Beruflich Dokumente
Kultur Dokumente
SPE,
and
F. Samaniego-V.,
SPE,
U. of Mexnm/Pemex,
and F. Rodriguez,
SPE,
U. of Mexrco/IMP
sP~
130S9
Summary.
The themetical basis for the pseudolimar-flow mcdel is established. It is demonstrated by use of an analytical model that
the lin--flow graph (p vs. W) cm be extended to the analysis of pressure data of fractured wells intersected by an intermediate- or
high-conductivity fractme (CP >57r). It appears that the fracture-conductivity effect during the pseudolinem-flow pecicd can be handled
as pseudoskin pressure drop that adds to the pressure drop cawed by tluid-Ioss damage. The combination of the pseudolineac-flow anafysis with other interpretation techniques is illustrated through examples of field cases.
Itrtroduetlon
During the last 2 decades, many analy$is methcnlsl-30 to iaterpret
pressure data to estimate both formation and hydraulic fracture pamneters have been prop.med. In adtiltion to type-awe analysis,
three of the most widely used graphical metbcds of interpretation
are the- linear-flow graph (p vs. J), the bilinear-flow graph (p
vs. t*), and the pseudoradial-flow gmph (p vs. log r). ft cannot
be overemphasized that each technique appIies to a specific flow
regime. Unfm-htmtely, the general pressure behavior of a fractured
well includes not only these flow periods but also several intermediate transition flow periods, as shown in Fig. 1. The only current
graphical analysis methcd avaitable to interpret pressure data falling in the transition periods is the typwrve-matching
technique.
The linear-flow mudel has km applied in the past to early
time/pressure data of iniiite-conducdvity fratires (for practical
purposes, a finite-conductivity fracture can be considered as an
Mtite-conductivity-fmcture case whmever the pressure dmp along
the fracture is negligible-i.e., C kr)a 31J2).This model assumes
uniform flux in the formation an i nezllmble
---~~ stom~e ca!mcitv
.
. within the fmctare.
The application of the li&r-tlow graph was extended empiricaffv to cases where the fracture has low or intermediate values
of c_onductivity.29 This r&@res either knowledge of the formation
permeability frum prefracture testing or a ti-ial-and-errur prueedure.
f.fa&y and Bandyop2dbyay30 used a uniform-flux Z@ticti
mudel to show that the pressure behavior of a ftite:cunductivity
vefiical fracture at intermediate values of tine can be described
by the pressure behavior of the War-flow mcdel plus an extra pressure drop that is a function of the fracture conductivity. They slso
a.wumed tit the fracture penetrates the formation completely in
the horizontal direction (X,lX$=1).
The purpose of this work is tu show through tie analysis of an
irnafydcd solution for tinite-conductivity fractures that the )ine2rflow mcdel maybe extended to cuses of h@ fracture conductivity
without assuming a uniform flux along the fracture. Furthermore,
analysis of field cases illustrates the application of the pseudolinearflow mcdel and its combmtiun with other interpretation techniques
tu obtaii reliible estimates for fructure and reservoir parametem
PseudrrIhrear.Flow
Model
ApWD=f(@,,CD).
Cq@ght
438
fiacmre ~nducti~v,
%.
2S ~s~~
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...(1)
!989Sccl@ly
ofPelroleurn
Eng!nws
For the pericd including the btiear flow, Transition Zupe 1, and
the hear flow, Q. 1 can be expressed in the bpkce space as2731
d3(pwD)=m cOril{@s/(cfl)2]
* }Acvzsl(cp)zl
x,
(2)
where d3 (pwD) is the Laplace transform of the dimensionless wellbore pressure and s is the Laplace space parameter.
The hyperbolic cotangent fiction may be mitten as an infinite
serie3.32
cool(x)
+[2&B&-1/(2n)!]
+ . . . . . . . . . . . . . . . . . . . . . . . . . ...(3)
Hence,
s(pwD)=T/sc@ti[s/(cp)21
+(~/3)[s/(cD)2]u
x (1/ti[s/(c*)Zl
1/45{@/(C@)2]
}3 + . .).
(4)
At large values of time, s-O and the series in RI. 4 can be approximated by use of two terms:
dXpwD)=m/2sK+m/3c@ . . . . . . . . . . . . . . . . . . . . . . . ...(5)
faplace inversion of this equation yields
pwD=~~(~/3C~).
. . . . . . . . . . . . . . . . . . . . . . ...(6)
The fist term on the right side cmresponds to the classic linearflow expression discussed extensively in the literature. The second term is a result of the fnite conductivity of the ti-acture. HauIey and Bandyopadbyay30 presented a similar equation; the only
difference is that their numerical constant in the secund term is unity
instead of the u/3 constant. It is obwousthat this equation also incfude.sthe infite-conductivity case for which the second term on
the right side goes to zeru.
