the 3025 100 56
condensate. The results show that by preventing the formation of 4025 100 54
5275 100 52
10,075 99 45
15,225 98 39
20,325 97 34
25,075 96 31
30,175 95 28
35,075 94 26
40,625 93 24
45,375 93 23
50,025 92 22
55,525 91 21
60,425 90 21
65,575 89 20
70,576 88 20
75,126 87 20
80,076 86 20
85,026 85 19
90,226 84 19
95,076 83 19
100,026 81 19
105,771 80 19
107,821 79 18
108,711 79 18
Fig. 13. ESEM spot analyses of reboiler debris.
 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40 35

6 is employed to find the wax appearance temperature (WAT) and

predict the wax appearance location in the sea pipeline. Table 5
illustrates the hydrocarbon composition of the sea pipeline.
5
Table 6 shows the variation of temperature and pressure along the
length of the sea pipelines.
Wax Weight Perccent

4 Fig. 14 shows the wax precipitation starts at 293 K. This

temperature is reached in the pipeline 94 km from the wellhead at
a corresponding pressure of 86 bar. This is shown in Fig. 15.
3

5. Conclusion
2
The results of this study for the South Pars gas processing plant
demonstrate that by preventing wax formation we could reduce
1 debris build-up significantly. The combination of wax and emulsion
in some systems could not be reduced by heat and de-emulsifier
alone. The prediction of wax precipitation as a function of
0
200 220 240 260 280 300 distance from the beginning in the sea pipelines could help us to
Temperature (K) prevent the wax formation before it becomes severe and forms
emulsion.
Fig. 14. Wax weight percent versus temperature in sea line. In order to predict the wax formation in the South Pars gas
sea pipeline, a thermodynamic model based on the multi-solid
phase model is used. For this purpose the methods for split-
then an obvious approach would be to apply a crystal modifier to ting the heavy fraction and predicting critical and fusion prop-
interfere with these interactions. Whereas the South Pars gas sea erties are compared and the best ones are selected. The results
line temperature is below the crystallization temperature (crucial show that the Al-Meshari method is the most accurate method
temperature) of the wax problem, the wax can form. The wax to split the hydrocarbon heavy fractions. Also, calculating
molecules can combine with the emulsion stabilizing the interface the fusion and critical properties by proposed equations (Eqs.
between water and gas condensate. So it is necessary to disperse (10)e(14)) and heat of fusion and heat capacity using Escobar
wax molecules. A de-emulsifier and crystal modifier should be correlations (Eqs. B-2-24 and B-2-25) could lead to the best
implemented together to resolve the emulsion and disperse the answers for critical and fusion properties of heavier compo-
wax molecules respectively. So a completely reasonable approach nents. Finally, the combinations of the best predicted methods
would be to include the crystal modifier as part of the emulsion are used for predicting the wax formation in five oil systems
resolution treatment. Another aspect of this treatment is that given by Pedersen et al. The consistency of these combinations
a crucial temperature would be expected, since the melting point is approved again.
of the wax should be approached when attempting to de-emulsify The result of this study is used to predict the WAT (Wax
this gas condensate. Appearance Temperature) in the sea pipeline to the South Pars gas
To predict the wax precipitation at the sea pipelines a large set of processing plant. The modeling shows that wax precipitation starts
available data were used. Also, the temperature and pressure at 293 K and 86 bar. Also, the temperature and pressure profiles
profiles along the pipeline were simulated. The result of this study along the pipeline were simulated. At this pressure and tempera-
ture the sea pipeline is 94 km away from wellhead.
Waxes may be the main issue, but the presence of other complex
such as emulsions, silt, corrosion by-products, and inorganic scales
70 can act as a binder, producing high bulk density debris, which
appears to be the primary problem material. Most of these mate-
rials can be incorporated into the wax, or conversely the waxes can
60
Phase Envelope of Sea Line WAT form in them. The challenge is to determine which problem to
attack first in order that the most efficient solution may be
50 implemented. Good production practices and the placement of
a treatment facility to remove emulsion, brine water and sus-
Temperature (C)

Total Length pended solid are very important and effective to enhance the
40 of Sealine performance of gas processing plant. Crystal modifier works as
(105 Kilometer)
a wax dispersant and prevents the agglomeration of heavy hydro-
30 carbons. In order to fight this problem, chemical demulsifier must
be injected as early as possible and in several points (wellhead,
upstream and downstream of separation equipment, located in
20 offshore).

