Professional Documents
Culture Documents
Introduction
Several correlations have been published that can be
used to predict pressure losses in vertical oilwell
tubing for the simultaneous, upward, concurrent, continuous flow of oil, water, and gas. Because of the
extreme complexity of multiphase flow, the proposed
correlations are by necessity highly empirical. The
validity of the correlations is then somewhat limited
to the quality and scope of the data upon which they
are based. Therefore, some correlations perform
quite well for cases in the range of the data used
in developing the correlation but fail for other
applications.
The purpose of this study is to ascertain the
accuracy of several pressure-loss prediction methods
in terms of flow variables familiar to the practicing
engineer. The results of the study should assist the
petroleum production engineer in selecting the most
accurate method for his problem and also should
indicate the general accuracy to be expected from
the methods.
The correlations included in the study are those of
Poettmann and Carpenter,! Baxendell and Thomas,2
Duns and Ros,3 Fancher and Brown, 4 Hagedorn and
Brown, 5 and Orkiszewski. 6 Each of these correlations
was proposed specifically for predicting pressure losses
in vertical oilwell tubing for the upward flow of multiphase well fluids.
The pressure-loss prediction methods were programmed for the IBM 360 computer and tested
against 726 well tests from field and experimental
Several methods for predicting pressure losses for upward, concurrent, continuous,
multiphase flow in vertical. ai/well tubing were tested against measured pressure losses
from 726 well tests. This comparison of well known pressure-loss correlations reveals
the relative strengths and weaknesses of each correlation and should be useful in
selecting satisfactory methods for various applications.
AUGUST. 1974
903
recommended for predicting pressure losses for multiphase flow, but for various reasons they were not
included in this study. The work of Griffith and
Wallis 12 was not considered separately because of its
application to only the slug flow regime. However,
it is partially included in the Orkiszewski method
along with the work of Griffith l l for the bubble flow
regime.
A correlation by Yocum 13 was not included
because it centered on the prediction of pressure
losses in horizontal flowlines. Baker and Keep14 did
not recommend their method for cases where there
was a likelihood of slip between gas and liquid.
The Tek15 correlation was not included in the
statistical study because it predicted unusually high
pressure losses for many of the well tests. A correlation by Hughmark and Pressburg16 was based on
data obtained using a relatively short tube. It was not
included in the study because it has been reported
that the correlation gives values for pressure loss that
are too high when applied to data taken from long
tubing strings. 7
The work of Gaither et alY was not included in
the statistical study, although it may perform well
for the small-diameter pipes from which the correlated data were taken.
An attempt was made to include those correlations
that have been considered most acceptable by oil
industry users for broad ranges of flow conditions.
Correlations that have not attained this degree of
general usage may be quite adequate for predicting
pressure losses in specific applications.
Programming of Methods
A separate main program was written for each
pressure-loss prediction method, with fluid property
correlations handled as subroutines. Subroutines were
written for calculating values of formation volume
factor and solution GaR, oil and water viscosity, oil
and water surface tension, gas viscosity, and gas
compressibility.
Calculations were made for changes in tubing
depth corresponding to assigned pressure changes
starting with the measured tubinghead pressure. When
values of fluid properties representative of the average
conditions in the tubing section were needed, the subroutines were entered with arithmetic average pressure and temperature. When temperature varied with
depth, the calculation was by trial and error to match
the given temperature gradient. The gravity of the
free gas and that of the gas dissolved in the oil phase
at higher pressures were both set equal to the total
produced gas gravity.
The values of pressure, temperature, and other
fluid conditions calculated for the exit of the first
tubing section w~re used as inlet values for the next
length. The calculation proceeded in this incremental
manner until the total depth was reached or exceeded.
The bottom-hole pressure was then calculated by
linear interpolation of the last two pressure/depth
coordinates. It should be noted that calculations starting at the measured bottom-hole condition and ending
at the tubinghead would not necessarily give the same
pressure drop.
The pressure increment sizes were selected to be
between 5 and 10 percent of the absolute pressure to
minimize the inaccuracies of averaged physical properties and at the same time to maintain a reasonably
short computation time. The pressure-loss methods
studied required only a fraction of a second on the
IBM 360/85 co~puter to calculate a pressure
traverse.
Certain extensions were made to some correlations
to cover extremes encountered in analyzing the well
test data. When a dependent variable in a correlation
TABLE I-MINIMUM, MAXIMUM, AND AVERAGE
VALUES OF WELL DATA
Minimum Maximum
8.760
1.049
5,082.0
1.0
788,000.
