Articulo - Contenido de Agua - Metodos Analiticos

CHERD-1599; No.
of Pages 15
ARTICLE IN PRESS
chemical engineering research and design x x x ( 2 0 1 4 ) xxxxxx
Contents lists available at ScienceDirect
Chemical Engineering Research and Design

journal homepage: www.elsevier.com/locate/cherd
Analytical methods to calculate water content in

natural gas
Lin Zhu a, , Luling Li a , Jia Zhu a,b , Li Qin c , Junming Fan a
a
Key Laboratory of Gas Process Engineering, School of Chemistry and Chemical Engineering, Southwest Petroleum
University, Chengdu 610500, PR China
b Suining Field of Sichuan Oil and Gas Mine, Petro-China Southwest Oil & Gas Field Company, Suining 629000, PR
China
c Carbon Black Plant of Luzhou, Petro-China Southwest Oil & Gas Field Company, Luzhou 646000, PR China
a b s t r a c t
This work presents a review on the main analytical methods for estimating water content in natural gas samples of
different types and with different contents. The analytical methods for sweet natural gas are mainly developed for gas
under conditions of temperatures of 223.15510.93 K and pressures of 0.1100 MPa. The calculation of water content
in sour natural gas should be calculated in a temperature range from 288.15 to 444.26 K and a pressure range from 0.5
to 69 MPa. By combining correction methods with analytical methods for sweet natural gas randomly and comparing
the results with the experimental value published in the literature, it was found that the simplied thermodynamic
mode & Bahadori correction method, the simplied thermodynamic mode & Mohammadi correction method, the
Sloan & BukacekMaddox correction method and the Bahadori & Khaled correction method could reproduce with
the least AAD%. Additionally, by comparing the results of these analytical methods with the experimental data
in literatures, this work presented the optimum analytical methods for sweet natural gas and sour natural gas,
respectively, at different temperatures and at different pressures.
2014 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
Keywords: Water content; Sweet natural gas; Sour natural gas; Analytical method; Evaluate; Combination mode
1.
Introduction
the consumption of dehydrate agents and for predicting the aqueous

dew points.
In the initial design of the facilities of production, transmission, and

processing of natural gas (NG), the dissolved water in the gas phase
of NG is an essential factor to be considered. The main forms that the
dissolved water exists in are liquid water, gas hydrates and ice. The formation of a liquid phase may cause corrosion or/and two-phase ow
NG can be divided into two categories: sweet natural gas (gas whose
acid concentration is less than 5% of the gas mixture) and sour natural
gas (gas whose acid concentration is more than 5% of the gas mix-
problems. Additionally, a decrease in the temperature (or an increase
ture) (Zhu, 2008). Many related approaches have been conducted for
the estimation of water content in NG. To sum up, the methods for
sweet natural gas are composed of three main types: charts plotted with
in the pressure) will make the problems caused by the gas hydrates
or ice more worse. The acid gases contained in NG may generate acid
liquor, which would integrate with free water, resulting in the corrosion
limited experimental data, thermodynamic models based on phase

equilibrium, and empirical or semi-empirical correlations developed
with limited application. Because an increase in the amount of lique-
of pipeline, instruments, valves etc. Consequently, to prevent corrosion and to avoid the formation of ice or gas hydrate, it is necessary
to remove saturated water from the acid gas concentration by dehy-
ed H2 S or CO2 signicantly enhances the solubility of water, presence

of acid gases (i.e. hydrogen sulde and carbon dioxide) makes the water
dration facilities before transmitting and processing. For engineers,

accurate prediction of water content is the foundation of calculating
content increase obviously (Carroll, 2002; William et al., 2012). Therefore, some corrections including charts and correlations should be
applied, when it comes to analyzing water content in sour natural gas.
Corresponding author at: No.8 Xindu Avenue, Xindu District, Chengdu City, 610500, PR China. Tel.: +86 15884505696.
E-mail address: zhulinswpi65@gmail.com (L. Zhu).
http://dx.doi.org/10.1016/j.cherd.2014.05.021
0263-8762/ 2014 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
Please cite this article in press as: Zhu, L., et al., Analytical methods to calculate water content in natural gas. Chem. Eng. Res. Des. (2014),
ARTICLE IN PRESS
CHERD-1599; No. of Pages 15
Nomenclature
W
T
P
C
a
b
A
B

S
d
F
V
V
M
y
K
x
R
AAD
AD
a
b
Z
NG
MIM
STM
MTM
water content in mg/Sm3 (288.15 K, 0.101 MPa)

absolute temperature (K)
total pressure (MPa)
parameter in Eq. (1)
parameter in Eqs. (2), (4), (5), (6), (9), (23), (24),
(32) and (35)
parameter in Eqs. (4), (6), (29), (30), (31) and (35)
parameter in Eqs. (7) and (8)
the mole fraction of brine in water
the average relative molecular mass
the correction factor between sweet natural gas
and sour natural gas
the average molar volume over the pressure
interval P to Psw or P to P0
the molar volume
molecular mass
the mole fraction in the gas phase
the standard equilibrium constant
the mole fraction in liquid phase
the universal gas constant
the average absolute deviation
the absolute deviation
equation of state parameter
equation of state parameter
the compressibility factor
natural gas
modied ideal model
simplied thermodynamic model
modied thermodynamic model
Greek symbols

the chemical potential
fugacity coefcient
Superscripts
o
the standard situation
the equivalent situation
equi
l
the liquid phase
Subscripts
water
H2 O
H2 S
hydrogen sulde
carbon dioxide
CO2
CH4
methane
NHC
the non-hydrocarbon gases such as H2 S, CO2
HC
the hydrocarbon gases
sweet natural gas
sweet
sour natural gas
sour
sw
the water in saturation state
the critical situation
c
0
the reference situation
the component i
i
the mixture gas
mix
r
the reduce situation
In order to describe the water content of gas in equilibrium with

hydrocarbon, several charts have been presented, some of these being:
the McKettaWehe chart (GPSA, 1998), the Campbell chart (Campbell,
1991), the Katz chart (Khaled, 2007) etc. Of the available candidates,
the McKettaWehe chart is the most popular, with surprising accuracy

for sweet natural gas containing over 0.7 (mole fraction) of methane
(GPSA, 1998). Furthermore, some correction charts, i.e. the Wichert
chart (Wichert and Wichert, 1993), the Campbell chart (Campbell, 1991),
and the Robinson chart (Robinson et al., 1978), have been developed for
estimating water content of sour natural gas. However, because of the
difculty in obtaining accurate data and the need for interpolations,
charts cannot be widely used.
Based on the phase equilibrium, several thermodynamic models
are available for the estimation of water content in sweet natural gas.
Some models utilize the equality in the activity coefcient of water in
different phases, and others refer to the equation of state. Additionally, with the different models applied in liquidvapor, hydratevapor,
icevapor and liquidhydratevapor regions, thermodynamic models
always have high accuracy. However, these models are too complicated
to be performed by simple tools (Mohammadi et al., 2004a; Mohammadi
and Richon, 2007).
By matching the existing data to the equations, researchers
have obtained some empirical or semi-empirical correlations that
are simple, convenient and operated with high degree of accuracy.
Consequently, these correlations remain popular among engineers.
Increasing numbers of correlations such as the Bahadori method
(Bahadori et al., 2009) and the Behr method (Behr, 1983) have been
reported. Nonetheless, the presence of large amounts of heavy hydrocarbon may make the correlations, based on methane, (i.e., not a
heavy hydrocarbon), such as the Behr method (Behr, 1983) and the
SharmaCampbell method (Sharma and Campbell, 1969), have lower
accuracy. In general, most of the correlations are used for the sweet
natural gas containing few heavy hydrocarbons at the applicable conditions. Unfortunately, due to the shortage of consistent experimental
values, these correlations also need further verication for low temperature conditions.
Summarily, with the difculty of reading accurately, the need
for interpolation of charts as well as the complexity of thermodynamic models, the brief and non-robust analytical methods are
always popular because of their high accuracy, speediness, convenience, and their programmable nature. Analytical methods generally
include empirical correlations and semi-empirical correlations, as well
as some simplied thermodynamic models. With the limitation of
application, analytical methods are not available for every case, and
a comprehensive report about the optimal methods at given conditions is still absent. As the analytical methods for sour natural
gas are modied versions of the analytical method for sweet natural gas, the choice of which sweet natural gas analytical method
should be used, together with the correction method to estimate water
content in sour natural gas remains a controversial problem at all
time.
The main aim of this work is summarizing the main analytical
methods of estimating water content in natural gas by category and
choosing the optimal analytical method of sweet natural gas for every
correction method to accurately predict the water content in sour natural gas. Then, taking the results and the applicable range of each
analytical method into consideration, the optimum analytical method
for sweet or sour natural gas at specic temperature and pressure is
focused on.
2.
Analytical methods for sweet natural gas
The analytical methods for estimating water content in sweet

natural gas mainly consist of three catalogs, namely correlations originally regressed from chart data, correlations
originally regressed from experimental data, and equations
referring to the calculation of the phase equilibrium in
waterhydrocarbon systems. In this work, a review of the
analytical methods for estimating water content in sweet
natural gas is presented and is detailed in the following
sections.
ARTICLE IN PRESS
Table 1 Constants C in Eq. (1).