94 Kilometer
10 from wellhead Appendix A. Splitting methods

Katz Method (Katz, 1983):

0
78 83 88 93 98 Katz correlation is represented in a mathematical form as
Pressure (bar) follows:

0:25903i
Fig. 15. Phase envelope of the sea line and the Wax Appearance Temperature (WAT). Zi ¼ ZC þ 1:38205e (A-1)
7
36 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40

Ahmed Method (Ahmed et al., 1985): MC þ h

Ahmed et al. presented a method to split Cþ b¼ 7
(A-12)
7 to a number of a
pseudo components. The following expressions are used to esti-
mate pseudo-component mole fractions:
M h
y ¼ (A-13)
" þ þ# b
Mðiþ1Þ Mi
Zi ¼ ZC þ þ (A-2)
7 Mðiþ1Þ Mi In previous equations, “Mi” is the molecular weight of compo-
nent “I”, “G” is gamma function, “h” is origin of the distribution and
“a” is a parameter which determines the shape.
Where:

Appendix B. Correlations for estimating critical and fusion

Zi ¼ Zðiþ 1Þ
Zði 1Þ (A-3)
properties and heat capacity

Notes:
Miþ ¼ MCþ7 þ Sði 7Þ (A-4)
1. In all the following equations the unit of temperature and
In crude oil systems for “i” smaller and equal to 8; S ¼ 16.5 and pressure are Rankine and psia respectively.
for “i” larger than 8; S ¼ 20.1. 2. The specific gravity is defined relative to water at 60 F.
In gas condensate systems for “i” smaller and equal to 8; S ¼ 15.5 3. API ¼ ð141:5=SÞ 131:5.
and for “i” larger than 8; S ¼ 17.
Pedersen Method (Pedersen et al.,):
Pedersen et al. presented the following correlations: B.1 Critical properties

The critical properties (Tc, Pc) are evaluated according to the

Zn ¼ eAþBMn (A-5) following correlations:
Cavett (1962) correlations:
Where A and B are determined from the following two equations:
Tc ¼ 768:071 þ 1:7134  ðTb 459:67Þ
N
0:10834  10 2
 ðTb 459:67Þ2
eAþBMn
X
zcþ ¼ 0 (A-6)
6
459:67Þ3
7
n¼7 þ 0:3889  10  ðTb
2
(B-1-1)
0:89213  10  ðTb 459:67Þ  API
XN
M AþBMn Mcþ þ 0:53095  10 5
 ðTb 459:67Þ2 API
e 7
zcþ ¼ 0 (A-7)
n¼7
Sg n Sgcþ 7 þ 0:32712  10 7
 ðTb 459:67Þ2 API2
7

Al-Meshari Method (Al-Meshari, 2004):

Al-Meshari splits the new plus fraction Cþ
7 according to the log Pc ¼ 2:829 þ 0:9412  10 3
ðTb 459:67Þ
original experimental extended heavy fraction using the gamma
0:30457  10 5
ðTb 459:67Þ2
distribution model. The Gamma distribution model is based on the
Gamma Distribution function, which is a skewed probability þ 0:15184  10 8
ðTb 459:67Þ3
density function (p(M)). He used the following correlations to 0:20876  10 4
ðTb 459:67Þ API (B-1-2)
estimate pseudo-component mole fractions:
þ 0:11048  10 7
ðTb 459:67Þ2 API
0:4827  10 7
ðTb 459:67Þ API2
ZMi
zi
¼ PðMÞdM ¼ P0 ðMi Þ P0 ðMi 1Þ (A-8) þ 0:1395  10 9
ðTb 459:67Þ2 API2
zþ
7
Mi 1
Kesler and Lee (1976) correlation:
Where:
Tc ¼ 341:7 þ 811S þ ð0:4244 þ 0:1174SÞ Tb

