. .

SPE 15028
SPE
Variable-Rate Reservoir Limits Testing
by T,A. Blasingame and W.J. Lee, Texas A&M U.
SPE Members

Copyright 1986, Society of Petroleum Engineere

This paper wae prepared for presentation at the Permian Basin Oil& Gas Recovery Conference of ttre Society of Petroleum Engineers held in Midland,
TX, March 13-14, 1988.

This paper was selected for presentation by an SPE Program Committee following raview of information contained in an abstract submittad by the
author(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the
author(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Papers
presented at SPE meetings are subject 10 pubfica!ion review by Editorial Committees of the Society of petroleum Engineers. permission to coPY. is
restricted to an abstract of not more thdn 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment of where
and by whom the paper ie presented. Write Publications Manager, SPE, P.O. Box 833838, Richardson, TX 75083-3838. Telex, 730989, SPEDAL.

ADSTRACT focus on a particular rate scheme, we develop a

general varisble-rate approximation that should
This paper presenta a new method of estimating give accurate rasults for typical production
drainage area size and ahape from production data situations,
(bottom-holepressures and flowrate). The method
fa a rigorouslyderived approximationfor variable- Without a variable-rate’solution we would have
rate flow in a closed reservoir. This method uae the more tedious material balance methods that
requires a graph of Aplqm va. the superposition require average reservoir presaurea to estimate
plotting function (which is easily calculated by reservoir pore volume. This would require the well
hand). The slope and intercept of the graph are to be shut-in, which results in Ioat revenue.
used to provide the desired estimatea of drainage However, with the new method, the reservoir pore
area size and shape. volume and shape can be estimated directly from the
production data, without shutting-inthe well.
The method that we propose ia an approxima-
tion, however it has baen proved to be very The problem of variable-rate flow in bounded
accurate for the constant rate, constant preaaure, systems is51imitedin the literaturet$~ work by
exponential rate, logarithmic rate, hyperbolic Earlougher and the “stabilized flow” ‘ methods
rate, sinusoidal rate, and discrete rate casea. (which uae average reservoir preseurea). Though
The method also gives acceptableresults for square both approached give acceptable results for their
wave rate and random rate caaea. specific application, Earlougher’a caae is not
realt5tic and the “stabilized flow” methods again
The new method Is derived for the time after require the well to be shut-in for average
the initial pressure transient has reached the reservoir pressure determinations. This suggeets
outer boundary. The changea in flowrate cauae the need for a general solution for variable-rate
additional transients, but we aaaume that this flow in a bounded reservoir.
effect ia negligiblewhen compared to the influence
of the outer boundary. Therefore, if the change in In the “Descriptionof the New Method” section
flowrate does not dominate the influence of the we will present the general variable-ratesolution
outer boundary, the new method should give accept- and the reservoir characteristics which can be
aisle results. Also, at present, this method ia derived from it. Also, we will verify the general
only derived for single-phaseflow of a liquid of variable-rate equation (Eq.(2)) using analytical
small and constant compreaaibility. and finite-difference emulation. Then a
step-by-stepprocedure for applying our method and
INTRODUCTION a complete example will be ahown in the “Method of
Application”section. Finally, we will present the
The purpose of this paper is to present a derivation of the exact solution for variable-rate
simple, but accurate meti-tod
of predicting reservoir flow, in a bounded circular feseivoir and the
drainage area size and shape f-~m variable-rate approximate solution for variable-rateflow in any
productiondata. Previous works have dealt with shape reservoir in the Appendix of this report.
constant or cyclically constant rate and constant
bottom-hole pressure production. ii summary of DESCRIPTIONOF THE NEW NETHOD
;~t:ue:r~ds is shown graphically in Figure 1.
defined the cyclically constant or In the Appendix, we derive an equation
square wave rate case in FLgure 2. Rather than describing the behavior of the bottom-hole
pressure, p f, as a function of time for a single
well produc%g in a bounded reeervoirwith a

Referencesand illustrationsat end of paper;

cvli
““ ,
 Variable-RateReser ir Limits Teatinz SPE 150U

varieb~e rate. This solution assumes the Eq. 2 suggests that a graph of Ap/qm vs. ~ will be
following: a straight line of slope