C1
21.58610805
C2
C3
C4
C5
C6
1.280044975
4808.426205
113.0735222
40377.6358
3.8508508 102
Table 2 Constants a in Eq. (2).

P/MPa
a0
30.0672
27.5786
27.8357
27.3193
26.2146
25.7488
27.2133
26.2406
26.1290
24.5786
24.7653
24.7175
0.1
0.2
0.3
0.4
0.5
0.6
0.8
1.0
1.5
2
3
4
2.1.
a1
0.1634
0.1435
0.1425
0.1383
0.1309
0.1261
0.1334
0.1268
0.1237
0.1133
0.1128
0.1120
a2
P/MPa
4
1.7452 10
1.4347 104
1.4216 104
1.3668 104
1.2643 104
1.1875 104
1.2884 104
1.1991 104
1.1534 104
1.0108 104
1.0113 104
1.0085 104
Correlations regressed from chart data
These correlations are generally from regressing the data of

McKettaWehe chart (GPSA, 1998), a chart considering the
effects of gas gravity and brine. With the nonlinear and complexity characters of the chart, most of these correlations are
failed to be included in the whole chart, except the method
proposed by Ning et al. (2000).
2.1.1.
Sloan method
For the low temperature region, Sloan (Kobayashi et al., 1987;

Sloan, 1998; Sloan et al., 1976) reported a method by providing
the variable water content as a function of T and P, as shown
by Eq (1):
= 16.02 exp C1 + C2 ln P +
C5
C3 + C4 ln P
2
+ 2 + C6 (ln P)
T
T
where the water content (WH2 O ) is in mg/Sm3 , the temperature (T) is in K, and the pressure (P) is in MPa. The subscript
sweet refers to sweet natural gas, and the constants C1 to C6
are listed in Table 1. This correlation is accurate for the temperature ranging from 233.15 to 323.15 K and for the pressure
ranging from 1.4 to 13.8 MPa.
Eqs. (2) and (3) conducted by Ning et al. (2000) illustrated the
correlations depending on the main chart and the correction
chart of McKettaWehe chart, respectively.
WH2 O,sweet = (714.855 + 1.1T + 369.673d 1.42Td)
My
i i
28.966
a2
0.1232
0.1128
0.1172
0.1133
0.0973
0.0946
0.0851
0.0881
0.0834
0.0773
0.0845
0.0838
1.1618 104
1.0264 104
1.0912 104
1.0425 104
8.4136 105
8.1751 105
7.0353 105
7.4510 105
6.9094 105
6.1641 105
7.1151 105
7.0494 105
and y are the molecular mass, and the mole fraction of component i in gas, respectively. In order to calculate the coefcients
a0 , a1 , a2 depending on the pressure, an integrated table has
been presented in Table 2. It should be noted that this correlation has taken the effect of gas gravity into account, and it
is more suitable for 0.1 MPa< P <100 MPa.
2.1.3.
Khaled method
Concerning that water content in NG tends to rise with

increasing temperature or decreasing pressure, Khaled
(2007) proposed the following equations regressed data of
McKettaWehe chart, Campbell chart and Katz chart.

5
a
i=1 i
T i1
5
i=1

bi T
i1
(4)
where a and b are parameters listed in Table 3. The applicable

temperature and pressure range of this correlation are from
310.93 to 444.26 K and from 1.38 to 68.95 MPa, respectively.
2.1.4.
Bahadori method
Bahadori (Bahadori et al., 2009) stated the water content was

a function of pressure and temperature and proposed a new
alternative to predict the water content in sweet natural gas,
as shown by Eq. (5):
3 3
Ning et al. method
(1 0.02247S) exp(a0 + a1 T + a2 T 2 )
a1

(1)
d=
26.8976
25.1163
26.0341
25.4407
22.6263
22.1364
20.4434
21.1259
20.2527
19.1174
20.5002
20.4974
5
6
8
10
15
20
30
40
50
60
70
100
WH2 O = 16.02
WH2 O,sweet
2.1.2.
a0
(2)
WH2 O,sweet = 10
i=0
a T
j=0 ij
j (log P+3)i
(5)
where a is the constant presented in Table 4. This method can

be used to estimate water content of sweet natural gas for
temperature ranging from 288.15 to 393.15 K and for pressure
ranging from 1 to 15 MPa.
Table 3 Constants a and b in Eq. (4).

(3)
where S and d are the mole fraction of brine in water, and the
average relative molecular mass, respectively. The average relative molecular mass d could be calculated by Eq. (3), where M
a1
a2
a3
a4
a5
706,652.14
8915.814
42.607133
0.0915312
7.46945 105
b1
b2
b3
b4
b5
2893.11193
41.86941
0.229899
5.68959 104
5.36847 107
ARTICLE IN PRESS
investigated the following equation:

a10
a11
a12
a13
a20
a21
a22
a23
2.2.
19.46736072
0.062742134
0.000238447
5.6739 107
30.07101715
0.171087078
0.000226883
3.12335 108
15.1343742
0.102010483
0.000216579
1.348433 107
1.955498725
0.013966992
3.26487 105
2.42011 108
a30
a31
a32
a33
a40
a41
a42
a43
Correlations regressed from experimental data
+A5 ln P + A6 (ln P) + A7 (ln P)

2
A8 + A9 ln P + A10 (ln P) + A11 (ln P)

T3

(7)
2.2.3.
Kazim method
According to the experimental data of water content at different temperature conditions, Kazim (1996) put forward an
analytical expression for calculating the water content in
sweet natural gas, as shown in Eq. (8)
Zhu et al. Method
Similar to the method of Khaled, Zhu et al. (Zhu, 2008; Zhu

et al., 2003) had reported an equation as below by regressing
water content data in hydrocarbon gas mixture at experimental condition. Due to the restriction of experimental data, the
applicable range of this method is a little complicated, as
exhibited in Fig. 1.
7
i

a (T 273.15)
i
WH2 O = 101.325
i=0
7
bi (T 273.15)
(6)
i=0
where a0 a7 , b0 b7 are constants listed in Table 5.
2.2.2.
A2 + A3 ln P + A4 (ln P)
T2
Coefcients A0 A11 were obtained on the basis of tting

dew point of natural gas versus the water content value and
showed in Table 6.
Except the method of Zhu et al. (Zhu, 2008; Zhu et al., 2003),
most of correlations (such as Behr method (Behr, 1983) and
Kazim method (Kazim, 1996)) regressed experimental data of
water content in equilibrium with methane.
2.2.1.

WH2 O = A0 exp A1 +
Behr method
Based on the experimental data of water content in methane

for pressure ranging from 1.379 to 20.679 MPa, Behr (1983)
WH2 O,sweet = 16.02A1 A1.8T459.67

2
Ai =
4
ai,j
145.0377 P 350 j1
600
(8)
(9)
j=1
where A is a parameter linked with pressure and calculated

by Eq. (9). The constants ai,j of Eq. (9) are listed in Table 7. This
approach is suitable for temperature up to 355.37 K and for
pressure ranging from 2 to 8.3 MPa.
2.3.
Equations referring to the calculation of phase
equilibrium in water-hydrocarbon systems
To avoid the limitation of application, a method based on the
calculation of phase equilibrium is a good choice. Combined
with the saturated vapor pressure of liquid water and ice,
Fig. 1 Comparison of the applicable range of each analytic method.

ARTICLE IN PRESS
Table 5 Constants a and b in Eq. (6).

a0
4.65925
b0
4.67351 101
a1
3.37802 101
a2
1.11426 102
a3
2.04372 104
a4
1.91021 106
a5
1.56275 108
a6
1.99046 1010
a7
-1.23039 1012
4.60019
8.68387 103
-4.65719 103
9.32789 105
2.06031 106
-4.78943 108
2.37537 1010
Table 6 Constants A in Eq. (7).

