You are on page 1of 6

The Predicted Performance of a Gas-Condensate System

Washington Field, Louisiana

JAMES E. BERRYMAN SOHIO PETROLEUM CO.


JUNIOR MEMBER AIME OKLAHOMA CITY, OKLA.

T. P. 4490

ABSTRACT usually single phase flow, the pressure-depletion per-


formance is primarily a function of the phase behavior
A method is presented for calculating the pressure of the reservoir fluid.
depletion performance of a gas-condensate reservoir
from the original reservoir fluid composition using the The properties of a gas-condensate system of major
importance are: (1) compressibility factor at reservoir
laboratory determined pressure-production history and
temperature and pressure; (2) dew point pressure; (3)
retrograde liquid accumulation as guides. Equilibrium
vaporization ratios for this calculation were modified composition changes of the vapor phase with pressure
depletion; (4) recovery of original in-place hydrocar-
from published data until the calculated dew point
pressure and retrograde liquid accumulation agreed bons at any pressure; and (5) retrograde condensate ac-
with the values observed in the laboratory. cumulation.
The calculated results are compared with the pro- These properties are most generally determined by
duced gas composition determined by low temperature a pressure depletion-type experiment on recombined
analysis during constant volume laboratory experiments separator gas and liquid samples in the laboratory.
of pressure depletion behavior. This comparison shows Ideally the pressure-depletion performance can be cal-
that the calculation procedure is adequate for deter- culated from the composition of the original reservoir
mination of gas phase composition. fluid if the reservoir temperature and initial pressure
Data from field separation tests in low temperature are known.""" It has been reported " . . . compari-
separators are presented which verify the applicability son of such computations with depletion-type ex-
of published equilibrium vaporization ratios for a wide periments on recombined samples in the laboratory have
ranRe of separator conditions. yielded results that appeared to be entirely adequate.'"
The success of such calculations is determined by the
INTRODUCTION equilibrium vaporization ratios chosen, especially those
for the higher boiling point hydrocarbons.
A reliable prediction of the pressure depletion per- It is the purpose of this paper to present a method
formance of a gas-condensate reservoir is necessary in for computing the change in composition of the vapor
determining reserves and evaluating field separation phase during pressure depletion by production using
methods. The predicted performance is also used in laboratory determined properties of the system to guide
planning future operations and studying the economics the calculation and to illustrate the use of the method
of projects for increasing liquid recovery. Because the with data from the Cockfield "D" sand of Washington
flow of fluid in a volumetric gas-condensate reservoir is field, St. Landry Parish, La. Results of field tests and
Ori,::dnal manuscript received in Society of' Petroleum Engineer!" laboratory analysis are included. As a result of this res-
off!.ce on July 16. 1956. Revised manuscript received on Jan. 11.
19<17. Paper presented at Petroleum Branch Fan Meeting- in L()~
ervoir fluid study, a cycling program has been evaluated,
Angeles, Oct. 14-17, 1956. the field has heen unitized, and the cycling started in
D}scussion of this paper is invited. Discussion in writin~ t::~ July, 1956.
('oples) may be sent to the offices of the Journal of Petroleum
1'echnologll. Any discussion off4?red after De('. 31, ]fHi7. shoull]
he in the form of a new paper. I f{t'fen;'n~('s given at end of papel'.
SPE-698-G
PETROLEl'M TRA:\"SACTIONS, AIME J()~
FIELD SAMPLING the volume required for the pressure depletion experi-
ment.
During testing of the first three wells completed,
The average dew point pressure was 56 psi higher
companion separator gas and liquid samples were col-
than the original average BHP of 4,580 psig. For the
lected for laboratory analysis. Production test data ob-
two samples the deviation between the laboratory dew
served during the sampling were very similar and the
point pressure and the recorded BHP of the well at the
results for a typical well are summarized in Table 1.
time of sampling was 16 and 90 psi. The data indicate
Test data for three flow rates were obtained for each
that the reservoir fluid was at the dew point under
well over a two or three day period. The wells were
original reservoir conditions and all calculations are
shut in at night. Total gas production while on test was
based on this conclusion. The visual determination of
less than 3 MMcf and the producing rates which varied
the dew point was aided by marked color changes in
from test to test were in the range of 0.7 to 3.6
the gas-condensate system.
MMcf/D. Production through calibrated equipment per-
The compressibility factor was calculated from pres-
mitted frequent determinations of the separator gas-
sure-volume measurements over a range in pressure
condensate ratio. When this ratio had remained con-
from 200 psi below the dew point to 400 psi above.
stant for two hours under the conditions of the largest
The average compressibility factor of 0.929 at initial
