Professional Documents
Culture Documents
By
W. John Lee, Humble Oil & Refining Co., Robert R. Harrell, Mobil Oil Co.,
and William D. McCain, Jr., Mississippi State U., Members AIME
@ Copyright 1972
American Institute of Mhbg,MetaUurgical, and Petroleum Engineers, Inc.
This paper was prepared for the Northern Plains Section Regional Meeting of the Society of
Petroleum Engineers of AIME, to be held in Omaha, Neb., Mq 18-19, 19T2. Permission to copy is
restricted to an abstract of not more than 300 words. Illustrations may not be copied. The
abstract should contain conspicuous acknowledgment of where and by whom the paper is presented.
Publication elsewhere after publication in the JOURNAL OF PETROLEUM TECHNOLOGY or the SOCIETY OF
PETROLEUM ENGINEERS JOURNAL is usually granted upon request to the Editor of the appropriate
journal provided agreement to give proper credit is made.
Discussion of this paper is invited. Three copies of any discussion should be sent to the
Society of Petroleum Engineers office. Such discussion may be presented at the above meeting
and, with the paper, may be considered for publication in one of the two SPE magazines.
bility of better deliverability estimates than included in the equation. Further, the equatior
the back-pressure test, but this method also pro contains a turbulence constant, B, which charac-
vides little additional information about the terizes the degree of turbulent or non-darcy
well. flow near the wellbore.
A third method of testing gas wells uses Equation (1) is the basis for the conven-
pressure drawdown tests3. These tests CSI’I pro- tional analysis of a drawdown test: It suggests
vide estimates of permeability-thickness product that if we plot pwf2 vs. the logarithm of time,
and extent of well damage, in addition to deliv- a etraight line should result, and the slope of
erability estimates. Unfortunately, drawdown the line should be related to permeability-thicl
tests must be run at strictly constant rate if ness product. Unfortunately, use of this equa-
the drawdown data are analyzed in the usual way. tion to model a pressure drawdown test requires
‘il
...._
BVeLL m~~~ ...lGnvt matolv
w..”..,..-.--, $ ~oIM~~n~ rate drawdown the drawdown test to be run at an absolutely COY
tests are inconvenient and, at times, virtually stant rate. Winestock and Coipitts~ anaiysis
impossible to perform. One can always use super showed that if the rate varies during a drawdovn
position in analysis attempts, but superposition test by even a few percent, eq. (1) does an ex-
requires tedious arithmetic, and as a practical tremely poor job of modeling the test.
matter, forces much of the arithmetic required
in analysie to be performed on a computer. Winestock and Colpitts suggest that the COI
ventional pressure drawdown equation for tran-
In an attempt to fill this gap in gas well sient flow in a gas well be rearranged as fol-
testing technology, Winestock and Colpitts4 in lows :
1965 proposed a technique for analyzing variable
rate drawdown tests. They claim that their me- rl
thod offers all the information theoretically Pe2 - Pwf2 - @tj2
possible with the drawdown test, and in additior =
qg(t)
they claim that the test data are easily ana-
lyzed.
1
+2s * . . (2
-)
The testing program which Winestock and ● ●
L1
bility-thickness product changed from one re-
+Bq(t)2. . . . . . . . . . . . . . . . .(1)
gion of the reservoir to another.
. . ...*
+*J L
The pressure at the drainage radiuS:hPe, Ap(t) = pressure drawdown at time t in vari-
~r~
and the permeability-t”hickness product, roll, =kl~ rate test
u..-
found from an analysis of the buildup and draw-
down tests, and the skin factor, S, is found AP(t) ‘ pressure drawdown caused by productio
—
from the drawdown tests. Equations used to cal- at unit rate
culate these quantities are summarized in Appen-
dix A. q(t) = ~$n ;~s~ime t in variable rate draw-
is negligible when compared to the simulated variable rate drawdown tests vs.
the logarithm of flowing time, t. Note that
o ● AP(t) “Pi” is used instead of “pe”; in our simulated
qt
++ — tests of “new” reservoirs the quantities are
identical. Now if the Winestock and Colpitts
and there clearly must be some magnitude of rate method is valid, a straight line should appear ox
change with time beyond which the Winestock and such a plot, and the slope of this line should
:olpitts method will “overcorrect” drawdown data. properly reflect formation permeability-thickness
k evaluation of the drawdown test analysis pro- product used in the calculations. Further, the
cedure is thus required before we accept its total drawdown in pressure at any time should
validity. agree with that predicted by the equation
~i2 - ~wf2 .
