Вы находитесь на странице: 1из 20

Society of Petroleum Engineers

SPE 25664

Transient Testing of Limited-Entry Wells in Communicating


Layered Reservoirs
M.D. Abbaszadeh, Nobuhiko Tomita, and Tatsuo Tanaka, Japan Nat!. Oil Corp.
SPE Members

Copyright 1993, Society of Petroleum Engineers, Inc.

This paper was prepared for presentation at the SPE Middle East Oil Technical Conference & Exhibition held in Bahrain, 3-6 April 1993.

This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper,
as presented, have not been reviewed by the Society of Petroleum Engineers and are sUbject to correction by the author(s). The material, as presented, does not necessarily re11ect
any pos~ion of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society
of Petroleum Engineers. Permission to copy is restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment
of where and by whom the paper is presented. Write Librarian, SPE, P.O. Box 633836, Richardson, TX 75083·3836, U.S.A., Telex, 163245 SPEUT.

ABSTRACT Pressure transient testing in the form of single well


buildup/drawdown test or multiple well interference/pulse test
This paper presents both theory and field application for well can be used as a tool to estimate layer properties in stratified
testing of single or dually completed limited-entry wells in reservoirs. Generally, pressure transient behavior is
layered reservoirs. A comprehensive analytical solution is complicated by limited-entry effects because wells are partially
developed which considers anisotropic layers with crossflow completed in the layers. Substantial amount of theoretical
among layers in the formation. Features of the model are and numerical work has been done to study the effect of
discussed and sensitivity studies are presented to illustrate restricted entry on transient well tests in single layer
effects of contrast in layer properties, dual completion, and reservoirs with and without vertical pressure support.l-7
geometry of limited-entry wells on transient responses. These studies form the basis for better understanding of the
Some behavior of interference tests for different well limited-entry well effects in layered reservoirs.
completion geometries are also studied.
Hantush and JacobS have presented solutions for pressure
An interpretation method is presented which characterizes distribution in a two-layer bounded aquifer for steady-state
high and low permeability layers from pressure transient data. flow to a well draining one of the layers. Russell and
A field pressure buildup test from a layered reservoir, where a Prats9,lO summarize practical aspects of formation crossflow
high permeability layer is between two low permeability between layers; while, Deans and Gaoll and Gao12 provide
layers and the well is partially completed in a low rigorous numerical and semi-analytical formulation of the
permeability zone, is analyzed by the interpretation method of crossflow when layers are fully open to flow. Kazemi and
this paper. Properties of high and low permeability layers are Seth 13 study combined effects of anisotropy, stratification,
computed from this single buildup test crossflow and restricted flow entry on transient buildup tests
through numerical simulations. Javandel and Witherspoon 14
present an analytic solution for a partially penetrating well
INTRODUCTION
that is open in either layer of a two-layer system with
Characterizing layered reservoirs and understanding the crossflow at the interface between the two layers. Their
solution, however, assumes that the permeability anisotropy
behavior of fluid flow transfer among layers is of interest to
ratio is the same in both the layers. A more general
reservoir engineers in efficient exploitation, proper
investigation of transient behavior of limited-entry wells in
implementation of recovery processes, and better management
stratified systems has been performed by Yeh and
of hydrocarbon fields. Often wells are completed in a fraction
of formation to either preferentially deplete the most Reynolds. 1S,16 They use a numerical simulator and present
economically productive zones or delay/reduce gas or water a set of semilog type curves which combine pressure
coning. When the most prosperous zones are near depletion, drawdown responses of restricted-entry wells in a variety of
the tighter and less productive layers may be opened for well/reservoir configurations. Kuchuk 17 presents an
additional recovery. For some reservoir geologic settings, interesting analytical solution for a single horizontal well in a
however, there may be no other alternative but to produce the stratified reservoir with crossflow in the formation. His
reservoir through less productive zones because large volumes solution technique offers a procedure for solving transient
of hydrocarbon are deposited in these zones. problems of multilayer reservoirs.