21.
12,458.
918.
5,140.
104.
20.
50.0
80.0
9.5
1.00
0.600
0.90
2,856.
180.0
330.0
56.2
1.15
1.400
20,000.00
0.26
140.00
Arithmetic
Average
1.878
395.2
3,993.
1,252.
518.
93.9
126.0
34.4
1.03
0.850
905
Data Source
Number
of
Tests
General Description
Poettmann and
Carpenter1
49
Fancher8
83
One field well with gas injection; 2%-in.00 tubing; medium-high flow rates; 95
percent water cut; low pressures.
346
Baxendell and
Thomas 2
25
Orkiszewski 6
22
Espanol Herrera23
44
Field data on 2%- and 2'l's-in.-OD tubing; mostly low flow rates; light oils;
high pressures.
U. of Tulsa
Data Group 1
56
Field data on 2%- and 2'l's-in.-OD tubing; mostly low rates; low to medium
pressures.
101
Field data on 2%- and 2'l's-in.-OD tubing; gas wells making low to medium
amounts of water; medium pressures.
Hagedorn 10
camach0 24
906
tions and the way in which they are used limit the
ability of a multiphase flow pressuure-loss correlation
to make an accurate prediction. Therefore, the
pressure-loss prediction method, made up of a combination of pressure loss correlation and fluid physical
property correlations, must be considered as a package when tested against measured pressure losses.
Validation of Programming
Because of the intricacy of some of the correlations
and the complexities of the programming involved,
pressure losses calculated by the computer program
for each pressure-loss prediction method were compared with calculated pressure losses available from
another source for the same well test data. Close
agreement between the calculated pressure losses
from this study and the independent calculation
indicates that the programming was done correctly.
Minor differences in calculated answers from different sources are to be expected because of the use
of different physical property correlations and programming techniques. However, it is difficult to
Correlation Used
Comments
Pseudocritical temperature
and pressure of hydrocarbon gases
Katz25 ,26
The correlation becomes more in error as the nonhydrocarbon content of the gases is increased, but may still be accurate to about 2% when only small amounts of nonhydrocarbons are present. z-factors for reduced temperatures less
thari 1.05 were calculated for reduced temperatures of 1.05.
Solution GOR
Lasater2s,29
Standing30,31
Carr et al. 32
Oil viscosity
Water viscosity
Beal iS
Data are for a specific gravity of 1.0. The curve was extrapolated to higher temperatures. No attempt was made to account for the effects of salt content, dissolved gas content,
or pressure.
'
Baker34,35
Data of Ref. 36
Data for pure water against air were used. Surface tension
was assumed to vary linearly with temperature and have
values of 69.56 dyne/cm at 104F and 58.9 dyne/cm at
212F.
AUGUST, 1974
by linear interpolation
the methane-water sys74F and above 280F
280F, respectively.
907
generalize as to exactly how much deviation in calculated pressure losses could be expected because of
those differences.
The programs developed for the Poettmann and
Carpenter, Baxendell and Thomas, Fancher and
Brown, and Hagedorn and Brown methods were
found to reproduce the pressure losses calculated by
each method quite closely. Comments on comparisons
for each method are given below.
"Results for the 427 well tests in the range of the correlation.
Method
Poettmann and Carpenter
Baxendell and Thomas
Fancher and Brown
Duns and Ros
Hagedorn and Brown
Orkiszewski
"no-slip" static head
Average
Percent
Standard
Difference Deviation
- 107.3
195.7
- 108.3
195.1
5.5
15.4*
1.3
8.6
53.5
36.1
50.2"
26.1
35.7
33.0
Data Group
Orkiszewski data on heavy-oil wells
(22 well tests)
Orkiszewski
Orkiszewski Results of
Results
This Study
Prediction Method
Ros
Orkiszewski
Results of
Results
This Study
-1.2
10.4
-12.0
13.1
+22.7
18.7
-42.4
19.7
+ 16.4
41.4
-29.7
42.0
-2.1
11.1
-3.7
41.1
+2.3
20.0
-13.9
39.5
+8.7
12.7
+6.0
11.6
-1.0
12.0
+3.7
13.7
+5.8
12.4
-4.8
15.1
-13.0
22.2
+22.8
24.7
908
Discussion of Results
The statistical results for the prediction methods
applied to all 726 well tests are given in Table 5.
Definitions of percent difference, PD, arithmetic
average of percent differences, APD, and standard
deviation of percent difference values from the average percent difference, SD, are given below.
