New Skin and Wellbore Storage Type

Curves for Partially Penetrated Wells

Fikri J. Kuchuk, * SPE, Sohio Petroleum Co.
Paul A. Kirwan, * * Sohio Petroleum Co.

Summary. An analytic solution is derived for the transient pressure behavior of a partially penetrated well when wellbore
storage and skin effects are significant. New type curves, generated by this solution, enable the engineer to analyze well test data
from partially penetrated wells and to discriminate between near-wellbore damage and pseudoskin caused by partial penetration.

Introduction
Partial completion of wells to inhibit gas and water coning has been make the selection of a straight line very difficult; at worst, it can
common practice in the petroleum industry for many years. Numer- represent a breakdown of the radial flow assumption, which is an
ous studies, predicated on a variety of assumptions, have inves- integral part of the Horner method. Some of these problems were
tigated the theoretical pressure response and productivity of such recognized and discussed in two case studies 7 ,8 of buildup test in-
wells. However, very few of these studies have addressed the prob- terpretations in Prudhoe Bay field in Alaska.
lem of interpreting early-time transient pressure data from a par- The purposes of this study are (1) to present an analytic solution
tially penetrated well when wellbore storage is significant. that describes the transient pressure behavior of a partially penetrated
The value of early-time analysis for fully penetrated wells has well flowing at a constant rate in an infinite reservoir, and that takes
been discussed by several authors. 1-5 There are a number of rea- into account wellbore storage, skin, and permeability anisotropy;
sons why a reliable model for interpreting early-time data is espe- (2) to present a set of new type curves for partially penetrated wells
cially needed in partially penetrated wells. The transition period in the Gringarten format; (3) to compare the above results with
preceding radial flow in a partially completed well can display previous studies and to emphasize certain salient differences; and
markedly different characteristics from its fully completed coun- (4) to show by example how the new type curves may be used to
terpart, even if the apparent total skin in the two systems is identi- estimate the various components of skin in a saturated reservoir.
cal. In other words, pseudoskin, caused by partial penetration, does
not have the same effect on early-time data as damage skin, which
is conceptually associated with an infinitesimally thin region around Previous Work
the wellbore. This is an important observation that provides the Streltsova 6 used the McKinley format to present type curves for
practicing engineer with means of identifying the various compo- partially penetrated wells with wellbore storage. As in the original
nents oftotal skin, and of verifying conclusions drawn from analy- McKinley approach, 3 all calculations were made for zero wellbore
sis of data obtained after radial flow is established if such data are skin and a single value of the diffusivity parameter, krltjJp.c t • on
available. the basis that this parameter exerts less influence on pressure
In many instances, however, radial flow is not established with- response than transmissivity, krhlp..
in the duration of a well test. When wellbore storage is significant, The Streltsova solution was obtained by assuming a uniform-flux
the time it takes to reach the end of the afterflow effects is depend- condition at the wellbore and using the McKinley wellbore storage
ent on the total skin of the system. Because of the large total skin (afterflow) algorithm.3 Bilhartz and Ramey 9 and Gringarten and
apparent in many partially completed wells, the time to the end of Ramey 10 have argued that a uniform-potential (infinite-conduc-
afterflow can be prohibitively long. Cost or operational restrictions tivity) inner-boundary condition is a better representation for the
can make it impractical to run tests long enough to determine radi- partially penetrated wells than a uniform-flux condition. The differ-
al flow. In these circumstances, early-time analysis is the only ap- ence in early-time pressure behavior resulting from these different
proach available. inner-boundary conditions will be seen later when the Streltsova
Analysis of early-time data may prove to be of particular value solution is compared with results from this study.
in interpreting results from tests in saturated or near-saturated reser- In 1975, Gringarten and Ramey 10 investigated the infinite-
voirs. Pressure drawdown in a flowing well causes the evolution conductivity boundary condition in some detail. By considering the
of solution gas within the vicinity of the wellbore. Oil mobility then flux distribution along the perforated interval in a partially penetrated
becomes a function of distance from the well. The Horner analy- well after radial flow is established, they were able to define an
sis, if feasible, will yield the total skin resulting from well bore "equivalent average pressure point." This represents the location
damage, completion geometry, and the modified transmissibility in the vertical direction at which, after radial flow is established,
around the wellbore caused by the presence of free gas. Only by the pressure obtained by assuming uniform flux at the wellbore is
including an interpretation of early-time data can one hope to iden- equal to the expected wellbore pressure for a uniform-potential con-
tify the different components of total skin. dition.
Moreover, the problem oftest interpretation in saturated reser- Bilhartz and Ramey 9 used a finite-difference model to inves-
voirs is often confounded by the presence of a gas cap, which can tigate the combined effects of wellbore storage, skin, partial penetra-
cause two complicating factors. First, a large standoff that inhibits tion, and vertical anisotropy on transient pressure behavior,
gas coning can lead to very low penetration ratios. Second, if a assuming uniform potential at the wellbore. These numerical simu-
well is in direct communication with a gas cap, a Horner plot of lation results were in good agreement with the analytic solution
buildup data may exhibit a pronounced curve flattening that may presented here.
adversely affect the semilog straight-line period. 6 At best, this can Chu et al. 11 recently presented a method for analyzing transient
pressure data dominated by wellbore storage and skin. They sug-
gested that available type curves for fully penetrated wells can be
'Now with Schlumberger·Doll Research. used for partially penetrated wells by use of a simple coordinate
"Now with BP Petroleum Development. transformation. The implications of this transformation are discussed
Copyright 1987 Society of Petroleum Engineers later.