In this paper, we present general analytical solutions for


transient pressure distribution in a multilayer reservoir with
References and illustrations at end of paper single and dual completion limited-entry wells. The solutions
consider anisotropic layers with crossflow among layers in the
547
2 TRANSIENT TESTING OF LIMITED-ENJRY WELLS IN COMMUNICATING LAYERED RESERVOIRS SPE 25664

fonnation. The nature and significance of interlayer crossflow


are investigated. Sensitivity studies are presented to (8)
understand effects of contrast in layer properties. dual
completion. and geometry of limited-entry wells on transient
responses. Behavior of interference tests for different well t _ 0.0002631U (9)
. completion geometries are also studied. These sensitivity P - r.'-
O'
studies help better identify conditions under which well tests r
provide maximum information about layers and assist in rp =- (10)
proper design of tests. Finally. a field pressure buildup test rO'
from a. str~tified reservoir with distinct contrast in layer
properties IS analyzed to illustrate the applicability of the
model presented in the paper. (11)

MATHEMATICAL FORMULATION
(12)
First. we develop solutions for a multilayer reservoir where
only one of the layers is open to flow. Special cases of this
more general formulation with application to interference (k I k,).
l
Tz; = • (13)
testing are also provided. Next, we consider a two-layer (kllk,)..
reservoir where both layers are partially perforated. This latter
case is more complicated because in addition to pressure.
contribution to total flow rate from each layer also should be The solution of these system of equations is derived in
calculated. Appendix A. using the method of source and Green's
functions. Briefly. first a solution for an instantaneous point
source is obtained by a combination of Laplace transformation
Multilayer Reservoir With Single Completion with respect to time and Hankel transformation of order zero
with respect to radial distance. The point source solution is
Fig. 1 shows the configuration of the well and reservoir integrated along the perforated interval to yield the solution
system considered for this case. The reservoir consists of N for an instantaneous limited-~ntry line-source well.
layers of infinite extents with the limited-entry well located in Integration of this last instantaneous solution with respect to
the nth layer. Each layer has uniform properties and is time yields pressure distribution in the reservoir for a well
anisotropic with permeability anisotropy ratio of kzi/kri . produci."g ~t a constant flow rate. The final result is given by
a combmauon of Eqs. A-28 and A-30. The solution indicates
Layer properties are different from each other, except for
initial pressure which is the same for all layers. The regular that bottomhole pressure is a function of vertical location z at
the wellbore. This nonuniform pressure distribution at the
assumptions of single-phase fluid with small compressibility
wellbore is due to the assumption of uniform flux made in
also apply. Pressure diffusivity equations and boundary
formulations. A good approximation to the infinite-
conditions for the system of Fig. 1 are:
conductivity wellbore solution can be obtained by averaging
pressure at rp=1 over the open interva1.7 ,18 Such averaging
i=I•..•N (1) yields.

1
PwDa = Sb v..3( zbD -ZaD )
()PPI =0 at Zp =0 (2)
Ozp
{~U..P"[P"_le-2VIIIDII-1(eVlllbD -eVlllaDt +2

PPi =PPi+1 at zp =Zpi (3)


-2 cosh V..(ZbD _ZaD)]+U.. [p..e-2VnIDn

4. ()PPi = 4- ()PPi+1 at Zp =zpj (4) (1- cosh V..(ZbD - ZaD )}+(e-VlllbD _e-VnlbD )2]
I Oz
P
1+1 OzP
+V..(ZbD - ZaD)+ l_e-vn (lbD- l aD)} (14)
()PPN =0 at Zp =ZpN (5)
dz p where. Vi and Un are defined by Eqs. A-17 and A-29.
PPi=O at t p =0 (6) respectively. The p terms are given by simple recursive
relationships of Eqs. A-25 and A-26 of Appendix A. For the
special case of a three layer and a well completed in the most
rp ()pp_1 =_ hp_ Zdl) S Zp S ZbD (7) bottom layer (n=I). Eq. 14 reduces to
Or:P '0->0 ZbD - zdl)

and dimensionless groups are defined as,

548
M. D. ABBASZAPEH. N. mMITA AND T. TANAKA SPE 25664

-PwDa (a, s) = 3 1 { 2A
-'2""""':,;;,t.-::..I.--
sbvI (ZbD - ZaD) e '1 DI - A

[sinh(vlzbD)-sinh(vlz..o)t +~(e-'1lbD -e-'1lbDt


(21)