PD = e:..pm - e:..pc 100 percent,
e:..pm
n
APD =
SD =
L PD i
~i=-.:1=---_
-/--.--:....i~--=l_(_PD_i_-,.-A_P_D_)_2
"
n- 1
where
APe = calculated pressure
40
'"
u
15 20
'"
'"
I
o FANCHER-
is
I-
0
15
u
'"
0'"
'"
'-"
ffi
-20
>
<
-40
,p.
BROWN
,.....
o~~
\0
\~
HAGEDOR
BROWN
DUNS"ROS
o
ORKISZEWSKI
60
z
0
;::
<
~
c
40
'"
<
c
z 20
;:!:
'"
1. 049
L 380
1. 610
1. 995
2.376 2.441 2.992
. (311
(175)
1140l
,(228)
(411
(60)
(46)
VALUES OF INSIDE TUBING DIAMETER, INCHES
909
LU
HAGEDORNBROWN
r:s
20
...""
LU
g OI-""""---------;H-.""".~:::__I_____7L---=-~---___4
""
"LU
LU
<:>
~ -20
LU
;::
-40
...
~- 40
~
~
c
c
<
""c
z
o
20
" 0/t::.-~
/ ~EDORN-BROWN
::e'o ~
6-
-~LS/
FA:BROWN
DUNS-ROS
ORKI SZEWSKI
t::.-;::?" "'~_O
o
OL......;*;;---~:;;--__:::.l.;;;_---::::_l:::--,.,...L------l
13-20
25-31
(20)
(54)
(1m
(52)
(49)
II>
LU
20
...""
Ci
o FANCHERBROWN
LU
0
!Z
LU
""
"LU
LU
-20
""
t::.
t::.
~o-:----0
~.*
' \
LU
>
<
HAGEDORNo BROWN
\-DUNS-ROS
-40
...
z-
c
c
""
<
c 20
z
<
III>
ORKISZEWSKI
/0
/'0."",'"''
-/\
60
;:::
< 40
t::.
-/0
/
/~
L. ~O
~~
HAGEDORNBROWN
oFANCHER_ BROWN
DUNS-ROS
o'-----;;;;-';-,,;;---;;:;;;-'=;---;=I:=:--:=~:-::-::-----l
20-400
(2201
(226)
(167)
(113)
RANGES OF PRODUCING GAS/LIQUID RATIO, SCF/STB
(NUMBER OF WELL TESTS IN EACH GROUP IN PARENTHESES)
910
the diameter that must be bridged to form and maintain a slug of liquid or gas. Fig. 1 shows a tendency
for some of the correlations to overpredict pressure
losses for the 2.992-in. tubing. Note that the Hagedorn and Brown and the Orkiszewski correlations do
quite well for 1- to 2-in. tubing.
Produced-On API Gravity
It has long been recognized that crude-oil properties
such as surface tension and viscosity can often be
correlated with their specific gravities. Even though
the properties of produced oil vary from the properties of oil flowing in the tubing, the produced-oil
gravity would still show some trends with prediction
accuracy. Fig. 2 indicates that the Hagedorn and
Brown method overpredicts pressure loss for heavier
oils and underpredicts pressure loss for lighter oils.
The Orkiszewski method seems to improve as oil API
gravity increases. No trend is indicated by the standard deviation plot.
Producing GLR
The producing GLR is an indication of the actual
GLR in the tubing; but, of course, the actual GLR
depends heavily upon tubing pressure as well as other
factors. The Hagedorn and Brown and the Orkiszewski correlations generally tend to show more scatter
in prediction accuracy and also to overpredict pressure loss for higher producing GLR's (see Fig. 3).
Also, the Duns and Ros correlation overpredicts as
GLR is increased. The Fancher and Brown correlation tends to underpredict pressure losses, yet outperforms other correlations at high producing GLR's.
This was also observed by Camacho. 20
Producing WOR
The presence of water in the production string can
have several effects upon the pressure gradient. Most
obvious is that because water is usually more dense
than the oil and gas phases it tends to increase the
pressure gradient. However, accurate representative
values of physical properties for a liquid made up of
a mixture of water and crude oil are not always possible to obtain by averaging separate water and oil
properties. This is because water-oil emulsions can
form that do not follow the averaging rwes.