Pwd = Pw/Pi
and the agreement was again found to be excellent dimensionless normalized plotting function,
Note that gas compressibility, Cg, has been re- fd ‘ (1 - pwd2 “dqd2)/qd
placed by I/pi in the logarithm term.
td ~ax be 10, 25, 50, and 75 percent of the 24,400 should be 5.45 frOmW. (8); the computed
‘d =
value from the simulated drawdown test was 5.53,
absolut~ open flow to check whether the magnitud
which is excellent agreement.
of the rate affected validity of the Winestock-
Colpitts method. The rate was allowed to declin
Figure 4 shows results from another simu-
from a high starting value to its final value in
lated drawdown test, but with S = 1.0and
the following way: Dimensionless rateY qdY varie
B’ = 1000. At td= 62,500, the rate, Cld,was
with dimensionless time, td, according to the
0.0609, and the rate was declining at 10 percent
-., cycle.
equation qd = a - b log(t+-O.l). Note that or Cn%s Eiid r~te over nsrh Iogio
-..-..
t ~ , ~~~ = 62,500 for rde = 1000; thus the constan
In this simulated test, pwd2 does not Plot
0.1 in the logarithm above is completely negli-
as a straight line against td; in fact, over the
gible for all but earliest times. The constant
“b” was chosen such that the decline in rate, wa range of times plotted, Pwd 2 is increasing dur-
O, 10, 25, 50, 75, 100, and 200 percent of the
ing the drawdown test. The cause of the increas(
f<nal rate per loglo cycle, and the constant “a”
, -..ql.* L.+ desired in sand-face pressure is the decrease in rate --
was cnoSeilsu~l~ ~,,=.~ ~e121~ ~~ve the
.--,~mtifim=l analysis would predict a
d note tY1atG“LLJ=...AU----
final value at td =td max. This variation in rat’ meaningless “negative” permeability if applied tf
- ,-. -..- s a test such as this.
of change or r~ow ra~e ..-- d...6..~-
-o~m-.A to ~he~k hOW
w~~
Figure 3 shows results of a typical simu- Figure 5 shows the effect of the magnitude
lated drawdown test from our investigation. The of rate decline on the nomalized drawdown test
plot is in terms of dimensionless variables: The plotting method. Curves are shown for constant
nomalized plotting function, fd, sand-face pres- rate, rate declining at 50 percent of the final
rate (at td = td max = 62,500), rate declining at
sure squared, pwd2, and rate, qd, against dimen- s
100 percent , and-rate declining at 200 percent.
sionless time, td. For this particular test,
At each rate decline magnitude, final rate mag-
skin factor, S, and turbulence constant, B, were nitudes from 10 percent to 75 percent of the ab-
zero. At td = td max = 62,500, the rate, Cld,wa solute open flow rate were used; and this final
rate magnitude proved to have almost no effect
25 percent of abs~lute open flow, or qd(tmax) ‘
of fd.
0.0390. Rate was declining at 25 percent of thi
finai rate over eackt IOgIo cycle; e.g.$ st The slope and position of the line for con-
td = 6250, qd = 1.25(0.0390) = 0.0488 and at stant rate agrees exactly with the predictions
of Eq. (8); for rate declining at 50 percent per
= 625, qd = 1.5(0.0390) = 0.0585.