546 SPE Formation Evaluation, December \987

 CLOSED UPPER BOUNDARY

0,75

0,74

Zo
OPENED 0,73
INTERVAL

0,72

0.71

CLOSED
INTERVAL O. 70 '::---'-----'----''--'--'--'---'-:-_ _--'-_ _'------'----'---'--'-~.LJ

'w CLOSED L'OWER

_00 30 100 1000

BOUNDARY
WELLBORE
RADIUS Fig. 2-Location of equivalent average pressure point at the
wellbore, at which the pressure Is computed from the uniform-
Fig. 1-Representation of a partially penetrated well. flux solution.

Mathematical Model Dimensionless radius:

The well model shown in Fig. I is considered to be partially com- lir
pleted in a horizontal, homogeneous, vertically anisotropic forma- rD=-'
tion. The flow of a slightly compressible liquid with a constant rw
viscosity is assumed. The top and bottom of the formation are Dimensionless effective average pressure point:
assumed to be closed. A uniform-potential condition (infinite con-
ductivity) is assumed at the wellbore. The dimensionless pressure • Z
•
solution for this system without wellbore storage and skin was ZD=-'
presented by Gringarten and Ramey 10 (Eq. 13) as hw
Applying the Laplace transformation to Eq. 1 yields

PD(tD, rD, zi»= J

o
rID [ 2
1+- L:
00 I 2
_e-(mrblh wD ) T
trb n=1 n
l[
PD(S, rD' zi»=- Ko(rD.[;)+-
2 L: -1
00

s rb n=1 n
2 dr

[rD~s+C:: yJ], ...... (2)

Xsin(n7rb) cos(n7rbzi» ] e- rvl4T - , .................. (I)
2r xsin(n7rb) cos(n7rbzi) Ko

where the dimensionless variables are as defined here. where

s = Laplace image space variable and
Dimensionless wellbore pressure: Ko = modified Bessel function of the second kind of zero
order.
Dimensionless pressure values can be calculated from Eq. 2 with
the Stehfest 12 numerical Laplace transform inversion algorithm.
Results were checked vs. values given in Refs. 9 and 10. There
Dimensionless time: is good agreement between the numerical Laplace solutions from
this study and the solutions from Ref. 10, especially for tD > 3,000,
where the two results were identical to four significant figures. How-
ever, the summation given in Eq. 2 converges very slowly because
of its alternating nature. Thus, either a few thousand terms must
be summed or a method for the acceleration of convergence such
as Euler's transformation has to be used.
Penetration ratio: Dimensionless wellbore pressure with storage and skin for par-
tially penetrated wells can be obtained with the superposition
theorem that is given by van Everdingen and Hurst!3 and Agarwal
hw et al. 2 as an integro-differential equation of a convolution type:
b=-.
h

Dimensionless wellbore length:

+S[l_CDdPwD(tD)], ..................... (3)

dtD

SPE Formation Evaluation, December 1987 547

 100r--------------------,r--------------------.--------------------~

b 0
0.5
hwO = 250
COe 2S 10 15
ANAL YTiC SOLUTION

'i b 0.5
Q.

ui
hwO = 500
a: 3
J
<f)
10
<f)
w
a:
Q. 10
<f)
<f)
W b 0 0.5
...J
Z hwO 1000
o 0

iii
z
w
c Oe 2S = 10 2
::E
2i

10 100 1000

Fig. 3-Comparison of numerical and analytic uniform potential solutions.