A0
16.02
A6
0.70827
A1
15.01438824
A2
1,397,805.38
A3
1.20055
A4
0.12061
A5
0.55096
0.09904
129,261,863
1.29214
0.25962
0.01738

j
Temperature ranges
310.93 K T 355.37 K
T < 310.93 K
1
2
3
4
a1,j
a2,j
a1,j
4.34322
1.35912
6.82391
3.95407
1.03776
0.02865
0.04198
0.01945
10.38175
3.41588
7.93877
5.8495
these correlations agree well with the experimental data, even

at low-temperature conditions. In these methods, the saturated vapor pressure is regarded as an essential factor, which
can be calculated as below (Mohammadi et al., 2004b)
Psw = 106 exp
73.649 7258.2
T
ln T + 4.1653 106 T 2
a2,j
1.02674
0.01235
0.02313
0.01155
As the report pointed out, this modied ideal model can get
a great agreement with experimental values for temperature
ranging from 273.15 to 477.59 K and for pressure up to 14.4 MPa.
2.3.3.
Simplied thermodynamic model (STM)
Depending on the phase equilibrium system, thermodynamic

model always has high accuracy. However, it is too difcult to
be performed by handheld calculations. To simplify the thermodynamic model, Mohammadi et al. (Mohammadi et al.,
2004a, 2004b) developed a simplied thermodynamic model
to estimate water content, considering the physical effects
directly, which was shown in Eq. (14):
7.3073

(10)
5 T 2 102.5115496
Psw,ice = 1.31579 101032.5576407/T+51.0557191log T0.0977T+7.035810

where Psw , subscript liq and ice correspond to the saturated
vapor pressure of water, and the liquid water, and ice, respectively.
2.3.1.
Model of saturated vapor pressure
In a waterhydrocarbon system of NG, hydrocarbon is normally saturated with water at reservoir conditions. Therefore,
this approach (Wang, 1994) assumes the partial water vapor
pressure equivalents to the saturated vapor pressure of water.
Related to the partial pressure law of Dalton, the water content
can be expressed by the following equation:
WH2 O,sweet =
2.3.2.
761900.42 Psw
P Psw
(12)
The ideal model is always used to estimate water content

at low-temperature conditions while the calculation accuracy
is not satised at high temperature. A new modied ideal
model (MIM), as the extension of ideal model, was presented
by Mohammadi et al. (2005) and expressed by Eq. (13).
WH2 O,sweet
761900.42 Psw
=
exp
P
761900.42 Psw exp

WH2 O,sweet =
11.81479 P0.92951
T

(13)
(PP
sw )VH2 O
RT
(14)
H2 OP
where H2 O and VH2 O are respectively the fugacity coefcient of water in the gas phase calculated by Eq. (15), and
the average molar volume of condensed water over the pressure interval PPsw . The calculation of VH2 O varies from the
different formations of water in NG and calculated by Eqs.
(16) and (17). R is the universal gas constant which equals
to 8314 m3 MPa mol1 K1 . Additionally, as reported in other
references, this correlation is accurate only for temperature
ranging from 273.15 to 377.59 K and for pressure up to 13.8 MPa.
Modied ideal model (MIM)
(11)
H2 O = exp
(0.06930.905)
T

P+
0.3179
T
0.0007654 P2
(15)
VH2 O,Liq = 4.501 102 6.710 104 T + 1.784 106 T 2
(16)
VH2 O,ice = 19.655 + 0.0022364 (T 273.15)
(17)
ARTICLE IN PRESS

T/K
CO2
a0
299.8167
310.9278
327.5944
344.2611
2.3.4.
2.202797
2.496936
2.930297
3.18733
a1
1.058120906
1.036484877
0.990489109
0.924518418
T/K
a2
a0
0.3427
0.3103
0.2400
0.2139
299.8167
310.9278
327.5944
344.2611
377.5944
410.9278
Bukacek method
3.1.1.
Unlike the above methods, by amending the ideal model

(Carroll, 2002, 2003), Bukacek also provided a calculated
method integrated with the method of estimating saturated
vapor pressure of SaulWagner (Bukacek, 1955; Carroll, 2003),
which is expressed as below
WH2 O,sweet =
B = 10
761900.42 Psw
+ 16.016 B
P
1713.26/T+6.69449

Psw = Pc exp
=
a1
2.314489
2.544338
2.890856
3.241253
3.773303
4.277016
0.678582201
0.70422576
0.799408896
0.896948037
1.030222461
1.255345485
0.3004
0.3046
0.3319
0.3646
0.4232
0.4897
BukacekMaddox correction method
Maddox and Bukacek (Maddox et al., 1988) proposed a correction approach to complete the component contribution model.
This correlation is expressed as Eq. (23) and it is accurate under
the condition of a pressure range from 0.7 to 20.7 MPa, and a
temperature range from 300.15 to 411.15 K for H2 S and from
300.15 to 344.15 K for CO2 .
log WNHC = a0 + a1 log P + a2 (log P)
(23)
(19)
Tc (7.85823 + 1.83991 1.5 11.7811 3 + 22.6705 3.5 15.9393 4 + 1.77516 7.5 )

T
Tc T
Tc
a2
(18)
(21)
where Tc and Pc are the critical temperature of water and the

critical pressure of water, respectively. It is reported that the
error of this method is within 5% for temperature ranging
from 288.15 to 511.15 K and for pressure ranging from 0.1 to
69 MPa.
3.
H2 S
Analytical methods for sour natural gas
When acid gases are contained in the NG, the analytical methods for sweet natural gas may have lower accuracy due to the
higher solubility of water in acid gases, compared with hydrocarbon, and some corrections should be made to the above
methods to correct for the additional water content. General
correction methods are composed of three aspects: corrected
by Maddoxs component contribution model, corrected by an
equivalent mole fraction of Robinsons H2 S model, or corrected
by directly taking the effect of acid gas on the physical properties into account.

(20)
where subscript NHC stands for H2 S or CO2 . The temperature

dependent coefcient a is given in Table 8. The contribution
of sweet natural gas, as the reference (Maddox et al., 1988)
reported, can be predicted by McKettaWehe chart.
3.1.2.
Bahadori correction method
Bahadori (Bahadori et al., 2009) designed a new correction

method as below to provide good information about the equilibrium water content in sour natural gas for a larger range
of conditions, including temperature ranging from 298.15 to
393.15 K and pressure ranging from 1 to 15 MPa. Meanwhile,
Bahadori also developed a method introduced in Section 2.1.4
for the estimation water content in sweet natural gas, which
is always used together with this correction method.
WNHC =
3 3
i=0
j=0
aij Pj T i
(24)
where a is constant listed in Table 9.
3.1.
Correlations corrected by Maddoxs component
contribution model
Maddoxs component contribution model (Carroll, 2003;
Maddox et al., 1988) assumes water content in sour natural
gas is the sum of contribution made by sweet natural gas, H2 S
and CO2 . Thus, it can be expressed as follows:
WH2 O,sour = yHC WH2 O,sweet + yCO2 WCO2 + yH2 S WH2 S (22)
where y is the mole fraction, and the subscript HC, sour
refers to hydrocarbon and sour natural gas. This model is
applicable only for the mixture gas whose acid gas is less than
0.4. The contribution of sweet natural gas can be calculated by
an appropriate correlation or chart discussed as before. Then
the estimation of water content in acid gas is the most critical
step as below
Effective water
content of H2 S
a10
a11
a12
a13
a20
a21
a22
a23
a30
a31
a32
a33
a40
a41
a42
a43
2.28473455055 106
4.10134356738 105
0.0233870949722 106
1.55604987964 102
2.28569862175 104
4.28913302821 103
2.84221095695 102
4.35332471889
76.4984711141
0.0149326955228 103
1.11025652987
2.34671115643 102
0.0857373579771
1.73248422725 102
1.41055829053 103
3.56668823871 105
Effective water
content of CO2
1.51705929796 106
2.53953558137 105
0.0218779871207 106
7.00199672753 102
1.54692319504 104
2.75431390822 103
2.51540890528 102
8.37305127559
52.8444857343
9.97118369879
9.5940533224 101
3.30375581283 102
0.0605410197916
1.20630929063 102
1.21662951101 103
4.31605047324 105
ARTICLE IN PRESS
Table 10 Constants a and b in Eqs. (29)(32).
P < 10.34 MPa

10.34MPa20.68 MPa
P > 20.68 MPa
a0
4.095 102
a1
1.82865639 103
a2
1.93733 101
b0
3.59 101
5.16 102
1.04
b1
7.46 104
2.84 102
5.48 102
b2
4.7282 104
1.25249 102
23.6857
P < 10.34 MPa:
3.2.
Correlations corrected by an equivalent mole
fraction of Robinsons H2 S model
Analyzing the experimental data of water content in CO2 and

H2 S, Robinson discovered the saturated water content of CO2
was 0.75 times equivalence to saturated water content of a
comparable amount of H2 S under same conditions and developed an equivalent mole fraction of H2 S model calculated by
the below expression (Robinson et al., 1980).
F = ln
1
b0 + Requi (b1 + b2 P)