flow rate the samples were taken. Comparison of the
gas-cond~nsate ratio tests of the sampled wells with reservoir conditions gives an original gas in place value
of 1.50 Mcf/bbl of hydrocarbon pore space.
tests of other wells indicated that the properties of the
While determining the pressure-production perform-
reservoir fluid were constant throughout the reservoir.
ance, the equipment is operated at a constant volume
LABORATORY ANALYSIS with the two cells connected together. Gas phase ma-
terial is removed from the top of the windowed cell by
In the laboratory two of the companion separator gas production into a measuring system. The retrograde
and liquid samples were recombined in their produced liquid accumulation is determined by pumpi?g mercury
ratio and the pressure depletion performance of each into the windowed cell and measurmg the mcremental
sample was measured. The third sample was not anal- volume of mercury required to raise the gas-retrograde
yzed because the results obtained from the two sam- liquid contact and the mercury-retrograde liquid con-
ples compared favorably. tact to window level.
The laboratory equipment consists of two equilibrium depleting the pressure on the system by production with
cells contained in an air bath; a double barrel, power- the cell at a constant volume. This dew point pressure
driven, volumetric mercury pump; and auxiliary equip- agreed with what previously obtained in the variable
ment for measuring pressure, temperature and gas phase volume study. A sample of the fluid produced in re-
production. One of the cells is equipped with windo~s ducing the cell pressure to the dew point pressure was
for visual observation of the fluids in the cell. ApproXI- collected for low temperature fractional analysis. Other
mately 18 per cent of the total volume of the cell is samples of the vapor phase material produced during
below the window level. The cells are arranged so that pressure depletion below the dew point were obtained
they may be used either separately or as a single unit for analysis at pressure intervals of approximately 600
and either as variable or constant volume equipment. psi. A sample of the condensed liquid remaining at ap-
The equipment is operated as a variable volume sys- proximately 500 psig was also analyzed. The retrograde
tem when determining dew point pressure and compres- liquid accumulation was measured at intervals cor-
sibility factor of the recombined sample at reservoir responding to about 300 psi pressure reductions. Meas-
temperature. Sufficient quantities of the recombined urement of the volumes of vapor phase material pro-
sample are charged to the windowed cell to give pres- duced enabled the determination of the pressure-produc-
sures several hundred psi above the dew point pres- tion performance of the sample. This performance was
sure. With the sample at reservoir temperature and reported at per cent of hydrocarbons in place at the
confined to the volume above the window in the win- dew point produced vs pressure. The compressibility
dowed cell, the pressure is lowered by removing mer- factor, the pressure-production performance and the
cury from the second ceIl. The volume of the system retrograde liquid accumulation for one of the pressure
is increased until condensing liquid may be seen in the depletion experiments are shown in Fig. 1.
windowed cell. Material balance checks on total hydrocarbons han-
Pressure-volume measurements are recorded at regu- dled and on individual components were made for the
lar intervals for calculation of the compressibility fac- system. Only small differences existed when comparing
tor. After observation of the dew point the volume of the total material charged to the equilibrium cell with
the system is increased further to obtain data for one the amount produced and remaining at low pressure.
or two compressibility factor determinations below the Material balances on individual components using the
dew point. When the desired amount of data is ob- results of the low temperature analyses and the pres-
tained, all the recombined sample is returned to the sure production performance were not nearly so good.
windowed cell and the space it occupied is adjusted to The data obtained from the low temperature analyses
TABLE I-TEST DATA DURING SAMPLING were smoothed by changing the composition at each
Test Date Aug. 29, 1952 analysis point by trial and error until material balance
Flowing Tubing Pressure 3,075 psig
Flowing Tubing Temperature 93°f checks were obtained on each component. The smoothed
Separator
Separator
Pressure
Temperature
530 psig
74°f
values for one sample are compared with individual
Separator Liquid Production 12.12 bbl/hr data points in Fig. 2.
Separator Gas Production 133.4 Mcf/hr @ 15.025 psia and 60 0 F
Separator Woter Production o The change in composition of the produced gas dur-
Separator Gos~Condensote Ratio 11,007 cu ft/bbl
Stock Tank liquid 9.38 bbl/hr ing the various pressure depletion steps was used to
Separator Gas-Stock Tank calculate the recovery of the individual components.
Liquid Ratio 14,222 cu ft/bbl
Specific Gravity of Separator Gas 0.688 For the two samples, the recoveries for butanes and
Gravity of Stock Tank Condensate 57" API
lighter were 87.6 and 86.0 per cent, respectively, of