where Eq. 4 is a general solution that may be applied to a range of reser-

PwD = dimensionless wellbore pressure with wellbore voir engineering problems. It provides a means of incorporating
storage and skin effects and the effects of wellbore storage and skin if the solution for the
PD = dimensionless wellbore pressure for the constant-rate constant-rate case with zero wellbore storage and skin is known.
case without storage and skin effects. Eq. 4 with Eq. 2 gives the dimensionless wellbore pressure for
a partially penetrated well flowing at constant rate when wellbore
The equation for dimensionless wellbore storage is storage and skin are significant. The computation of z~ was given
by Gringarten and Ramey, 10 and is shown as a function of hwD
5.6146C in Fig. 2.
Eq. 4 was transformed from the Laplace image space by use of
27r</>ct hr2w ' the numerical Laplace inversion technique described by Stehfest. 12
Results are compared with the numerical solutions presented by
where S=skin effect.
Bilhartz and Ramey 9 and found to be in very good agreement. Fig.
Replacing the PD term in Eq. 3 yields the dimensionless pres- 3 compares results obtained from Eq. 4 with results from an in-
sure for partially penetrated wells with wellbore storage and house, numerical simulator that incorporates the inner boundary
skin. 14 However, the result is an iterative convolution integration
conditions defined by Ref. 9.
that takes a considerable amount of time to be evaluated. Much of
this complexity can be eliminated if the substitution is made in
Laplace image space. 13 ,15.16 The Laplace transform of Eq. 3
yields Estimation of Pseudoskin Resulting From Partial Penetration.
A large-time solution can be obtained from Eq. 4 by allowing S
SPD(S)+S to approach zero (tD ..... oo). Thus,
PwD(S) = ..................... (4)
s{l +CDS[SPD(S)+S]}
I [ S 2 00
PwD(s)=- -[16 In(s)-ln(2)+'Y]+-+- ~
TABLE 1-COMPARISON OF PSEUDOSKIN VALUES S b 7rb n=1 n
FROM THIS STUDY WITH STRELTSOVA 18
AND BRONS/MARTING 17

hWD =50 • ) Ko (n7rb)]

xsin(n7rb)cos(n7rbz D - - , .... : ........ (5)
Relative Relative hwD
Error Error
b This Study Ref. 18
----- ~ Ref. 17
----- ~
0.1 34.911 36.067 -3.3 34.898 0.03 where 1'=0.5772 ....
0.2 14.646 15.222 -3.9 15.686 -7.1 Inversion of Eq. 5 yields
0.4 4.891 5.169 -5.7 4.692 4.0
0.6 1.871 2.033 -8.6 2.949 -57.6
0.8 0.537 0.614 -14.3 1.034 -92.5 S 2 00

hWD= 100 PwD(tD)=Y2[ln(tD)+0.80907]+-+- ~

b 7rb n=1 n
0.1 41.440 42.256 -2.0 41.137 0.73
0.2 17.563 17.970 -2.3 18.453 -5.1
0.4 6.002 6.196 -3.2 5.732 4.5
0.6 2.377 2.487 -4.6 3.411 -43.0 . •
xsm(n7rb)cos(n7rbzD)K (n7rb)
0.8 0.737 0.782 -6.1 1.207 -63.7 o - - ............... (6)
hwD

548 SPE Formation Evaluation, December 1987

 30 TABLE 2-COMPARISON OF Po's FROM THIS STUDY
WITH REF. 18. SOLUTION FOR b = 0.3 and h wO = 200
28
Po Relative
26 This Study Ref. 18 Error

24
to/CD (uniform potential) (uniform flux) ~
0.6901 0.669 0.664 0.74
6.901 5.472 5.489 -0.31
22 69.01 16.911 16.997 -0.50
690.1 18.663 19.122 -2.47
20

18 presented in Ref. 17 were obtained by assuming uniform flux at

Q.
til the wellbore and then computing the wellbore pressure at a location
C 16 three-quarters of the distance along the open interval. The values
:ii! presented in Ref. 18 were also obtained by assuming uniform flux
II
0 14 at the wellbore, but the resulting solution was then integrated along
"::s the open interval to obtain an average wellbore pressure. It can be
IL 12 seen from Table 1 that the pseudoskin values from Ref. 18 are con-
sistently higher than the values obtained from this study. This is
10 a direct consequence of the difference in boundary-condition as-
sumptions; the uniform-flux solution gives larger values of dimen-
8 sionless pressures at large times than the infinite-conductivity
solution. This is confirmed by Table 2, which compares dimen-
6 sionless pressures from the Streltsova* solution with values obtained
from Eq. 4. In fairness, it should be noted that for many practical
4 purposes, the absolute differences in pseudoskin values between
the studies are small.
2
Effects of Pseudoskin and Damage Skins on Transient Behavior.
oL-~~~ __L-~~~~~~~ Fig. 5 shows the dimensionless pressure solutions for two well-
o 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 defined systems. The first is a well with a penetration ratio, b, of
Penetration ratio, b 0.3, a dimensionless wellbore length, hwD' of 1,000, and a damage
skin of 5.0; the second is a fully completed well with a damage
Fig. 4-Pseudoskln from study as a function of band hwo'
skin of 31.96. The total skin associated with the first system is also
31.96 (St=S/b+Sp ), so the two systems have identical values of
total skin.
The first term in Eq. 6 is the familiar large-time approximation As might be expected, the two solutions converge at small and
of the line-source solution. We can therefore immediately deduce large times. However, there is a significant difference in the tran-
that the pseudoskin resulting from partial penetration, Sp' is given sition period preceding radial flow. The difference between the two
by systems is more pronounced in the log/log pressure derivative plot
that was proposed by Bourdet et al. 19 than in the log/log pressure

~ !.. Sin(mrb)COS(mrbz"1»Ko(mrb) .
plot, as can be seen from Figs. 5 and 6. The early-time behavior
Sp=.2.. ........ (7) of partially penetrated wells deviates from a unit slope earlier and
~b n=l n hWD reaches radial flow later than its fully completed counterpart. Thus,
pseudoskin resulting from partial penetration does not have the same
Fig. 4 shows pseudoskin from Eq. 7, expressed as a function effect on pressure behavior as damage skin, conceptually associat-
of penetration ratio, b, and dimensionless wellbore length, h wD ' ed with an infinitesimally thin region around the wellbore.
Table 1 compares values of pseudoskin from Eq. 7 with values ob- 'Personal communication, T.D. Streltsova, Exxon Production Research, Houston,
tained by Brons and Marting 17 and Streltsova. 18 The values Feb. 1982.