Radisl flow into the well over the net pay

m = 0.2339 — $.......**o**.,$*.Q..*.(4)
thickness; vr $h;tA
Homogeneousand isotropicporous medium;
Uniform net pay thickness;
and intercept
Porostty and permeabilityconstant
(independentof pressure);
Fluid of small and constant compressibility; b = 70.6 *ln 4A ‘ . .........*..(5)
vr
Constant fluid viscosity; eyC r
Aw
Small pressure gradienta;
Negligible gravity forces; and
will result. Since we know that transient snd
Any rate schedule.
stabilized flow alternate for each rate change, we
can suggest that for a small rate change stabilized
As shown in the Appendix, the complete
flow will dominate. However for large rate changes
solution for a circular drainage area is
the infinite series in Eq. (1) (i.e., tranaient
r flow) dominatea and Eq. (2) la not valfd. This
means that transient data will not lie on the
Ap = Pi - pwf= 141.2 *[qm [ln~- ~ +
w straight line predicted by Eq. (2). Therefore,
these outlying points should not be used when
obtaining the slope and tnterceptof the graph.
rw’
—2+ S] + 2 m(O.0002637)* Qm Because Eq. (2) is an approximation and we
2re t know that stabilized flow will not be achieved for
all rate changea, we must fnveatigate empirically
Xn rw the accuracy and applicability of Eq. (2). Our
investigationused the complete solution (Eq. (1))
~ Jo( re ) and a fintte-di.fference reservoir simulator.
-2:(qj- ‘9$
qj-1) n=l finite-differenceformulationwaa fully implicit.
j=l
Xn 2J0 ‘(Xn) The simulator modeled radial, single-phase flow of
a single-phaae Liquid of amsll and constant com-
pressibility. The simulator was
so~:;l!$ ~
EXP(-Xn2nt0.0002637
@wtA
#-(t-t j-1)) ] ..... ...(1) comparison with the analytical
transient and paeudosteady-stateflow in a bounded
circular reservoir at a constant producing rate.
If we assume that the infinite series in eq.
(1) is negltg.iblethen we call this “stabilized Earlougher5 developed a method to determine
flow”. Though each rate change introduces a new resemoir drainage area for wells with rate
hiatorfea that are cyclically constant or square
transient that keeps the well from reaching true
pseudosteady-state flow, this does not keep the wave, as shown in Figure 2. He proved that the
well from exhibitinga pseudosteady-state-ltke flow alope of a p f vs. time graph provided an estimate
regime that we call “stabilized flow”. Also, we of pore voi%me. However, he did not give an
use the term ‘istartof stabilized flow” for the explicit method for estimating reservoir drainage
shape for this case. Earlougherused superposition
time to reach paeudosteady state for a constant
rate. This gives us the lower bound for the atart to generate this data, therefore, to verify our
of stabilized flow. However, a large rate change method for the situation he studied we reproduced
will cause the infinite seriee to dominate eq. (1) hia examplea using a aimilsr approach. The mein
and stabilized flow will not exist, though after differencebetween our examples and Earlougher’sis
some time the infinite series will become neglig- that he simulated a well centered in a square
ible and stabilized flow will again exist. What reservoirand we simulated a well in the center of
this means ia that each new transient will a circular reservoir. This difference does not
eventually die off and stabilized flow will domi- distort the interpretation or comparison 0.?
nate, so there will be alternating periods of both reaulta.
tranaient and stabilized flow. Therefore; if we
neglect the infinite series in eq. (1) (i.e.,
assume atabili.zedflow) and ge~$ralizethe drainage “Table 2 shows the reservoirdata for the cases
area shape by using the Dietz comparing reaulta from the new method with results
the following variable-rate approximation can ‘k:
‘hape ‘actor’ using Earlougher’smethod. Figure 3 is a graph of
written for stabilizedflow: rate vs. time for the three caaes that were
simulated using Eq. (l). The first case ia the
square wave rate caae, which aervea as the basis
*). o,+ln ~ 4A ~ + 0.2339 *E ,(2) for the Earlougher analyaia technique. The second
qm t caae is a sinusoidal rate caae and though
e cArw
Earlougher did not model this exact case, he did
where suggest that hia method was applicable to seasonal
rate chsnges. This case ahowa the rate demand
schedule for approximatelyone year. The third and
7
~=lqcytj-l) .- ‘
final caae in this compariaop is for a completely
t= randomrate-schedule from a single well.-
Qm/~ i . ...,,...,.....(3)
%n
“ .

-...1
IF. wvfa
-- T. A. Blasin~ame & W. J. Lee 3

Earlougher gave a completely random rate case in the graph. A least-squarea fit gives
ais work, though it could not be realistically more accurate results than “eyeball”
wialyzedusing his method. fits.

The conventional. p f vs. time plot is ahown in 5. Estimate reservoir size from the slope,
Figure 4 and we see tha% only the square wave rate mvr, of the graph using Eq. (4).
case can realistically be analyzed using
Earlougher’s me~hod. Figure 5 showa our method 6. Estimate the reservoir ahape factor, C
with Aplqm va. t. Note that the square weve rate from the intercept,b , of the graph an#
case gives two straight lines of the same slope and Eq. (5). Skin fa~?or and reservoir
two intercepts. Since there is no theory for permeability must be known (from a
interpreting the two intercepts we simply took an pressure buildup teat, for example) to
8rithmsticaverage. The sinusoidaland random rate make this estimate.
cases essentially line up on the same trend, which
ia what we expect since all results are for the Table 1 is a summary of plotting and analysis
same reservoir. Table:3 shows the relative error techniques for reservoir limits tests. Figure 1
in slope and intercept “for each of these three ahowa the type of graph for esch caae. In each
cases. The slope fs directly proportional to pore caae, the slope provides an estimate of drainage
volume while the intercept is exponentially area size, A, and the intercept provides an
proportional to the shape factor, C . Slopes and eatimate provides an estimate of shape factor, C .
intercepts were determined from the %est straight- A numerical value of shape factor allows us 40
line in a least squares sense paaaing through the estimete drainage area shape by uatng a table of
linear portion of the data in each caae. Table 3 shapa fac~~rs and shapes, such as the one presented
showa our method is more accurate than Earlougher’s by Dietz. The new method for analyzing variable-
for determining slope and intercept, and thus for rate production includes the other methods as
determining drainage area size and shape from special cases, similar results will be obtained
variable-rateproductiondata. whether the new general method or the older methods
for special cases are used.”
klenext investigatedthe effect of the type of
rate decline on the reaulta obtained using our new EXAMPLE - Random Rate Production
method. We modeled aix cases With our
fintte-difference simulator. Table 4 glvea the In this example we simulated a random rate
reservoir properties used in the simulations. The decline in a bounded circular reservoir with the
typea of rate decline scheduleswere constant rate, analytical.solution. This is the seinecase shown
logarithmic decline, exponential decline, earlier in Figures 3, 4 and 5. Also, the perttnent
hyperbolic decline, discrete (stair-step)decline, reservoir data is given in Table 2. The production
and the rate schedule resulting from constant data is given in Table 6.
bottom-hole pressure production. Figure 6 shows
rate vs. time for each case i?hileFigure 7 shows Ustng least squares on the straight-line
pwf vs. time for each of these caaes. Only the portion of Figure 9, the following slope and
constant rate case can be snalyzed using Figure 7. interceptwere obtained,
Figure 8 stiowaAptqm VS. ~. Note that virtually
all of the points fall on the same line, and again = ‘5 ailSTBIDlhr
;*::;:;:-lP
this ia expected because all cases are for the ssme ::: = , psi/STB/D
reservoir data. Error analysis resulte are given
In Table 5. This table shows that the new method The reservoir drainsge area, A, Is estimated
had small error in all cases, which suggests that from Eq. 4,
the introduction of transients due to changes in
rate has llttle effect on the stabilized flow
‘m
Vr =’0.2339 —
solution (Eq. (2)). +h~tA

APPLICATIONOF METHOD therefore,

We suggest the following procedure to obtain 0.2339 —

A=
the best reaulta from our variable-rate test +hc~mvr
analysis technique.
(1.0) .
1. Meaaure or estimate both bottom-hole = 0.2339
pressures and flowratea as functions of (0.15)(100)(5x10-6)(1.269x10-5)
time.