VOL. 210, J957


so
- TABLE 2-COMPARISON OF RESERVOIR flUID COMPOSITIONS
laboratory Mathematically Recom,bined

V ~
Componen_t_ Recombined Sample Field Test Sample
Methane ----~- ·112:'09"----
<T
/' Ethane
Propane
5.53
3.95
5.56
4.31

\
Iso·butane 1.11 1.12
N-butane 1.38 1.44
Iso·pentane .65 .71
N·pentone .64 .51
Hexanes 1.10 .83

RETROGRADE,~
LIQUID ACCUMULATION
Heptanes pi us 3.33
100.00
3.43
100.00
<II",
JU TABLE 3-FIELD TEST DATA-LOW TEMPERATURE SEPARATION
bl « 20
O:J
0:0.
<
\ Test Oate
Free Water Knockout Pressure
Oct. 14, 1954
3,100 psig
mZ Free Water Knockout Temperature 90°F

~ 10 \ LTX
l TX
Pressure
Top Temperature
750 psig
40°F

\
l TX Bottom Temperature 85°F
low Pressure Separator Pressure 85 psig
~ow Pressure Separator Temperature 75°F
o 1,753 psig to check the equilibrium vaporization ratios
<II
Z
used in evaluating methods of field separation. Gas and
o liquid compositions calculated using the equilibrium
m
~ 80
...... 0.98 N
vaporization ratios of Winn are compared to the actual
U
o "\ COMPRESSIBIL~
i
> analysis in Table 5. Unfortunately, this series of tests
~
II: f-
o FACTOR "- ::;
> was not made through calibrated equipment and the
X 60 0.94 ~ field gas-liquid ratio is not available for checking against
~
~-
"'U <II

<
J
"'a:0. the calculated ratio. A plot of the published K factors
~ vs the normal boiling point is shown in Fig. 3 for a
0. ::1
:; 40 -- 0.90 S pressure of 1,306 psig and temperature of 63 The 0

PRODUCTlON~
,
...o test data are also shown. The results of these test data
f-
Z i'" are also shown. The results of these tests justified the
",20
use of the published data in the study of field separation
~
U
0:
"'0. over a wide range of operating pressures.

1000 2000 3000


PRESSURE - PS I G
4000

FIG. l-LADORATORY PRESSURE-DEPLETIO:-i PERFOR:\IANCE.