10'

0
C

-
i5
0
a.
~

1:1

- Partially penetrated
o Fully penetrated 10°
- - Partially penetrated
........ Fully penetrated

10°4-----~--~~r_--~~~~-.--------~~
100 10' 102 103 10'

Fig. 5-Difference between the behavior of partially and fully Fig. 6-Difference between the behavior of derivatives of par-
penetrated wells. tially and fully penetrated wells.

SPE Formation Evaluation, December 1987 549

 1000

PENETRATION RATIO :: 0.10

Zo = 0.7101

a.
.
c

u.i
a:
::J

'"'"w
a:
Q.
5.6146C
'"ffi 10 Co 2TTQlc1hr!
....I
Z
o
iii
zw
::E
o

1 L -_ _~~~~~_ _~~~~~L-_ _~~~~~_ _~~-L~~~_ _~~~~~

1 10 100 1000 10,000
10 _ Kh llt
DIMENSIONLESS TIME CO - 0.000295 ~ C

Fig. 7-Example type curves for b=0.1 and hWD =500.

This means that the application of a linear coordinate

transformation 11 to type curves for fully penetrated wells will not
WELL C-1 yield correct skin and kh values for partially penetrated wells.
Moreover, the skin and kh values will deviate more from the cor-
~~ rect values as the penetration ratio decreases and dimensionless well-
bore length increases.
TOP OF
FORMATIOf'!, _ ~
--- Wellbore Storage and Skin Type Curves
~ for Partially Penetrated Wells
? -8500
Eq. 4 was used to generate new type curves for partially penetrated
~ wells. For fixe<! values of penetration ratio and dimensionless well-
-8550 bore length, dimensionless pressure was plotted as a function of
tD/CD and the dimensionless parameter group, (CDe 2S ). An ex-
-- --- -----GOC ample of the type curves for b=O.l and hWD =500 is shown in
-8600 Fig. 7.
Use of the CD e 2S group as a correlating parameter was first pro-
posed by Earlougher and Kersch. 4 It was incorporated into type
-8650 curves for fully penetrated wells by Gringarten et al., 5 who en-
hanced its usefulness somewhat using If/CD as the type-curve ab-
scissa. As suggested by Chu et al., 1 CD e 2S conforms as a
-8700 correlating parameter in partially penetrated systems as well as it
) does in the fully penetrated case. That is, for a fixed-value CD e 2S ,

( -8750
any combination of CD and S yields almost the same dimension-
less pressure for a given value of ID/CD ,

r> -8800
PERF'O. INTERVALS
It is important to note that the skin term, S, is the damage skin
and not the total skin of the system, which is Sp +S/b. This is a
necessary distinction because the different components of skin have
different effects on pressure behavior, as illustrated in the previ-

142'
1 < -885<§(f
~B- ------"20'-
====):2':= ous example.

1- , ------ The Start of Radial Flow. Bilhartz and Ramey 9 suggested that
~ -'-8~C 121-----4"'-
the start of a semilog straight line for a partially penetrated well
1- -t-----~--
~O--------20'- could be estimated from
Z 1_
~\'~~ -8950 ==:3E= -====-==-~8'
-------- ~
tD=CD(60+3.5S t ), ................................ (8)

~
\HEAVY TOTAL 82'
where St=S/b+Sp '
OIL TAR
Eq. 8 and the "llh log cycle rule" both underestimate the actual
time required for pressure in a partially penetrated well to attain
a semilog straight line. The loci of points marked "approximate
start of radial flow" in Fig. 7 represent the time at which the semi-
log slopes are within 10% of the theoretical slope after radial flow
is established. These time limits are useful for analyzing test data
Fig. 8-Cross section through Well C-1.
from a bounded system, for determining when semilog analysis may
be valid, and for designing well tests. However, the log/log deriva-
tive plot is a better tool than the 1Vo log cycle rule or 10 % rule,
550 SPE Formation Evaluation, December 1987
 10000r------------r-I-----------rr-----------'Ir-----------,------------.