+VI(ZbD - z..o)+e-'1(lbD- 1aD) -I} (15) The Hankel inversion of dimensionless flux is obtained from
Eq. 16 by replacing the pressure term by the crossflow flux
term. Cumulative dimensionless crossflow rate as a function
and ~1 is obtained from Eq. A-25 with ~3=1. For a two-layer of radial distance is obtained by integrating crossflow flux as,
reservoir, the solution is obtained by letting k z3=O, or ~2=1,
in Eq. 15. The single layer solution is derived for kz 3=O and (22)
k z2=O which result in ~1=1. For this latter case, the Hankel
inversion can be performed analytically. The Hankel
inversion of Eqs. 14 or 15 for rp=1 is19:
INJEREERENCE lEST IN A TWO-LAYER RESERVOIR
(16) Pressure at an observation well in a two-layer system can also
be derived from the more general Eq. A-30. The wellbore
For calculational purposes, the Laplace inversion is performed pressure at an observation well is approximated by averaging
numerically using the Stehfest algorithm20 and the Hankel Eq. A-30 over the open interval at that well. When active and
integral is computed with the Gaussian quadrature integration observation wells are both in layer I, the pressure at the
formula with varying number of quadratures to achieve a observation well is given by,
desired accuracy.

WELLBORE STORAGE AND SKIN


- r as-
PD.Obs(D'
1 {4A A
')-2Sb VI3(,Z bD -z,aD) e2'1'ID -
Although wellbore storage and skin are not explicitly included [sinh VlzbD - sinh VI zaD ][sinh VIZ' bD -sinh VIZ' aD]
in the model, they can be added to the model through the well-
+(e-'1 I ' bD _ e-'1" aD )(e-VllbD - e-'1'aD) + 2VI.:1%1D
known convolution scheme in Laplace space. 21 The result
is, +e-'1 4z2D + e-VI4l3D _ e-vII'al-" all_ e-'1I'bl-I ' bll} (23)

c
PwD ( D'
S)-
-.[
SPwDa+ S / b (17)
S l+sCD(sPwDa
_ +S/ b)] where, &1=Zb-za, dz2=zb-z'a, &3=z'b-za for z'»>zb, z'a<za;
Az 1=zb-z'a, AZ2=z'b-za, Az=zbD-z'a for z'b>zb, z'a>za ;
The Stehfest routine is now applied to Eq. 17, using either AZI =z'b-z'a, AZ2=z'b-za, AZ3=zb-z'a for z'b<zb, z'a>za ;
Eqs. 14 or 15 for the PwDa term. dzl=Z'b-za, dz2=z'b-za, AZ3=Zb-Z'a for z'b<zb, z'a<za. The
subscript D has been dropped in these relationships for
CROSSFLOW IN THE FORMATION simplicity.

The flux distribution at the interface of layer 1 and layer 2 When the active well is in layer 1 and the observation well is
when the perforated layer is at the bottom (n=l) is obtained in layer 2, the solution is:
from the Darcy's law as follows,

at z= zl (18)

In terms of dimensionless parameters,

(19)
and the Hankel inversion formula for both Eqs. 23 and 24 is
given by Eq. 16 with appropriate terms for pressure.