"" 5000 CALCULATION OF PRESSURE DEPLETION
PERFORMANCE
The general differential vaporization method of cal-
the amount in place initially at the dew point pressure.
The pentanes plus in the two original samples aver- culating the effect of retrograde condensation upon the
aged 66.1 bbljMMcf and their recoveries to a final cell gas phase composition outlined by Allen and Roe" was
pressure of 500 psig were 48.0 and 48.9 per cent of that followed. A much more detailed explanation of the
initially in place at the dew point pressure. 1000
80.0
0 - _0 a ------a- i-

FIELD TESTS 60.0

In Oct., 1954, nine months after the start of gas 40.0


sales from the field, samples of the liquid and gas stream
were taken at the low temperature separator, the low
pressure separator and the stock tank. Recombination 20.0
of these samples resulted in a fluid composition that
has been accepted as best representing the reservoir fluid
of the field. A comparison of the fluid compositions 10.0

determined from the laboratory recombined sample f- B.O

and the field test samples is shown in Table 2. A &um-


maryof the field test conditions is presented in Table 3.
Z
'"
u
0:
6.0
0-
• . 0
oc.
c._
Stage separator flash calculations at the actual oper- '"0. 4.0
o -r - - -
ating temperatures and pressures of the field installa- a:
\ , • ~OC7·
<
J
q
,/. oc.
tion were made to compare calculated recovery with ::>
2.0
0 0
0
./
actual recovery and to compare calculated composi-
U
'"0J V-
tions with the actual analysis. The calculated stock tank
recoveries in bbljMMcf of sales gas using the labora-
tory determined composition and the field test composi-
::1

1.0
0.8
0 ___

0
'I:-~-
/ 0
_0--
o~ -:::::i~5
C.

g
tion were 80.6 and 78.7, respectively. Actual average 0.6 \ \

production for nine months was 78.6 bbl/MMcf. The \, 0

reservoir fluid composition obtained from the field test 0.4


analysis was used to calculate the theoretical compo-
sitions at the various sample points. These data are pre-
I
sented for comparison with the actual analysis in Table 0.2

4. The equilibrium vaporization ratios used were those


presented by Winn'.
In July, 1955, another series of separator gas and 0 1'------- I
o 1000 2000 3000 4000 5000
liquid samples from a high pressure separator were col- PRESSURE - PSIG
lected at separator pressures of 1,026, 1,306 and FIG. 2·--L~BORATORY ANAI.YSIS OF EXPE/lIME:\TAL PROOlJfl:O GA~.

101
TABLE 4--COMPARISON OF CALCULATED AND ACTUAL ANALYSIS 10.0

"
COMPOSITIONS FOR LOW TEMPERATURE SEPARATOR TEST
LIQUID

'"
GAS
Actual Mol
Per Cent
High Pressure Separator
Calculated Mol
Per Cent
Actual Mol
Per Cent
Calculated Mol
Per Cent

~ '" "'"'"
~
~~.,
90.11 89.97 18.14 19.17

'"
Methane
Ethane 5.34 5.39 7.30 6.79

~
Propane 3.07 3.14 14.19 13.57
Iso-butane .59 .55 5.36 5.66
N-butone .48 .58 9.07 8.30 ~

~ ,~ "'"l'- ~
~
Iso-pentane .13 .15 5.36 5.17
N-pentane .09 .09 3.83 3.90
Hexane .07 .05 6.89 7.07 0",
Heptane plus .12
100.00
.08
100.00
29.86
100.00
30.37
100.00 o 0
:f",~ f'\
r-.....
low Pressure Separator
Methane
Ethane
Propane
62.69
15.38
14.44
59.40
14.78
16.18
.96
3.26
12.86
1.70
3.32
12.44
>=
«
a:
z
~ '" ~,h
~ 1'-.,
1'-.,

t"-r--.
V
......v

""
3.50 7.55 6.60 o
Iso-butane
N-butane
3.09
2.81 3.88 10.36 10.22 >=
« /1/
Iso-pentane
N-pentone
.79
.32
1.04
.60
6.65
5.31
6.96
5.34
~