By multiplying Eq. 6 by CP, we obtain
PwD(Cj)=d%fD(%)2
fbis equation
indicates
+(d~j.
. . . . . . . . . . . . . . . . . . . . . (7)
that a
,:W
~:
100 L,,e
1 ..
-.
,.
BUN;AR-
,rJ-z,..,
Fig.
]~-z
,~o
15
,@
I,.(cr.],
104
,.6
10
Q
o
TRANS!-W~
FLOWPEQ!ODS
100T
~08
5J=Y
i
Ig. 3Linear-flow
fracture.
I
cm , 5T
qf~
twD
lg. 2-Finite.conductlvify
vertical
fracture
[n an infinite
mewolr.
i.:
conductivity fracture. Tfds is not so for ftite-conducdvity fractures, as shown in Figs. 4 and 5. These graphs present the flux
distibwion for CP equal to 5X and 10T and for values of time
when E+ 7 applws. The flux distribution tends to be mom uniform as flowing time increases. his information was obtained by
use of Cinco-Ley et als 16mcdeI. .4kbougb the pressure behavior
under these conditions is similar to the linear-flow behavior. it is
not the same because of the dtierence in flux distributions, Consequently we propose that the pressure behavior exhibited by fjdte.
conductivity fractures be called pseudolinear flow.
Fig. 6 shows a heavy line representing the pseudolinear-tlow behavior. The beginniig of thisflow period occurs at a value of
t.r/D(C@2 = Ii.e., one log cycle of time after the end of the
bdineir flow and two log cycles before the start of the apparent
one-half-slope straight line, The pseudolinm-flow wricd includes
part of the transition mne between the one-qmwter-slope straight
line and the on&Mf-sloIE straight linq therefore, a theoretical basis
exists (Eqs. 2 through 7) for the extension of the linear-flow graph
to cases of intennedtie and high fractore conductivities for pressure data falfing within the transition zone, which can be analyzed
only by type-curve matching with currentfy avadable metbcds.
0
0
?g. 4Flux
0.2
0.4
distribution
A+=mL&+bLf,
during
the
0.8
XD
pseudoilnear
1.0
flow;
. . . . . . . . . . . . . . . . . . . . . . . . . ..(11)
, . . . . . . . . . . . . . . . . . . . . . . . . . ..(12)
for gas wefls. FLfO, FLti AL@ and ALfi are unit conversion
factors.
Eq. 8 indicates that a graph of pressure data vs. the square root
of time during the pseudoliiea.r flow should exhibit a straight line
of slope mLf and intercept b=f, as shown in Fig. 7.
According to Eqs. 9 and 11 for oil wefls and Eqs. 10 aod 12 for
gas wells, the slope is inversely proportional to fmctuze area hrf
and the intercept is inversely proportional to C)/.xf; hence,
*
. . . . . . . . . . . . . . . . . . . . . ..(13)
Analysis
0.6
:,D =5T.
krf= (FLfo@/nI~f)(@@cJ
Pseudollnear-Flow
.. _.~
. . . . . . .. . . . . . . . . . . . . . . . . . ...(8)
~ . . . . . . . . . . . . . . . . . . . . . ..(15)
f . . . . . . . . . . . . . . . . . . . . . . . . . . . ..(16)
...............
. . . . . . (17)
439
::
:4
C,D= ]0 T
lfD
-.
FLOW
10+
10+
,.-2
,.O
,)2
t.,.(m),
,,)6
104
,~
I
o
Fig. 6-PseudolInear-flow
0.2
0,4
0.6
0.8
pressure
behavior.
1.C
XD
lg. 6-Flux
distribution
during
the
pseudolinear
flow
:m - 10W.
siooless form
~rtbpLf=(9/m)(bL,ImLf)
Z . . . . . . . . . . . . . . . . . . . . . . . . . ...(M)
and Limitations
PwD=q+%+(7/3cp).
. . . . . . . . . . . . . . . . . . . . ..(19)
of ffie defi-
. . . . . . . . .. (20)
L_l
............
(21)
()
bu a++
I
c
Fig. 7Linear-fiow
havior.
440
pressure bs-
.wEFormationEvdualion.Swemb=r
~~9
3
mLf=i.otiXlo4pSi2
Cp-l MSCF~-l~fA
2
mw
1
..
I
. .
.-
/,
0
} b~f
50
100
150
200
fi,hours
Fig. 9-Billnear-Now
TABLE 1TEST
example.
INFORMATION
Fig. 10Linear-flow
AND RESULTS
OF ANALYSIS
example.