ðM hÞ a 1
expf ½ðM hÞ=bŠg þ ð0:4669 3:2623SÞ  105 =Tb (B-1-3)
pðMÞ ¼ a
(A-9)
b G a
ln Pc ¼ 8:3634 0:0566=S
e y ya X
yj N  
P0 ðMi Þ ¼ (A-10) 0:24244 þ 2:2898=S þ 0:11857=S2  10 3 Tb
GðaÞ j ¼ 0 ða þ kÞ
Q
 
þ 1:4685 þ 3:648=S þ 0:47227=S2  10 7 Tb2
 
e y ya N
X yj e y ya 0:42019 þ 1:6977=S2  10 10 Tb3
P1 ðMi Þ ¼ (A-11)
GðaÞ j ¼ 0 ða þ kÞ aGðaÞ
Q
(B-1 e4)
 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40 37

 
6
ln Pbr 5:92714 þ 6:09648=Tbr þ 1:28862 ln Tbr 0:16934 Tbr
u ¼   For Tbr  0:8 (B-1-5)
6
15:2518 15:6875=Tbr 13:4721 ln Tbr þ 0:4357 Tbr

2 Tb
u¼ 7:904 þ 0:1352Kw 0:007465Kw þ 8:359 Tbr Tc ¼ (B-1-14)
0:02436  M 0:58


1:15 Exp 0:41966
ð1:408 0:1063Kw Þ (B-1-6)
þ For Tbr > 0:8
Tbr  
Pc ¼ 14:5  Exp 4:65757 0:13423  M 0:5 ðM < 300Þ
Where Pbr ¼ Pb/Pc, Tbr ¼ Tb/Tc and Kw is the Watson characterization
(B-1-15)
factor.
Twu (1984) correlation: For Naphthene:
h  
Tb ¼ exp 5:71419 þ 2:71579 ln M u¼ 0:3 þ Exp 8:25682 þ 5:33934  M 0:08 (B-1-16)

0:286590ðln MÞ2 39:8544=ðln MÞ   

i Tb ¼ 1:8  1028 Exp 6:9565 0:02239  M 0:666666
0:122488=ðln MÞ2 24:7522 ln M þ 35:3155ðln MÞ2
(B-1-17)
(B-1-7)
  
Tc ¼ Tb = 1:2 Exp 0:06765 0:13763  M0:35 (B-1-18)
"
3 7
Tc ¼ Tb 0:533272 þ 0:191017  10 Tb þ 0:779681  10 Tb2
 
# 1 Pc ¼ 14:5  Exp 12:3107 5:53366  M 0:1 ðM < 300Þ
10 0:959468  102
0:284376  10 Tb3 þ ðB 1 8Þ (B-1-19)
ðTb =100Þ13
For Aromatic:
Tb
j ¼ 1 (B-1-9)   
Tc Tb ¼ 1:8  1015 Exp 6:91062 0:02247  M 0:666666
 1 (B-1-20)
Pc ¼ 3:83354 þ 1:19629 j2 þ 34:8888 j þ 36:1952 j2
2 Tb
þ 104:193 j4 ðB 1 10Þ Tc ¼
1:03 Exp 0:29875 0:06814  M0:5

(B-1-21)

Riazi and Daubert (1987):  

Pc ¼ 14:5  Exp 9:77968 3:07555  M 0:15 ðM < 300Þ
q ¼ aTbb Sc expðdTb þ eS þ fTb SÞ (B-1-11)
(B-1-22)

Table B-1
The values for coefficients of Riazi and Daubert correlation.