.. ~
a:
o . " " "'A
1'--
- V/
~"'''
Hexane .30 .36 9.73 9.98
Heptane plus .18
-100:00
.26 ~
100.00
43.44
100.00 .
~ ~"'~ V/1J
100.00
t--.. 1- ~
~ Vj
~

.
:>
a:
method is presented by Organick and Meyer'. The
:J 0 .I ~-t.,-\( ./
necessary calculations were performed on a Model II ~
IBM CPC using the general purpose control panels
:>
a ./'" ['
developed by Rachford and Peaceman'. The computa-
'" !
tions of vapor liquid equilibrium were programmed for
the digital computer by the widely used "twenty ques-
"-
0iIGINAL DATA y !
tions" method described by Rachford and Rice".
"-
"-
"- I. ,,~/
~~
q'jl rh:ALUES
MATCH l.ABORATORY
uSE. TO

Equilibrium vaporization ratios or K factors from 1',, I'~V


,,~.~.'
M
~
LIQUID ACCUMULATION

the correlated data of Rzasa, Glass and OpfelI' were "


--
~
:.-~

used for the calculations. Starting with values for 194°F


and a convergence pressure of 5,000 psi a as the first
I
estimate, the factors were adjusted by trial until the ex- ,0100 1000 10.000

perimental dew point of 4,580 psig was matched by cal- PRESSURE - PSIG

culation. This occurred with a convergence pressure FIG. 4-EQU1LlURWM VAPORIZATION RATIO-K. AT
TEMPERATURE OF 192°F.
of 5,300 psig. The curve for the heptane plus fraction
was estimated using the molecular weight and specific After selecting the equilibrium vaporization ratios,
gravity as guides. The adjusted K values used in the cal- the calculation was started using 4,580 psig as initial
culations are shown plotted against pressure in Fig. 4. reservoir pressure and the composition as determined
Reasonable agreement exists between the selected K by the field tests of the low temperature separator in-
factors and other published data and points determined stallation as the composition of the reservoir fluid. The
from the laboratory analysis of this gas-condensate sys- main point of divergence from the reported method of
tem. computation was the use of laboratory pressure-produc-
10.0
tion performance (per cent of in place hydrocarbons
produced vs pressure in Fig. 1), to determine the amount
of gas phase material removed or produced from the
system during a pressure depletion intervaL

~ Use of the laboratory data eliminates the need of


obtaining compressibility factors for the vapor phase
1.0 ""- .
~
TABLE 5-COMPARISON OF CALCULATED -t-ND ACTUAL ANALYSIS
COMPOSITIONS AT VARIOUS SEPARATOR CONDITIONS
GAS liQUID
z Actual Mol Calculated Mol Actual Moll Calculated Mol
o
~N
a: o. I
N" •
u

,",c
.
u Methane
Ethane
Per Cent

91.28
90.05
4.26
5.29
Per Cent
Test l-Seporafor Pressure 1,026 psig-Separator Temperature 53°F
Per Cent

25.91
7.40
Per Cent

26.48
7.45
o

"'~f
Propane 2.84
3.13 11.28 12.53
Q.
'-u Butane 1.09
1.12 10.93 12.61
«
> oc
..
u
Pentane
Hexane
Heptane plus
.32
.26
.08
.07
.13
.08
6.62
5.46
32.40
7.96
6.12
26.85
li

K ioo.oo' 100:00
::> "iOQ.OO 1 00.00
a:cD Test 2-Separafor Pressure 1,306 psig-Separator Temperature 64°F
:::i Methane 90.86 89.87 33.80 31.53
:; Ethane 4.15 5.26 8.00 7.51