2. Calculate ~PIqm and the plotttng

function,~ (Eq. (3)). = 2.458x108 ft2

3. Plot Apl~ VS. ? on Carteaian coordinate

graph paper.
= 5643 acres
4. Determine the slope, mvr, and the inter-
cept, bvr, of.the best straight line on
—

Qa*
 . .

Variable-RateReservoir Limits Testin!z SPE 15028

The reservoir shape factor, CA, is estimated consta~t pressure caae, STB/Day
from Eq. 5, (std m Id)

4A b=
b cr Pi - 70.6 ~ln 4A
2’
vr “ 7Q”6#lneYc ~ 2 Y
Aw e cArw
interceptof pwf vs. t Grsph for
therefore, constant rate case, psia (kPa)

4A . 70.6~ln 4A z , intercept
b
CA = b kh vr Y
eCr
eyr 2 EXP(*) Aw
w . of AP/~ VS. ~ graph for general
v~riable-ratecase, psi/STB/D (kPa/std
= 4 (2.458x108) m /d)
2
(2.542x10-1)(100)(100)]
‘1”781)(0”5)EXP( To 6(1 0)(2 o, B= Liquid3format~n volume factor, RB/STB
. . .
(ream /atdm)
= 33.46 Reservoir shape factor, dimensionless
CA CA =
Our estimate for the reservoir shape factor, Pore space compressibility,pai‘1 (kPs-l)
Cf =
CA, iS slightly high, this is becauae if we
exponentiate a small error in the intercept, bvr, Gas compressibility,psi‘1 (kPa-l)
it becomes a large error in the reeervoir shape c&? =
factor, CA. Therefore, a better comparison would co = Oil compressibility,psi-l(kPa-l)
be that of the input and calculated intercepts.
This is shown in the summary table below. Cs + Coso + CWSW + Cf, total
Ct = ~g
summary of Results compressibility,psi‘1 (kPa-l)

Variable Input This Work % Error c= Water compressibility,psi-l(kPa-l)

w
A 5760 acres 5643 -2.30x10~ h= Net pay thickness,ft (m)
31.62-I 33.46 5.83x10-I
:A 2.542x10-1 -4.27x1O
2.553x1O .
vr J ~ ~ productivityindex~ STBIDIPSi
SUMMARY AND CONCLUSIONS P - Pwf (std m3/dlkPa)

we have de~eloped a method of determining Jo = Zero order Bessel function of the first
drainage area size and ahape for wells with kind
variable-rateproductionhistories. This method is
an approximation that givea very good results for First order Bessel function of the first
‘1 =
cases where the rate changea are small. However, kind
large rate changea only dominate the stabilized
flow solution (Eq. (2)) until the tranaient intro- k= Effective formstionpermeability,md
duced by that rate change becomes negligible.
Therefore, the new methad should be considered k
m= ‘0”001439@pctAln 4A ‘
accurate for any production period so long aa the Cp
reservoir outer boundary ia being felt by the Y2
pressure response. e cArw
slope of log ~ va. t Graph for constant
The condition of the resrvoir can aleo pressure case, cycleihour (cyclelhr)
prescribe the applicability of the new method in
that if any of the assumptions concerning the m=
reservoir are violated, the method may not work. cr -0”2339 *’ ‘lOpe ‘f ‘wfva” t ‘raph
Specifically, if water influx, solution gas for constant rate case, pailhour (kpa/hr)
evolution, multi-phase flow, or reservoir heter-
ogeneitiea exist then this method should not be
used. -0.2339 k, slope of Pwf VS. t graph
‘Sw = t
NOMENCLATURE
for square wave rate case, psilhour
(kPa/hr)
A= Reservoir drainage area, ft2 (rn2.)
m= , slope of Aplqm vs. E
vr 0“2339 ~hicA
kh (pi-pwf)
b . graph, pai/STB/D/tm) (kPa/atdm3 /D/hr)
Cp 70.6 BVln 4A
Y2 Ap = pi- pwf, pressure drop, psi (kPa)
e cArw
intercept of log”q#s. t graph for .—

!lM
-- .
 . .

QPE
.,.”
1K07R
. ----
T. A. Blasin!aame
–.––
& W. J. Lee .

;=
Average reservoir pressure fOr outer 4= Porosity, Fraction
boundary of reservoir,re, psia (kPa)
Integer Subscripts
kh
- P“f)s
PJ) = 141.2 qBP (Pi j= Rate counter
dimensionlesspressure
m= Number of rates up to time t
Pi “ Original formationpressure, psia (kPa)
n= Infinite series counter in Musket’s equation
Pwf = Flowing bottom-holepressure, psia (kPa)
Other Subscripts
Pr = Pressure at radius r from the center of
the well, psia (kPa) Cp = Constant pressure caae

Fr = Average reservoirpreaaure for radius r cr = Constant rate case

from the center of the well, psia (kPa)
Sw = Square wave rate case
q“ Liquid flowrate, STB/D (std m3/d)
vr = General variable-ratecaae
?= Average liqufd flowrate3forsquare wave
rate case, STB/D (atd m Id) REFERENCES

Liquid flowrate at3tirnat, variable-rate 1. Jones, P.: “Reservoir Limits Tests, Oil and
%=
case, STB/D (std m /d) Gaa J. (June 18, 1956) 184.

