SOCIETY OF FETROLEUM ENGINEERS CF AIME
6200 North C,?ntral Expressway Dallas, Texac K206 THIS IS A PREPRINT  SUBJECT TO CORRECTION
=ER
SPE
4629
Decl
ine
Curve
Analysis
By
Using
Type
Curves
M. J. Fetkovich, Member AIME, Phillips Petroleum Co.
@ Copyright 1973 American Institute Mining,Metallurgical, PetroleumEngineers, of and Inc.
This paper was prepared for the 48th Annual Fall Meeting of the Society of Petroleum Engineers of AIME, to be held in Las Vegasj Nev,j Sept. 30Ott. 3, 1973. Permission to copy is restricted to an abstract of not more tham 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment of where and by whom the paper is presented. Publicat.icn 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 prcvijed agreement to give proper credit is made. Discussion of this paper is izv5ted. 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. ABSTRACT This Faper shows that the declinecurve analysis approa~h does have a solid fundamental basis. The exFonent.ial decline is shown to be a longtime soluticm of the constantpressure caie. The constantpressureinfinite and finite reservcir solutions are placed on a common dimensionless curve with all the standard “empirical” expmential, hyperbolic, arid harmoni> declinecurve equations. Simple combinations of material balance equations and new forms of oil well rate equaticms fo: solutiongas drive reservoirs illustrate under what circumstances speoifis values of the hyperbolic de:line exponent (l/b) or “b” should result. curve matching is demonstrated. This paper demonstrates that declinecurve analysis not only has a solid fundamental base, but provides a tool with more diagnostic power than has previously been suspected. The type curve approach provides unique eolutions upon whish engineers can agree, or shows when a unique solution is not pssible with a type curve only. INTRODUCTION
Fiatetime declinecurve extrapolation is one of the oldest and most often used tools of the petroleum engineer. The various methods used have always been regarded as strictly empirical and not very scientific. Results obtained for a well or lease are subject to a kglog type curve analysis can be performed Gn declining rate data (constantterminal wide range of alternate interpretations,mostly as a function of the experience and objectives pressure case) completely analogous to the logof the evaluator. Recent efforts in the area log type curve matching procedure presently of declinecurve analysis have been directed teing employed with constantrate case pressure towards a ~urely computerized statistical transient data. Production forecasting is done approach. Its basic objective being to arrive by extending a line drawn through the ratetime at a unique “unbiased” interpretation. AS data cverlain along the uniquely matched or best pointed out in a com rehensive review of the theoretical type curve. Future rates are then literature by R2msa , “in the period from simply read from the real time scale on which 1964 to date, (1968 , several additional papers !? the ratetime data is plotted. The ability were published which contribute to the underto calzulate kh from declinecurve dzta by tyFe standing of declinecurves but add’litt.le new ..eferen:es illustrations at end o? yapr, and
JIECL.INE CURVE ANALYSIS USI technology”. A new direction for declinecurve analysis was given by Slider2 with his development of an overlay method to analyze ratetime data. Because his method was rapid and easily apFlied, it was used extensively by Ikunsayin his evaluation of some 200 wells to determine the distribution o the declinecurve exponent term “b”. J GentryIs Fig. 1 displaying the Arps’4 exponential, hyperbolic, and harmonic solutions all on one curve could also be used as an overlay to match all of a wells’ decline data. He did not, however, illustrate this in his example application of the zurve. The overlay method of SMder is similar in principal to the loglog tyFe zurve matching procedure presently being employed to analyze constantrata Fressure buildup and drawdown data 5$’. The exponential decline, often used in decliaecurve analysis, can be readily shown to be a longtime solution of the It followed then constantpressure zase1013 that a loglo~pe curve matching procedure could be develcFed to analyze declinezurve data.
q
TYPE CURVES —
SPE 4%
nd for b = 1, referred tc as harmonic decline, e have
y=.—
(3) [l+;it] “ “ “
q
i
,
q
A unit solution (Di = 1) of Eq. 1 was eveloped for values of “b” between O and 1 n 0.1 increments. The results are lotted as set of loglog type curves (Fig. 17 in t~~s f a declinecurve dimensionless rate ES&l ‘Dd i
q q q
*
q
q
“(k)
nd a declinecurve dimensionless time ‘Od =Dit . . . . . . . . (5)
This pap?r demonstrates that both the analytical 3onstantpressure infinite (early transient pericd for finite systems) and finite reservoir solutions ~an be placed on a common dimensionless loglog type curve with all the “ standard “empirical exponential, hyperbolic, and ha~monic decline curve equations developed by Arps. Simple combinations of material balance equations and new forms of oil well ratt equations from the recent work of Fatkovich~ illustrate under what circumstances specific values of the hyperbolic decline expnent “b” should result in dissolvedgas drive reservoirs Loglog type curve analysis is then Ferformed using these curves with declining rate data completely analogous to the loglog type curve natcliing procedure presently being employed with constantrate case pressure transient data
From Fig. 1 we see that when all the basic leclinecurvesand normal ranges of “b” are Iisplayedon a single graph, all cur~es coincide =0.3. Any md become indistinguishable at t Iata existing prior to a t,~dof O. till ,&P~T 9 ;O be an exponential decline regardless of the ;rue value of b and thus plot as a straight lin< )n semilog paper. A statistical or leastSquaresapproach could calculate any valus of J between O and 1. \NALYTICALSOLUTIONS (CONSTANTPi’iSSURET A [NNER BOUNDARY) Constant well pressure solutions to predizieclining production rates with time were first pklished in 1933 b l;iore,Schilthui.s and were pssented {u::stlO, and HurstlI . fiesults fcr infinite and finite, slightly compressible, ~in.glephase lane radial flow systems. The p rf2SUltS were presented in graphical fcrm in Lerms of a dimensionless flow rate and a ji,mensionless ime. The dimensionless flow rat t ~D ,?anbe ex~ressed as
BASIC EQUATIONS AIISl RATETIWW4U ATIONS Nearly all conventional declinecurve analysis is based on the empirical ratetime 3quations qi.ven by Arps4 as
%++&+&v””””(’)
and the dimensionless time tu as t = D 0.00634 kt !J~~trw2 . . . . . (7)
For b = 0, we aan obtain the exponential declin aa.uation from E@. 1
The original publication did not inulude tabular values of q and t . For use in this paper infinite SOIU?? ion va ues were obtained from Ref. 15, while the finite values were obtained from Ref. 16. The infinite solution, and finite solutions for r /r from 10 to ant!~b. 100,000, are Tlot.ted on
Fif$~.w2a
4JLJ_
.
‘
“
“
“
“
q
D:t
‘(2)
,PEf+629
M. J. FE’I In terms of the empirical exponential ~ec~necurvej Eq. 2, Di is then defined as
Most engineers utilize the constantpressure solutionnot in a single constaRtFressure problem but as a series of constantIressure step functions to solve water influx prcblems using the dimensionless cumulative production Q~ (13). The relationship between {Q and qD is ~ d(QD) —“=QD dtfi
u
k@w.