000066 ~
1000 I- -

",;
"-
;
"-
<l

100 I- -
PRESSURE DATA FROM PBUI

TYPE CURVE

10L-__________- L__________ ~L_ _ _ _ _ _ _ _ _ _ _- L__________~__________~

0.1 10 100 1000 10000

Time, ~t, min

Fig. 9-Type-curve match of the pressure data from PBU 1.

as pointed out by Bourdet et al. 19 to determine the start of semi- The second assumption is readily defensible for the particular
log straight line. As shown by Fig. 6, depending on band h wD • physical system from which the well test data are to be analyzed
it may take an additional half cycle to reach the semilog straight here. The first assumption is less easy to defend because one ex-
line for partially penetrated wells. pects a monotonic increase in oil mobility (a monotonic decrease
in the gas saturation) in reality with distance away from the well-
Field Example. The example selected here is a Prudhoe Bay well. bore rather than a constant mobility within a region where free gas
The data for this well were first presented by McKinley et al., 8 is liberated as pressure drops. The first assumption's validity is to
who identified it as Well C-l. It was described in Ref. 8 as a "gas- a large extent dependent on the shape of the oil/gas drainage rela-
cap well with poor shale protection. " This description is somewhat tive permeability relationship.
misleading because (1) the buildup surveys show no evidence of For interpretation of this particular field example, the above two
direct communication with the gas cap, and (2) the shales shown assumptions seem to lead to useful results. Generally, of course,
at 8,818 and 8,777 ft [2688 and 2675 m] in the cross section in these assumptions do not have to be made for an undersaturated
Fig. 8 are correlatable with surrounding wells. In fact, we will at- reservoir. Thus, the use of new type curves for partially penetrated
tempt to show that the buildup data are consistent with the flow wells is the same as the use of the Gringarten et al. 5 type curves.
rate (see Fig. 10) that comes from below the shale at 8,818 ft [2688 The pressure drop due to the total skin (the skin estimated from
m]. either Homer or MDH methods) is given as
The first buildup test (PBU 1) was carried out June 18, 1978,
2 months after the well started production. The average reservoir 141.2q/LB
pressure at this time was just above the bubblepoint, while the bot- !:J.ps= St . ................................ (9)
(kh)j
tomhole flowing pressure of the well was approximately 1,300 psi
[8963 kPa] below the bubblepoint. A reduction of oil transmissi-
bility caused by liberated solution gas is therefore to be expected The total skin is made up of three components: damage skin
in the near-wellbore vicinity. We will call this the "mobility effect." (removable), partial penetration skin (within the inner-annular
The Homer analysis of the first test gives a permeability-thickness region), and skin resulting from the reduction of the oil mobility
product of 29,020 md-ft [8.7 md·m] and a total skin of 6.2. Be- around the wellbore. Thus,
cause the average reservoir pressure was above the bubblepoint at
the time of this test, it is assumed that the kh value obtained from
the Homer analysis reflects the transmissibility of the formation
beyond the region influenced by the mobility effect. This physical
system evidently does not conform well to the assumption of uni-
form horizontal permeability distribution in the mathematical model
PERF'D
on which the type curves are based. To use the type curves for a INTERVAL 6/20178 4/15179 5/8/79 12/2179 12/28179
real problem, it is necessary in this instance to make two more as- A
sumptions.
1. The physical system behaves like a composite system wherein B

the inner-annular region has a uniform permeability-thickness prod-

c
uct of (kh)wb and the outer region has a corresponding
permeability-thickness product of (kh)j (formation permeability- o
thickness product). Note that (kh)wb refers to the permeability-
thickness product of the inner-annular reduced-permeability region
100
over the whole formation thickness. In other words, the thickness 1
PBUl
0 "Flow
PBU2
1 " % Flow
PBUS
of this inner-annular reduced-permeability region is the same as
the whole formation thickness. PBU3 MATRIX
ACID
2. The radial flow is achieved within the inner-annular region
(Le., partial penetration effects are resolved within this region) for Fig. 10-Flowmeter surveys for Well C-1.
the system considered in this example.