Cumulative liquid production at ~ime t, 2. Jones, L.G.: “Reservoir Reserve Tests,” ~

Qm= Pet. Tech. (March 1963) 333-337.
variable-ratecase, STB/D (atd m /d)

r = Distance from center of the well, ft (m) 3. Matthews, C.S. and Russell, D.G.: Pressure
Buildup and Flow Testa in Wells, Monograph
rfre, dimensionlessdistance Series, Society of Petroleum Engineera,Dallas
‘D = (1967) 1.
r . Drainage radiua of the well, ft (m)
e 4. Earlougher, R.C., Jr.: “Estimating Drainage
r = Wellbore radius, ft (m) Shapes from Reservoir Limits Tests,” J. Pet.
w Tech. (Oct. 1971) 1266-1268.
-s
r’= rwe , effectivewellbore radius, ft (m)
w 5. Earlougher, R.C., Jr.: “Variable Flow Rate
s . Skin factor, dimensionless ReservoirLimits Testing,” J. Pet. Tech. (Dec.
1972) 1423-1430.
s . Gas saturation,fraction
g 6. Cox, D. O.: “Reservoir Limit Testing Using
so = Oil saturation,fraction Production Data,” Log Analyat (March-April
1978) 13-17.
Sw = Water saturation,fraction
7. Ehlig-Economidea,C. A. and Ramey, H. J., Jr.:
t . Flowing time, hr “Tranaient Rate Decline Analys%a for Wells
Produced at Constant Pfessure,” Sot. Pet. Eng.
t! = Dummy variable of integration,hr J. (Feb. 1981) 98-104.
—
8. Ehlig-Economides,C. A. and Ramey, H.J., Jr.:
{tq(t’)dt’ “Pressure Buildup for Wells Produced at a
z= , plotting function for
qm Constant Pressure,” Sot. Pet. Eng. J. (Feb.
1981) 105-114.
variable-rateteats, hr
9. Fetkovich, M. J.: “The Isochronal Testing of
0.0002637 ‘t , dimenaionleaatime Oil Wells,” paper 4529 presented at the 1973
‘D = ~~trw2 SPE Annual Technical Meeting and Exhibition,
Las Vegas, Nev., Sept. 30-Ott. 3, 1973.

0.0002637 — , dimensionlesstime 10. Muakat, M.: Physical Principles of oil

‘DA = @~A
baaed on drainage Production, McGraw-Hill Book Co., Inc., New
area York (1949).
Vr = n(r2-rw2)~h, radial pore volume, ft3(m3)
11. Dfetz, D. N!: “Determination of Average
Xn = Root of first order Bessel function of Reservoir Pressure From Buildup Surveys,” ~
the first kind (i.e., J1(xn) = O) Pet. Tech. (Aug. 1965) 955-959.

Y“ 0.577216, Euler’s constant 12. Aziz, K. and Settari, A.: Petroleum Reservoir
Simulation, Applied Science Publiahera, — New
P= Liquid viscosity, cp (Pas) York (1979). ‘-
QGG
..”
 . ●

Variable-RateReservoir Limits Testing SPE 15028

13. Musket, M.: The Flow of Homogeneous Fluids = 141.2 *[q1 PD(t) +
Through Porous Media, McGraw-Hill Book Co., Pi - Pr
Inc., New York (1937).
(q* - ql) pD(t - t~) + (q3 - q2) pD(t-t2)+..](A-3)
~!Effectof oil Production
14. Templaar-Lietz,W.:
Rate on Performance of Wells Producing from Eq. (A-3) suggests a general relation for n rate
More Than One Horizon,” Sot. Pet. Eng. J. changes. Writing this general relation in sumsa-
(March 1961) 26-31. tion notation gives
15. Collins, R. E.: Flow of Fluids through Porous
Materials, Reinhold Publishing Corp., New
- Pi - Pr = 141.2 W;(q
kh ~=1 ~ - qj-l) PD(t - tj-l)(A-4)

16. van Everditigen,A. F. and Hurst, W.: “The Eq. (A-4) is s general equation which models
Application of the LaPlace Transformation to arbitrary rate changea in a producing well,
Flow Problems in Reservoirs,” Trans., AIME specific flow regimes such as tranaient or
(1949) 186, 305,324. pseudoateady-atate can be modeled with the
appropriate pD(t) function in Eq. (A-4). At this
17. Matthewa, C. S., Brons, F., and Hazebroek, P.: point we will develop a relation for variable-rate
“A Method for Determfnatlon of Average flow in a bounded reaervotr,using Eq. (A-4). This
Preaaure in a Bounded Reservoir,” Trans., AIME requirea that we know pD(t) for the specific
(1954) 201, 182-191. eometry modeled. Matth~~s, Brons, and
;::yo~l?
and Earlougher, ——
et al. give methods
18. Earlougher, R. C., Jr., Ramey, H. J., Jr., of determining PD(t) for varioua reservoir
Miller, F. G., and Mueller, T. D.: “Pressure geometries. However, the most convenient geometry
Distributions in Rectangular Reservoira,” ~ ia a w 1 centered in a bounded ciruclar reservoir.
Pet. Tech. (Feb. 1968) 199-208; Trans., AIME, MuakatfJ give this solution as
Vol. 243.
2
19. Odeh, A. S. and Jones, L. G.: “Preaaure ‘D + s + 21Tt
Drawdown Analyais, Variable-Rate Caae,” ~ pD(rD? ~) = ‘ln ‘D ‘~ ‘~ DA
Pet. Tech. (Aug. 1965) 960-964; Trans., AIME,
234. a Jo (xnrD) EXp(-Xn2TtDA) .
-2X ...(A-5)
20. Wineatock, A. G., and Colpltta, G. P.: n=l Xn2 Jo2(Xn)
“Advancea in Estimating Gaa Well Deliver-
ability,” J. Cdn. Pet. Tech. (July-Sept.1965)
where
111-119.
(A-6)
$...****.*.-.*•.***~***.**~
21. Ramey, H. J., Jr., and, Cobb, W. M.: “A ‘D = ‘fre
General Pressure Buildup Theory for a Well in
Xn are the positive roots of Jl(Xn) = O .(A-7)
a Closed Drainsge Area,” J. Pet. Tech, (Dec.
1971) 1493-1505,
= 0.0002637~ ,..............(A-B)
‘DA
APPENDIX $BCtA