D: == A N pi
q
q
q
q
q
(U)
(8) “ “ “ “
q
In terms cf a dimensionless time for ieclinecurve analysis we have from Eqs. 5 and 15
“
“
Fetkcvich’7 Fresented a simplified approacl to water influx calculations for finite systems Defining N that gave results whi,>hcompared favorably variables,pi an: ‘qi) ‘n ‘em ‘f ‘eservoir with the more rigorous analytical sonstantpressure solutions. Equation 3 of his Fa~er, Tl(rrw2) @ ct h pi . . (17) e ~or a constantpressure pwf, can be written as N 5.615 B pi =
q q q q
‘~=[ :l ‘ t
‘“)
Jo (pi  Pwf) q(t) = [“[T . e but qi =Jo (pi  pwf)
q
. t.
,
.
(9) and khpi (A)= ‘=i’max IJJ*3AJ3 ~nb [() rw  1 2 .* (18)
1
1
.
.
.
q
(10) The declinecurve dimensionless the, of reservoir variables, becomes
0.00634 kt
in terms
‘m Substituting I@. 11 into 10 we zan writs
=
@ B ct rw2
J.
.
.
(19)
Now substituting Eq. 10 and 12 into 9 we obtain
qit
Pwf
l— 01
Pi
NPi
q
* (13)
..**
(20)
Equation 13 can be considered as a derivation of the expmential decline equation in terms of reservoir variables and the constant~ressure imposed on the well. For the same well, different values of a single constant oackpressure p f will always result in an exponential decfine i.e., the level of backpressure does not change the type of decline. For p f= O, a more realistic assumption for a wel!fon true wideopen decline, we have
To obtain a declinecurve dimensionless rate qw in terms of q~
~=;[(q(p]t
q, . .
,U)
.
L
DECIJNE CURVE ANALYSIS USING TYPE CURVES The published values of qD and tD for the infinite and finite constantpressure so~utions were thus transformed into a declinecurve dimensionless rate and time, q using Eqs. 20 and 21. Fig. 3 @ :;l:”?%f the newly defined dimensionless rate and time, qm and tm, for various values of re/rw. AL the onset of depletion, (a type of pseudosteady state) all solutions for various values of re/rw develop exponential decline and converge to a single curve. Figure 4 is a combination of the constantpressure analytical solutions and the standard “empirical” exponential, hyperbolic, and harmonic declinecurve solutions on a single dimensionless curve. The exponential decline is common to both the analytical and empirical solutions. Note from the som~site curve that rats data existing only in the transient period of the constant terminal pressure solution, if analyzed by the em@rical ArPs approach, would require values of “b” much greater than 1 to fit the data. SOLUTIONS FROM RATE AND MATERIALBA LANCE EQUAIIONS ‘he method of combining a rate equation and material balance equation for finite systems to obtain a ratetime equation was outlined in Ref. 17. The ratetime equation obtained using this simple approach, which neglects early transient effects, yielded surprisingly good results when compared to those obtained using more rigorous analytical solutions for finite aquifer systems. This rateequation materialbalance approach was used to derive some useful and instructive declinecurve equations for solutiongas drive resewoirs and gas reservoirs. RATE IJJUATIONS Until recently, no simple form of a rate equation existed for solutiongas drive reservoir shut–in pressure. Fetkovich~ has proposed a simple empirical rate equation for solutiongas drive reservoirs that yields results which compare favorably with computer results obtained using twophase flow theory. The proposed rate equation was given as
=
SPE L625
(l/b) in terms of the back pressure curve slope (n) and to study its range of expected values. Also, the initial decline rate Di can be expressed in terms of reservoir variables. One further simplification‘lsed in the derivations is that pwf = O. For a well on decline, pwf will USUallY be ~intained at or near zero to maintain maximum flow Pates. Equation 23 then becomes
The form of Eqs. 23 and 23A could also be used to represent gas well behavior with a pressure dependent interwell permeability effect defined by the~R/FW). lhe standard form of the gas well rate equation is usually given as
‘i% =
Cg
(j5R2pwf2)n
.
.
.
. (24)
I
MATERIAL EAL!NCE OQUATIO’.; Two basic forms of a material balan$e equation are investigated in this study~ p~ is linear with N or G , and ~..2is linear w~th N or Gp (See ‘Figs.p5a and ~). The linearp~R relationship for oi2 is
I and for gas
Equation 25 is a good apFroXhnation for totally undersaturated oil reservoirs, or is simply assuming that durinR the decline vs. N can be approximated by 12!@@% a straigh line.p For gas reservoirs, Eq. 26 is oorrect for the assumption of gas compressibility (Z) =1. In terms of j5R2being linear with cumulative production, we would have kLi2 N () Pi NF+~Ri2 . “ (27)
q.
J?
‘i
()
ij
(~R2 pwf2)n
.
.
.
(23)
Ki ~R2=q
where n will be assumed to lie between 0.5 and 1.0. Although the above equation has not been verified by field results, it offers the opportunity t.odeflna t,hedeeljw expnqn{.
‘I%isform of equation results in the typical shape of the pressure (~ ) VS. cumulative ~roductfon (Np) relaticn%ip of a s@lutfonFas
3PE 4629
M. J. FE! )VICH
5
drive reservoir as depicted in Fig. 5b. Applications would be more appropriate in nonpmrated fields, i.e., wells are produced wideopen and go on decline from initial production. This would more likely be the case for much 18 of the declinecurve data analyzed by Cutler obtained in the early years of the oil induetry before proration. RATETIME EQUATIONS, OIL WELLS Ratetime equations using various combinations of material balance and rate equations were derived as outlined in Appendix B of Ref. 17. Using Eq. 23A and @. 25 the res~ting ratetime equation is
Clo(t)
exponent (l/b) is 2.0 and 3.0 m b = o.5~ ad 0.333 respectively. This range of “bE values fits Arpsl ftidings using Cutlorls declinecurve data. He found that over 70 percent of the values of “b~ lie in the range Ramsay found a different O~b ~0.5. distribution of the value of “b” analyzing modern rate decline data from some 202 leases. His distribution may be more a function of analyzing wells that have been subject to proration and are better represented by the assumptions underlying the ratetime solution given by Eq. 28, i.e., ‘ver the decline period% ‘s” ‘pWas linear DECLINECURVE ANALYSIS OF GAS WELLS Declinecurve analysis of ratetime data obtained from gas walls has been reported in only afewinstances19S 20. Using Eq. 24 with ~wf =0, and Eq. 26, the ratetime equation for a gas well is
‘= [+J+,1%
—.