(20)
Two-Layer Reservoir With Dual Complt!tion

Fig. 2 shows the configuration of the well and reservoir


Using the pressure distribution in the reservoir from Eq. A-
system considered for the dual completion case. Appendix B
30, the dimensionless flux in Hankel and Laplace spaces is,
provides the formulation and solution for pressure transients
and transient layer flowrates. Again, the method of source and
Green's functions in conjunction with linear transformations
of Laplace and zero-order Hankel is used. Briefly, solutions
549
4 TRANSIENT TESTING OF LIMITED-ENlRY WELLS IN COMMUNICATING LAYERED RESERVOIRS SPE 25664

for individual layers are first obtained in analogy to the Yeh.Reynolds Correlation
method of Appendix A. The layer solutions are then set equal
to each other at the wellbore to yield a common wellbore Yeh and Reynolds1S,16 defined the following dimensionless
pressure. The effect of differing layer skin factors are parameters that they found to correlate pressure transients of
augmented to the skin-free solutions through the use of the many well/reservoir configurations in multilayer reservoirs
van Evedingen's skin concept. 22 The final solution is: with arbitrary completion locations:

(25) _ 1 (- ) S
PwD =PwD --in hwD - - (26)
Ii Ii
where, the dimensionless layer flow rates are given by Eqs. B- - tn
34 andB-35. The terms ~. U2 • V;. ~ are defined by Eqs. tD =-=t-
1&;0
(27)
B-19. B-20. B-24. and B-25.
where. fl and ii..w, are functions of layer properties and
VERIFICATION OF SOLUTIONS completion geometry. defined in Refs. 15 and 16. We
investigated the applicability of these correlating parameters
The analytical solutions were verified by comparing results for both the single-layer and two-layer reservoir parameters of
with those from the commercial finite-difference numerical Tables 1 and 2 of Ref. 15. We found that Yeh-Reynolds
reservoir simulator, ECLIPSE. The r-z model of the correlation is only an approximation. In fact. the
simulator was utilized with a total of 100 logarithmically- dissimilarity of derivative curves of Fig. 3 suggests that they
spaced radial gridblocks. The number of gridblocks in the can not be collapsed into one single curve by simple shifting
vertical direction varied from 30 to 100. depending on well's of the time axis. as suggested by Eq. 27. Based on this
geometry, with small gridblocks placed near the interface of investigation. we decided not to utilize the correlating
layers and near the perforated portions. Wells were completed parameters of Yeh and Reynolds. Instead. we chose to
in aquifer in order to minimize discrepancies between the perform sensitivity studies for the effect of various parameters
results of analytical and numerical simulations due to effects on well test responses to gain a better understanding for the
of varying oil formation volume factor with pressure. significance of various reservoir parameters and well
completion geometry.
Fig. 3 compares numerical results with the analytical solution
for a two-layer reservoir and three cases of: a) the well is fully Single Well Drawdown Behavior
completed in the high permeability layer, b) the well is
partially completed in the low permeability layer. and c) the This section investigates the influence of contrast in layer
well is partially completed in a two-layer reservoir with the parameters. presence of a low permeability barrier. location
same layer properties. As Fig. 3 indicates the agreement and geometry of the open interval on pressure transients.
between analytical and numerical simulations is good for all Two and three layer reservoirs with properties listed in Table
cases. These simulations verify the validity of analytical 1 (case b) are used. These sensitivity studies provide a better
solutions. The simulation study also helps determine the understanding for the behavior of pressure transients in more
needed accuracy in numerical Laplace invertor and numerical complicated multilayer systems.
integration of Hankel integrals. It is determined that
satisfactory results are generated with only 6 terms in the Effect of Contrast in Layer Permeabilities: Figure 4 shows
Laplace inversion algorithm. the effect of contrast in permeability. kr 2lkrl. on drawdown
pressure derivatives in two and three layer reservoirs. The
The markedly different transient behaviors of Fig. 3 indicates two-layer reservoir parameters correspond to case b of Table 1.
that the contrast in layer properties and the location of For the three-layer reservoir case, the geometry and properties
perforation governs characteristics of pressure transient tests. of the third layer are identical to layer 1 in order to isolate the
When the well is in the high permeability layer. the response impact of layer 2. As Fig. 4 shows there are two semilog
is essentially identical to that from a naturally fractured straight lines (corresponding to flat portions of derivatives
reservoir with transient interporosity flow model and matrix curves) for all cases. The early-time data represents a radial
being of slabs configuration. 23 flow confined to the perforated horizon in layer 1 with the
derivative level being equal to 0.51b. where b is the
penetration ratio. Permeability of layer 1 can be obtained
RESULTS from this early-time data. Late-time data corresponds to radial
flow in the entire formation and can be used to obtain
There are several parameters related to layer properties and permeability of layer 2. The level of second plateaus is equal
location of perforations that affect pressure transient response to O.5kr 1h l/Dcrh for both two-layer and three-layer reservoirs.
of limited-entry wells in layered reservoirs. It would be As permeability of layer 2 increases. transient responses of
desirable to obtain appropriate variables that would correlate the two-layer and the three-layer systems become similar. and
many of the transient responses into a limited number of essentially indistinguishable from each other for
manageable curves. Such an attempt has been made by Yeh kr 2lkrl> 1000. Thus. a high permeability layer adjacent to a
and Reynolds 1S ,16 in constructing their semi-log type test zone will absorb pressure transients and the influence of
curves. other layers can not be seen on a test. At the limit for a very
large contrast in layer permeabilities. the response becomes
similar to that from a limited-entry well in a single layer
550
M. D. ABBASZAPEH. N. IDMIIA AND T. TANAKA SPE 25664