In this appendix, we will derive the general

A = me2 ............................(A-9)
variable-rate equation for stabilized flow in a
bounded circular reservoir. We will then use the
Combining Eqna. A-5, A-6, A-8, and A-9, and letting
~t;~f reservoir pressur~4approach presented by
r . rw givea
and Templaar-L5etz to derive a variable-
r r
rate equation for stabilized flow in any bounded 3
reservoir. pD(rwtt) “ ln< ‘~ ‘fi+s
e
When Duhamel’s theorem is applied to the
constant rate solution
‘low rete, the reau~t ,s~,l% cOntinuOuslY changing + 2n(0”0002637) @_k~tA

dpD(t-t’)
Xn rw
Pi - Pr ‘p {tq(t’) dt dt’.(A-1)
= 141’2= (y)
COJ kt
o EXP(-Xn2 m(o.oo02637)$pctA
—)
@@y@ the convolution theorem to Eq. (A-1) ‘2n~lx2J;(X)
gives non . ..................(A-1O)

~ ft pD(t-t’) ~dt’..(2)2) Combintng Eqna. A-4 and A-10 gives

Pf-Pr= 141”2 kh O

To model discrete
(A-2) the result i~%te Changea’ ‘e ‘iscretize ‘q” ‘p = “ - ‘Wf = 141”2~ ‘~[’n;-~ +
-.

Qcc!
.

SPE 1 5028 T. A. Blasineame & W. J. Lee 7

2 definition of dimensionlesspressure we obtsin the

r
w + S] + 2 IT(O.0002637)
* Qm following equstion relating pressure and time for
t constant rate flow:
2re2
r
Xn rw Pr=P~- 141.2*[ln $. $+ —
r2+s
Jo (— 2re2
re )
-2 f (q - qj-l) f
j=l j n=l x 2J 2
(Xn) Xr
no
Jo (~)
kt ?
Exp(-X~~(O.0002637)~(t-t ~-1))1 $.........(A-11) + 2“(0”0002637)$PCtA - 2 ~=1 X2;
@DctA Jo (Xn)
n

where EXF&Xn2n(0.0002637)& t)]. ..............(A-15)

.
1.

% = tw’) ‘t’ =&j% -tj-l)”*’”””””*(A-12) Solving Eq. (A-n) for p= (i.e., rw = r) gives
sn equation relatinz -. pressure and tine for
~A;.~l now develop a plotting function fr~~ variable-rateflow:
Eq. - . Following the Odeh and Jones
derivation of the plotting function for r
2
variable-rate transient flow, we divide both side: Pr = Pf -141.2 ~kh[qm[ln<-~+ —+s]
of Eq. (A-n) by th final flow rate, qm. Letting t 2re2
= Qm/qyields

+ 2Tr(0.0002637)
~
r r; @pctA ‘m
Q= u1n:-3+—
141.2 kh[ + s]
4
% w 2re2
Xr
Jo (~)
: -2jYl(qj - qj-l) :=1 ~ 2J 2:X )
+ 0“2339 @h~tA .
non
Xnrlq
Jo (~) EXP(-Xn21r(0.0002637)
*(t-t j-l))1(A-16)
t
-282.4 .-#j:l
‘v (q~ - ‘j-l) ~
% n=l X
n 2Joe2(xn) l’akingthe time derivative of each term in Eq.
I (A-15) Yields.
EXp(-Xn271(0.0002637)&(t-tj-1)) ~,........(A-I3) Xnr
t 3P . Jo(~e )
& Cr “ -0.2339* [1+ nil
For a usable plotting function to be t J02(Xn)
developed, the infinite series in Eq. (A-13) must
be negligible. This is true for the constant rate
2
case at pseudoateady-atate and was shown to be ‘t )1 .............(A-17)
‘Xp(-xn ‘(0”0002637)$pctA
approximately true for the variable-rate caae at
bounded reservoir flow conditions using simulated
examplea in this paper. Therefore-, we can ap
approximate Eq. (A-13) by neglecting the infinite %B
&)vr = -0.2339$-[ 1+ ~=1 ‘qj ‘qj.l) nil
series. t
r rw 2
$! z 141.2#[ln~-~+ —2+ s]
r
m w 2r Jo(y) 2
e -+t-tj-l))]
‘Xp(-xn ‘(0”0002637)$pccA
Jo2(Xn) . ...*...(18)8)
+ 0.2339 ~hftA ~ . ...........(A-14)
When stabilized flow occurs (i.e., when the
bounded reservoir terms dominate the tranaient
Eq. (A-14) is the variable-rate approximation
terms) the infinite ‘series in Eqns. (A-17) and
for stabilized flow for a well centered in a
(A-18) become negligible. Therefore, the two
bounded ctrcular reservoir.
derivativesbecome
We will now use Dietz’~’ approach for
constant rate flow in a bounded reservoir of ap
-) cr = -0.2339* ,......19)........(A-l9)
general shape. This requires 8Pr/8t for both the
at t
-constant--rateand the--variable-rate-caaes.. If.we..
combine Eqna. (A-5), (A-6), (A-8), (4--9),and the
,7
*W{
 . .