q
q
(y
C@) A unit solution, (qoi/N .) = 1, ofEq. 28 is plotted as a loglog ty~ curve for various values of n, Fig. 6, in terms of the (For deel.inecurve dimensionless lttie t =0, q .%”(q .) . these de~ivations with p For the limiting range o~fbackpr%ure %%? sloFes (n) of 0.5 and 1.0, the Arps empirical declinecurve exponent (l/b) is 2.0 and 1.5 respectively or “b” =0.500 and 0.667 a surprisingly narl;owrange. respectively To achieve an~pnential decline, n must be equal to zero, and a harmonic decltne In Fractical applications, requires n+. if we assume an n of 1.0 dominates in solutiongas (dissolvedgas)drive reservoirs and ~ vs. N is l%near for nonuniquely defined rate?ime data, we would simply fit the ratetime data to the n = 1.0 curve. On the Arpsj solution type curves, Fig. 1, we would use (l/b) =20rb =0.667. The ratetime equatior obtained using Aq. 23A and Eq. 27 is qo(t)
1
1
——
q
“’
(3C)
for all backpressure zurve slcFes wh~re n > 0.5. For n = 0.5, the exponen~ial decline is obtained
C&) ‘gi
.e

*t ()
q~
..0.
q
(31)
The unit solutions of Eqs. 30 and 31 are plctted as a loglog type curve cn Fig. 8. For the limiting range of backFressure curve slopes (n) of 0.5 and l.oj the AW declinecurve expment (l/b) is 00 and 2, or b = O (exponential) and 0.500 respectively. The effect of backpressure on a gas well is demonstrated for a backprassure curve slope n =1.0 on Fig. 9. The backpressure is expressed as a ratio of Pwf/Fi* Note that as pwf + p. (Ap+O) the type curva approaches , ex’ymentlalldecline,the liquid case solution. Whereas backprsssure does not change the type of decline for tha liquidcase solution it does change tha type of decline in this case. Using the more familiar rate and material balance equations for gas wells, we can obtain the cumulativetime relationship 5;’integrating the ratettie equations 30 and 31 with Gp= /t o qg(t)dt . . . . . (32)
—=

q
q
(29)
‘oi (~)t+r p
The unit solution of Eq. 29 is plotted as a loglog type curve for various values of n, Fig. 7. This solution results in a complete reversal from that of the previous one, n =0 yields the harmonic decline and n + .W gives the exponential decline. For the limiting range of backFressure curve slopes (n) of 0.$ and 1.0, the declinesurve
r
DECLINE CURVE ANALYSI USING TYPE CURVES For n > 0.5 we oktain
SPE 4629
common point on both sheets are recorded.
&
1

[1 + (2n1)
+t]e ()
&,..”
:’
q
(33;
and n = 0.5
Lag1og type curves of Eqs. 33 and 34 couk be prepared for convenience in obtaining cumulative production.
Recent papers by Agarwal, et.al.5, E#mey6, Raghavan, et.al.7 and GringartenS et.al. ~ have demonstrated or discussed the application and usefulness of a typecurve matching pror.edure interpret constantrate pressure to bu~ldup and drawdown data. van Poolen21 demonstrated the application of the typec.urveprocedure in analyzing flowrate datz obtained from an oil well producing with a constant pressure at the well bore. All of his data, hcwever, were in the early transient period. No depletion was evident in his examples. This same typecurve matching procedure can be used for declinecurve malysis. The basic steps used in typecurve matchin~ declining ratetime data is as follows: 1. Plot the actual rate versus time data in any convenient units on loglog tracing paper of the same size cycle as the type curve to bs used. (For convenience all type curves should be plotted on the same loglog scale so that various solutions can be tried.) 2. The tracing paper data curve is placed over a tyFe curve, the coordinate axes of the two curves being kept parallel and shifted to a position which represents the best fit of the data to a type curve. More than one of the type curves presented in this paper may have to be tried to obtain a best fit of all the data. 3. Draw a line through and extending beyond the ratetime data overlain along the uniquely matched type curve. Future rates are then simply read from the realtime scale on which the rate data is plotted. 4. To evaluate declinecurve constants or reservoir variables, a matoh point is se~ezted anywhere on the over?.app?ng ~rt$on of the curves and the coordir,ates this of
5* lf none of the type curves will. reasonably fit all the data, the departure curve method 15$ 22 should be attempted. This method assumes that the data is a composite of two or more different declinecurves. After a match of the late time data has been made, the matched curve is extra@lated backwards in time and the departure, or difference, between the actual rates and rates detemi.ned from the extrapolated curve at coiresponding times is replotted on the same loglog scale. An attempt is then made to match the departure curve with one of the type curves. (At all times some consideration of the type of reservoir producing meck~tism should be considered.) Future predictions should then be made as the sum of the rates determined from the two (or more if needed) e:krapolated curves. TYPE CURVE MATCHING EXAMPLES Several,examples will be presented to illustrate the method of using type curve matching to analyze typical declining ratetime data. The type curve approach jmwides unique solutions upon which engineers xm agree, or shows when a unique solution is not possible with a type curve only. In the event of a nonunique solution, a most probable solution can be obtained if the pI’OdUCi.ng mechanism is known or indica~ed. ,ARPS’HYPKRBOLJC DECLINi EXAMPLE Fig. 10 illustrates a type curve match of Arpsl example of hyperbolic decline4. Every single data point falls on the b =0.5 type curve. This match was found to be unique in that the data would not fit any other value of “b”. Future producing rates can be read directly from the realtime scale on which the data is plotted. Ifwe wish to determine q. and D.. use the match mints indicated on “ Ftg. 10 *: follows @= ~ j 1000BO~ % 30,303 BOPM
qm
=0.33
=
qi =
1000 BOPIVI = 0.G33
‘Dd
=12.0
= Dit = Di 100MO.