q a b c d e f
4 3
MW 581.96 0.97476 6.51274 5.43076  10 9.53384 1.11056  10
4
Tc ( R) 10.6443 0.81067 0.53691 0.00051747 0.5444 3.5995  10
Pc (psia) 6.162  106 0.4844 4.0846 0.004725 4.8014 3.1939  10 3

Riazi and Al-Sahhaf (1996) for ParraffineNaphthenee Pan et al. (1996a) modified Riazi and Al-Sahhaf correlation (Riazi
Aromatic species: for parraffins: and Al-Sahhaf, 1996) for heavy component of PNA species: for
parraffins:
 
u ¼ 0:3 þ Exp 3:06826 þ 1:04987  M0:2 (B-1-12)
Pc ¼ 14:5ð0:67909þ22:179 Expð 0:0028417MÞÞ ðM > 300Þ
(B-1-23)
  
0:666666
Tb ¼ 1:8  1070 Exp 6:98291 0:02013  M
(B-1-13) For Naphthene:

Pc ¼ 14:5  ð2:58854 þ 27:6292 Expð 0:0044951  MÞÞ ðM > 300Þ (B-1-24)

38 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40

For Aromatic:

Pc ¼ 14:5  ð4:85196 þ 42:9311 Expð 0:0056193  MÞÞ ðM > 300Þ (B-1-25)

B.2 Fusion properties and heat capacity Nichita et al. (2001) proposed the total enthalpy
(fusion þ transition) correlation for normal alkane:
The fusion properties (Tf, DHf) and heat capacity (DCP) are
evaluated for paraffin species according to the following DHit ¼ DHif þ DHitr (B-2-15)
correlations:
Won’s correlations (1986, 1989) for normal alkanes For Mi > 282

20172 DHif ¼ 0:1186Mi Tif (B-2-16)

Tif ¼ 374:5 þ 0:02617Mi M  450 (B-2-1)
Mi
DHitr ¼ 0:0577Mi Ti1
tr
(B-2-17)
32362
Tif ¼ 411:4 M > 450 (B-2-2)
Mi tr
Ti1 ¼ 366:39755 þ 0:03609Mi 2:08796  104 =M (B-2-18)
DHif ¼ 0:1426Mi Tif (B-2-3) For Mi  282

1 DHit ¼ 0:1777Mi Tif (B-2-19)

DHitotal ¼ DHif þ DHitr (B-2-4)
2
Coutinho et al. (2002) suggested that when wax precipitates
Pedersen et al. correlations (1991a,b) for all species of hydro-
from a real crude oil, the normal paraffins solidify to form an
carbon (paraffinenaphtheneearomatic) (Al-Meshari, 2004)
orthorhombic solid solution. For n-paraffin with a carbon number
4 of 42 or more, the normal melting point and the enthalpy of fusion
DCpi ¼ 0:3033Mi 4:635  10 Mi T (B-2-5)
can be obtained (Coutinho et al., 2002):

DHif ¼ 0:5148  0:1426Mi Tif (B-2-6)

h  i
Tif ¼ 421:63 193611:63 exp 7:8945 N 10:07194
Chung (1992) improved Won’s correlation (Won, 1986, 1989) as (B-2-20)
follows:
4 < N < 42
DHif ¼ 0:9  Tif ðMi Þ0:55 (B-2-7)
DHif ¼ 1000 0:00355N 3 þ 0:2376N 2


3:6209N þ 18:5391
Ericksen et al. (1993) proposed correlations for non-normal
(B-2-21)
hydrocarbons:

80Ci N  42
 
f f
Ti;non normal ¼ Ti;normal 90 (B-2-8) (B-2-22)
100 þ Ci DHif ¼ 1000ð3:7791N 12:654Þ
 
DHif ¼ 1:14  0:1426 Mi Tif (B-2-9) Where “N” is the carbon number.
Ji et al. (2004) applied the correlation according to difference
The correlation proposed by Lira-Galeana et al. (1996) is used to between odd or even carbon number for normal paraffin.
calculate fusion temperature for asphaltenes and iso paraffins and Escobar-Remolina (2006) developed multi-conceptual
fusion enthalpy for iso paraffins. properties:

Tif ¼ 333:45 419 exp½ 0:00855Mi Š (B-2-10) 1576:33

Tif ¼ 370:12 þ 0:002403Mi
MWi
Dhfi ¼ 0:0527Mi Tif (B-2-11)
139:67 expð 0:008546Mi Þ (B-2-23)
Pan and Firoozabadi (1997) proposed the correlation to estimate
the enthalpy of fusion for aromatic species:
DHif ¼ 0:107834Mi Tif (B-2-24)
Dhfi ¼ 11:2Tif (B-2-12)
4
DCpi ¼ 0:2557Mi 4:4245  10 Mi Tif (B-2-25)
Pan used Lira-Galeana’s correlation (Lira-Galeana et al., 1996) to
calculate fusion temperature. Chen et al. (2007) correlate the fusion enthalpies of normal
Marano and Holder (1997) proposed the enthalpy of the solid alkanes based on carbon number: n-Parraffins with odd carbon
state transition as a function of carbon number: number:

DHtr ¼ DHtotal DHf (B-2-13) 7 < Cn < 21

(B-2-26)
DHtotal ðkJ=moleÞ ¼ 3:7791Cn 12:654 (B-2-14) DHif ¼ 0:5754Mi Tif
 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40 39

n-Parraffins with even carbon number: Table C-1

Comparison results, absolute average relative error, AARE, for finding the Accurate
8  Cn < 22 Set of Correlations.
(B-2-27)
DHif ¼ 0:8064Mi Tif Absolute average relative error, j((Cal-Exp).Exp)j using six different set of
correlations compared to the base case

n-Parraffins with all carbon number: Fluid # Main set Set 1 Set 2 Set 3 Set 4 Set 5 Set 6
1 0.38 2.52 0.57 3.31 3.71 3.81 5.62
21  Cn < 38 10 0.42 4.91 0.49 3.90 10.42 9.11 9.54
(B-2-28) 11 0.34 4.55 1.07 8.05 2.69 3.94 8.41
DHif ¼ 0:4998Mi Tif 12 0.39 0.28 0.51 0.4 0.58 0.72 2.33
15 0.14 0.50 0.31 1.07 0.81 1.25 2.16