FOR CASE 1U
Reservoir Data
5,100
1,275
0.0027
56
0.085
725
0.0254
p{, psia
:tim~:a
h: ft
*OR
pd. Cp
Cg,, psi1
Results
C,lxt, md-fflft
1.465 XI0-4
Type
Curve
Analysis
0.103
131.2
13.5
x,~k, ft-md *
C,W
rnd x-n
Pseudolinear flow
0.0965
140.5
13.0s
13.6
>107T
50.00328
%
k, md
C,, md-ff
2236
>2,291
xl. ff
TASLE 2TEST
Siiinear Flow
INFORMATION
AND RESULTS
OF ANALYSIS
FOR CASE 2=
Res6tvoir Data
q, Mscf/D
7,350
tp,hours
2,640
118
0.1
0.023
690
0.0252
h, ft
fp
~ OR
Pg. OP
ct. psi-f
PM, pda
Analysis Results
C,, md-ft
x,wk ft.md k
C,/x,, md-tift
cm
Type Curve
14/3
58.9
0.3978
15.9
0.129 x10-3
1,320
Bilinear Flow
154
0.4;85
16.74
Pseudolinear Flow
137
54.6
0.427
17
)02
HLil\lIIIl;l
107 t.0_lo5
,.-2
1_10:2___l
,.-1
Fig. 1I-lype-cttrve
LO-l_
I
matching
L_
=(1,491.2x
=9.65 x10-2
lxf){fi(t)M1+Pci[tx fD(cD)21M}
,%
2.637x 10-4 X 1
( 0.35 X0.0264 X1.465 XIO-4X1.2X10-3
and C#i=xfl(Cf/Xf)
1x725)/(56x2x
105)
md,-timl.
s 0.tM328 md,
X104)2
cf2 c*/dtlhzi
and Xf>Xfl/_
ft [698 m].
graph for this example. The hiieactlow straight line includes ~ints up to ==40 hours; it Iras a slope
m =8.3x104 psi2/cp-Mscf-D-hr A [139.3 X 1012 kpazfpa. s-std
mY.d.h] and goes through the origin. Application of the bilinearflow-anatysis equation yields
Fig. 9 shows the bti-tlow
Cfl=[Fb~qT/mbf
Eq. 18 can be used to check whether the proper straight line has
been drawn. The stwt of the pseudolinear-tlow period occurs at
kS(Cf/xf)(lllOr)
22,291
_.
At [ho. rsl
=0.103
pl
102
Cf/xJ=ALkqT/hbLf
@hDcjD)Mlh(PFdM
fi=(cf
(+#c,) kh]z
2
444.75 X 1 x725
X 1.465 X 10-4)*
x56
]01
for
_I__l_
t.tD(cfD)2
=FLtiqT/mLf h(@cJ
Case 2. This example represwm Wefl A in Ref. 27. A pressurebuifdup test was run in an MI-IFgas well for 120houm after a flow~g @d
of 2,1W ho~. The information for this testis presented in Table 2.
Fig. 11 shows the application of the type-curve-matching technique. As can be seem, both bilinear and pseudoliiear-flow pericds are present in the test. 71e w curve md bilkwflow
graphs
are shown here; the analysis and a canplek discussion are presented
in Ref. 27.
Fig. 12 indicates a welldefied bibw-flow
straight line that
goes through the origin of the graph. Fig. 13 presents the linearflow graph and shows that pressure data exhibit a straight-line portion where slope mLf =5.65x107. psi2/cp-ti% [26.86X 109
kpa2~a.S.h!4J md intercept bLf= 15 x 107 psizkp 171.3x 109
kpa2/Fa.s]. Application of the pseudolinear-flow @YS~ Yield3
40.925x7,350x690
x+=
5.65x107X 118 x(O.1 xO.02.32X 1.29 X 10-4)X
AA7
md%t
10-4)*
SeWernber 1909
,0
t
.,~~
;
40
mL,=563xf07psi2c7
F
!L
02468
m,
lg. 12BObtear-llow
1.s==7example.
and Samaniegl
1,491.2 [418.361
.4Lfi = unit c0nver3i0n factor for linear flow of oil,
147.86 [1928.9]
b = intercept
bf = fracture width
bpL = intercept on Cartesian pseudolinear graph of
c! = 3YS@ItI total
Cf =
C@ =
f=
F =
Ftig =
Fig. 13Linear-flow
V.SZ7
graph
10
how,
for Cinco-Ley
and %maniegc
example.
FM.