Fig. 4 indicates that the correlations do a better
job of predicting pressure losses when little or no
water is present. Where the produced liquid is predominantly water, the correlations seem to overpredict pressure loss and exhibit more scatter in the
prediction errors. Duns and Ros did not recommend
using their prediction method with averaged physical
properties for the liquid phase, as was done in this
work, but instead recommended that two calculations
be performed, one with oil and one with water, each
at the same total liquid production rate. The gradients were then to be averaged to get an estimate for
the combined water and oil gradient.
Superficial Velocity of Produced Liquids
The definition of superficial velocity of produced
liquids used in this study is simply the production
rate of oil and water (in STBjD) divided by the crossJOURNAL OF PETROLEUM TECHNOLOGY
sectional area of the tube times the appropriate constant to give units of feet per second. This quantity
is not normally used by petroleum production engineers but was included to show tr~nds based upon
the liquid-production-rate/tubing-area ratio. It is interesting to note that, as displayed in Fig. 5, the
pressure loss predictions generally improve as the
superficial velocity of produced liquids increases.
Also, although the average percent differences for
predictions at low superficial velocities for the Hagedorn and Brown and the Fancher and Brown correlations are quite low, there is still considerable scatter
in the prediction errors at low superficial velocities.
Apparently, as production rates increase, the gas and
liquid have a tendency to flow more predictably or
uniformly than at lower rates. The poorer predictions at lower liquid superficial velocities could be an
indication of a more cyclical or noncontinuous type
of flow, which would be expected to be more elusive
of being predicted by correlating techniques based
on continuous flow.
Use of Figures 1 through 5
Figs. 1 through 5 should be used cautiously for selecting a prediction method for a particular set of flow
variables. The trends in the statistical values they
show may not always be caused entirely by the particular flow variable against which the statistical
values are plotted because the data may have a statistical correlation between different flow variables.
For example, test data for wells having high produced GLR's were also those for wells having tubing
diameters between 1.995 and 2.992 in. Therefore,
the trends could be caused either by the range of
tubing diameter or by produced GLR, or by both.
Within these limitations, the figures may be used as
a guide to eliminate or select a particular correlation
in the absence of other information.
40,.--------.------,---.-----.-----.--------,
~~
e:::::==--==-~
O~H~GEDORN-BROWN
-
20 -
as
'-'
15
D..
~ORKISZlWSKI
e--e
'"
~~NCHER-BROWN -
~ -20 f-
15
;c
e e DUNS-ROS
-40f-
60
#-
z"
FANCHER-BROWN
~4O
~
%ORKI SZlWSKI
(;t;
~
/~
H
~O#~~
",C:>f/ / 0
~~/
.~~
'"
'"
IX
~20
z
;!:
lI'l
6_t:;
HAGEDORNBROWN
(8)
(9)
(92)
(325)
40
.----.-1----.1---..--1--'1------,
20
I-
....
(.)
Conclusions
1. A comparison study such as this one depends
a great deal on the quality and range of basic well
data. Different well data may result in different conclusions as to the method having the best over-all
performance.
2. Inaccuracies of fluid physical property correlations for predicting volumes and properties of downhole fluids have an unknown effect on the statistical
results. Each pressure-loss prediction method, which
combines a pressure loss correlation and fluid physical property correlations, must be considered as a
unit when tested against measured pressure losses.
3. No single pressure-loss prediction method was
found superior to all others considered for all ranges
of producing well flow variables. Over all, the Hagedorn and Brown method, when tested against all 726
well tests, performed best.
4. Figs. 1 through 5 may be used with caution to
eliminate or select methods for particular ranges of
producing well flow variables.
5. The inability of the methods to consistently predict producing well pressure gradients with accuracy
indicates the need for more research in the area of
vertical multiphase flow.
AUGUST, 1974
as
a:::
....
::::
7;::::><-
I-
as
(.)
....
....
C>
~
....
>
o
o
a:::
c..
-20 -
t::. ORKISZEWSKI
DuNS-ROS H~~O:N- _
<
-40-
l/<
:z
o
BD
DUNS-ROS
e
RKISZEWS~
FANCHER-I::.......... "
40 BROWN o~o
~. HAGEDOR~~~
BROWN
~~
V'l
0-.4
.4-1.0
1.0-2.0
>2.
(219)
(221)
(154)
(132)
RANGES OF SUPERFICIAL VELOCITY OF PRODUCED LIQUIDS, FTISEC
Fig.
~Statistical
911
Acknowledgments
We wish to thank Amoco Production Co. for releasing
this information for publication. Also, we wish to
acknowledge the assistance of Phil Reed of the Continental Oil Co. Research Dept. for his help in computer program comparisons.