Cn  38
(B-2-29)
DHif ¼ 0:6741Mi Tif
List of symbols
Where (Tf[¼]K), (DCP[¼] cal/mol-K) and (DHf[¼]cal/mol)
Edminster (1958) proposed a simple equation for calculating the AARE Absolute Average Relative Error
acentric factor of pure compounds. This equation is a function of Cn carbon number
critical pressure and critical temperature and normal boiling point Cp specific heat capacity
and has been applied successfully to petroleum fractions. Cs number of solid phases
C number of component
f fugacity
PC
0  1
log F feed mole numbers
3 B 14:7 C H enthalpy
u ¼ B C 1 (B-2-30)
7 @ TC A i counter of component
1
TB j counter of component
K equilibrium constant
M molecular weight
Appendix C. Methodology for finding the accurate set of Mþ7 Average molecular weight of the Cþ7 components
correlations for fusion and critical properties and acentric Mb Bounding molecular weight
factors of SCN groups Mbi Bounding molecular weight of the ith fraction
Mbi 1 Bounding molecular weight of the (i 1)th fraction
The critical properties and acentric factor of each component is N number of moles
required to determine the physical and thermodynamic properties P pressure
of fluids in equation of states. Fusion properties used in stability P (M) Probability density function of M
analysis are related to the solid precipitation or dissolution at P0 (M) The cumulative probability density function i.e. the
different conditions. Seven different sets of correlations were used integral of p(M)
to calculate the critical and fusion properties and acentric factor for P1 (M) Integral of the M times the Gamma distribution
the single carbon number groups in the extended composition. The (probability density) function
wax weight percent is calculated and matched with the experi- R gas universal constant
mental data using these properties. The results of six sets of s solid moles
correlations are compared to the base case. The main set is the set S specific gravity
of correlations selected in this work. SCN Single Carbon Number
Main set: Tf, Hf (Eqs. 10, 11 and B-2-24), Tc & Pc (Eqs. (12)e(14)), T temperature
u-Edminster (1958) (Eq. B-2-30). V volume
The other six sets of correlations are as follows: WAT Wax Appearance Temperature
x feed composition
1 First set: Won (1986,1989) e Tf, Hf (Eqs. B-2-1eB-2-4), Twu (1984) z feed mole fraction
e Tc & Pc (Eqs. B-1-7eB-2-1), u-Edminster (1958) (Eq. B-2-30) Zf liquid mole fraction
2 Second set: Lira-Galeana et al. (1996) e Tf, Hf (Eqs. B-2-10, B-2- zþ7 Average molecular weight of the Cþ7 components
11), Twu (1984) e Tc & Pc (Eqs. B-1-7eB-2-1), u-Edminster zi Bounding molecular weight
(1958) (Eq. B-2-30) zi/zþ
7 Bounding molecular weight of the ith fraction
3 Third set: Escobar-Remolina (2006) e Tf, Hf (Eqs. B-2-23, B-2-
24), Twu (1984) e Tc & Pc (Eqs. B-1-7eB-2-1), u-Edminster Greek letters
(1958) (Eq. B-2-30) D variation
4 Fourth set: Lira-Galeana et al. (1996) e Tf, Hf (Eqs. B-2-10, B-2- g activity coefficient
11), Riazi and Al-Sahhaf (1996) e Tc & Pc (Eqs. B-1-13eB-1-15), f fugacity coefficient
u-Edminster (1958) (Eq. B-2-30) u acentric factor
5 Fifth set: Lira-Galeana et al. (1996) e Tf, Hf (Eqs. B-2-10, B-2-11), a Alpha; The shape parameter of the Gamma Distribution
Riazi and Al-Sahhaf (1996) e Tc & Pc & u (Eqs. B-1-12eB-1-15) b Beta; A composite parameter of the Gamma Distribution
6 Sixth set: Escobar-Remolina (2006) e Tf, Hf (Eqs. B-2-23, B-2-24), G Gamma function
Riazi and Al-Sahhaf (1996) e Tc & Pc & u (Eqs. B-1-12eB-1-15) gþ7 Average Specific Gravity of Cþ
7 components

As it is shown in Table C-1 using the Main Set could lead to the Superscripts
minimum absolute average relative error for predicting the wax L liquid
weight percent. S solid
40 M.R. Rahimpour et al. / Journal of Natural Gas Science and Engineering 10 (2013) 25e40

tr transition Escobar-Remolina, J.C.M., 2006. J. Fluid Phase Equilib. 240, 197e203.