ap where
qmB
=) -0.2339 w ,................(A-2O)
at vr = Vr = m(r2 - rw2) @h . ...............(A-30)

Only the rates in Eqna. (A-19) and (A-20) dVr= 2Trr@hdr . ...............(A-31)
differ. Note that there fs no dependence on shape
for either equation. Combining Eqns. (A-29) - (A-3A) and solving
for the average reservoir pressure, pr,
Dietzll developed an average reservoir
pressure relation and a reservoir shape relation
for constant rate flow based on Eq. (A-19). We cr=Tl(r2-r1W2) +h [2T@hpwf {rrdr
will.now derive similar relations for variable-rate w
flow using Eq. (A-20). The general requation of
radial flf~ or the diffusivity equation was given
by Muskat as ~B!J r ?-r2
+ 2m0h(141.2) ~ fr(ln$ - w )r dr]
4N.ICt w w 2r 2
Q(+) . 0.0002637k apr

~
. ,..........(A-21)
e. . . . (A-32)

Completing the integrationin Eq. (A-32),

Combining Eqns. (A-20) nd (A-21) yields
2(141.2) qmBIJ 2
ap ir = Pwf + [ ~ln~
a(r_
F a; ) s -28,.4 ~ . ...........(A-22) (r2-rw2)kh w
khre2

(r2-rw2) (r2- r 2,2

IntegratingEq. (A-22) yields — - owl ● . . . . . . . . . . . . . . (A-33)
4 8re’
ap qmB P
c1
-141.2 — r+ ..........(A-23)
+= T*
khre2 Simplifying Eq. (A-33),

qmBP
Applying the outer boundary condition, pwf + 141.2~[ A1n3-+
Fr =
(r2-rw2) ‘w
apr
— = O at r=r . ..................(A-24)
at e
(r2-rw2)
Combining eqns (A-23) and (A-24) Q 1. ........................(A-34)
4reL
apr
~41 z qmB~ ~ 1
. .........(A-25)
r“ kh –-~
r “ If we let r = re and note that re2 >>r W* Eq.
e (A-34) becomes
qmBp
IntegratingEq. (A-25) yields F= pwf + 141.2~ [ln~ -~] . .....(A-35)
w
qmB u
2
—(lnr -— Eq. (A-35) givea us a relation for the average
‘r = 141’2 kh
2r2) + C2 ‘ ““”””””(A-26) reser~fir pressure, p, du;ing stabilized flow.
e
Dietz gave the followi~g :elation for the average
reservoir pressure, p, for constant rate
Applying the inner boundary condition,
pseudosteady-ststeflow
Pr = Pwf at r= r . ....................27)27)
w qBV r
Combining Eqna. (A-26) and (A-27) ; = pwf+141.2~[ln~- +1 . .......(A-36)
w

Note that if we solve for the productivity

~Bp r ‘2 - ‘W2 ~ ~A-281
Pr = pwf + 141.2= (ln~- index, J, using Eqna. (A-35) and (A-36) we obtain
W 2re2 “ the following equality

Introducingthe average reservoir pressure, ;=, for +).s kh

J (A-37)
any radiua r, = ~-pwf cr p-pwf)vr r
141.2Bp[lnS- ~]
r 6
;r = +r {w Pr dvr . ...............(A-29)
Eq. (A-37) proves that the productivity
indices for constant rate pseudosteady-atateand
variable-rate stabilized flow are equal. Now we
will-uaeEq.- (A-35)-and-theapproxitite----
.

u
A.
l~n9n
..”.”
T.
---- A.”
. Blasinzame& W. J. Lee
.“-
9

20
variable-ratetransient flow relation to develop 2.246 A
. .................(A-44a)
a general reservoir shape factor, CA” 26he CA = bvrkh
approximatevariable-ratetransfent solution is EXP( ) rw2
70,6 BU
qmBB Eq. (A-35) can also be put in a
. ........(A-38) general
Pwf = Pi - 70-6 kh ln(~t) form by use of the shape factor, CA.
ey D

Combining Eqr.s. (A-35) and (A-38) for Pi-P ~ = Pf+ qmB1’

70.6Tln 4A z . .,.....45)45)
yields w
eyC r
Aw
qmBv
Pi -; — in (CA tDA). ..........(A-39) And finally, we will expreas Eq. (A-42) In
= 70”6 kh
dimensionlessvariables as
mere the reservoir shape factor, CA, is defined
for a circular reservoir as P~ =~ln ~4A 2+2ni ...............(A-46)
DA “
eCr
Aw
3/2
. 4re
. .......................(A-40)
CA eY Where ~ is substituted for t in tDA. Eq.
(A-47) is the same as the const
presented by Ramey snd Cobb~Pt ~~~~p~”’{ti~
Slp -~=o, so-pi = ~ UP to this
substitutedfor t.
~~. cAt&$ref;r~, for Pif > lICA. This
‘DA
meana that, for tDA >1/c , p ca~be predicted using
Eq. (A-35). Also, sif?ce the constant rate and
variable-rate solutions are essentially the same
then the shape factors predicted for constant rate
bare also valid for variabl.e-rsteflow. Therefore,
we can generalize Eq. (A-14) for other shapes by
using the Deitz shape factor, CA.