.,. = unit conversion factor for btiear flow of oil,
44.1 [34.971
FLfi = unit conversion factor for linear ffow of gas,
40.925 [11 .498]
FLfO = unit conversion factor for Iinear flow of oil, 4.064
[53.01]
h = formation thickness. ft M
k = permeabfity, md
$(p.D) = Laplace transform of dimensionless weflbore
pressure
m = slope
IIZPL= S1OF of stight liie for pseudoliiez flow,
psizkp pcPa2/Pa.s]
p = pressure, psi m]
Ap = pressure change, psi P@
APWD = dimensionless pressure drop
~ = ~eff floW rate, STI+V [stock-tank m31dJor
MscflD [Std m3/d]
~ = sti fa~or or fapkice space variable
t = time, hours
tX,D = dimensionless time
T = reservoir temperature, R KI
x = variable used in Eq. 3
XD = dimensionless variable
j+ = ~xte~
mdls
f = fia*@ fdf-length in the z dii~fion
a = pressure (or pseudopressure) unit convemion
f3.ctOr
% = Pr~~
(OI pseudopressure) unit conversion
factor, 1,424 [399.5]
~. = p~s~~ (or psedopressme) nit mnvetsion
factor, 141.2 [1S42]
i3 = time unit conversion factor, 2.637 x10-4
g = fluid viscosity, cp pa.s]
.$ = p6r0sity, fraction
+ = real gas potential, psizlcp Ma2/Pa.s]
A* = real gas potential change, Psizkp &Paz/Pa.s]
Subscripts
b = bilinear
bf = biliiear flow
D = dmensiordess
e = external
f = fracture, flowing
g=ga3
h = hydrocarbon
443
~ = ~ti~
L= linear
Lf = Jinear flow
M = match
~=oiJ
p = production, prcducing
PL = pseudoJinear
s = dmage
w . wellxxe.
,.
Snpmaipt
= average
1983) 219-30.
H. and Samaniegc-V., F.: Transient pressure Analysis
27. Cinmley,
Referwrcss
1. S@t, J,O.: The Effect of Vertical Frmiwres cm Transient Pre%ure
B&vim of Wells,,s JPT @cc. 1%3) 1365-69; Trans., AR@ 22S.
2. Russell, D.G. and Tmitt, N.E.: Transient Pressure Behavior in Vertically FraduredP.&xvoim,JFT(&I. 1954) 1159-70; Trans.,AIME,
231.
3. Lee, W.J. Jr.: AmJysis of HydmuJicaflyFracturedWellswithFrwsure BuildupTess,>, pqer SPE 1820 presented at the 1967 SPE Annual Meeting, Housmn, Oct. 1-4.
4. Clark, K.K.: eTransient Pressure Testing of Fmtwed Water Jnjt%tion Wetls,v JPT (June 1968) 639-43; Trans., AJME. 243.
5. Milfheim. K.K. and Cichwuicz, L.: iTesdng and Analyzing LowPermeabilitv Fractured Gas Wells, JPTITeb. 196S) 193-9ti Trans..
AfME, 243-.
6. Watfwdmrger,R.A. andRamey, H.J. Jr.: Welf Test Jmerpretation
of Vertically Fractured Gas WeIfs,, .JPT (T&Y1969) 625-3Z Tram.,
AJME, 246.
7. van EverdinSen, A.F. and Meyer, L.J.: %dysis of Buildup Curves
Obtained After Well Treabnent, JPT(April 1971)513-24 Trans.,
AIME. 251.
8. Raghavan, R., Cady, G.C., and Ramey, H-J. Jr.: kWeUTest Analysis for VeniwJJy Fractured WeJls, JPT(Aw. 1972) 1014-21JTrans.,
AJME, 253.
9. Gringanen, A.C., RameY,H.J. Jr., amAffaghavM, R.: UnsteadyState Pressure Distributions Created by a Well With a Single JrdiniteConducdvily VerdcaI Fracture, SPEJ (Aug. 1974) 347-Q Trans.,
AJME, 257.
10. Gringarten, A.C,, Sanmy, H.J. S,., amdR@avan, R., ,A@ied PreswreAnalysis for Frac.turedW.US,,,JPT(IUIY 1975) 887-92; Trm.,
AJME, 259.
11. P.amey, H.J. Jr. and Gringarten, A. C.: &EHect of 3tigb-Volume Ver-
444
T coth{(I/cfD)[2/(
lA@ +C2hr)(sfi)ll ~ ,
scp(l/cfi)[2/(uw)
. (A-l)
+(zhr)(sfi)] *
. . (A-3)
=m3
= Pa.s
= m
= m3
= pmz
E+(Y3 == J&a
E-01
= kPa-l
E+OI
= kPa2
E+WI = K
.Cunvwabnwor 1sexact.
SPEFE
Ofigl.ti2FEnmnutipt r=61vdIorrewuvi
Sept.16,1934.Paperacceptedforp.bllca
12059) flint
In Hou8M.,
W-19.