References
1. Poettmann, F. H., and Carpenter, P. G.: "The Multiphase Flow of Gas, Oil, and Water Through Vertical
Flow Strings with Application to the Design of Gas
Lift Installations," Drill. and Prod. Prac., API (1952)
257.
2. Baxendell, P. B., and Thomas, R.: "The Calculation of
Pressure Gradients in High-Rate Flowing Wells," J.
Pet. Tech. (Oct. 1961) 1023-1028; Trans., AIME, 222.
3. Duns, H., Jr., and Ros, N. C. J.: "Vertical Flow of Gas
and Liquid Mixtures in Wells," Proc., Sixth World
Pet. Cong., Frankfurt (1963) 11,451-465.
4. Fancher, G. H., and Brown, K. E.: "Prediction of
Pressure Gradients for Multiphase Flow in Tubing,"
Soc. Pet, Eng. J. (March 1963) 59-69; Trans., AIME,
228.
5. Hagedorn, A R., and Brown, K. E.: "Experimental
Study of Pressure Gradients Occurring During Continuous Two-Phase Flow in Small-Diameter Vertical
Conduit,s," J. Pet. Tech (April 1965) 475-484; Trans.,
AIME,'234.
6.0rkiszewski, J.: "Predicting Two-Phase Pressure Drops
in Vertical Pipe," J. Pet. Tech. (June 1967) 829-838;
Trans., AIME, 240.
7. Brown, K. E.: Gas Lift Theory and Practice, PrenticeHall, Inc., Englewood Cliffs, N. J. (1967).
8. Fancher, G. H.: "An Investigation of Continuous Multiphase Flow in Vertical Tubing Strings," MS thesis U.
of Texas at Austin (1962).
'
9. Ros, N. C. J.: "Simultaneous Flow of Gas and Liquid
as Encountered in. Well Tubing," J. Pet. Tech. (Oct.
1961) 1037-1049; Trans., AIME, 222.
10. Hagedorn, A. R.: "Experimental Study of Pressure
Gradients Occurring During Continuous Two-Phase
Flow in Small Diameter Vertical Conduits," PhD dissertation, U. of Texas at Austin (1964).
11. Griffith, P.: "Two-Phase Flow in Pipes," paper presented at Special Summer Program, Massachusetts Institute of Technology, Cambridge (1962).
12. Griffiith, P., and Wallis, G. B.: "Two-Phase Slug Flow,"
J. Heat Transfer, Trans., ASME (1961) 307.
13. Yocum, B. T.: "Two-Phase Flow in Well Flowlines,"
Pet. Eng. (Nov. 1959)'B-40.
14. Baker, W. J:, and Keep, K. R.: "The Flow of Oil and
Gas Mixtures in Wells and Pipelines: Some Useful
Correlations," J~ Inst. of Pet. (May 1961) 47, No. 449,
162.
15. Tek, M. R: "Multiphase Flow of Water, Oil, and
Natural Gas Through Vertical Flow Strings," J. Pet.
Tech. (Oct. 1961) 1029-1036; Trans., AIME, 222.
16. Hughmark, G. A, and Pressburg, B. S.: "Holdup and
Ptessure Drop With Gas-Liquid Flow in a Vertical
Pipe," AIChE Jour. (Dec. 1961) 677.
17. Gaither, O. D., Winkler, H. W., and Kirkpatrick, C. V.:
"Single- and Two-Phase Fluid Flow in Small Vertical
Conduits Including Annular Configurations," J. Pet.
Tech. (March 1963) 309-320; Trans., AIME, 228.
18. Beal, C.: "The Viscosity of Air, Water, Natural Gas,
Crude Oil and Its Associated Gases at Oil Field Temperatures and Pressures," Trans., AIME (1!146) 165,
94-115.
19. Hatchek, E.: The Viscosity of Liquids, G. Bell and
Sons, London (1928).
20. Multiphase Flow Through Vertical or Horizontal Pipes
(With Segmenting Features), Program 1302, Users
Reference . Manual, D. R. McCord and Assoc., Inc.,
Dallas (Dec. 1970).
21. Production Research Div., Research and Development
Dept., Continental Oil Co.: Personal communication
(Nov. 19, 1971).
912
APPENDIX
Extensions to Correlations
Figure numbers and symbols referred to are those defined by the original publication for each correlation.
Poettmann and Carpenter
The friction factor for values of the Reynolds number numerator (abcissa of Fig. 4) less than 0.80 and
more than 170 were taken as 100 and 0.001, respectively.