K
a 0.0 I Propane 2.89 3.17 11.31 11.73
..... Butane 1.38 1.19 10.21 11.41
o F'IEI.D TEST DETERMINATIONS
Pentane .36 .31 6.71 7.19
- AF'TER WINN .19 5.74 5.65
Hexane .08
Heptane pi us .17 .12 24.23 24.98
100.00 100:-00 100.00 100.00
Test 3-Separator Pressure 1,753 psig-Separator Temperature 77°F
Methane 89.59 89.36 41.83 38.85
0.00 I Ethane 4.19 5.24 7.62 7.48
-300 -200 -100 o 100 200 300 Propane 3.33 3.28 11.85 10.42
Butane 1.68 1.36 8.32 9.70
NORMAl. BOII.ING POINT - OF. Pentane .55 .40 5.12 6.11
Hexane .29 .15 4.79 4.87
FIG. 3-COMPARISON OF PUBLISHED "K" VALUES WITH Heptane plus .37 .21 20.47 ~
FIELD TEST DATA. -100:00
100.00 100.00 100.00

VOL. 210. 1957


from pseudo critical temperature and pressure correla- on the calculations. The volumes computed using the
tions. Another change was the assumption that the estimated properties for the heptane plus fraction are
compositions of the produced gas during an interval was shown primarily for comparison with the volume cal-
equal to the composition of the vapor phase at the start culated using constant properties for this fraction. The
of the pressure depletion step. effect of this change is relatively small and no attempt
An improved method of determining the composition was made, nor believed to be necessary, to change the
of the gas to be removed from the system would be to over-all calculation so that these volumes would match
assume a composition equal to the arithmetic average the laboratory data.
composition of the vapor phase at the start and final Original laboratory data of the change in composi-
pressure of an interval. This computation properly per- tion of the vapor phase during pressure depletion are
formed would require iterative procedures to determine shown in Fig. 2. The calculated data are plotted in Fig.
the average composition of fluid removed in each step. 5. For a better comparison, the calculated and labora-
Limited machine storage and programming for as auto- tory data for the individual components are plotted
matic a computation as possible eliminated the latter separately in Fig. 8 so there will be no overlap and the
method for use with the CPC although it will be used differences may be easily seen. The composition from
when the computation is programmed for the magnetic the low temperature separator installation which was
drum calculator now available. used as a starting point in the calculation agrees very
The first pressure reduction interval was 280 psi and 100,0
all subsequent intervals represented 300 psi steps. The 80,0
.r,,_
calculation was continued to a final pressure of 400 60.0
psig. The volume occupied by the liquid phase at thc
end of each pressure depletion step was determined 40.0
from the phase composition by the method of Standing
and Katz'.
20.0
During the first prediction the equilibrium vaporiza-
tion ratios for the heptane plus fraction were used as
originally drawn. The molecular weight and specific
10,0
gravity of the heptane plus fraction in the liquid phase

--
8.0
were assumed to remain constant at 144 and 0.788, re- I-
~ 6.0
spectively. The calculation procedure was then repeated u
Cz
a: c.
just as before except that the K factor for the heptane ~ 4.0
plus fraction was changed by trial until the calculated a:
~C7+
~
volume of condensed liquid agreed closely with the «
-'
:::> "- -:./'-
C.
laboratory observed data. The properties of the heptane
plus fraction were assumed to remain constant. The final
values of K factors for heptane plus used in this calcu-
lation are shown in Fig. 4. The calculated change in
u 2.0
w
-'
o
::;

1.0
........'s,.
_? ......
c.

-
C._
composition of the vapor phase is plotted in Fig. 5. 0.8

There is little difference in the calculated change in 0,6 '- ./


.= <-=
~-
composition for the methane through pentane frac-
tions in the two computations performed. The changes 0.4
'"
- - CALCULATED BY CHANGING K Of C7~
in composition for the hexane and heptane plus frac- - - CALCULATED BY USING ORIGINAL ESTIMATE
tions calculated using the original K factors are also Of K fOR C7+
0.2
shown in Fig. 5.
The calculated volumes of retrograde condensate for
the two computations are shown in Fig. 6 along with o. I
the curve of the laboratory observed data redrawn fwm o 1000 2000 3000 4000 5000
PRESSURE - PSIG
Fig. 1. The fourth set of data on Fig. 6 represents the
FIG. 5-CALCULATED CO~IPOSITION OF VAPOR PHASE.
volumes calculated using an estimate of the change in
properties of the heptane plus fraction with pressure '" 50

