Академический Документы
Профессиональный Документы
Культура Документы
Abstract
The compressibility factor of natural gases is a parameter that is used for various engineering
purposes in the petroleum industry. The Standing and Katz correlation, among other methods, is a
widely accepted method of determining z-factor manually from charts for natural gas of either
known or unknown composition. A major setback is that, for computer-based applications, it is not
convenient to obtain z by this means. This paper highlights the limitations of the other direct z-factor
determination methods and then presents a new approach of computing z-factor based on the
Standing and Katz correlation, which eliminates the limitations observed with the other methods. A
set of z-factor equations were developed by regressing data (in different ranges of pseudo-reduced
temperatures and pressures) obtained from the Standing-Katz and Brown et al correlations using a
Visual Basic program. Results obtained from this approach were checked against those from the
other correlations and were found to be more accurate than the other methods, with average absolute
error in z less than 0.1%. A subroutine for calculating z-factor could easily be incorporated into any
window-application program written in MS Excel, MATLAB or visual Basic, using equations
developed in this approach when determining the properties of natural gas, estimating gas reserves,
sizing oil and gas separators, designing gas transmission pipelines, and pressure traverses in pipes
for multiphase flow conditions. A standard z-factor table may also be computed using the set of
equations for all ranges of pseudo-reduced temperatures and pressures.
1. Introduction
The z-factor comes into play for various engineering purposes, which include estimation of
gas reserves, design of oil and gas separators, design of pipelines for the transmission of produced
gas, among others.
A number of these tasks or procedures have been developed in such a way that makes it
necessary to employ the services of a computer in order to accomplish them in reasonable time.
Carrying out such tasks by hand would make it lengthy, tedious and as such time consuming. An
example of such tasks or procedures is the determination of pressure traverses in pipes for
multiphase flow conditions.
The Beggs and Brill method of calculating pressure traverses, for instance, is one that
requires the gas compressibility factor. This method, involving about 21 steps, is an iterative one
wherein a pressure drop is obtained at the end of each iteration using, among other data, an initial
assumed pressure drop. If the difference between the initial and calculated pressure drops is
substantial, the iteration is repeated with the calculated pressure drop in each iteration serving as the
assumed pressure drop for the next iteration. This process is continued until the difference between
the assumed and calculated pressure drops is small. Arriving at a value for the final pressure drop
typically requires a number of iterations.
What this means is that the working or operating pressure changes with each successive
iteration making it a necessity to obtain, for each iteration, all data that are pressure dependent one of
which is the gas z-factor. It is quite evident from the foregoing that manually obtaining z from the
chart and entering the value into the computer, for each iteration, is rather inconvenient as it would
undoubtedly slow down the computation process.
Programming such tasks as the Beggs and Brill method (Brown and Beggs, 1977) for
calculating pressure traverses in pipes for multiphase flow conditions cut down on the amount of
time required for the calculation. Such reduction in computation time could be increased if a means
was devised to incorporate the determination of gas compressibility factor into the program thus
eliminating the need to manually obtain it from the chart for successive iterations. How can this be
accomplished?
This paper reviews existing literature and presents a new approach for determining z-factor
for computer-based applications. Three other correlationsthat can be programmed for
useconsidered in this paper are those of Hall and Yarborough, Beggs and Brill, and Drankchuk
65
and Abou-Kassem. The limitations that make them unfit for use for engineering purposes requiring
precision are highlighted.
2. Background
There are various correlations available for the calculation of gas compressibility factors.
Using these correlations or equations of state (EOS), one can program the computer to solve directly
for z. The correlations or equations of state considered for such purpose are Standing & Katz (1959),
Hall and Yarborough, Beggs and Brill, and Dranchuk and Abou-Kassem.
Standing and Katz Correlation
Since z is a function of the gas pseudo-reduced temperature (Tpr) and pressure (Ppr), it is
necessary to first determine the pseudo-critical temperature (Tpc) and pressure (Ppc) of the gas and
subsequently use these to obtain the pseudo-reduced temperature (Tpr) and pressure (Ppr).
For natural gas of known composition, the pseudo-critical pressure and temperature can be
determined from Kay's mixing rule (Bradley, 1987) which gives these properties as:
Ppc = yi Pci
(1)
Tpc = yi Tci
(2)
Where Ppc = pseudo-critical pressure of gas mixture, Tpc = pseudo-critical temperature of gas
mixture, Pci = critical pressure of component i in the gas mixture, Tci = critical temperature of
component i in the gas mixture, and yi = mole fraction of component i in the gas mixture
For a gas whose complete analysis is not known, a correlation developed by Brown et al can be used.