SPE Formation Evaluation, December 1987 551

 less the penetration ratio is low. It should also be noted that the
TABLE 3-ANALYSIS OF PBU 1 FOR
corollary is true-namely, that this approach cannot be used to es-
DIFFERENT ASSUMED VALUES OF b
timate k z.
Type-Curve The best type-curve match for b =0.7 was obtained with the curve
Match (kh)wb Total Skin defined by CD e 2S =10 4 . The match point was dp= 100 psi [689
b C De 2S (md-ft) S Sp + Sib kPa], PD=0.88; .:it = 10/60 hours, t DICD=22.9.
- - - --- !E.. From the pressure match, we obtain
0.3 10 3 46,982 -0.04 13.7 13.8
0.4 10 3 35,301 0.002 8.5 8.5
0.5 10 3 27,774 -0.02 5.5 5.5 PD
0.6 10 4 26,736 1.12 3.6 5.5
(kh)wb=141.2qJ.lB-
.:ip
0.7 10 4 22,842 1.14 2.2 3.8
0.8 10 4 20,246 1.13 1.2 2.6 0.88
=141.2x 13,500 x 1.02 x 1.335 x - -
0.9 10 6 24,659 3.40 0.5 4.3
100
=22,842 md-ft [6.9 md·m].
where SM is defined as the skin effect resulting from reduction of
the oil mobility. From the time match, we obtain
Substituting Eq. 9 in Eq. 10 yields
O. 000295 (kh) wb .:it
(kh)! C=-----
SM=Sr---(Sp+Slb) . ......................... (11) J.I tDICD
(kh)wb
Fig. 9 shows a conventionallog/log plot of the pressure data from 0.000295 x 22,842 10
this test. These buildup data were matched vs. the new type curves .--------x---
for assumed penetration ratios from 0.3 to 0.9. For this process, 1.02 6Ox22.9
emphasis was placed on obtaining a highly accurate match of the
data in the time interval marked "match region." The endpoint =0.048 bbl/psi [0.001 m 3 /kPa].
of this interval (120 minutes) corresponds to the approximate start
of radial flow, estimated from the type curves. Up to that time, The dimensionless storage constant, CD, is given by
it is assumed that the pressure behavior reflects the reduced per-
meability in the vicinity of the wellbore caused by the mobility 5.6146C
CD = - - -
effect. 27rcf>crhr~
The starting point of the match region (8 minutes) is the point
at which errors in shut-in time cease to be significant. This requires 5.6146XO.048
amplification. Because of the finite time required to shut in a well,
27r xO.22 x 15 x 10 -6 x 142 X (0.3)2
it is not uncommon to find that the actual shut-in time is later than
the reported time. If this is not accounted for, a log/log plot can
give a spurious indication of decreasing wellbore storage. =1,017.03.
There was sufficient character in the defined match region to ob-
tain a reasonable match for each penetration ratio considered. The The damage skin is calculated from
results of the analyses are shown in Table 3. We will illustrate the
analysis technique for one example, a penetration ratio of 0.7. CD e 2S
The dimensionless wellbore length, hwD' was estimated as S=O.5ln--
CD
hw 80r,;:
~
hwD=-~ ~=- xv5 =596.
rw k z 0.3 =0.51n---
1,017.03
We will therefore use type curves developed for hWD=500. Note
that the analysis is relatively insensitive to small errors in hwD un- =1.14.

TABLE 4-COMPARISON OF NEW TYPE-CURVE ANALYSIS

WITH EARLY-TIME ANALYSIS FROM REF. 8 ON WELL C-1

Early-Time
Analysis
Horner From
Analysis New Type Curves Ref. 8
q J* (kh)f* (kh)wb (kh)wb
Test Date (STB/D) (STB/D/psi) (md-ft) b hWD (md-ft) q. (md-ft) .!!.L
- - --- 13,500 10.0 29,020 0.7 500 22,842 0.86 10,300 0.64
PBU 1 6/18/78
PBU 2 4/13/79 2,300 5.0 13,227(*?) 0.7 500 18,131 0.55 6,400 0.37
Matrix Acidizatlon with HCI
PBU 3 5/7179 8,800 10.2 22,396 0.7 500 19,119 1.2 19,000 1.0
PBU 4 12/2/79 4,500 2.2 3,788 0.7 500 3,565 1.2 3,600 0.22
Matrix Acidization with 12-3 Mud Acid
PBU 5 12/28/79 12,700 8.7 18,746 0.7 500 15,335 1.2 18,200 1.0
Reanalysis of PBU 4 and PBU 5 with Corrected Penetration Ratios
PBU 4 12/2179 4,500 2.2 0.1 250 22,495 (very low)
PBU 5 12/28/79 12,700 8.7 0.5 500 21,488 0.99

• Because of small differences in interpretation, these values are slightly different from those reported in Ref. B.
"Test of insufficient duration for the establishment of radial flow.

552 SPE Formation Evaluation, December 1987

The pressure drop associated with the damage skin, Ilps, is given PBU 4, and that even after the second matrix acidization, the topmost
by set of perforations remained closed. PBU's 4 and 5 were reana-
lyzed with the new type curves, but suitably adjusted penetration
141.2qp.B S ratios were used this time. These results are also shown in Table 4.
Ilps= x- Calculation of Ej for these last two tests takes into account not
(kh)wb b only the calculated damage skin, but also the effect of the additional
pseudoskin caused by the changed penetration ratio. For example,
141.2x13,500xl.02x1.335 1.14 the reanalysis of PBU 5 gave a damage skin of -1.58. The pseu-
doskin, however, because of a penetration ratio of 0.5, is 5.5 (from
22,842 0.7 Fig. 4), whereas the expected pseudoskin, because of a penetra-
tion ratio of 0.7, is 2.2. The additional skin effect caused by the
= 185 psi [1275 kPa). upper set of perforations being closed to flow is Sp(0.5)-Sp(0.7).
Hence,
Sp is obtained from Fig. 4. For h wD =500 and b=0.7, Fig. 4
yields Sp =2.2. SM can be estimated from the approximate rela-
tionship defined in Eq. 11,

141.2x 12,700x 1.02 x 1.355 (-1.58 )

= --+5.5-2.2
21,488 0.5

29,020 ( 1.14) = 15.9 psi [110 kPa),

=6.3--- 2.2+--
22,842 0.7
and the Ej is given by
=1.44.
(P-pwj)-IlPS
The flow efficiency, Ej , of this well may be calculated in the or- Er--"""'----
dinary way. Hence, (P-Pwj)

(P-pwj)-Ilps =0.99.
Ej = ---=-------.:...-=-
(P-Pwj) This suggests that the additional skin effects caused by the top set
of perforations being closed is largely offset by the stimulated per-
=0.86. formance of the remaining perforated intervals.