~~
z o a
depletion of the system and the liquid phase compo-
sitions determined by changing the K factors for the
o
III
a:
«
! Q~---'"
o
~
~

0" ~

heptane plus factor. The properties used in this latter


case were estimated from the change in molecular
u
o 40
a:
",.
-+ Q

0
~
~

"\ ~
,.." ~

weight and specific gravity of the produced gas observed :0::


~

in the laboratory experiments. The estimated values for w

\\
u 30
the molecular weight of the liquid and vapor phases «
J
0.
and the laboratory observed data are plotted in Fig. 7. ~
...o 20
COMPARISON OF CALCULATIONS WITH ...u -
LABORATORY SAMPLE
\
LABORATORY WORK J;
A ORIGINAL CURVE fOR K OF C7·

One of the assumptions for the calculations was that


J;

a:
w
10 f---o CHANGING K Of C7+-CONSTANT
C7+ PROPERTIES
1\
\
0.
the calculated retrograde liquid accumulation should be o CHANGING K OF C7~-CHANGING
equal to that observed in the laboratory. The original r~ PROPERI'ES -I
laboratory data used as guides in the calculations are 1000 2000 3000 4000 5000
shown in Fig. 1. The volume of retrograde liquid ob- PRESSURE - PSIG
served in the laboratory are com!,ared in Fig. 6 with FIG. 6-COl\IPARISON OF CALCULATED RETROGRADE LIQUID TO
the computed volumes for the three conditions imposed LABORATORY OBSERVED DATA.

PETROLEI'''' TR'\;\~ACTI()N~. \1\lE 106


--
160
0
z
0
~ 150
- 20'-----·
i ~ETHANE

0
«
a:
............... ~ PHASE '!:~~~~~
----1-- - - - - - - - - - - .
4~---- -----.~-----
\L
+ 0
S
\L
140
7
0 0[7
l-
X

"
W 130
:t
/ ______________

6~-------.----. ----
---" ___________________

--~--------.------'----
PROPAN:'_

a:
«
...J
:>
0 120
0............
r-.-~ _0/
k VAPOR I PHASE
4 -... ---+-----------'__ ..
2---~=====~====~=~
_ L '- - -

'"...J
0
- ESTIMATED MOLECULAR WEIGHT
I
~
0 LABORA~ORY DETER,MINATIONS ,
• PENTANE
04"---------·----~------·--~--
1000 2000 3000 4000 5000 2()-----=------- ------- -~ _________ ~ --- -,..--.0----

PRESSURE - PSIG
FI(;. 7-_\luu:ClJLAR WEIGHT OF
.'1.>;0 VAPOR PHASES.
C; L, IAQI III " _._~I _ _~:F--r-;
:;1_. ___ .~
i' _-+-- _____ _

well with the laboratory data. The propane fraction is 00------ ~---__.________-----_, ~-.

an exception and for this component it appears that 401--- - - ---i--- - ---~---.--,- - - - .- - - -

the molecular per cent used was high. The data for 20r-- _.- .

the pentane and heptane plus fractions show that both 1Ot----· -
08~-·--.I----
-~