This correlation, presented in graphical form, relates the pseudo critical temperatures and pressures
of naturally occurring systems with their specific gravities (Katz et al, 1959). Having determined the
Tpr and Ppr, z may be obtained from either the Standing-Katz (Fig. 2) or the Brown et al chart (Fig. 3)
Hall-Yarborough Equation
The equation given by Hall and Yarborough (Ikoku, 1984) is given below:
0.06125Pprte 1.21t
y
(3)
66
where
t = Tc / T, and y = the reduced density which is obtained as the solution of the equation:
y 0.06125Pprte
1.2 1 t 2
y y 2 y3 y 4
(4)
Where:
0.5
0.36Tpr 0.101
0.066
10 T pr 1
D 10
and
Where
(5)
pr = 0.27Ppr/(zTpr);
; and c
4
pr pr
A 1 A
2
10
11
pr
2
pr
Tpr
The Beggs and Brill method, while being quite accurate for certain ranges, is not applicable
when Tpr < 0.92. In determining the value of the temperature dependent term A, it is necessary to
evaluate the square root of (Tpr 0.92) which would mean an imaginary root when Tpr < 0.92.
Also, for some values of Tpr and Ppr, the temperature and pressure dependent term B, gets so
large that evaluating eB results in an overflow of values. Negative values for z were sometimes
obtained from the method for some values of Tpr and Ppr
The Dranchuk and Abou-Kasem method, for the most part, gave good results for z, but the EOS
involves the use of an iterative method such as the Newton's method, necessitating an
assumption before convergence would occur. Once convergence is obtained the final value is
given as the calculated z factor. It was, however, observed that in some instances different initial
or assumed values of z resulted in convergence to different values at the end of the iteration thus
resulting in different final values for z for the same set of Tpr and Ppr values. There were even
cases where using certain initial values for z resulted in a negative value for compressibility
factor. So despite its accuracy, this method for obtaining z factor may not be incorporated into a
design program since it is not possible to predict or determine when such erroneous values may
result.
Therefore, another method was sought that would give values for the gas compressibility factor
without the limitations highlighted above.
3. Theoretical Development
This approach is based on the Standing and Katz method which is generally accepted as the
industry standard and were developed from data collected on methane and natural gases (Bradley,
1987). In addition to the Standing-Katz charts, the charts by Brown et al for low-pressure systems
were also used.
68
The compressibility factor charts are essentially curves with the gas compressibility factor, z,
being a function of the pseudo-reduced pressure Ppr. These curves appear on the charts for various
values of Tpr the pseudo-reduced temperature.
In this method, pseudo-reduced pressure values were selected and regressed with
corresponding z values obtained from the charts to give equations that expressed z as a function of
Ppr. This regression process had to be carried out for each pseudo-reduced temperature value on the
chart.
In a bid to ensure that the regression process gave rise to equations that were as accurate and
reliable as possible, two regression exercises were carried out for each set of values. These two
different exercises were carried out in such a way that they yielded two different equations one
linear and the other quadratic. Both equations expressed z in terms of Ppr. The equations were of the
form:
z = A(Ppr) + B
(6)
z = A(Ppr)2 + B(Ppr) + C
(7)
one can obtain accurate z values for this range of Ppr via interpolation. A summary of the equations
derived from the regression processes and by means of which z values may be obtained is found in
Table 3
As was the case for the z-factor, various values of specific gravity were regressed with the
corresponding values of pseudo-critical temperature and pressure respectively setting the specific
gravity as the independent variable. This process yielded two (2) equations of the form:
Xpc = A (g) 2 + B (g) + C
(8)
Where Xpc = pseudo-critical constant (temperature or pressure), g = specific gravity of natural gas
and A, B, C are constants
Dune and Oriji (2007) regressed data obtained from the Standing and Katz method and
obtained the following equations:
Tpc = 158.01 + 342.12(g) 16.04(g)2
(9)
70
value of 1.2, the program obtains a value for z by determining values for z at Tpr = 1.4 and Tpr = 1.5
when Ppr =1.2 and then interpolating between these values to give the appropriate value for z.
5. Results and Discussion
Results obtained from the computer program were analyzed to ascertain their level of
accuracy. In order to determine the accuracy of the compressibility factor values obtained from this
method, z values obtained from it were compared in Table 4 with those obtained from the
correlations by Hall and Yarborough, Beggs and Brill, Dranchuk and Abou-Kassem and manual
reading of the appropriate compressibility factor charts (Standing and Katz) for several pseudoreduced temperature and pressure values.
Table 4: Compressibility Factor, Z, for Various Methods.