.Di=
~= 100 Mo
0.12 MO.l
The data could have also been matched ustng the ty~s ~urves on Pigs. 6 and 7. h both cases the mat~h would have oeen obtained with a backpressure curve slcFe n = G.5 which
5PE 4629
M. J. FETKOVICH
7
the ex~nential interpret.ation gives a total life of 285 MONTHS, the harmonic ll@ MONTHS. ‘his points out yet a further advantage of the type curve approach, all pssible alternate pi“ interpretations zan be conveniently placed on one curve and forecasts made from them. A The fact that this example was for a lease, statistical analysis would of course yiald a a group of wells, and not an individual well single answer, but it would not necessarily be raises an imprttint question. Should there be the correct or most probable solution. Cona difference in results between analyzing eazh sidering the various producing mechanisms we well individually and summing the results, ~ou>d select. or simply adding all wells production and analyzing the total lease production rate? a) b = O, (exponential), if the reservoir is Consider a lease or field with fairly unifomn lib!: n is similar for or highly undersaturated. reservoir properties, eazh well, and all wells have been cn deckine b) b = O, (exponential),gravity drainage with at a similar terminal wellbore pressurey Fwf~ no free surface28. for a sufficient period of time to reach FSeUdCc) b = 0.5 gravity drainage with a free steady state. According td Matthews et.al.23 Iiat(Fseudo) steady state the drainage volumes surface58. in a bounded reservoir are proportional to the d) b = 0.667, solutiongas drive reservoir, rates of withdrawal from each drainage volume.” (n=l.O) if~Rvs. Npis linear. It follows then that the ratio q./N . will e) b = 0.333, solutiongas drive reservoir, be identical for each well and th$isthe sum of (n=l.0) if~k2 vs. Npis approximately the results from each well will give the same linear. results as analyzing the total lease or field production rate. Some rather dramatic illusFRACTURED WiLL EXAMPLE trations of how rapidly a readjustment in drainage volumes can take place by changing Fig. 12 is an example of type surve the production rate of an offset well or matching for a well with declining rate data drilling an offset well is illustrated in a available both before and after stimulation. Similar drainage volume paper by Marsha. (The data was obtained from Ref. 1.) This readjustments in gas reservoirs have also been type problem usually presents some difficulties demonstrated by Stewart25. in analysis. Both before and after frac. loglog plots are shown on Fig. 12 with the after For the case where some wells are in frac. data reinitialized in time. These before different portions of a field separated by a and after loglog plots will exactly overlay fault or a drastic permeability change, each other indicating that the value of “b” readjustment of drainage volumes proportional —— did not than e for the well after the fracture to rate cannot take place among all wells. The ratio ./N . may thenbe different for different treatment. tThe before frac. plot can be considered as a type curve itself and the group %#f hlls. A total lease or field after frac. data overlayed and matched on it.) production analysis would then give different results than summing the results from individual Thus all the data were used in an attempt to de~e “b”. When a match is attempted on the well analysis. A stilar situation can also Arps unit solution type curves, it was found exist for production from stratified reservoirs that a “b” of between ~.6 and 1.0 could fit 26, 27, (nocw~~flow)o the data. Assuming a solutiongas drive, a match of the data was made on the Fig. 6 ARPS’ EXK)NENTIAL DECLINE EXAMPLE type curve with n = 1.0, b = 0.667. Fig. 11 shows the results of a type curve Using the match Feints for the before analysis of Arpsl example of a well with an frac. data we have from the rate match mint, apparent exponential decline. In this case, there is not sufficient data to uniquely 1000 BOPM establish a value of “b”. The data essentially qw “0.2.43 = w= qoi fall in the region cf the type curves where all q’oi curves coincide with the expnmtial solution. As shown on Fig. 11 a value ofb =0, 1000 BOPM = 4115 BOPM qo~ = 1.0 (harmonic) appear to (~pnential) orb= q 243 fit the data equally well. (Of course all values in between would also fit the data.) From the time match point, The difference in forecasted results from the two ewtreme interpretationswould be great in qoi t = = 0.60 = 0+115 BOPM)(1OO MO % later years. For an economic limit of 20 BOP?4, r N pi pi 5.sequivalent to b = 0.5. Match pints determined from these curves could have been ;sed to zalculate~ and ~/Npi and finally
()

“
.
...
““.”,,

+Ws
/
N
pi =
f4115 BOPM) (1OC MO.) 0.60
=
685,833 BBL
then
q ~=
proprtion to the flowrate, the ratio qoi/Kpi should have remained the sane as that obtained prior to treatment or 0.006000 MO.l. This then would have indicated an N of pi N ‘J&J 130PM —1 .006000 MO. = 1,243,833 BBL
N ...4
4115 BOPM 685,833
= .006000 M0.1
pi =
Now usin~ the match points for the after frac. data we have from the rate match Pint, q~ =0.134 = ~ “Oi qoi . 100C BOPM 0.134 = . 1000 BOPM qoi
Actual in~rease in reserves as a.rssult of’the fracture treatment a~pea,rsto lie between ths two extremes. Based on the msthod of al,alysis used, the astual increase in reserves attributable to the fracture treatment is 198,109 BBL, (660,442 BBL  462,333 BBL; STRATIFIED r’SiRVOIR EXAMll& This example illustrates a method of analyz5ng declinecurve data foi layered a (nocrossflow) or stratified reservoir using type ~urvese The Qata is taken from i@f. 18 and is for the East Side Colinga Field. Ambrose2? presented a cross seztion of the fiel~ showing an upper and lower oil sand se~arated by a continuous blazk shale. This layered description for the field along with the predictive equation for stratified reservoir presented in Zef. 29 led to the idea of using the departure curve method (differencing)to analyzed dezlint?curve data. 27 After Russell and Prats the production rate of a well (or field) at p;eudosteady state producing a single phase liquid at the same constant wellbore pressure, (Fwf = O for simplicity), from two stratified layers is
7463 BOPM
From the time matsh pint .1*13 . + () pi
q.
~ .m6W(loo pi
MG.~
‘Dd
N
pi
= (7463 BOPM) (loo MOJ 1.13
= 660,4.42 BBL
then
q oi
‘= N pi
7463 BOPM = .011300 MC.l 660,w2 BBL
We can now zheck the two limiting conditions to be Considered following an inzrease in rate after a well stimulation. They are: 1. Did we simply obtain an acceleration of production, the well~ reserves remaining the same? Did the reserves hcrease h direzt proportion to the increase in producing rate as a resu t of a radius of drainage readjustment 3 ? BefGre treatment, N . was fOUnd to be 685,833 BBL. Cumulative F~%duction determined from the iatedata prior to stimulation was 223,500 BBL. N then at the time of the fracture treatm~i%t is
2.
J%
qT(t) = qi~ e
~Npi 1
)
t
+qi2e
qi — N
() t
pi 2
q
.(35)
‘r
qT(t) =ql
(t) +q2
(t)
.
.
q
.