(29)
10.34 MPa P 20.68 MPa:
F = exp(b0 + Requi (b1 + b2 P)
(30)
P > 20.68 MPa:

equi
2S
yH
= yH2 S + 0.75 yCO2
(25)
where the superscript equi refers to equivalent H2 S. This

equi
model is suitable for mixture gas with yH S less than 0.4 (mole
2
fraction) (Carroll, 2003; Robinson et al., 1980).
On the basis of the equivalent mole fraction of H2 S model,
Wichert (Wichert and Wichert, 1993) proposed a correction
factor F combined water content of sweet natural gas and
water content in sour natural gas directly.
equi
, T, P)
2S
F=
b2
b0 + Requi b1 +
P
2
(31)
In the above equations, the coefcient Requi relates to T and

the equivalent mole fraction of acid gases is calculated by Eq.
(32).
Requi = 1/(a0 +
T(a1 + a2 /
equi
))
2S
(32)
yH
F = f (yH
(26)
where a and b are constants listed in Table 10.
WH2 O,sour = F WH2 O,sweet
(27)
3.3.
Correlations corrected by taking the effect of acid
gas on the physical properties into account
where the subscripts sour and sweet respectively refer to

sour natural gas and sweet natural gas, and f corresponds to
a function. There are mainly two methods to estimate the
correction factor F as following.
3.2.1.
Mohammadi correction method
Considering correction factor F as a variable of the function of

T, P and also the equivalent mole fraction of acid gas, Mohammadi (Mohammadi et al., 2004b, 2005) developed the following
equation to calculate F.
equi
2S
F = 1 + yH
0.03185 T
T0
0.01538 TP
0.02772 P
+
T0 P0
P0
3.3.1.
(28)
In the above equation, T0 and P0 are reference pressure

(0.101 MPa) and reference temperature (273.15 K), respectively.
The Mohammadi correction method should be used in a temperature range from 310 to 420 K and a pressure range from
0.5 to 35 MPa. To estimate water content of sour natural gas
by this way, according to Mohammadi et al., 2004b, the water
content in sweet natural gas should be rstly calculated by
McKettaWehe chart or the modied ideal model.
3.2.2.
The difference in water content between sweet natural gas and

sour natural gas results in a difference in the physical properties of the NG. Researchers quantied these effects using some
simple correlations and developed analytical methods for sour
natural gas. All of the analytical methods are based on the calculation of phase equilibrium in waterhydrocarbonacid gas
system and do not require the estimation of the water content
in sweet natural gas in advance.
WH2 O,sour =
761900.42 Psw,sour (1 ysalt yH2 S yCO2 )

P Psw,sour (1 ysalt yH2 S yCO2 )

T
Psw = Pc exp f
Khaled correction method
With different expressions in different pressure conditions,

Khaled (Khaled, 2007) provided a new approach for calculating
the correction factor F on general condition as well as on high
temperature or pressure conditions. This correction method
with the same applicable range as Khaled method is always
used together with to estimate water content in sour natural
gas, and is expressed as below
Wang et al. method
In waterhydrocarbon system, the partial water vapor pressure referred to the saturated vapor pressure of water. Based
on this theory, researchers designed Eq. (12) to estimate water
content in sweet gas. Similarly as in waterhydrocarbonacid
gas system, Wang et al. (Wang, 1994) proposed the following
corrected equations.
T
Tc
Tc
Tc
T
= 7.21275 + a 0.745
(33)
T
Tc
(34)
2
+ b 0.745
T
Tc
c
(35)
where Tc , Pc and the subscript salt are the critical temperature of water (647.3 K), the critical pressure of water
(22.12 MPa), and the salts, respectively. The constants a, b and
c applied to T and Tc are equal to 3.981, 1.05 and 3, respectively,
T/K
P/MPa
Exp. data/(mg/Sm3 )
Predictive value/(mg/Sm3 )
14.6021
18.0607
250.847
415.673
304.788
246.447
278.702
367.243
795.834
14.5432
25.6841
268.901
1048.63
445.585
324.565
268.945
1419.91
308.413
1887.71
399.944
3292.19
406.531
509.728
845.300
5605.50
553.534
Bahadori
Zhu et al.
839.3556
498.6448
410.1247
303.3382
249.4215
1358.927
289.2065
1828.822
382.9775
3246.476
361.1246
466.2574
841.8197
5580.316
493.7654
14.8818
242.393
1028.684
437.9177
327.8632
272.8369
1428.40
304.246
1930.20
385.204
3436.41
349.264
441.705
806.736
5919.36
466.121
14.2327
23.2376
175.224
412.662
281.239
206.388
202.383
265.260
641.584
18.234
27.090
415.983
301.591
279.275
Mohammadis gases mixture (94 mol% methane + 4mol%ethane + 2mol% n-butane)

6.020
239.237
304.0291
269.345
288.15
6.000
540.949
531.5104
544.365
297.93
10.05
318.474
362.7582
335.939
297.95
303.10
1.068
3276.17
2335.575
3.661
303.00
967.614
1031.461
1092.023
303.13
17.56
338.284
406.1704
17.50
592.000
637.4745
594.6245
313.14
1.924
4906.64
4657.062
5585.577
4933.883
322.00
4.174
2323.80
2390.901
2152.614
2456.676
321.80
2.915
5500.92
333.00
6945.25
5701.111
347.30
0.567
52,494.9
51478.98
4.115
7542.81
347.20
6791.777
7760.225
361.00
4.599
11580.9
12809.3
12081.38
0.954
53180.7
361.20
62839.75
302.61
536.191
378.3929
3052.526
1037.932
361.0989
596.8615
4819.883
2466.908
5629.553
49774.16
7734.233
12065.65
52386.34
327.1308
543.2643
381.0635
3226.095
1050.534
349.1989
563.7447
5029.846
2487.332
5826.277
51855.76
7912.374
12165.24
54351.49
280.2468
498.4195
283.2702
1041.408
202.3638
343.6993
3879.429
2438.266
5341.693
7506.226
11537.76
300.8260
523.2586
1045.680
2375.413
5373.133
7350.37
12581.51
CH4 + H2 O
240.00
254.00
268.15
282.98
288.15
288.15
288.15
288.15
293.11
293.11
298.11
303.11
303.13
308.12
313.11
313.11
313.12
3.450
6.900
1.500
1.000
4.000
6.000
8.000
1.000
10.00
1.000
10.00
1.000
17.56
17.50
10.00
1.000
25.00
9.88746
16.63987
225.0804
1125.402
497.7619
334.8355
253.3722
1430.868
271.7042
1929.26
378.6174
3536.977
230.3565
282.9065
840.0322
5900.322
503.2154
AAD%a
the corresponding method cannot be used in this situation.
a

N

(WExp WPre )/WExp
AAD% = 100
N
i=1
Khaled
Behr
215.719
287.491
Kazim
Model of Psw
MIM
STM
Bukacek
5.98233
12.26818
202.063
925.677
324.962
216.610
162.446
1301.51
177.830
1782.05
241.052
3242.80
184.143
244.801
561.981
5657.37
224.813
963.972
376.916
268.981
215.601
1353.67
250.457
1851.01
337.526
3357.41
773.971
5831.62
962.146
376.489
269.124
215.840
1349.50
247.909
1843.28
330.728
3336.32
737.138
5783.53
290.952
1890.51
382.742
3403.44
360.640
462.659
828.651
5879.18
492.176
215.8531
397.4085
237.5425
3022.463
876.6712
4621.751
2109.166
5180.601
50086.19
6872.304
10723.53
52095.43
267.8381
483.0809
325.4228
3127.808
983.9658
4857.855
2347.206
5533.208
50563.79
7420.8
11498.79
52765.8
537.8864
378.2609
3196.166
1051.559
360.6383
588.9837
4995.013
2482.544
5741.59
50931.13
7782.981
12122.81
53411.67
215.8906
397.5418
237.5722
3033.762
877.4689
184.1431
321.5448
4647.219
2113.789
5211.205
53518.93
6923.733
10848.38
55764.28
ARTICLE IN PRESS
Ning et al.
Sloan
Table 11 The Comparison of the estimating value of predictive approach and experimental data for sweet natural gas (Mohammadi et al., 2004b, 2005).
ARTICLE IN PRESS
Table 12 The optimum method for sweet natural gas at different conditions.
The range of T
The range of P
Suitable method
223.15273.15 K
273.15288.71 K
P0.00054053T 0.419T + 108.1492T 9274.1