Tpr
Ppr
This
Hall-
Beggs &
Dranchuk &
Standing &
Approach Yarborough
Brill
Abou-Kassem
Katzs
0.75
0.048
0.9512
1.0069
N/A
0.9546
0.951*
1.30
0.020
0.9950
0.8840
0.9977
0.9969
0.9967*
1.62
0.065
0.9970
1.0005
0.9959
0.9950
0.9956*
1.20
1.500
0.6740
0.1605
0.6759
0.6532
0.6730*
1.80
0.900
0.9562
1.0036
0.9599
0.9550
0.9700
2.00
1.200
0.9718
1.0041
0.9690
0.9621
0.9700
2.60 13.000
1.2693
1.3987
0.8222
1.2732
1.0500
2.88
6.000
1.0615
1.1436
-0.4410
1.0607
1.0620
1.76
3.500
0.8776
0.9739
0.8727
0.8818
0.8768
1.25 10.200
1.1824
1.0038
N/A
1.1818
1.1825
*Values were obtained from z-factor charts developed by Brown et al (Bradley, 1987; Fig 3)
The extent to which results from this approach and the other methods deviated from those
obtained from the Standing-Katz method for a given Tpr of 1.20 can be seen from Table 5. A plot of
absolute error versus pseudo reduced pressure for the different methods were made (Fig. 1). A look
at Table 5 as well as the plot (Fig. 1) reveals that the results from this approach are actually more
accurate and in harmony with those obtained from the Standing-Katz and Brown et al methods than
those from the other correlations considered.
A measure of the degree of accuracy of the various methods is seen when one considers the
error in computing z-factor from the various methods. These errors are computed with the results
obtained from the Standing-Katz and Brown et al charts serving as the reference. Whereas errors
greater than 1% were obtainable (in some cases) with the 3 correlations considered, the error
71
associated with this approach do not exceed 0.2%. Indeed, the average error associated with this
approach, for the Tpr and Ppr values indicated, is less than 0.1%. This is much better than the
averages of the other correlations which are in excess of 0.5%. Thus utilizing z values from this
approach yields results that are accurate, reliable and acceptable.
Table 5: Absolute Error in z, % Deviation from Standing & Katz for Tpr = 1.20
z-factor
Absolute Error in z, %
Ppr
HallYarboro
ugh
Beg
Dranchuk
gs &
&
Brill
AbouKasem
0.0
0.9934
48
0.5
0.99
0.9904`
23
0.8934
0.90
This
Appro
ach
Standi
ng
&
HallYarboro
ugh
Beggs
Dranchu
&
k-
Brill
Abou-
Katz
This
Appro
ach
Kasem
0.2424
0.0505
0.0202
*
0.8951
0.2889
0.5444
0.000
0.8027
0.2822
1.5092
0.1227
0.6532
0.4309
2.9421
0.1486
0.5181
0.5199 0.520
0.3462
4.2692
0.3654
0.0192
0.5632
0.5661 0.567
0.9171
1.1464
0.6702
0.1587
0.7937
0.7897 0.790
0.0759
1.9367
0.4684
0.0380
0.9870
0.9902 0.989
0.4247
0.9707
0.2022
0.1213
26
0.9
0.7999
0.81
27
1.5
0.6576
0.67
59
2.5
0.5218
0.49
78
3.5
0.5618
0.56
05
6.0
0.7906
0.80
53
8.0
0.9848
0.99
86
72
10.
1.1963
1.20
2
13.
1.1959
1.1952 1.195
0.1088
1.1967
0.0753
0.0167
93
1.4602
N/A
1.4550
1.4565 1.458
0.1509
N/A
0.2058
0.1029
1.6452
N/A
1.6358
1.6432 1.643
0.1339
N/A
0.4382
0.0122
0
15.
0
3.75
Hall-Yarborough
Beggs & Brill
2.25
1.5
0.75
0
0
10
11
12
13
14
15
16
Ppr
0.6714
1.196
0.6792
0.0691
*Values were obtained from z-factor charts developed by Brown et al4 (Fig 3)
Figure 1: Absolute Error in z, % Deviation from Standing & Katz for Tpr = 1.20
73
6. Conclusion
A comparison of results from this approach with those obtained from other methods shows
that the results from this approach are more reliable and accurate. The method is therefore fit for use
for all computer-based applications that require the gas compressibility factor.
A look at the summary of equations obtained and utilized by this method reveals that the
equationsfor the various ranges of Tpr and Ppr are numerous. Programming these equations for
use is therefore recommended as the most practical way of using this method.
References:
Bradley, H. B.: Petroleum Engineering Handbook, SPE, Richardson, TX, chap. 12, chap 20, 1987.
Brown, K. E. and Beggs, H. D.: The Technology of Artificial Lift Methods Vol. 1, PennWell Books,
Tulsa, OK, 1977, p 85.
Dune, K. K. & Oriji, B. N.: Alternative correlation for the computation of critical temperature &
pressure. Global Journal of Engineering Research, Calabar, vol. 5, no. 1, 2007, pp 69-74,
Golan, M. and Whitson, C. H.: Well Performance, D. Reidel Publishing Co., Dordrecht, Holland,
1986, pp. 17 21.