(36)
N
pi
=685,833 EBL
The total production from both layers then is simply the sum of two separate forecasts. Except for the special case of the ratio q /N ~ being equal for Loth layers, the sum o}tio expnentials will not in general result 223,500 BBL =462,333 i3BI in another exponential. In attempting to match the ratettie data to.a type curve, it was found that the late time data can be matched to the exponential (b =0) type ourve. Fig. 13 shows this matcki of the late the data designated as layer 1. With this match, the curve was extrapolated backwards in time and the departure, or difference, between the actual rates determined from the ext~apolated curve was replotted on the same loglog scale. See TAB~ 1 for a
If only accelerated pradution was obtained and after the reserve” remained the same, q~ F pi the fracture treatment should have been 7&63 BOFM f+62,333BBL = o.016J_42 MO.l
Actual (qoi/Npi) after treatment was 0.011300 MO.l. If the resarves increased in direct
SPE L629
M. J. FETKOVICH
9
summary of the departure curve results. The difference or first departure curve, layer 2, itself resulted in a unique fit of the exponential type curve, thus satisfying Eq. 35 which can now be used to forecast the future production. Using the match points indioated on Fig. 13 to evaluate qi and Di for each layer the predictive equation becomes
or
qT(t)
= 58,824 BOPY e
50sGOOBO~
 (o.2oo)t +
.*.*. ..*
e  (O*535)t
(37)
where t is in years. 30 Higgins and Le~htenberg named the sum of two expnentials the double semilog. They reasoned that the degree of fit of empirieal data to an equation increases with the number of constants. This interpretation is not claimed to be the only interpretation pX3sib16 for this S3t of data. A match with b =0.2 can be obtained fitting nearly all of the data points but can not be explained by any of the drive mechanisms so far discussed. The layered concept fits the geologis description and also offered the opportunity to”demonstrate the departure curve method. The deFarture curve method essentially places an infinite amount of combinations of type curves at the disposal of the engineer with which to evaluate ratetime data. EFFECT OF A CHANGE IN BACKPRESSURE The effe~t of a change in backpressure is best illustrated by a hypothetical single well problem. The reservoir variables and conditions used for this example are given illTable 2. The analytical singlephase liquid solution of Fig. 3 is used to illustrate a simple graphical forecasting superposition procedure. The inverse procidure, the departure or differencing method can be used to analyze declinecurve data affected by baskpressura changes. After Hurst12, suparpsition for the constantpressurecase for a simple single pressure change can be expressed by kh q (t) = 141.3(@) [ kh (Pwfl  Pwf2) + ‘103(~)[’n(~)~] ‘~(tod’~~) In ($) $] ‘M(t~l
Up tc the time of the pressure zhange pwf2 at t~l the well production is S~PIY he q foret ql  depicted on Fig. U. as a function of time is simply kde by evaluating a single set of match points using the reservoir variables given in Table 2. At Pwfi and re/r’ =100 w t=lDAY; t =0.006967 Dd = 2.02
ql(t) = 697 BOpD; q~
Plot the rate 697 BOPD and time of 1 day on loglog tracing paper on the same size cycle as Fig. 3. Locate the realtime points over the dimensionless time pints on Fig. 3 and draw in the re/r~ curve of 100 on the tracing paper. Read flow rates as a function of time directly from the real time scale. When a change in pressure is l~de to pwf2 at tl, t equal zero for the accompanying zhange in rate q29 (really a Aq for superpostion)j this rate change retraces the qm vs. t~ curve and is simply a constant fraction of ql =ql
[
q2
Pwfl  pwf2 Pi  ~wfl
1
 1 day after the rate change is ::U:? :; t q2 = 697 BOPD = 221 BOPD 1000 Psi  50 Psi 4000 psi  1000 psi
(pi  Pwfl)
The total rate qT after the pressure shange is q = ql + q2 as depicted in Fig* M* All rates o T flow for this example were read directly from the curves on Fig. 11+and summed at times past the pressure change pwf2. The practical application of this exam~le in declinecurve analysis is that the departure or difference method can be used on ratetime data affected by a change in
) ,
lWf?T. TN17 “...4.1.

flIIRVR “...v
“.,.,
AN?ITY.STS

IISTNC
““J.
..”
TVPT7
. ..
PIIT+VF<
““.
.”
“.
.sPr?
6+”t.,~
J. L’)G
backpressure. The departure curve represented by q2 on Fig. 14 should exactly overlay the curve represented by q . If it does in an actual field example, ~he future forecast is correctly made by extending both curves and summing them at times beyond the Fressure change. CALCULATION OF kh FROM DLCLINECURVE DATA
wallbore radius r ‘ (obtained from the buildup analysis) is u%d to obtain a good match between buildup ant declinezurve ~alculated. kh. ‘. TYPE CUFWdS l~iiKNOkJNRESERVOIR AND FIJJ~ PROPER71EL —
All the type surves so far discussed wera developed for de~linecurve analysis using Pressure buildup and decliaecurve data some necessary simplifying assumptions. For spe~ifi,creservoirs, when PVT data, reservoir were available from a highpressure, highly undersaturated, lowpermeability sandstone variables, and backpressure tests are reservoir. Initial reservoir pressure was available, typs surves could be Eenerated fcr various relative permeability wrves and estimated to be 5750 psia at 9300 ft. with ba>kpressures. These curves developt~ for a bubblepoint pressure of 2841 ~sia. Two fieldwide pressure sarveys were condu?tsd while a given field ~iouldbe more accurate for analyzing flecline data in thak field. Conthe reservoir was still undersaturated. lable 3 summarizes vl,ereservoir properties and ventional material balance programs or more basic results obtained from the rressure kuild sophisticated simulation models could be used UF analySiS on %azh Well. Note that nearly to develop dimensionless sonstantpressure 31 all wells had n8ga.tive shins as a result of type zurves as was done bv Levine and Pratts (See their FiF. 11). hydraulis fracture treatments. Also, a~~earing on this table are results obtained CONCLUSIGNS from an attenpt to calculate kh using declinecurve data available for eazh of tk,e wells. Declinezurve analysis not only has a Ten of the twefitytwo wells started on solid fundamental base, but providss a tool with more diagnostic power than has previously desline when they were first placed cn been suspe~ted. The ty~!e:urve ap~roach ~:roproduction. As a result, the early production vides unique solutions u~on whi~h engineers desl:ne data existed in the transient pmiod can agree, or shows when a unique solution is and a ty~e curve analysis IlsingYip. 3 was matched to one of the r@/’rwst:ms. Gther wells not possible with a tyre zurve ocly. in ths event of a nonunique sclution, a most ~robable listed on the table, kh5ra an r ~rw mat:h is solution zan be obtained if the produ~ing not indiczted, were Frorated wefls and began meshanism is known or indicated. their de:line several montk,safter they were first put on production. For the de2lineNO~NCLATU~, zurve determination cf kh, the reservoir pressure existing at the beginning of decline for =l{~giro~al of detline >urve exponent each well was taken from the pressure history lb (1/b~ match of the two fieldwide pressure surveys, The zonstant bottom hole flowing ~ressure for ~ = Formation volume factor, res. vol.,’ the wells ranged between 80G and 900 psia. surface vol. A type curve mat~h using deeline~urve data to talculate kh for well No. 13 is illustrated on Fi~. 15. A type ?urve matsh using pressure buildup data obtained on this same well is illustrated on Fip. 16. The constantrate type curve of’Gringarten et.al.8 for fractured wells was used for mat~hing the pressure buildup data. The buildup kh of47.5 rnd.ft. cmmpares very well with the kh ofl+O.5 MDFT determined by using the ratetime dezlinezurve data. In general, the comparison of kh detarnrine d from deulineturve data and prsssure buildur, data tabulated on Table 3 is sur~risingly good. (The pressure buildup analysis was performed independently by another engineer.) One fundamental observation to be made from the results obtained on wells where a match of re/rw was not ~ssible is that the effective Ct (J= D! ~= G G h Jo J; k, n N F p = Total compressibility, psil Gas well ba~kpressure :urve coefficient 1 = Initial Decline rate, t =Natural logarithm base 2.71828 = Initial pasin~dase, surface measure = CUmu].ative gas production, SUrfaCe measure = l%i~knsss, ft. = Productivity index, STK BHL/DAy/pS~ = Productivity index ’(backpressuresurve coefficient) STK BBL/DAY/(psi)2n = Effective permeability, md. = Expnent of baskpressure curve = Cumulative oil ~roduction, S’fKBBL
SPE L629
M. J. FE!
)VICH 7.