Esmaeilzadeh, F., FathiKaljahi, J., Ghanaei, E., 2006. J. Fluid Phase Equilib. 248, 7e18.
f fusion
Firoozababadi, A., 1999. Thermodynamics of Hydrocarbon Reservoirs, first ed.
McGraw-Hill, New York.
Subscripts Ghanaei, E., Esmaeilzadeh, F., FathiKaljahi, J., 2007. J. Fluid Phase Equilib. 254, 126e137.
Javanmardi, J., . Nasrifar, Kh, . Najibi, S.H., . Moshfeghian, M., 2006. J. Appl. Therm.
C critical
Eng. 26, 1812e1819.
F feed Jeirani, Z., Lashanizadegan, A., . Ayatollahi, Sh, . Javanmardi, J., 2007. J. Nat. Gas
f fusion Chem. 16, 293e300.
i component number Ji, H., Tohidi, B., Danesh, A., Todd, A.C., 2004. J. Fluid Phase Equilib. 216, 201e217.
Katz, D.L., 1983. JPT 35, 1205e1214.
j component number Kesler, M.G., Lee, B.I., 1976. J. Hydro. Proc. 55, 153e158.
n component number Dorset, D.L., Snyder, R.G., 1996. J. Phys. Chem. 100, 9848e9853.
sub sublimation . Lira-Galeana, C., . Firoozababadi, A., . Prausnitz, J.M., 1996. J. AICHE 42, 239e248.
Marano, J.J., Holder, G.D., 1997. J. Ind. Eng. Chem. Res. 36, 1895e1907.
tot total Mei, H., Kong, X., Zhang, M., 1999. SPE 5165.
tr transition Nichita, D.V., Goual, L., Firoozababadi, A., 2001. J. SPE Prod. Facil. 16, 250e259.
vap vaporization Pan, H., Firoozabadi, A., 1997. SPE Annual Technical Conference and Exhibition, San
Antonio, SPE 38857.
Pan, H., Firoozababadi, A., Fotland, P., 1996. SPE 31640.
Pan, H., Firoozababadi, A., Fotland, E., 1996. SPE Annual Technical Conference and
References Exhibition Held in Denver, Colorado, USA, SPE 31640.
Pedersen, K.S., 1995. SPE Prod. Facil. 10, 46e49.
Ahmed, T.H., Cady, G.V., Story, A.L., 1985. SPE 14266. Pedersen, W.B., Hansen, A.B., Larsen, E., Nielsen, A.B., Ronningsen, H.P., 1991.
Al-Meshari, A.A., 2004. Phd thesis, New Strategic Method to Tune Equation of-State J. Energ. Fuels 5, 908e913.
to Match Experimental Data for Compositional Simulation, Texas. Pedersen, K.S., Skovborg, P., Ronningsen, H.P., 1991. J. Energ. Fuels 5, 924.
Cavett, R.H., 1962. Proc. of 27th API Meeting, San Francisco. pp. 351e366. Pedersen, K.S., Thomassen, P., Fredenslund, A. In: Chorn, L.G., Mansoori, G.A. (Eds.),
Chen, W., Zhao, Z., Zhang, X., Wang, L., 2007. J. Fluid Phase Equilib. 255, 31e36. C7þ Fraction Characterization, Taylor & Francis New York Inc, New York, 137 p.
Chueh, P.L., Prauznits, J.M., 1967. J. AICHE 13, 1099e1107. Rahimpour, M.R., Jamshidnejad, Z., Jokar, S.M., Karimi, G., Ghorbani, A.,
Chung, T.H., 1992. SPE 24851. Mohammadi, A.H., 2012. J. Nat. Gas. Sci. Eng. 4, 17e28.
Coutinho, J.A.P., 1998. J. Ind. Eng. Chem. Res. 37, 4870e4875. Riazi, M.R., Al-Sahhaf, T.A., 1996. J. Fluid Phase Equilib. 117, 217e224.
Coutinho, J.A.P., Andersen, S.I., Estenby, E.H., 1995. J. Fluid Phase Equilib. 103, 23e39. Riazi, M.R., Daubert, T.E., 1987. J. Ind. Eng. Chem. Res. 26, 755e759.
.Coutinho, J.A.P., .Knudsen, K., .Anderesen, S.I., .Stenby, E.H., 1996. J. Chem. Eng. Sci. South-Pars Gas Company, Asalooye, Iran, 2012.
51, 3273e3282. Twu, C.H., 1984. J. Fluid Phase Equilib. 16, 137e150.
Coutinho, J.A.P., Edmonds, B., Moorwood, T., Panski, R.S., Zhang, X., 2002. SPE 78324. Vafaie-Seftti, M., Mousavi-Dehghani, S.A., 2000. J. Fluid Phase Equilib. 173, 65e80.
Edminster, W.C., 1958. J. Petrol. Refiner. 37, 173e179. Won, K.W., 1986. J. Fluid Phase Equilib. 30, 265e279.
Erickson, D.D., Niesen, V.G., Brown, T.S., 1993. 2184. Won, K.W., 1989. J. Fluid Phase Equilib. 53, 377e396.