‘d cycle, the slope ia still essentially correct,
but the line is shifted upward somewhat. This
Note that pwd2 does not plot as a straight
means that a small error would result in esti-
line against td; thus, in this case, it is not mating skin factor from such data. For rate de-
clining at 200 percent per cycle, the data for
possible to estimate the permeability-thickness
t > 104 deviate noticeably from the desired
product from the plot of pwd2 vs. td. Note, how
s1!raight line, but even here, the Slcwe at earl-
; eht ~~~-e
ever, that fd does plot ss a straAa... ier times, and thus permeability estimates, are
of acceptable accuracy. In any event, results
against td; further, the slope of this line is
for larger magnitudes of decline in rate are pro-
1.15, which is preciseiy the slope thzt shculd b bahIY of more academic than practical value -- if
observed if Winestock and Colpitts method were should usually be possibly to keep rate changes
exact, as shown by writing Eq. (2) in dimension- within the 50 percent decline rate curve, and for
less form: this decline, normalizeddrawdown test piotting
appears to be of high accuracy.
1. Manual of Back PPessures Testing of Gas The analysis procedure for the buildup and
Wells, Interstate Oil Compact Commission,
drawdown tests closely follows the procedure
Oklahoma City (1962). suggested by Carter, Miller, and Riley3 for con-
stant rate drawdown tests.
SPE 3872 W. JOHN LEE, ROBERT R. HARRELL. and WILLIAM D. McCAIN, JR. 7
kb.- m
. . . . . . . . , ●(A-2) Details of Mathematical Analysis
/0.014 kt \
-~ln . . . . . . ..** (A-3)
[\+Bgcgrw* ,I
Values of pwf and qg(t) used in this eq- This equation was solved with the initial and
uation must be at a time, t, at which wellbore boundary conditions
storage effects have ceased and boundary effect:
have not yet appeared.
p=piforallratt=o. . . . . . ..(B-2
Skin factor can also be calculated from a
buildup test, of course, but one must be cau-
tious: The equation ordinarily used to calculate qg = Oor~=Oatr=refort>O . ..(B-3
skin factor for a buildup test assumes 6 ~ ~;
the relationship between the apparent skin fac-
tor~ S’, as calculated from.? buildup test and qe0. = q,(t) Orqg(t) =* ~
the true skin factor, S, isiL ii
atr= rwfort>O . . . . . . . . . . . .(B-4
8 EVALUATION OF A GAS 1 .L TESTING METHOD SPE 387
apd2
—= O aty = ln(re/rw) for t>O . . (B-7)
ay
L
apd2 + )i+l,n
(Pd2)i-1 n-2(~d2)i,n+(pd2
—=qdaty=Ofort>O . . . . . (B-8)
(AY)2
ay 1
(Pd2)i_ln+l - 2(pd2)i n+~+(pd2)i+~n+~ For the quantity (pd)i n+l,2 we used the ap-
proximation 9
(AY)2
=
~2(i-l)Ay
(pd2)in+~- (pd2)in (pd)i,n+~/2 = ~ (pd)’~,n+~+ (Pd)i,n
[ 1
(pd)i,n Atd . . ...0 .* (B-1:
‘)
. (B-9)
(pd)’i,n+~ is our first approximation to
● ✎ ✎✎✎☛ .0.0
where
I I I I I I I I I I /
I I I
20 50 100 200
TIME (hours)
I 1/ kh(convontionol) 39 ~ 118md. ft.
I
10 20 50 I 00 200
TIME (hours)
I-
d---
z-
R
0
i=
v
z
?
DIMENSIONLESS TIME, td
2
N.
.0
103 10’ 10=
DIMENSIONLESS TIME, td
oCONSTANT RATE
VDECLINING RATE:50 PERCENT/CYCLE
A DECLINING RATE: 100 PERCENT/CYCLE
—IJDECLINING RATE:200pE RcENT/cycLE
I ! 1!11 I I ! II I I I Ii
1(32
.- 1(33 I 04 105
‘DIMENSI(JNLESSTIME, td -