11 Raghavan, R., Cady, G. V. and Ramey, Analysis for H. J., Jr.: i,~;ellTest Vertically Fractured Walls””,J. Pet. Ta?&~ (hlgo 1972) 10L4. Gringarten, A. C., Ra!!ey, H. tJ.~Jr. and. Analysis for Raghavan, R.: tgpres~ure Fractured ~lells”,paper SPE 4051 presentad at the 47th b.nnual Fall lfeethin, San Antonio, Texas, (Gzt 811, 1972Y . MeKiriley, M.: “Wellbore Transmissibility R. from AfterflcwDominatedPressure Buildup Data”, J. Pet. Tesh. (July, 1971) 863. Moore, T. V., Schilthuis, R, J. and Hurtitj w. : tl~e Deter~~natiOn of Permeability from Field Data”, Bull..ApI (~ky, 1933) ~ 4. Hurst, W.: “Unsteady Flow of Fluids in , (Jan., 1934) Oil lieservoirs” &sics. ~, 20. Hurst, W.: “l!ater Influx into a ileservoir and Its Application to the Zquaticm of Volumetric Balance”, —— Trans., ?l~’~(1943) ~, 57. van Everdingen, A. F. and Hurst, ~~.:“The AFpli2ation of’the Laplaee Transformation to Flow Proble.ns :eszrvcirs”, Trans. in AIME (1949) 186, 3050 Fetko\fi~h, J.: “The Isochronal Testing M. of Oil v~ells”,Papr SPE 4529 presented at the 48th Annual Fall Meetinz, las Vepas) !~~vada,(S~~t. 30G~t. 3, 19?3). Ferrt~,,J.$,Knowles, D. B., i3rown, ~:.J~.~ and Stallman$ R. W.: “’Ihecry Aquifer of Tests”, g. S. Gecl, Surv., ‘:later9FF1% 2 Pap3r 1.53~, 109. Tsarevi~h, K, A. and Kuranov, I. J’.: l’f~a~culation the I’1ow l:atesfor the of Ce:lt3r ‘Jellin a Circular Seser’mir Under Jlastie Conditlcns”, Problems @f Rgservoiy Hvdrods’nami~s Part.I, Leningrad, m 934
q
N
Fi
Cumulative oil production to a reservoir shutin pressure of O, STK BBL Initial pressure, psia Reservoir average pressure (shutin pressure), psia Bottomhole f’lowin~ pressure, psia Initial surfase rate of flGw at t = O Initial wideoFen surface flow rate at pwf = O Surface rate of flow at time t
Fi h Fwf % (qi
8.
9.
q(t, q~ ‘Dd
10. Dimensionless rate, (Eq. 6) Decline wrve dimensionless rate, (Eq. &) lXLmensionless mxnulative production External boundary radius, ft. Wellbore radius, ft. 12* Effective wellbore radius, ft. Time, (Days for tD) Dimensionless time, (@. 7) Decline nrve dimensionless time, (Eq. 5) Porosity, fraztion of bulk volume Vis20sity, cp. :4. REFERENCES 1. Ramsay, H. J., ~Jr.:“The Ability of Rat@Time )ecl.ine Curves to Predizt Future Prcxiuction Rates”, M.S. T?_lesis3 of U. Tulsa, Tulsa, Okla. (1968). Slider, H. C.: ‘IASimplified Method of Hyperboli~ Decline Curve Analysis”, J. Pet. Tech. (March, 1968), 235. Analysis”> Gentry, R. W.: llDe~linecurVe J. Pet. Tech. (Jan., 1972) 38. of Arps, J. J.: IiAnalysis Decline Curves”$ Trans. AIME (1945) &@, 228. Ramey, H. J., Jr.: “ShortTimeWell Test Data Interpretation in the Precence of Skin EffeZt and Wellbore Storage”, J. Pet. Tech. (Jan., 1970) 97. Agavwal, R., A1Hussainy, Il.and Ramey, H. J., Jr.: llAnInvestigation of ~~elibore Storage and Skin Effect in Unsteady Liquid Flow: I. Analytical Treatment:j Sot. Pet. Eng. J. (Sept., 197G) 279. 17. 15. 13. 11.
QD r e r w ~f tw ‘D % ~ P
2.
16.
3.
4.
5*
~etkovi~h, M. Jo: “A Simplified Approach to ~!at:?rnflux CalculationsFinite I Aquifer Systems”, J. Pet. Tech. (tJuIY~ 1971) Xl4. of W. W., Jr.: llEstimation Lhderflr.lttlg]~, ground Oil Reserves bv Oilwell P~duction Curves”, Bull., USBM”(1924)~ Stewart, P. R.: “LowPermeabilityGas
Well Porfomnanee at Constant. Pre9sure”,
18.
6.
19.
J. Pet. Tech.
(Gept. 1970) 1149.
2
DECLINE CURVE ANALYSIS USING TYPE CURVES
SPE 4629
20.
Gurley, J.: 11Aproductivity and EsOnOnd.C Projection MethodOhio Clinton Sand Gas Walls”, Paper SPE 686 presented at the 38th Annual Fall Meeting, New Orleans, La., (Oct. 69, 1963). van Poollen, H. K.: “HOW to Analyze Flowing WellTest Data... with Constant Pressure at the Well Bore)’, Oil and Gas Journal (Jan. 16, 1967).
27.
Russel, Q. G. and Prats, M.: “Performance of Layered Reservoirs with Crc)seflowSingleCcimpressibleluid Case”, SOS. F P@. Eng. J. (March, 1.962)53. — Matthews, C. S. and Lefkovits, H. C.: llGra~~,y ~rainage Performance of Depletion‘l&pe Reservoirs in the Stripper Stage”, Trans., AIME (1956) ~, 265. Ambrose, A. W.: !!underg~und ~enditions in Oil Fields”, Bull., USBM (1921) 195, 151. Higgins, R, V. and Lechtenberg, H. J.: !!Merits ~e~line Equations Based on of Production History of 90 Reservoirs”, Paper SPE 2450 presented at the Rocky Mt. I@ional Meeting, Denver, Colo., (My 2527, 1969). I&wine, J. S. and Prats, M.: “’TheCal:u~ated performance of SolutionGas Drive Reservoirs”, Sot. Pet. Eng. J. (s3pt., 1961) 142.
28.
21.
29.
22.
Withers~on, F. A., Javandel, I., Neuman, S. P. and Freeze, P. A.: “Interpretation of Aquifier Gas Storage Conditions from Water Pumpinp Tests”, Monopraoh AGA~ New York (1967) 110. Matthews, C. S., Brons, I!.and Hazebroek, P“ “A Method for Determination of A&age Pressure in a Bounded Reservoir”, Trans., AIME (?954) 201, 182. of hlarshjH. N.: llMethod Appraising Result! of Production Control of Oil Wells”, ~. API (Sept., 1928) 2@, 86. tewart, P. R.: “Evaluation of Individual Gas Well Reserves”, Pet. Eng. (Mayj 1.566) 85.
30.
23.
310
24.
25.
ACKNOtilEIXXMENT I wish to thank phillipS Petroleum Co. for permission to publish this FaFer.
26.
——
Lefkovi%, H. C. and Matthews, C. S.: NOTE : Tle author has a limited quantity of Curves to Gratity“Applisaticm of Desline full size type curvas with grid suitable Drainage Reservoirs in the Strippr Stagel’ for actual use which are avzilable on Trans., AIME (1958) ~, 275. written request. _——. . . ..