Several interesting points arise from Table 3. First, we note that

for assumed penetration ratios of 0.6 to 0.8, the best matches were Conclusions
obtained in each instance on curves defined by CDe 2S = 10 4 . Simi- 1. An analytic solution is developed that describes the pressure
larly, for penetration ratios of 0.3 to 0.4, the best matches were behavior of a partially penetrated well with an infinite-conductivity
all obtained on curves defined by CD e 2S = 10 3 • Thus, contrary to condition at the wellbore flowing at constant rate when wellbore
what one might first assume, changing the penetration ratio does storage and skin are significant.
not result in a mere rebalancing of the skin components so that 2. A uniform-flux condition at the wellbore gives rise to higher
(Slb+Sp ) and (kh)wb both remain constant. In fact, unless the computed values of pseudoskin caused by partial penetration than
curve shape demands that a discrete jump be made from one CD e 2S does the infinite-conductivity condition.
group to another, (Slb+Sp ) and (kh)wb will both increase with 3. Pseudoskin caused by partial penetration does not have the
decreasing penetration ratIo. same effect on early-time pressure behavior as damage skin, con-
This provides us with a coarse method of estimating the correct ceptually associated with an infinitesimally thin region around the
penetration ratio if radial flow information is available. In most un- wellbore. In a partially penetrated well, pressure deviates from the
dersaturated reservoirs, this would entail selecting the penetration unit slope on a log/log plot earlier and reaches the semilog straight
ratio that gave values of (kh)wb and total skin (Slb+Sp ) that were line later than its fully completed counterpart.
consistent with the values obtained from Homer analysis. In this 4. Type curves that may be used to identify the various compo-
instance, however, because of the mobility effect, we expect the nents of skin for partially penetrated wells are generated. If the final
(kh)wb value and the total skin term to be less than the correspond- semilog straight line is reached, the type curves may also be used
ing values of (kh)j and St obtained from the Homer analysis. Rela- to estimate the penetration ratio and hence the ultimate producing
tive permeability measurements suggest that at critical gas saturation, interval of the well.
relative permeability to oil is reduced to about two-thirds of its value
at zero gas saturation. From Table 3, it therefore seems that a
penetration ratio of 0.7 or 0.8 is most appropriate. This in tum im-
Nomenclature
plies that the producing interval for the well is between the shale b = penetration ratio
at 8,818 ft [2688 m) and the heavy~oil tar at 8,960 ft [2731 m). B = FVF, RBISTB [res m 3 /stock-tank m 3 )
Four subsequent buildup tests on this well (PBU 2 through PBU C t = system total compressibility, psi -\ [kPa -\)
5) were presented by McKinley et ai. 8 These tests were also ana- C = wellbore storage constant, bbl/psi [m 3 /kPa)
lyzed with the new type curves, assuming a penetration ratio of CD = dimensionless wellbore storage constant
0.7. The results of these analyses are shown in Table 4. The flow h = thickness, producing interval, ft [m)
efficiencies calculated from the type-curve analyses seem to track hw = perforated interval, ft [m)
the well's PI very consistently in the first three tests, but the values
hWD = dimensionless wellbore length, ft [m)
for (kh)wb and Ejcomputed for the last two tests (PBU 4 and PBU
5) do not seem consistent. J = productivity index, STBIDIpsi [stock-tank m 3 /d/kPa)
The explanation for this can be found from flowmeter data. Fig. k = permeability, md
10 schematically shows the results of five flowmeter surveys carried kr = horizontal (radial) permeability, md
out on Well C-l during the period of interest. They suggest that kz = vertical permeability, md
the three upper perforated intervals became closed to flow before (kh)j = formation permeability-thickness, md-ft [md' m)