Ikoku, C. U.: Natural Gas Production Engineering, John Wesley and Sons, N.Y., 1984.
Katz, D. L. et al: Handbook of Natural Gas Engineering, McGraw-Hill Books Co., N.Y., 1959.
Lee, J. and Wattenbarger, R. A.: Gas Reservoir Engineering, SPE, Richardson, TX, 1996, pp. 6 7.
Oriji, B. N.: Separator Design: A Computerized Approach, B.Tech Thesis, Rivers State University
of Science and Technology, 2003 pp 14 20.
Siler, B. and Spotts, J.: Special Edition Using Visual Basic 6, Que, 1998.
Nomenclature
e
74
-pressure, psia
Pc
Pci
Ppc
Ppr
-pseudo-reduced pressure
pr
-pseudo-reduced density
-temperature, oR
Tc
-critical temperature, oR
Tci
Tpc
-pseud0-critical temperature, oR
Tpr
-pseudo-reduced temperature
yI
0 Ppr 0.07
Selected Ppr
Corresponding z
0.010
0.9947
0.020
0.9893
0.035
0.9814
0.050
0.9732
0.060
0.9678
Resultant Equation: z = - 0.537647 (Ppr) + 1.000098
75
Table 2: Regression Data Obtained From Standing-Katz z Factor Chart for Tpr = 2.2
10 Ppr 15
4 Ppr 10
Ppr
Ppr
10.5
1.17
0.988
12.0
1.238
1.045
13.0
1.280
1.077
14.0
1.322
1.113
15.0
1.361
10
1.156
Resultant equation:
z = 0.04109 Ppr + 0.745537
Resultant equation:
z 0.001988Ppr2 0.003488Ppr 0.921786
76
Figure 2: Compressibility Factor Chart of Natural Gases After Standing and Katz 2
Range of Ppr
Equation for z
0.60
0 Ppr 0.016
z 3.527273Ppr 1.000075
0.65
0 Ppr 0.036
z 1.9Ppr 1
0.70
0 Ppr 0.07
1.3610345Ppr 0.9998172
0.75
0 Ppr 0.07
z 1.017619Ppr 1.000017
0.8
0 Ppr 0.07
z 0.815Ppr 0.99994
0.8
0.85
0 Ppr 0.07
z 0.65923Ppr 0.99995
0.90
0 Ppr 0.07
0.90
0.95
0 Ppr 0.07
Values
77
0.95
1.0
0 Ppr 0.07
1.0
1.0
1.0
1.0
1.0
1.05
0 Ppr 0.8
1.05
1.05
1.05
Values
1.05
1.05
7 Ppr 15
1.10
0 Ppr 0.07
1.10
1.10
1.10
7 Ppr 15
z 0.102012Ppr 0.189738
1.2
0 Ppr 0.07
1.2
1.2
1.2
1.2
1.2
8 Ppr 15
1.3
1.3
0 Ppr 1.6
1.6 < Ppr < 3.8
1.3
1.3
6 Ppr 15
1.40
0 Ppr 0.07
1.40
1.40
1.40
1.40
6 Ppr 15
1.50
1.50
1.10
1.10
Values
Interpolate Using Predetermined
Interpolate Using Predetermined
Values
Interpolate Using Predetermined
Values
2
Values
z 0.275172Ppr 1.000131
Values
Values
z 0.09333Ppr 0.243254
Values
z 0.086983Ppr 0.28228
z 0.12459Ppr 1.000008
2
z 0.077786Ppr 0.361643
2
78
1.50
1.50
1.50
z 0.071736Ppr 0.404538
1.50
z 0.0711Ppr 0.4254
1.60
0 Ppr 0.07
1.60
1.60
1.60
1.60
1.60
8.0 Ppr 15
1.70
1.70
0 Ppr 1.6
1.6 < Ppr < 4.5
1.70
1.70
7.0 Ppr 10
1.70
10 < Ppr 15
1.80
1.80
0 Ppr 1.6
1.60 < Ppr < 4.0
1.80
1.80
7 Ppr 11.2
1.80
1.90
1.90
1.90
6.0 Ppr 10
1.90
10 < Ppr 15
z 0.0536Ppr 0.6024
2.0
0 Ppr 0.07
z 0.031Ppr 1.000065
2.0
2.0
2.0
2.0
2.0
10.5 Ppr 15
2.2
2.2
2.2
2.2
z 0.04109Ppr 0.745537
2.4
2.4
2
2
z 0.067933Ppr 1.000132
2
79
2.4
2.4
9.5 Ppr 15
2.6
2.6
3.0
3.0
3.0
10.0 Ppr 15
80