I
(1)
(2)
(l)(2)
TotalField Rate?T
L3yer1 Lateql ~
Iate.qz 80FY
38WL
layer2
90, GO) 6L,,LW lb8,uc 36,(ALI 27, 5L0
2J ,250
52,(KKM L2,5W. 3k, 5W* 28, %@
:3,(LAN 18,(YW 15,cw*
2;5L.G 135(L 75(XI k5CG 2650 1250 yxl G
16,2S I?,L.CA 10,:Kl 8,Sk
6,%
1Z,5GW lb, 5(IJ 8,503 6,$’w 5,4X L,55G 3,W.6 3,2cL 2,:%
5,LM
L,55G
3,t3L& 3,):J.
7,75L
lre.s.re ,~:d1: } ‘.el:
!.0 . — 1 ? t!
J*C Win
..esu:t: M
s
%tzhed
—.
5.4 J?, Y
$
s
.
(,,?3 ?.65 3.71 3.Ll [.29 ?. G7
;. L1 3.7L
rw J1.
m. H. 1.7L.5 \6. ? L3,(. 2?.5
LL.L
rw
L.3 3.> lf; .3 ‘7.6 18.3 2.“ .L IL. (. 16.5 02.9 .2.(, 13.3 ;I. P I>.& 9.1 2L.2 5.? ?.7 2.1 &.1 2.2 11.2
10.5 $.$ 5.5 :L.3 IL.[ s.] 1(,, ? 11.5 .1,1 iC.2 5 .L
10.Y IG.b
l&.3 ?b.1’ 18.6 12.(. 1’7.5 2J..2
3 b 5 1> 7 F 9 IL 11 12 13 I& 15 16 II ia 19 26 21 22
IL..? :5.1
57. \ IL. e :t.6
:GL.? 31.3.2
IP.L JL.1. >!.l 19.2 1I.C 18.8
2ti.~ 2e.6 25.6
L,]> ;. K 1.W
3. ~,? 3.85 &.lLl 3.59 L.5
55.9 8.[, L7.5 2?J+.8 101.5 IL.3 ‘?7.2 65.1
Lf.I.5 88.1
8.8 G.& 3.\ 5.6 11.1 8.9
22.;
]r.1
3.:2 1.5G 2.11 5.L8 2.1$ 3. ?9
;9.2 2L.6 16.5 2.?.5
3$.1
116.L
1.: 8 7 6 5 4 3
2
0.1 9 8 7 6
:IF ,! u 2
: 3
2
O.OJ 6 7 6 5 4 3
L
E
Qlt) = Q, 1 l+ bD,,]~ [ Qlt) ‘= Q1 —; ~ D,t 1
RATE
 TIME
EQUATIONS ;FORb>o
FORb=O
0.00
lDd.
D1l’
r Fig. 1  Type curves for Arps empi.
cal
ratetime
decline equations, unit solution (Di .’1).
‘D
9 8 7 6 – 5 4 

8 7
– 6 – – – 5 4 3
3 
2 
0.1 9 6 7 6 5 qD 4 F 3
4 qD
L
\
\]
2
Bsl
0.01 — : 7 6 5 
6
\ ‘J !i”
1;
: 5
4
3
4 –
3 2 2 – ,\,ll,,\\,,,,,~l \
0.001 1
I 2
I 3
I
I
[
[
[
I
I 2
r 34
I
I
I
I
I
I I 2 34 ‘D 5678103 234 5678104
: 2 3
j:;~[,
011
4587810
567810]
Fig. 2A  Dimensionless flow rate functions olane radial system infinite and finite 10,11,15,16 ~uter boundary, constant presure at inner bmndary.
1.0 9 8 7 6 5
1 o“
23 I
456781O* [ I I I I [ I I
2 I
3 I
455781O’ I I I I I I
2 I
34 I 1
56781O’ I I I 1 I
2 I
34 I
567810’ I i I I I I
2 I
3 I I I I I I I
6 5 4 1 3
4 E 3 t 2
0.1 9 8 7
— 
1
2
fj – ‘%
6 5 4
6 5 – 4 qD
–
3
3 :IzIzzl 2
\
–
2
0.01 9 8 7
— –
0.01 9 8 7 6 5
6 5 4
4
3
3 2 2
1 I I I I I I I 2 I 3 I \llllll I 23 ,\ll, 4567810’ llll \ I 23 I i I I r i I
~ I 23
~
0.001 I@
I 2
I 3
L
I
I
1
1 I
I
0.001
4567810S
45678106
45678106
4567810’
‘D
Fig. 2B  Dimensionless flow rate functions plane radial system infinite and finite 10,11,15,16 outer boundary, constant pressure at inner boundary.
lo.t&
I
I
!
[
I
I
I
1
I
I
I
1
I
I
I I
I
I
I
I
I
[
I
I
[
I
I
I
I
I
[
I
i
i
I
I
I
I
I
1
21O,ooi) 4
1.0 i9 8 7 6 5 E 4 ‘Dd 3 \ L L 100,OQO
2 
I
0.1 —
9 8 7 6 5 4 
0.00634 kt
1 )
~;
qDd . ~’Dd
3 
1
I 3 I I I I I I 4567810
2 
.01 10’
L~~:’lllll 34 2
‘Dd
567810”’
567891
I 2
plane radial system, infinite and Fig. 3  Dimensionless flovjrate functions flor finite soter boundary> constant Pressure at inner bo~dary”
,Oln. : 7 6 5 4
23 I
4567s10~ 1 I I
1 ! 1
23 )
!
4567810’3 I 1 1 I I I
23 I
\
45e78~v~ \ I I
I I \
23 1
I
~56781 ! I 1
I
1 I
23 1
1
4567810 [ 1 I
23 1 1 I 1
456781O’ 1 1 I I I 1
23 I I
45678101 I I 1 ) 1 , 
.10 9 8 7 6 5 4 3
2 10.CC.J t Um
1: 0 7 6 5 4
!
9 8 1
2
1.0
...... .. ...
I
Qo.j
U(II
1
.FORb.0
Wd”~=e~
I
I2
01 .t
7 6 5 4 1 3 I ANALYTICAL TY~E_C#RVE SOI.UTION
\\\Y
\\YNsN\\\
3
j‘e 0> 00 00 6., 0
2
2
001 9 8 1 7 e 5 4 3 . 0 .
% 0 . O@ 0 .
001 9 IJ 7 6 5 4 3
2 2
I 0001 ,“.
.
I 2
I 3
I 1 I I I 4567810)
I
I
234
I 1 ( I 1,8
5678 10:
I
23
! ! I [I11 1
4 5.? 870
! .?3 ‘Od
1 1 I .356781
1 1 1
I
~
1 23 1 1 4S 1 1 1 I )\l 67810 1 l\ 23 1 \l 4s6 1 k 7 8 .O> 23 456761W
0 COT
Fig. 4  Composite of analytical and e~irical type curves of Fi&s. 1 and 3.
o
~i~.
5A  Graplical
Np OR Gp
representation of material balance equation,
NPi OR G
I
TR VS Np APPROXIMATELY BEGINNING OF DECLINE PRORATED WELL WHEN DECLINE BEGINS
LINEAR
o
Fig.