P13.8 MPa
13.8100 MPa
P13.8 MPa
13.815 MPa
15 MPa69 MPa
P13.8 MPa
13.815 MPa
1569 MPa
69100 MPa
0.11.38 MPa
1.3869 MPa
69100 MPa
0.11.38 MPa
1.3869 MPa
69100 MPa
0.169 MPa
69100 MPa
Zhu et al.
STM
Zhu et al.
STM
Bukacek/Bahadori
Bukacek
STM
Bukacek/Bahadori
Bukacek
Zhu et al.
Zhu et al.
Khaled
Zhu et al.
Bukacek
Khaled
Ning et al.
Bukacek
Ning et al.
288.71295.15 K
295.15310.93 K
310.93413.15 K
413.15444.26 K
444.26510.93 K
while Tc is up to T. On the contrary, a is equivalent to 4.33, b is

equivalent to 185 and c is equivalent to 5.
3.3.2.
Modied thermodynamic model (MTM)
In comparison to sweet natural gas, higher water holding of

sour natural gas mainly results in the high solubility of polar
molecules (such as H2 S, CO2 ) (Mohammadi et al., 2006; Zirrahi
et al., 2010). In terms of the main factors of acid gas solubility, researchers corrected Eq. (14) by taking these factors into
consideration (Zirrahi et al., 2010) and proposed a modied
thermodynamic model (MTM), shown as below
W H2 O,sour =
o
761900.42 KH
state are given in Appendix A (Salari et al., 2008; Zirrahi et al.,

2010).
o
log KH
= 2.209 + 0.03097(T 273.15)
2 O,sour
1.098 104 (T 273.15)

+ 2.048 107 (T 273.15)
2 O,sour
(1 xH2 S xCO2 xCH4 ) exp (P P0 )VH2 O,sour /RT
(37)

(36)
H2 O,sour P
o
where KH
, x and VH2 O,sour are the standard equilib2 O,sour
rium constant of water in sour natural gas calculated by
Eq. (37) (Zirrahi et al., 2010), the mole fraction of acid gases
or hydrocarbons in the aqueous phase estimated by Eq.
(38) (Salari et al., 2008), and the average partial molar volume over the pressure interval P0 P which is equivalent to
18.18 106 m3 /mol. The fugacity of water in the gas phase
can be calculated by the modied RK-EOS which is widely used
to express the deviation of a CO2 + H2 S + H2 O mixture from
ideal behavior. The details of the implemented equations of
l(0)
ln(yi 10 P/55.508 xi ) =
i
RT
ln i
(38)
l(0)
where i
and i are the standard chemical potential of
component i in aqueous phase and the fugacity of component i. Duan and his co-workers (Duan and Sun, 1992, 2003;
Duan et al., 2007; Ziabakhsh-Ganji and Kooi, 2012) provided a
method listed in Appendix B to calculate these parameters.
Table 13 The AAD% of the combination formulas.

Sour
Sweet
BukacekMaddox
Sloan
Ning et al.
Khaled
Bahadori
Zhu et al.
Behr
Kazim
Model of Psw
MIM
STM
Bukacek
5.394
8.037
9.409
10.15
9.731
10.22
8.547
10.73
10.84
7.428
9.339
Bahadori
8.456
13.38
7.513
8.433
8.023
17.75
8.455
13.59
19.98
7.311
7.555
Mohammadi
5.784
8.360
6.736
8.071
7.107
13.46
12.14
16.46
21.96
4.895
6.631
Khaled
7.389
9.906
6.404
5.502
5.825
20.41
11.86
21.92
21.81
7.050
6.6133
ARTICLE IN PRESS
10
4.
Discussions and results
Due to the non-independency of estimating water content

in sour natural gas, this work rst discussed the analytical
methods for sweet natural gas. Fig. 1 shows the applicable
range of the analytical methods, except for the model based
on saturated vapor pressure without reporting in reference.
As shown in Fig. 1, the analytical methods of sweet natural
gas are used for estimating water content in a temperature
range from 223.15 to 510.93 K and a pressure range from 0.1
to 100 MPa, with the majority of the method being available
for temperature ranging from 288.15 to 413.15 K and pressure
ranging from 1.38 to 13.8 MPa. There is still little research in the
literature commenting on analytical methods at lower temperatures and pressures. The Zhu et al. method, Ning et al.
method, Bukacek method and Khaled method have a larger
range of temperatures and/or pressures, compared to other
methods.
As the effect of hydrocarbons can be ignored, at
least to some extent, the water content of sweet natural gas is affected only by temperature and pressure.
Therefore, this work assumes the water content in sweet
natural gas is independent of gas composition and studies the systems of pure methane and gas mixtures
(94 mol%methane + 4 mol%ethane + 2 mol% n-butane), respectively. Table 11 shows the results of the analytical methods of
sweet natural gas and the comparison with the experimental
values of the references (Chapoy et al., 2005; Zirrahi et al., 2010;
Kondori et al., 2013; Eslamimanesh et al., 2011). As shown in
Table 11, due to the formation of ice or hydrate, not all of the
analytical methods are available at low temperatures and/or
pressures.
Table 11 shows the predicted values for a model based
on saturated vapor pressure and the modied ideal model
are closed. Nevertheless, as the shortage of the applicable
range, the modied ideal model has an AAD% (the percentage of average absolute deviation) (Mohammadi et al., 2004b)
of 10.351 which is less than the AAD% of 20.392 regenerated
by the model based on saturated vapor pressure. According
to the relationship between AD% (the percentage of absolute
deviation) of the model based on saturated vapor pressure,
the results show the AD% increases with increasing pressure.
Especially at high-pressure conditions, the AD% is extremely
unacceptable. It may be caused by the non-ideality situation under condition of high-pressure. The Behr method also
shows large deviations with an AAD% of nearly 20. Both
Ning et al. method and Kazim method need interpolation,
as the nonlinear function of temperature or pressure, these
methods may lead to the higher AAD% with wider applicable range. As shown in Table 11, Ning et al. method with
a larger applicable range is reproduced with a larger AAD%
of 16.930, compared to Kazim method. The other analytical
methods show good agreement with the experimental data
(Chapoy et al., 2005; Mohammadi et al., 2004b), especially
the Khaled method, produced with AAD% of 2.3186. However,
Khaled method is only available at the temperature range
from 310.93 to 444.26 K. When out of the applicable range,
the other method should be requested. Taking the applicable range of each approach into consideration, this work
makes a suggestion about the optimum method at specic
conditions and listed them in Table 12. In order to avoid the
inuences of selected samples, this work considers methods
within a 0.5 deviation from the optimum as being acceptable.
The analytical correction methods corrected by the component contribution model or the equivalent mole fraction
of the H2 S model should be used together with the methods
of estimating water content in sweet natural gas. Except the
BukacekMaddox correction method is reported partially, the
other methods have been reported the corresponding method
to estimate water content in sweet natural gas. The report
studying the optimum method of sweet gas for the correction
method to generate the less AAD% is absent.
To nd out the best combination modes, combining the
methods for sweet natural gas and the correction methods
randomly, then calculating water content in allusion to the
gas samples (GPSA, 1998; Lukacs and Robinson, 1963; Ng et al.,
2001) given in Table 14 is necessary. From Fig. 1, we could get
the applicable range of each combination method by nding
the common applicable range of the method for sweet natural gas and the correction method. As can be seen, due to
the restriction on the application of correction method, most
of the combination modes, expect some methods corrected
by the Khaled correction method, are only available in general situations. The comparison between the results of these
combination formulas and the experimental data is listed in
Table 13.
As shown in Table 13, there is an obvious difference among
the AAD%s of each combination mode, so choosing an optimum one is extremely important. When the BukacekMaddox
correction method is used, all the AAD%s among the experimental and predicted value lie in an acceptable ranging from
5.394 to 10.84. The other correction methods are sensitive
to the method for sweet natural gas, especially the Khaled
correction method which would regenerate AAD%s ranging
from 21.92 to 5.502. For these sensitive correction methods,
choosing an optimum method to estimate the water content
in sweet natural gas is extremely important. Studying the
results of each method for sweet natural gas using together
with each correction method, we can know the Behr method,
the model based on saturated vapor pressure and the modied ideal model reproduce with AAD%s of more than 10, no
matter with which correction method. On the contrary, the
results of the combination modes related to the Sloan method,
the Khaled method, the Zhu et al. method, the simplied
thermodynamic model and the Bukacek method are always
acceptable. As Table 13 shows, for BukacekMaddox correction method and Khaled correction method the optimum
method to estimate water content in sweet natural gas are the
Sloan method and the Bahadori method. Both the Bahadori
correction method and the Mohammadi correction method
reproduce the least AAD% combining with the simplied thermodynamic method. Table 14 lists the results of the four
combination modes, Wang et al. method and modied thermodynamic model with a comparison of experimental data
of methane + carbon dioxide gas mixture, methane + hydrogen
sulde gas mixture and methane + carbon dioxide + hydrogen
sulde gas mixture, respectively.
As it can be seen, the result of Wang et al. method is
unacceptable, with an AAD% of more than 40. The modied
thermodynamic model has a great agreement with the experimental value. Nevertheless, the calculation of this model is
so difcult that could not be accepted in industry, except for
the samples with more than 0.4 (mole fraction) of acid gases.
Similarly to the methods for sweet natural gas, these analytical methods also have limited application. As shown in Fig. 1,
the calculation of water content in sour natural gas could be
worked for temperature between 288.15 and 444.26 K and the
T/K
P/MPa
Exp. value/(mg/Sm3 )
Mole fraction
CH4
CO2
H2 S
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
Methane + hydrogen sulde gas mixture