Baxendell and Thomas
The lower limit on the Reynolds number numerator
and associated friction factor were taken as 100 and
0.8, respectively. The friction factor was taken as
0.0044 for values of the Reynolds number numerator
greater than 8,000,000.
Fancher and Brown
The < 1,500 GLR line (Fig. 6) was extrapolated to
the points (f = 7.0, dvp = 0.08) and (f = 0.002, dvp
= 170). The other two GLR curves were extrapolated linearly over the same range of dvp. The endpoint vall/es of f were used when dvp was less than
0.08 and greater than 170.
JOURNAL OF PETROLEUM TECHNOLOGY
PLy(l;.
to correct an apparent typographical error. 21
Original manuscript received in Society of Petroleum Engineers
office July 19, 1973. Revised' manuscript received Oct. 31, 1973.
Paper (SPE 4267) was first presented at the SPE-AIME 48th Annual Fall Meeting, held in Las Vegas, Nev., Sept. 30-0ct. 3, 1973.
Copyright 1974 American Institute of Mining, Metallurgical, and
Petroleum Engineers, Inc.
.
This paper will be printed in Transactions volume 257, which will
cover 1974.
Discussion
G. L. Chierici, SPE-AIME, AGIP Exploration & Production
G. M. Ciucci, SPE-AIME, AGIP Exploration & Production
G. Sc1occhi, AGIP Exploration & Production
The authors, Lawson and Brill, are to be commended
for trying to take a fix on a subject that is affected
by a multitude of proposed methods.
The following points deserve discussion.
Lawson and Brill have evaluated fluid properties
through standard correlations. This introduces errors
that, except for gas-water systems, are unpredictable
and mask the error inherent in the pressure-loss prediction method being evaluated. Since the authors
say that PVT data were available for some of the
cases, an additional comparison limited to those cases
in which actual PVT data were used in the calculations would have been of interest. There are serious
arguments that lead one to believe that calibration of
fluid characteristics against experimental PVT data
is a minimum requirement for the reliability of calculated pressure drops.
Moreover, all three-phase cases should have been
discarded from the comparison because none of the
methods considered is based on a theory or correlation accounting for three-phase flow. Empirical extensions to three-phase flow have been tentatively made
in some of the methods, only in order to handle cases
of low WOR's. Lawson and Brill's use of average
values for the physical characteristics of oil-water mixtures does not take into account slippage and emulsification phenomena, whose relevance has' been
pointed out by Ros1 and by Duns and Ros.2
The above considerations suggest that grouping the
results into three main categories - gas-oil, water-oil,
and three-phase - would make more significant the
comparison of the evaluated prediction methods.
In comparing the relative merits of the various
methods the authors should have eliminated the bias
the data bank has towards some of the methods. This
bias is originated by the inclusion in the data bank of
cases that had been used to develop the empirical
relationships related to some of the methods. Considering the gas-water cases only, for which the validity
of PVT data is not questioned, Table D-1 shows how
the comparison is affected by the bias contained in
the data bank.
References
1. Ros, N. C. J.: "Simultaneous Flow of Gas and Liquid
as Encountered in Well Tubing," J. Pet. Tech. (Oct.
1961) 1037-1049; Trans., AIME, 222.
2. Duns, H., Jr., and Ros, N. C. J.: "Vertical Flow of Gas
and Liquid Mixtures in Wells," Proc., Sixth World Pet.
Cong., Frankfurt (1963) 11,451-465.
AUGUST, 1974
Number
of
Cases
Average
Percent
Difference
Standard
Df:JViation
318
318
318
102
318
318
-174.0
-174.7
18.4
41.3
10.7
13.5
247.5
247.0
42.8
84.2
22.6
43.1
318
318
318
102
101
234
-174.0
-174.7
18.4
41.3
35.6
20.6
247.5
247.0
42.8
84.2
20.8
47.8
913
Authors' Reply
The reviewers have criticized the selection of well
test data used in the statistical analysis, the evaluation
of pressure-loss prediction methods for systems other
than gas-water, and the use of standard correlations
to generate needed values of fluid physical properties.
We agree that relying on standard fluid physical
property correlations introduces error into the analysis, but published oil and gas well test data having
complete fluid property definition are not available.
Limiting the analysis to mr-water test data eliminates
the problem of physical property estimation, but
weakens the analysis by greatly reducing the amount
and range of available well test data. Also, performing the analysis using oil and gas well test data is
supported by the common use of the methods for oil
914