SPE Formation Evaluation. December 1987 553

(kh)wb = permeability-thickness product of inner-annular 3. McKinley, R.M.: "Wellbore Transmissibility from Afterflow-
reduced-permeability region over the entire Dominated Pressure Buildup Data," IPT (July 1972) 836-972; Trans.,
AlME,251.
formation thickness, md-ft [md' m] 4. Earlougher, R.C. Jr. and Kersch, K.M.: "Analysis of Short-Time Tran-
p = average pressure, psi [kPa] sient Test Data by Type-Curve Matching," IPT (July 1979) 793-800;
flp = pressure drop, psi [kPa] Trans., AIME, 257.
PD = dimensionless pressure 5. Gringarten, A.C. et al.: "A Comparison Between Different Skin and
Wellbore Storage Type-Curves For Early-Time Transient Analysis,"
Pi = initial pressure, psi [kPa]
paper SPE 8204 presented at the 1979 SPE Annual Technical Conference
1lp s = pressure drop caused by total skin and Exhibition, Las Vegas, Sept. 23-26.
PwD = dimensionless wellbore pressure with wellbore 6. Streltsova, T.D.: "Pressure Transient Analysis for Afterflow-Dominatcd
storage and skin effects, psi [kPa] Wells Producing from a Reservoir With a Gas Cap," IPT(ApriI1981)
Pwf = bottomhole flowing pressure, psi [kPa] 743-54.
7. Brown, M.E. and Mao, M.-L.: "Pressure Buildup Analysis of Prud-
q = production rate, STB/D [stock-tank m 3 /d] hoe Bay Wells," paper SPE 9455 presented at the 1980 SPE Annual
flr = radial distance, ft [m] Technical Conference and Exhibition, Dallas, Sept. 21-24.
rD = dimensionless radius 8. McKinley, R.M. et al.: "Early-Time Pressure Buildup Analysis for
rw = wellbore radius, ft [m] Prudhoe Bay Wells," IPT(Feb. 1984) 311-19; Trans., AIME, 277.
s = Laplace image space variable 9. Bilhart, H.L. and Ramey, H.J. Jr.: "The Combined Effects of Storage,
Skin, and Partial Penetration," paper SPE 6753 presented at the 1977
S= damage skin SPE Annual Technical Conference and Exhibition, Denver, Oct. 9-12.
SM = mobility skin 10. Gringarten, A.C. and Ramey, H.J. Jr.: "An Approximate Infinite Con-
Sp = pseudoskin caused by partial penetration ductivity Solution for a Partially Penetrating Line-Source Well," SPEI
St = total skin (April 1975) 140-48; Trans., AlME, 259.
11. Chu, W.C., Reynolds, A.C., and Raghavan, R.: "On the Analysis of
t = time, hours
Well Test Data Influences by Wellbore Storage, Skin and Bottom Water
tD = dimensionless time Drive," IPT (Nov. 1984) 1991-2001; Trans., AIME, 277.
flt = running test time (units specified) 12. Stehfest, H.: "Algorithm 368, Numerical Inversion of Laplace Trans-
z = distance along perforated interval, ft [m] forms," D-5, Communications ojthe ACM (Jan. 1970) 13, No. 1,47.
ZD = dimensionless effective average pressure point 13. van Everdingen, A.F. and Hurst, W.: "The Application of the Laplace
Transformation to Flow Problems," Trans., A1ME (1949) 198,171-76.
'Y = 0.5772 14. Cinco-Ley, H. and Samaniego-V., F.: "Effect of Wellbore Storage and
f.t = oil viscosity, cp [Pa' s] Damage on the Transient Behavior of Vertically Fractured Wells, " paper
T = dummy integration variable SPE 6752 presented at the 1977 SPE Annual Technical Conference and
q, = system porosity Exhibition, Denver, Oct. 9-12.
15. Fair, B.W.: "Pressure Buildup Analysis With Wellbore Phase Redis-
tribution," SPEI (April 1981) 259-70.
Subscripts 16. Cinco-Ley, H. and Samaniego-V., F.: "Pressure Transient Analysis
D = dimensionless for Naturally Fractured Reservoirs," paper SPE 11026 presented at
f = formation the 1982 SPE Annual Technical Conference and Exhibition, New
i = initial Orleans, Sept. 26-29.
17. Brons, F. and Marting, V .E.: "The Effect of Restricted Fluid Entry
M = influenced by mobility effect on Well Productivity," IPT(Feb. 1961) 172-74; Trans., AIME, 222.
r = horizontal (radial) 18. Streltsova, T.D.: "Pressure Drawdown in a Well With Limited Flow
w = bottomhole, well Entry," paper SPE 7486 presented at the 1978 SPE Annual Technical
wb = wellbore vicinity Conference and Exhibition, Houston, Oct. 1-3.
wf = flowing 19. Bourdet, D. et al.: "A New Set of Type Curves Simplifies Well Test
Analysis," World Oil (May 1973).
Z = vertical

Superscripts SI Metric Conversion Factors

- = Laplace transform of bbl x 1.589 873 E-Ol m3
= average ft x 3.048* E-Ol m
md-ft x 3.008 142 E-04 md'm
References psi x 6.894 757 E+oo kPa
1. Ramey, H.J. Jr.: "Short-Time Well Test Data Interpretation of the Pres-
ence of Skin Effect and Wellbore Storage," IPT (Jan. 1970) 97-104. • Conversion factor is exact. SPEFE
2. Agarwal, G.R., AI-Hussainy, R., and Ramey, H.J. Jr.: "An Investi-
gation of Wellbore Storage and Skin Effect in Unsteady Liquid Flow: Original SPE manuscript received for review March 12, 1983. Paper accepted for publica-
I. Analytical Treatment," SPEI (Sept. 1970) 279-90; Trans., AIME, tion Aug. 29, 1983. Revised manuscript received Feb. 4, 1987. Paper (SPE 11676) first
249. presented at the 1983 SPE California Regional Meeting held in Ventura, March 23-25.

554 SPE Formation Evaluation, December 1987