10.300
0.9200
0.0000
327.15
6.9000
0.8300
0.0000
344.15
9.4300
0.7300
0.0000
344.15
2.4680
0.7900
0.0000
344.26
4.2120
0.8100
0.0000
344.26
6.3760
0.7100
0.0000
344.26
6.9620
0.8300
0.0000
344.26
9.5950
0.7300
0.0000
344.26
9.6160
0.8400
0.0000
344.26
0.0800
0.1700
0.2700
0.2100
0.1900
0.2900
0.1700
0.2700
0.1600
Methane + carbon dioxide + hydrogen sulde gas mixture

1.3790
0.7500e
0.0625
0.1875
322.04
1.3790
0.7500e
0.1875
0.0625
322.04
10.339
0.7500e
0.1875
0.0625
322.04
10.339
0.7500e
0.0625
0.1875
322.04
1.3790
0.7500f
0.1875
0.0625
366.48
1.3790
0.7500f
0.0625
0.1875
366.48
4.1360
0.7500e
0.1875
0.0625
366.48
10.339
0.7500f
0.1875
0.0625
366.48
10.339
0.7500f
0.0625
0.1875
366.48
Sloan &
BukacekMaddox
906.66
929.52
1340.9
891.42
883.80
615.62
615.62
4320.0
4259.0
1683.8
4411.4
3733.3
11429
7085.7
5257.1
4693.3
3961.9
3619.0
7136.7
6659.0
1425.5
1382.1
451,817
45,466
22,798
8517.33
7960.3
602.81
628.33
6717.6
6719.4
1475.5
1679.1
STM &
Bahadori
915.73
911.78
579.31
577.78
4387.6
4464.0
STM &
Mohammadi
586.60
634.94
4287.9
4434.8
Bahadori &
Khaled
655.02
655.09
4600.5
4600.6
Wang et al.
2268.82
2233.26
1700.52
1508.00
1484.39
441.787
397.090
3619.65
3252.14
MTMa
625.41
649.52
4539.2
4398.3
1613.1
4605.0
4100.5
10503
6727.2
5074.4
4598.3
4084.3
3756.4
1510.5
4478.3
3784.0
10973
6805.5
5047.0
4466.9
3757.5
3482.9
1647.6
4600.7
3636.0
10938
6882.1
4915.6
4589.5
3606.2
3600.1
1236.48
3374.61
2153.59
9083.39
5432.58
3136.78
3359.22
2125.80
2458.63
1690.4
4441.2
3678.2
11,742
7318.0
5546.1
4856.1
4156.9
3806.1
6204.9
6131.3
1525.1
1467.2
39,917
40,432
15,240
8588.8
8720.6
6842.5
6823.0
1319.9
1352.5
45,284
45,359
16,321
7697.2
7841.4
6692.1
6692.2
1298.8
1298.8
45,705
45,705
16,558
7874.9
7875.1
6075.3
6075.3
805.03
805.03
37,117
37,117
12,006
4760.0
4760.0
7287.3
6760.5
1329.1
1229.1
46,304
46,902
18,305
8532.8
7312.2
AAD%
a
b
c
d
e
f
The modied thermodynamic model.

Composition of this mixture is 86.5% methane, 6.2% ethane, 1.6% propane, 0.7% C4 + , 4.9% nitrogen, 0.1% helium.
Consists of this mixture is 84.9% methane, 3.6% ethane, 0.7% propane, 0.3% C4 + , 10.5% nitrogen.
Composition of this mixture is 93.6% methane, 3.4% ethane, 0.5% propane, 0.4% C4 + , 2.0% nitrogen, 0.5% helium.
This mixture composes of 95% methane and 5% propane.
This mixture consists of methane, ethane, propane, i-butane, n-butane with a molar ratio equal to 90:6:2.5:0.6:0.9.
ARTICLE IN PRESS
Methane + carbon dioxide gas mixture

0.5000
0.9983b
0.0017
273.15
0.5000
0.9827c
0.0173
273.15
0.5000
0.7488d
0.2512
273.15
1.5000
0.9983b
0.0017
288.15
1.5000
0.9827c
0.0173
288.15
13.800
0.8900
0.11
311.15
13.800
0.8000
0.20
311.15
6.9000
0.8900
0.11
344.15
6.9000
0.8000
0.20
344.15
Prediction value/(mg/Sm3 )
11
Table 14 The comparison of the estimating value and experimental data (GPSA, 1998; Lukacs and Robinson, 1963; Ng et al., 2001).
12
ARTICLE IN PRESS
Table 15 The optimum method for sour natural gas at different conditions.
The range of T
288.15299.82 K
299.82310 K
310310.93 K
310.93420 K
420444.26 K
The range of P
Optimum method (method for sweet & correction)
113.8 MPa
13.815 MPa
0.41.38 MPa
1.3813.8 MPa
13.820.7 MPa
0.513.8 MPa
13.815 MPa
1535 MPa
0.513.8 MPa
13.815 MPa
1535 MPa
1.3813.8 MPa
1.3869 MPa
STM & Bahadori/Bukacek & Bahadori

Bukacek & Bahadori/Khaled & Bahadori
STM & Bukacek-Maddox
Solan & Bukacek-Maddox
Ning et al. & Bukacek-Maddox
STM & Mohammadi
Bahadori & Mohammadi/Zhu et al. & Mohammadi
Zhu et al. & Mohammadi
STM & Mohammadi
Bahadori & Khaled/Zhu et al. & Khaled
Zhu et al. & Khaled
Bahadori & Khaled/Zhu et al. & Khaled
Zhu et al. & Khaled
pressure between 0.5 and 69 MPa. Table 15, by comparing the

AAD%s among these analytical methods, suggests the optimum method for sour natural gas with a mole fraction of acid
gases of less than 0.4 at different conditions.
5.
Conclusions
This paper presented a review of the mainly widely used

analytical methods of the estimating water content in NG.
Additionally, this paper classied and evaluated these methods, as well as suggesting optimum approaches at different
conditions.
The analytical methods of calculating the water content
in sweet natural gas are mainly divided into three kinds: correlations regressed from charts, correlations regressed from
experimental data, and equations referred to the calculation
of the phase equilibrium in the waterhydrocarbon system.
The analytical methods for sour natural gas always need correction, as methods are usually initially devised to calculate
the water content in sweet natural gas. The correction method
has three aspects: corrected by Maddoxs component contribution model, corrected by an equivalent mole fraction of
Robinsons H2 S model, and corrected by directly taking the
effect of acid gases on the physical properties into account.
The simplied thermodynamic model is the best choice
for the Bahadori and the Mohammadi correction methods to
estimate water content in sweet natural gas. Among these
methods used the method of the BukacekMaddox to correct
the correlations of estimating water content in sweet natural gas, the AAD% of Sloan and BukacekMaddox correction
method is the least with a value of less than 5.394. At the same
time, using the method of Khaled to correct, the best partner
is Bahadori method with an AAD% of less than 5.502%.
Comparing the experimental data of references with the
calculation results of these correlations, and taking the applicable range of each method into consideration, this article had
made a suggestion about the optimum method to estimate

water content in NG at different conditions and reproduce
with the less AAD%.
Limited to the experimental data in the low temperature
region, the analytical method of estimating water content of
natural gas at low temperature conditions is still absent. The
results presented are suitable for sweet natural gas at the temperature over 233.15 K and for sour natural gas at temperature
over 288.15 K. Therefore, when new methods for the calculating water content in low temperature are proposed, the results
developed here will be inaccurate.
Acknowledgements
The authors would like to thank the nancer, the supporters,
and co-workers in key laboratory of gas process engineering.
Appendix A. Equations of state for H2 O and

mix rules
The RedlichKwong equation of state (RK-EOS) is expressed
by:
P=
RT
a
+ 0.5
Vb
T V(V + b)
(A1)
where a and b are the intermolecular attraction and repulsion

listed in Table A1. For mixtures, amix and bmix are calculated
by standard mixing rules as follows, some of the calculated
results are also listed in Table A1.
amix =

i
bmix =
yi yj (ai aj )
0.5
(A2)
yi bi
(A3)
Table A1 Intermolecular attraction and repulsion of acid gases and water..

aCO2
bCO2
aCH4
7.54 108 4.13 105 T (Zirrahi et al., 2010)