N~Gp
5B Graphical representation of mteri al balance
Npi OR G
equation.
23
4567810z
1.0 9 8 7
7 6 5 – 4 EXPONENTIAL
–
–
6
– 5 4
– 3
3 2 co 2.667 2.000 1.714 0.1 9 8 7 — – 1.500 1,333 1.250 1.125 1.050 1,0Q5 2
o
0.375 0.500 0.583 0.667 0.750 0.800 0.689 0.952 i 0.995 3.G.D. $1 –B 7 –6 5 –4 –3
6 
“ g3 u 2 –
flATERIAL
BALANCE
EQUATION:
fiR=.
IATE
Pi — () Npi
NP+P,
2
EQUATIONS: IOIL & GAs): Qo = ‘oi ~R\ T) ([  , ‘R  ‘@j
11$
TwO.PHASE
}01
Cl. it)
0.01 9876
RATE.TIME
EQUATION
[PM
= 01: ~
= ~+1 P“(i) l–
2n+l zn
‘8 7 6
{
SINGLE.PHASE LKIUIO: q. =
5 4 3
JO6R  Pwf)
f
~: 2 i I I [ 2 [ 34 [ I I I I I I 23 I I 4567810 23 45678102 0.001 ~ 2
0.00 1 ,0.3
I 2
I 34
I
I
I
1 I
I
I 2
I 3
I
I
I
I
567810”’
4567810’
56781
‘od
‘(~)’
finite system with constant Fig. 6  Dissolved gas drive reservoir rate  decline type curves effects not included. pressure at inner boundary (pI+T= O @ rlJ). Early transient
1$’ .
87 6 5 4 3 SLOPE n 2“+ 1=;
o
2 0.5 0.6 0.7 0.8 0.9 0.1 9 8 7 6 ~..5 824 ,1 S3 u 2 1.0 10.0 100.0
1 .mo 2.000 2.200 2.404) 2.600 2.800 3.mo 21.000 201.000
0.454 0.365 0.357 0.333
7\
1.000
b
–
–
3
\
0.500 0.417
2
—
0.1
0.046 0.005
– –
7 6
—
MATERIAL BALANCE ‘R’= IATE (*) ‘“+””
EQUATION:
EQUATIONS:
TWOPHASE [OIL&GAS): 0.01 : 7 6 5 4 3 SINGLEPHASE RATE.TIME EQUATION fexwnentiall RATETIME EQUATION (“~
GO = Jo,
q. = O): ~
T
6R ~ ~R’ ‘W* ( )[ (t) t+l w)’ GO = JO (6R  Pw$ (t) “ — .(i)’ I 1 I I
,!/
\ %4 n 5 !1 – 2n+l 1 – – – – 8 7 6 5 4
LIQUID:
l~f
q. = 01: ~
2
\
0.00
1 23 I 1 I i 2 I 3 I 4 I I
\l,l
10”’
I
23
I
I I I I
4~6761 2 3 4567810 23 45676102
45678103
5678
‘~~
(%3’
Fig. 7  Dissolved gas drive reservoir rate  decline type curves finite s~stem with constant pressure at inner bwndary (ph~ = O @ rl,). E:rN transient effects not included.
6114,1,
.9
u
6
I
. . .
I
—
‘P td
UI.. 0000O oo1+v~g
r+
g
CJo
II
1
co
.,1
G
L4
I
1U3
1.0 8 7 6 5 4 
2
3
4567810+
2
3
4
567810”’
2
3
456781
2 I
3 I
4567810 I I I I 1 I
23 [ I
4 [
567810~,o I I I I I – – – – 9 8 7 6 5 4 3
3 – 2
‘t
\\
\
\2
0.1 L
\
\
\.F
%
o @.
0.1
9 8 7 6 5
0 ,%
\\\\ ; 7 6 5 4 3 
2 
I
O.CO1
,03
I 23
I
I
I
I
1 I
II
I
2
I
3
I I I 1I I
4567810’
~
I 2
I 34
I
45678101
i \\,\,,l,,J
4
i
“
7
2
~
:01
7
6
5
4
3
2
I
I
56/81
..
illl
\
I
0.001
234
567810
23
46678102
with back pressure finite system with constant Fig. 9  Gas reservoir rate decline type cur~es = effects not included and pressure at inner boundary (pVfl constant Z r,,). Early transie,;t z = 1 (based ~n gas uell back pressure curve slope, n = 1.
‘Dd
= 0,1
1000
EXPONENTIAL
—
I
MATCH POINTS EXPONENTIAL, t=loo b =O EXPONENTIAL, b= 1
MO, ;tDd=l.lo
t= looh.hO. ;tDd=l.60 J(t) = 100 BOPM ; qDd ‘ o.2~2 ECONOL~lC LIMIT 20 BOWVIm 1480 MO. EXPONENTIAL
q(t) = 100 BOPM ; qDd = o.212 ECONOMIC LIMIT 20 BOPM @ 285 L1O.
\
10
100
1 285 fI~O.
1000
–L 1480klo,
FFIACTURE TREATED
BEFORE FRAC
\l
AFTER FRAC MATCH POINTS
()
J
\, “
—
I
‘“h
\ \ AFTER FRAC
n =
BEFORE FRAC
n = 1.0, (b = 0.667)
\ \ 1.13 = o.134 \ \ \ \ \
1.0, (b = 0.667)
t = 100 MO. ;tD~ = 0.60 q(t) = 1000 BOPM ; qDd = 0.243
t= loo MO. ;tD,j=
q(t)1000 ‘ BOPM ; qDd
1
Fig. 12  Typecurve
10 t MONTHS
analysis of a st imulatw! well before
100
and after fracture
1
treatment.
1000
...
Ioo,ooo
Oq’= q’+q’
/
EXPONENTIAL
\
/
>
$ , 10,03(
$’
_ b=O, MATCH POINTS EXPONENTIAL LAYER 1 t=loy Rs. ;tDd =2.0 LAYER 2
qz=’+ql
Cl(t) = 1000 BOPY ;qDd = 0,017 LAYER 2 t= loy Rs. ;tDd =5.35 ; qDd = 0.020
q(t)1000 = BOPY
q
1,00
1
t  YEARS
~ig, 13  Tj’pecurie anal:.’si of .s layered s reser,mir
10
(no crossflm.:) t,:: ffcrcncin~. di
100
+ Q(2)
< * \ \* \ \ \ \
=2
\
10
100
lUW
.. .
t  DAYS
Fig. 14  Effect of a change in back pressure on decline using graphical supe~osition.
11
1
I
‘ Dd
I
I
I
i Fig. 15  Tj~ecurve xatck.ng e
F :,., ..>
:’: .
..
L~,,;; . . .. . .
Pressure
at
0.00634 kt
‘3L ‘ ~ucl ~Z
10 0.1
1
10
Iw
t
HRS.
Fig. 16  Typecurve ::atcking example f~r calculating Kb from pressure buildup A data, :Jell13, ~iel[i (type curve from Refl.9).