reservoir with constant pressure gas cap effect.' Studies with center of layer 1. Fig. 7 indicates that the early-time data is
contrast in layer porosities indicated that porosity effects were purely controlled by limited-entry effec~ and ~t its ~urati~n
only important for kr2/krl<10. is practically independent of the penetration ratio. During this
period the early radial flow is confIned to the horizon of the
Fig. 4 shows that duration of the transition region between perforated interval only. A spherical flow period develops for
the fIrst and the second radial flow is appreciable, about 2 to 3 small penetration ratios, as manifested by the -1/2 slope
log cycles. Thus, when the permeability contrast is large, it straight line for b=0.025. The combined effects of layering
would take a long time to reach the formation radial flow for and limited entry for all penetration ratios and permeability
direct calculation of permeability of layer 2. However, the contrasts are similar, and confIned to about half a log cycle,
transition region has distinguishable characteristics for high 5x105<tD<106. Based on comparisons of Figs. 6 and 7 it is
contrast in permeabilities of layer 2, enabling estimation of concluded that length of the perforated interval has a lesser
k2 even if the test duration is not very long. effect on well tests compared to the location of perforations.

Effect of Interbedded Low Permeability Barriec Usually there Effect of Permeability AnisQtropy: Two-layer reservoir
are low permeability, or semipermeable streaks, interbedded sensitivity studies showed that the effect of kzl/krl was to
artlong productive layers. Such tight sublayers influence displace the derivative curves proportional to the kzl/krl
vertical pressure communication in reservoir. Fig. 5 shows ratio. Therefore, expressing the dimensionless time in terms
pressure derivatives of the same two-layer reservoir of Table of kz I, the corresponding curves correlated into a single set
I, but with a 5 ft low permeability barrier situated between of curves. This allowed us to study the effect of kzz/krZ Qn
the perforated layer I and the high permeability layer 2. It is well tests. Fig. 8 shows the correlated transient responses for
seen that when the permeability of barrier is very small, layer two ratios ofkr2/krl=1O and 1000. The fIgure indicates that
I and 2 are essentially isolated and the response corresponds to vertical permeability of layer 2 is less signifIcant, especia:lly
that of a limited-entry well in a single layer reservoir with for higher contrasts in layer horizontal permeabilities. It is
reservoir thickness being equal to the thickness of layer 1. As only when kz2/krZ becomes sma:ller than 0.001 that the effect
the permeability of barrier increases, communication between
Qf vertical permeability of the more permeable layer becomes
layer I and layer 2 is established and pressure derivatives
important Such low ratios of kz2/krZ are less likely to occur
portray the average response of the entire formation. The
time at which substantial communication between layer 2 and for productive and high permeability zones.
the perforated layer I occurs is earlier for larger values of
barrier permeability, where a smaller pressure difference Formation Crossnow
between layers is required to cause crossflow. We found that
the ratio of barrier permeability and its thickness characterizes Crossflow among layers in the formation is controlled by two
the influence of barrier in Fig. 5. factors: contrast in permeability (more rigorously, diffusivity)
of layers and location of perforations. . C~n!I'ast in
The leaky-barrier model presented here is the more general diffusivities promotes crossflow from a low di~USl~lty la~er
case of previously considered semi-impermeable wall to a high diffusivity layer due to establishment of differential
pressure between layers. The wellbore boundary condition
models l l ,12,24-26, where the barrier is assumed to be of