--- .. - - - - - - - - -.-..- - . - - .. ----

~

declined more in the laboratory experiment that did the OS'- ._. ~. __ ~_=--=r.:-------~-_-J:-~_~_.- ~.:~ . _ PLUS ~
040-- 1000 2000 3000' 4000 !lOOO
calculated values.
The agreement between the laboratory observed data F](;. 8--CO~IPARISO'\ OF CALCCLATED i\Iou:cCLAR PER CDI'
and the calculated change in vapor phase composition WITH LABORATORY ANALYSIS.
is very good. The differences that do exist are due in
perimental data. This comparison shows that the phase
part to the fact that changing the K factor for the hep-
behavior can be calculated by the method of matching
tane plus fraction can not correct the values from pub-
calculated and laboratory observed retrograde liquid
lished data to the true K factors for the system. Ex-
volumes as accurately as the vapor phase composition
perience has also demonstrated the difficulty of sam-
may be determined in the laboratory experiment of pro-
pling and transferring to the fractionating column a
ducing vapor phase material and making low tem-
sample of gas containing small amounts of the pentane
perature analysis of the produced fluid.
plus fractions. The trend of experimental error in anal-
ysis of such samples is to yield low results for the
heavier components which is what the data shown in ACKNOWLEDGMENT
Fig. 8 indicate. The author wishes to express his thanks to the man-
The method of computation outlined has been ap- agement of Sohio Petroleum Co. and of Gulf Oil Corp.
plied to another laboratory sample and the agreement for permission to publish this paper. Special thanks is
of calculated and observed results was similar to the clue the Gas-Gasoline Div. of Sohio Petroleum Co.
agreement shown for the data presented in this paper. for supervising the field tests during which a great part
of the data was collected.
CONCLUSIONS
REFERENCE:;
The agreement of data from field tests, laboratory
analysis and engineering calculations and the relation- 1. Organick, E. I., and Meyer, H. I.: "The Stepwise Gas De-
ship of each part of the data to be predicted perform- pletion as Performed With the Aid of an IBM Model 604
Electronic Calculating Punch", paper presented at IBM
ance of a gas-condensate system has been presented. Pet. Con£., Endicott, N. Y. (Oct., 1953). " ..
It has been shown that samples collected during 2. Hoffman, A. E., Crump, J. S., and Hoeott, C. R.;, Eqtul!-
initial field tests of relatively short duration were rep- brium Constants for a Gas·Condensate System, Trans.
resentative of the reservoir fluid. Although the com- AI1\l£ (1953), 198, l.
~. Allen, F. H., and Roe, R. P.: "Performance Characteristics
position calculated from separator samples taken after of a Volumetric Condensate Reservoir", Trans. AIME
nine months of production has been accepted as being (1950). 189, 83.
more representative of the reservoir fluid, the compo- t Winn, F. W.: "Simplified Nomographic Presentation Hydro-
sition as determined from the early test samples falls carbon Vapor-Liquid Equilibrium", Chern. Eng. Progress
Symposium Series No.2 (1952).48, 12l.
within the normal range of accuracy expected when de- :J. Rachford, H. H., and Peaceman, D. W.: "A High-Speed
termining compositions by recombining the low tem- Ei<Yht·Di<Yit Floating-Point General Purpose System for the
perature analysis of liquid and gas samples in their ~I~del If CPC". paper presented at IBM Customer Admin-
produced ratio. istrative School. Endicott, N. Y. (Aug .• 1953).
Ii. Rachford, H. H., and Rice. J. D.: "Procedure for USf~ of
The actual stock tank production for a nine month Electronic Digital Computers in Calculating Flash Vapori-
period agrees with the liquid recovery calculated using zation Hydrocarbon Equilibrium", TrailS. AIME (1952l.
published equilibrium vaporization ratios. Field separa- 195, 327. .
tion tests at high pressure verify the applicability of i. Rzasa. M. J., Glass. E. D.. and Opfell, .I. ll.: "Prediction
of Critical Properties and Equilibrium Vaporization Con-
the published K factor data over a wide range of separa- stants for Complex Hydrocarbon Systems", Chern. Eng.
tor conditions. Progress Svmposium Series No. 2 (1952), 48, 28,
The calculated change in vapor phase composition K. Stamling, M. B., and Katz, D. L.: "Density of Crude Oil
Saturater! with Natural Gas", Tran." AIME (1942), 146,
with pressure depletion has heen compared to the ex-
159. ***
107 VOL. 210, 1957

You might also like