27.8 (Ziabakhsh-Ganji and Kooi, 2012)
4.9838 105 (Zirrahi et al., 2010)
aH2 OCO2
aH2 O-H2 S
aH2 O-CH4
bCH4
aH2 S
bH2 S
bH2 O
29.58 (Zirrahi et al., 2010)

5.1319 105 (Zirrahi et al., 2010)
30.01 (Zirrahi et al., 2010)
18.18 (Ziabakhsh-Ganji and Kooi, 2012)
aH2 O-CO2 -CH4

aH2 O-H2 S-CH4
aH2 O-H2 S-CO2
aH2 O-H2 S-CH4 -CO2
7.89 108 (Zirrahi et al., 2010)

2,978,194,521.95218264T (Zirrahi et al., 2010)
42,151.44P + 3.257690003 108 9205.39T + 29,371,125.57
(Zirrahi et al., 2010)
2.664509778 108 (Zirrahi et al., 2010)
3.823642554 108 (Zirrahi et al., 2010)
5.0856254151 109 (Zirrahi et al., 2010)
1.28744589983 1010 (Zirrahi et al., 2010)
The units of ai and bi are MPa cm6 K0.5 mol2 and cm3 /mol, respectively.
ARTICLE IN PRESS
13
Table B1 Parameters c in Eqs. (B1)(B3).
c1
c2
c3
c4
c5
c6
c7
c8
c9
c10
H2 S
CO2
CH4
42.564957
8.6260377 102
6084.3775
102.76849
6.871443 105
3.5665902 102
10.590768
8.4482895 103
0
0
28.9447706
0.0354581768
4770.67077
3.07405726
1.02782768 105
33.8126098
0.907301486
9.0403714 102
1.14934031 102
932.713393
8.3143711
7.2772168 103
2.1489858 103
30.092013
1.4019672 105
6.6743449 105
4.8468502 103
7.6985890 102
5.0253331 105
0
Table B2 Parameters a for CO2 , H2 S and CH4 respectively in Eqs. (B4)(B7).

H2 S
CO2
2
5.2386075 10
2.7463906 101
9.6760173 102
1.3618104 102
8.8681753 102
4.1176908 102
3.6354018 104
2.2719194 103
7.6962514 104
2.1948579 105
1.1707631 104
4.0756926 105
5.7582260 102
1.000
6.0000 102
a1
a2
a3
a4
a5
a6
a7
a8
a9
a10
a11
a12
a13
a14
a15
V
V bmix
bH2 O
V bmix
The i is dependent upon component i, temperature and

total pressure. Pitizer et al. developed several equations for
l(0)
H2 S, CO2 , CH4 to calculate i /RT which were given by:
RT 1.5 bmix
ln
Appendix B. Equation of state for supercritical

uids CO2 , H2 S and CH4
l(0)
2aH2 O-acid ln (V + bH2 O /V)
+ amix bH2 O
ln (V + bH2 O /V) bmix /bmix + V
l(0)
2S
H
RT
RT 1.5 b2mix
PV
RT
(A4)
RT
+ yCH4 yCO2 aH2 OCO2 CH4

+ yH2 S yCO2 aH2 OCO2 H2 S yH2 S yCH4 aH2 OCH4 H2 S
+ yCO2 yCH4 yH2 S aH2 OCH4 CO2 H2 S
(A5)
c3 + c4 P
c6 + c7 P
+ c5 T 2 +
+ c8 P
T
680 T
= c1 + c2 T +
c3 + c4 P
c6 + c7 P
+ c5 T 2 +
+ c8 P
T
630 T
+ c9 P ln T +
l(0)
CH
RT
aH2 Oacid = yCO2 aH2 OCO2 + yCH4 aH2 OCH4 + yH2 S aH2 OH2 S
= c1 + c2 T +
l(0)
CO
where aH2 O-acid is a parameter expanded as Eq. (A5) which

takes the interaction force between water and the acid gas
molecules for mixtures of acid gas into account and assumes
that yH2 O equal to 0. As H2 O,sour refers to the fugacity coefcient of water in the vapor phase, researchers always take the
maximum root of Eq. (A6) as the molar volume of mixture gas.
= c1 + c2 T +
RTb
mix /P amix
10 PT 0.5
+ b2mix
a b
mix mix
=0
10 PT 0.5
(A6)
c10 P2
(630 T)
(B1)
(B2)
c3 + c4 P
c6 + c7 P
+ c8 P + c9 P ln T
+ c5 T 2 +
T
T2
(B3)
where c is constant related to the component and listed in

Table B1.
The equations of state for CO2 , H2 S and CH4 proposed by
Duan et al. and expressed as following:
Z = Pr Vr /Tr = 1 +
V 2 RT
V3
V
P
8.72553928 102
7.52599476 101
3.75419887 101
1.07291342 102
5.49626360 103
1.84772802 102
3.18993183 104
2.11079375 104
2.01682801 105
1.65606189 105
1.19614546 104
1.08087289 104
4.48262295 102
0.754
7.7167 102
8.99288497 10
4.94783127 101
4.77922245 102
1.03808883 102
2.82516861 102
9.49887563 102
5.20600880 104
2.93540971 104
1.77265112 103
2.51101973 105
8.93353441 105
7.88998563 105
1.66727022 102
1.398
2.9600 102
The fugacity coefcient of water in sour natural gas is calculated by Eq. (A4)
ln H2 O,sour = ln
CH4
2
a1 + a2 /Tr2 + a3 /Tr3
a4 + a5 /Tr2 + a6 /Tr3
+
Vr
Vr2
a7 + a8 /Tr2 + a9 /Tr3
Vr4
+a13 (a14 + a15 /Vr2 ) exp
a10 + a11 /Tr2 + a12 /Tr3

Vr5
(B4)
a15 /Vr2
Tr3 Vr2
ARTICLE IN PRESS
14
where Z, Pr , Vr and Tr are the compressibility factor, reduced

pressure, reduced volume, and reduced temperature, respectively. a is evaluated from PVT and the saturation pressure data
of the component i. Their value is given in Table B2, respectively. Pr and Tr are dened as:
Pr =
Pi
Pc
(B5)
Tr =
T
Tc
(B6)
where Pc and Tc are the critical pressure and critical temperature of the component. For the component CO2 , Pc = 7.382 MPa, Tc = 304.41 K; for the component
CH4 , Pc = 4.641 MPa, Tc = 190.6 K; for the component H2 S,
Pc = 9.08 MPa, Tc = 373.6 K. Pi refers to the partial pressure of
component i.
Vr should be obtained through solving Eq. (B4) by substituting Eq. (B5) and Eq. (B6) into it, then the compressibility factor
Z could be calculated.
Substitute Z into Eq. (B7) to calculate the fugacity coefcient of component i
a1 + a2 /Tr2 + a3 /Tr3
Vr
a + a /T2 + a /T3 a + a /T2 + a /T3

7
6
8
9
+ 4 5 r
r
r
r
+
2Vr2
4Vr2
Z 1 ln Z +
i = exp
+ a10 + a11 /Tr2 + a12 /Tr3
5Vr5
+a13
a14 + 1 (a14 + 1 + a15 /Vr2 ) exp(a15 /Vr2 )

2Tr3 a15
(B7)
where i refers to component i such as CO2 , H2 S and CH4 . The

constant a is same as the parameter of Eq. (B4).
References
Bahadori, A., Vuthaluru, H.B., Mokhatab, S., 2009. Rapid
estimation of water content of sour natural gases. J. Jpn.
Petrol. Inst. 52, 270274.
Behr, W.R., 1983. Correlation eases absorber-equilibrium-line
calculations for TEG-natural gas dehydration. Oil Gas J.,
9698.
Bukacek, R.F., 1955. Institute of Gas Technology Research Bulletin,
8, Chicago.
Campbell, J.M., 1991. Oil & Gas Training &Consulting, second ed.
Petroleum Industry Press, Beijing.
Carroll, J.J., 2002. The water content of acid gas and sour gas from
100 F to 220 F and pressures to 10000 psia. In: The 81st
Annual GPA Convention, Canada.
Carroll, J.J., 2003. Natural Gas Hydrates. Gulf Professional
Publishing, New York.
Chapoy, A., Mohammadi, A.H., Tohidi, B., Richon, D., 2005.
Estimation of water content for methane + water and
methane + ethane + n-butane + water systems using a new
sampling device. J. Chem. Eng. Data 50, 11571161.
Duan, Z., Sun, R., 2003. An improved model calculating CO2
solubility in pure water and aqueous NaCl solutions from 273
to 533 K and from 0 to 2000 bar. Chem. Geol. 193,
257271.
Duan, Z., Sun, R., Liu, R., Zhu, C., 2007. Accurate thermodynamic
model for the calculation of H2 S solubility in pure water and
brines. Energy Fuel. 21, 20562065.
Duan, Z., Sun, R., 1992. The prediction of methane solubility in