If we neglect r 212r 2 in Eq. (A-14) and uae

the effective wellbo~e r~dius to model the skin
effect, we obtain,

2
r
= 70.6~ln *,2 + 0.2339 ~ .(*-41)
~h~tA
E re
w

Now we wfll include the reaervofr shape factor, C ,

relation (eq. (A-JO)) in the argument of tie
natural logartthm:

l@z706BP
● @n - ‘ 0“2339 * ; “‘A-42)
%
Aw

Eq. (A-42) is the general stabilized flow

equation which servea as the basia f~r our analysis
technique. If we plot pl~ vs. t on Cartesian
coordinate graph paper, Eq. (A-42) dictatea the
following slope and intercept.

m = 0.2339 ~ . . ............. .(A-43)

●
vr

BP In , .................(A-44)
b = 70.6= Y4A2
vr
e cArw

Solving Eq. (A-43) for the reservoir drainage

area, A, yields

A = 0“2339 ~h~tmvr “ “*”””’”””””””””*”*””(A-43a)

Solving Eq. (A-44) for the reservoir ahape

-factor,CA, .ytelda.. —
$aP.a
“..
 TASLE I TABLE 2

SWMNARYOF RESERVOIR LI!SITS TEST AND .4NALYSI S TECHNIQUES sYSTEN PROPERTI H FOR CASES SIN1OATED ANALYTICALLY

TEST METHOD Geometry Single Well Centered in a

Bounded Circular Reservoir
Constant Rete (pwf VS. t Graph)
Ama, A 5760.0 Acres (9 mf2)

Re8ew0ir drainage radius, re 8937 ft

A - -0.2339 &Bm Net pay thickness, h 100.0 ft

t cr
Wellbore radius, rw 0.5 ft
2.246 A
CA “ kh (Pi - bcr) ~ 2 Reservoir permeability, k 100,0 md
EXP( )W
70.6 q BV
Reservoir porosity, @ 0,15

Total pore volume, thA 3.76x109 ft3

Constant Pressure (log q vs. t Gtaph)
Pluid viscosity, P 2.0 Cp
b -1
B Total compressibility, Ct 5X1 O-6 psi
A - -0,1016 ~
$hct(pi - Puf)
‘“CP
Initial Pressure, pi 2000.0 psia

Formation Volume Factor, B 1.0 R81STB

2.246 A
CA = kh (Pi - Pwf)
EXP(
70.6 bcpBIJ ‘r;
TABI.E 3

ERROR ANALYSIS POR CASES SIMW,ATED ANALYTICALLY

2.246 A
CA “ ~xp( -0.001439 k
tiIcp@ctA ‘r:
New Nethod New Method Earlougher Method
Case Slope Error, Z Intercept Error, !4 Slope Error,?i

square !-save Rate5 (pwf vs. t Cr@ph)

Square wve rate 2.145x10° 2.678x10° 6.529x10°
(8arlougher’s Method)
s Ire Wave Rate 4.276x10-’ -I.243x10-2 >>100%
A- -0.2339 .*
t 8W Xandom rate 2.069%10° -4. 270x10-’ 8.296x101

Not applicable from this method TABLE 4

CA =
sYSTfM 21ROPERTI IX FOR CASES MODELED WITH
THE FINITE- fiJFFERENCE SIMULATOR

General Variable-Rate (AP/qm vs. ; Graph)

Geometry Single Well Centered in a
Bounded Circular Resmwoir
A= -0,2339 ++
t Vr
Area, A 40.0 Acre8

Reservoir drainage redius, r 74s ft

e
2.246 A
CA =
khbvr 2 Net plly thickness, h 30.0 ft
‘w
‘Xp( 70.6 BU )
Wellbore radiua, rw 0.2 ft

Reservoir permeability. k 1.0 md

For alI methods, usc rw’ = rwe
-s for a non-zero skin factor.
Resetwoir porosity. ‘$ 0.30

Total pore volume, @hA 15.7X106 ft3

oil viscosity, P O,k CP

-1
Totril compressibility. Ct 15xlo-f’ psi

Initial pressure, pi 4800.0 psia

“..

TAELE 5
T.4i!LE 6
s3rRoR ANALYSIS FOR CASES MODELEDWITH THE
FINITE-DIFFERENcE SINOLATOR SIMULATED PRODUCTION DATA

~Pl~
Case Slope Error, % Intercept Error, Z Time Flowrate Cum. Prod. Pressure F
(hrs) (STB/D) (STB) (psia) (hrs) (psi/sTB/D)
Constant Rate -1. R97X10-4 -1. 3S8XID-2
0.7200E 03 O. lSOOE 04 0,4500E 05 0.1607E 04 0.7200E 03 0.2619E 00
Logarithmic Rate Decline -9.218x10-3 2.008x10-2 0. lkfIOE 0.4 0. 7000E 03 0,6600E 05 0. 1800E 04 0, 2262E 03 0.2855E 00
0.2160E 04 0. 1900E 04 0. 1230E 06 0. 1480E 04 0. 1553E 04 0,2732E 00
Exponential Rate DeclirIe -3.232 x10-5 9.277 %10-’ 0.2880E 0.4 0, 1200E 04 0, 1590E 06 0. 1644E 04 0.3180E 04 O.2959E 00
0.3600E 04 0.2500E 04 0.2340E 06 0.)294E 04 O.2246E 04 0. 2820E 00
Discrete (Stair-step) 0.4320E 04 0. 4500E 04 0, 3690E 06 0, 7459P, 03 0. 1968E 04 0.2786E 00
0, 5040E 04 0.4000E 03 O.381OE 06 0.1775E 0. O.2286E 05 0. 5616E 00
Rate Decline 9.910X10-3 -1 .706x10-2 O.576OE 04 0.3750E 04 0.4935E 05 0.9022803 0,3158E 04 O.2927E 00
0.6480E 04 0.4000E 04 0,6135E 06 0. 7970E 03 0.3681E 04 0.3007E 00
Hyperbol lC Rate Decline -1 .697x10-2 5 ,263x10-2 0.7200E 04 0.4200E 04 0.7935E 06 0,7077E 03 0,4225E 04 0,3076E 00
0.7920E 04 0. 3900E 04 O.8565E 06 0. 7483E 03 0.5270E 04 0.3209E 00
-4
Constant Pressure -1 .898x1C 2.513 x10-’ 0,8640E 04 0. 1250E 04 0.8940E 06 0. 1408E 04 0.1716E 05 oe4734E 00
0. 936oE 04 0. 1500E 04 0.9390E 06 0. 1336E 04 0.1502E 05 0.4420E 00
O.1OO8E 05 0. 5000E 03 0,9540E 06 0.1585E 04 0.4579E 05 O.8287E 00
0. I080E 05 0. 4000E 03 0,9660E 06 0. 1609E 04 0.5796E 05 O.9766E 00
C ‘152E 05 0.6000E 04 0.1146E 07 0.1392E 03 0.6584E 04 O.31OIE 00
0.1224E 05 0. 5500E 04 0.1311E 07 0. 2050E 03 0,5720E 04 0.3263E 00
0.1296E 05 0.4000E 03 0, 1323E 07 0.1491E 04 0.7938E 05 0. 1270E 00
0. 1368E 05 0. 5200E 04 0.1479E 07 0.2411E 03 O,6826E 04 0.3382E 00
Special Case Results: 0. 1440E 05 0. 5350E 04 0. 1639E 07 0. 1485E 03 0.7354E 04 0,3465E 00
0.1512E 05 0.4500E 04 0.1774E 07 0.3200E 03 0.9464E 04 0.3733E 00
Constant Rste 0. 1584E 05 0, 3400E 04 0.1876E 07 0.5707E 03 0. 1324E 05 0.4203E 00
(Pwf VS. t Graph) 5.128x10-4 5.717%10-3 0, 1656E 05 0.2200E OL 0. 1942E 07 0,8561E 03 0,2119E 05 0, 5199E 00
0.1728E 05 O. 1500E 04 0.1987E 07 O.1O22E 04 0.3180E 05 0.6517E 00
Constant Pressure
(Log ~ vs. t Graph) -9.953 X1 O-1 -7.2 ZIX10-1

b<cr
Pwf

‘cr IIIQ

.
t t
(a) Constant Rate (b) Constant Pressure

b
Pwf clpfqm

.
msw
m
b vr
vr
/
m .
s\$
/

(c) Square-Have Rate (d) General Variable-Rate

Fig. I-lyp[oel reservoir Ilmlts taat performance.

 m

I 10- Start of Stabilized Flow

❑
■

i ● Sinusoidal Rate
E m
● Random Rate 8
-i ~~o~
E ●****9*
Laa$s z AAA ●0 8
●
m
●

●’ ● ‘.
●
m

● 9*

●
● ●
●

I
8
●* “** ‘ ■ ●
‘w, I&t2 I At~=At2
Q
●9
1 I 1 1 1 I t 1 1 1 I I i 1 I 1 I t I I 1 I I I 1

I I 0 2k00 5$00 ldOOO 1?500 lE!OOO 1~500

t TIME, HOURS
Fig. 2-Two.rale syelic or aquara wave rate tsow histow. W. 3-Rata hlatorias for casaa shoulaled analytic+

200&
# ●***e.*

4
m ● ●
8*
●**
● S.
.*’
%0 ‘
.
~m
.s
150 ● 4.:*, ■ m
q....””’” .9

‘d
8
w 0.
1. .s
●

LEGEND
●-u.. .
●

4
S*
LEGEND ..a@** ● Square Wave Rate
! aO 8
.~”’#@
A Square Wave Rate ● Sinusoidal Rate
●
●
Ii 8 .m ● Sinusoidal Rate ● Random Rat:
●
~ a.
% Random R*te 9

1
k
50

01
k Start of Stabilized F1OW
8
w “ .
m

0.0
4 & Start of Stabilized Flow

1 1 1 I I 1 I I
I I [ 1I 1I I I I 1 I 11 ( I iI I I1 I I 11 I I I I 8 I i 1 I I 1 1 1 1 I 1 I 1

‘0 2400 5$00 7k00 lLk300 1~500 1dOOO 1~500 ‘0 5b00 1E1000 1 dOOO 2EJ000 2d000 3d00B

TIME, HOURS 7, HOUR5

Fig. a-ApJq. va. ~ curves !or caaes slmulatad analytiwlly.
Fig. 4-Pmasum drawdown data for cases al?nulatad analytically. .

1
 12s_ sOOO_

k Start of Stabilized Flow ~ Start of Stabilized Flow

(6)

10m-
400 (5)
—.
0
1 Constant Rate
& 2 Discrete (Stair-Step) Rate a
w 7 3 Logarithmic Rate
. UJ
4 Hypertmlic Rate
w
5 Constant Pressure . 300L7L
E S Exponential Rate L
K I

J
z
z -.... .—

200a_ 1 Constant Rate

2 Discrete (Stair-Step) Rate
2s

5 Constant Pressure

0- I 1 I ) 1 1
1000
( 1 t t 1 t I I 1 t I 1 # 1 1 I I I I I
1 I I I I I I I 1 I I I I I
0 5b00 ] dam 1km 2!&?00 2d000 0 5h00 1dOOO 14000 2d000 243E10

TIME, HOURS TIME, HOURS

Fig. 6-Fkte Matorias for finksNffarenca simulation caaaa. Fig. 7—Prasaure drawdown data for fmWdiierenca simulation caaas.

25 0“%
●

LEGEND
20 A Exponential Rate

0.4_
a Discrete (Stair-Step) Rate ●
15 .-’
a = 1.269 x 10-5 psi/STB/O/hr
Fi AA
o-

10
●’A
G \b ,2542X 1,-I p~i,s,e,o

& a.z Vr
3

I
-i
w
k Start of Stabilized Flow
Start of Stabilized Flow
0
0.El
1 1 1 I 1 1 i I I 1 I i I 1 I I I I1 I I I I 1
‘0 5h00 1dGOO 1!4000 2dm0 2Am0 3d000 w
al
~ , ‘I%:R5 ~, HOURS
Fig. 8-AjVqm Va. iOuwaa
fos
Nnwifferenca
aansslwoss
Casaa. Fig. 9-Example graph of Aplqm V% i for ramdom rafesaae aimulatad analytically.