natural waters to high ionic strength from 0 to 250 and from 0
to 1600 bar. Geochim. Cosmochim. Acta 56, 14511460.
Eslamimanesh, A., Mohammadi, A.H., Richon, D., 2011.
Thermodynamic consistency test for experimental data of
water content of methane. Am. Inst. Chem. Eng. 57,
25662573.
G.P.S.A, 1998. Engineering Data Book, eleventh ed. Gas Processors
Association, Tulsa.
Kazim, F.M.A., 1996. Quickly calculate the water content of
natural gas. Hydrocarb. Process. 75, 105108.
Khaled, A.A.F., 2007. A Prediction of Water Content In Sour
Natural Gas. Kingsaud University.
Kobayashi, R., Song, K.Y., Sloan, E.D., 1987. Petroleum Engineering
Handbook: Chapter 25: Phase Behavior of Water/Hydrocarbon
Systems. Society of Petroleum Engineers, Richardson, TX.
Kondori, J., Javanmardi, J., Eslamimanesh, A., Mohammadi, A.H.,
2013. Thermodynamic consistency test for isobaric
experimental data of water content of methane. Fluid Phase
Equilib. 347, 5461.
Lukacs, J., Robinson, D.B., 1963. Water content of sour
hydrocarbon systems. SPE, 293297.
Maddox, R.N., Lilly, L.L., Moshfeghian, M., Elizondo, E., 1988.
Estimating Water Content of sour Natural Gas Mixtures.
Laurance Reid Gas Conditioning Conference, Norman.
Mohammadi, A.H., Chapoy, A., Richon, D., Tohidi, B., 2004a.
Experimental measurement and thermodynamic modeling of
water content in methane and ethane systems. Ind. Eng.
Chem. Res. 43, 71487162.
Mohammadi, A.H., Chapoy, A., Tohidi, B., Richon, D., 2004b. A
semiempirical approach for estimating the water content of
natural gases. Ind. Eng. Chem. Res. 43, 71377147.
Mohammadi, A.H., Chapoy, A., Tohidi, B., Richon, D., 2006. Gas
solubility: a key to estimating the water content of natural
gases. Ind. Eng. Chem. Res. 45, 48254829.
Mohammadi, A.H., Richon, D., 2007. Semiempirical method for
determining water content of methane-rich hydrocarbon gas
in equilibrium with gas hydrates. Ind. Eng. Chem. Res. 47,
451458.
Mohammadi, A.H., Sameiyan, V., Tohidi, B., 2005. Estimation of
water content in sour gas. In: Proceedings of the SPE
Europec/EAGE Annual Conference, Spain.
Ng, H.J., Chen, C.J., Schroeder, H., 2001. Water Content of Natural
Gas Systems Containing Acid Gas, 174, GPA Research Report,
Tulsa.
Ning, Y.N., Zhang, H., Zhou, G., 2000. Mathematical simulation
and program for water content chart of natural gas. Chem.
Eng. Oil Gas 29, 7577 (in Chinese).
Robinson, J.N., Moore, R.G., Heidemann, R.A., Wichert, E., 1980.
Estimation of the Water Content of Sour Natural Gas.
Laurance Reid Gas Conditioning Conference, Norman.
Robinson, J.N., Wichert, E., Moore, R.G., 1978. Charts help
estimate H2 O content of sour gases. Oil Gas J., 7678.
Salari, H., Hassanzadeh, H., Gerami, S., 2008. On estimation the
water content of CO2 in equilibrium with formation brine. In:
The SPE Gas Technology Symposium 2008 Joint Conference,
Canada.
Sharma, S., Campbell, J.M., 1969. Predict natural-gas water
content with total gas usage. Oil Gas J., 136137.
Sloan, E.D., 1998. Clathrate Hydrates of Natural Gases, second ed.
Marcel Dekker, New York.
Sloan, E.D., Khoury, F.M., Kobayashi, R., 1976. Water content of
methane gas in equilibrium with hydrates. Ind. Eng. Chem.
Fund. 15, 318323.
Wang, J.Q., 1994. The simple method of calculating water content
in natural gas. Chem. Eng. Oil Gas 23, 192193 (in Chinese).
Wichert, G.C., Wichert, E., 1993. Chart estimates water content of
sour natural gas. Oil Gas J., 6164.
William, A.F., Musa, S.A., Ghati, M., 2012. Effect of Sour Gas on
Water Vapor Solubility in Natural Gas. Department of
Petroleum Engineering and Applied Geophysics.
Zhu, L., 2008. Natural Gas Processing Engineering, second ed.
Petroleum Industry Press, Beijing (in Chinese).
ARTICLE IN PRESS
Zhu, L., Bai, J., Wang, Z.H., 2003. Formulated calculation method
of water content in natural gas. Nat. Gas Ind. 23, 118120 (in
Chinese).
Ziabakhsh-Ganji, Z., Kooi, H., 2012. An equation of state for
thermodynamic equilibrium of gas mixtures and brines to
allow simulation of the effects of impurities in subsurface CO2
15
storage. Int. J. Greenhouse Gas Control 11 (Supplement),

S21S34.
Zirrahi, M., Azin, R., Hassanzadeh, H., Moshfeghian, M., 2010.
Prediction of water content of sour and acid gases. Fluid
Phase Equilib. 299, 171179.

Articulo - Contenido de Agua - Metodos Analiticos

Diunggah oleh

Informasi Dokumen

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Articulo - Contenido de Agua - Metodos Analiticos

Diunggah oleh

Hak Cipta:

Format Tersedia

CHERD-1599; No.

Contents lists available at ScienceDirect

Chemical Engineering Research and Design

Analytical methods to calculate water content in

the consumption of dehydrate agents and for predicting the aqueous

In the initial design of the facilities of production, transmission, and

problems. Additionally, a decrease in the temperature (or an increase

limited experimental data, thermodynamic models based on phase

ed H2 S or CO2 signicantly enhances the solubility of water, presence

dration facilities before transmitting and processing. For engineers,

CHERD-1599; No. of Pages 15

chemical engineering research and design x x x ( 2 0 1 4 ) xxxxxx

water content in mg/Sm3 (288.15 K, 0.101 MPa)

In order to describe the water content of gas in equilibrium with

the McKettaWehe chart is the most popular, with surprising accuracy

Analytical methods for sweet natural gas

The analytical methods for estimating water content in sweet

CHERD-1599; No. of Pages 15

chemical engineering research and design x x x ( 2 0 1 4 ) xxxxxx

Table 1 Constants C in Eq. (1).

Table 2 Constants a in Eq. (2).

Correlations regressed from chart data

These correlations are generally from regressing the data of

For the low temperature region, Sloan (Kobayashi et al., 1987;

Concerning that water content in NG tends to rise with

where a and b are parameters listed in Table 3. The applicable

Bahadori (Bahadori et al., 2009) stated the water content was

Ning et al. method

where a is the constant presented in Table 4. This method can

Table 3 Constants a and b in Eq. (4).

CHERD-1599; No. of Pages 15

chemical engineering research and design x x x ( 2 0 1 4 ) xxxxxx

investigated the following equation:

Table 4 Constants a in Eq. (5).

Correlations regressed from experimental data

+A5 ln P + A6 (ln P) + A7 (ln P)

A8 + A9 ln P + A10 (ln P) + A11 (ln P)

Zhu et al. Method

Similar to the method of Khaled, Zhu et al. (Zhu, 2008; Zhu

where a0 a7 , b0 b7 are constants listed in Table 5.

Coefcients A0 A11 were obtained on the basis of tting

Based on the experimental data of water content in methane

WH2 O,sweet = 16.02A1 A1.8T459.67

where A is a parameter linked with pressure and calculated

Fig. 1 Comparison of the applicable range of each analytic method.

CHERD-1599; No. of Pages 15

chemical engineering research and design x x x ( 2 0 1 4 ) xxxxxx

Table 5 Constants a and b in Eq. (6).

Table 6 Constants A in Eq. (7).

Table 7 Constants a in Eq. (9).

these correlations agree well with the experimental data, even

Simplied thermodynamic model (STM)

Depending on the phase equilibrium system, thermodynamic

Psw,ice = 1.31579 101032.5576407/T+51.0557191log T0.0977T+7.035810

Model of saturated vapor pressure

The ideal model is always used to estimate water content

761900.42 Psw exp

Modied ideal model (MIM)

VH2 O,Liq = 4.501 102 6.710 104 T + 1.784 106 T 2

VH2 O,ice = 19.655 + 0.0022364 (T 273.15)

CHERD-1599; No. of Pages 15

chemical engineering research and design x x x ( 2 0 1 4 ) xxxxxx

Table 8 Constants a in Eq. (23).

Unlike the above methods, by amending the ideal model