tends to favor crossflow from a nonperforated layer to a
small thickness and negligible capacity, with communication
perforated layer due to pressure drop created near wellbore by
between the overlying and underlying layers occurring by
fluid production in the perforated layer. Therefore, when a
pseudo-steady state interflow mechanism governed solely by
high diffusivity layer is .partially op~n t? flow, ~e ~wo
pressure difference between layers. In contrast, the model of
crossflow mechanisms act 10 the same direction, resulting 10 a
this paper considers a two-dimensional transient fluid flow
large crossflow from a low diffusivity layer to a high
inside the low permeability region, and may be more realistic.
diffusivity layer. This situation is similar to the crossflow
phenomenon from matrix to fIssures in n~tu~l~ fractu«:d
Effect of Location of Perforated Interval: The effect of
reservoirs. On the contrary, when a low dlffuslvlty layer IS
location of the perforated interval is illustrated in Fig. 6 for open to flow the two crossflow mechanisms are in t~e
the two layer reservoir of case b in Table I and a fIxed opposite direction and crossflow reversal ma.y occur. Iha~ IS,
penetration ratio of b=0.2. Three different positions of in regions around wellbore crossflQW IS from a high
perforations in layer I are considered: well at the ~uom, well diffusivity layer to a low diffusivity layer due to lower
in the middle, and well just below layer 2. The 10fluence of pressure in the low diffusivity layer. In reg~ons .f8! away from
permeability contrast has also been superimposed on this the wellbore, fluid crossflows from a low diffuSlvlty layer to a
graph. Fig. 6 indicates that the location of the ?pen interval high diffusivity layer and then travels towards p~oducer al~g
has a considerable effect on pressure transients for all the high diffusivity layer. Therefore, the entire formatiQn
permeability ratios, especially when the layer is comple~ contributes to flow through the high diffusivity layer. This
adjacent to the high permeability layer, zafhl=0.8. For thiS latter crossflow mechanism is more applicable when vertical
latter case, the duration of partial penetration effect is. the permeabilities are not very sma:ll and communicatiQn between
smallest and transients are mainly controlled by layenng. layers Qccurs easily. Fig. 9 is a conceptualization of
Because of vastly different shape of the curves in Fig. 6, we crossflow in a two-layer reservoir with well location in either
conclude again that the correlating parameters proposed by the high diffusivity layer 2 or the low diffusivity layer 1.
Yeh and Reynolds will not satisfactorily collapse the curves
of Fig. 6 into one single curve. Figure 10 shows distance-weighted interlayer dimensionless
flux, rpvcD, as a function of dimensionless radial distance,
Effect of Penetration Ratio: Figure 7 shows the effect of the rp, for the two-layer reservoir of Table l-case b. The open
length of perforations as a function of krz/krl for the two interval is in the low permeability layer between za=270 ft
layer reservoir of case b in Table I for completions in the
551
6 TRANSIENT TESTING OF LIMIIED-ENJRY WELLS IN COMMUNICATING LAYERED RESERVOIRS SPE 25664

and zb=350 ft. Fig. 10 clearly illustrates that crossflow near consider the two-layer reservoir of Table 2 with two restricted-
the well is from the high permeability layer 2 to the low entry wells. It is not our intentions to provide a detailed
permeability layer 1, as indicated by negative values. Away investigation of multiple well testing in layered systems.
from the well, reversal in the direction of crossflow occurs and Rather, the objective is to illustrate the significance of
fluid flows from the low permeability layer 1 to the high layering and partial penetration, and address issues for further
permeability layer 2. As contrast in layer permeabilities research.
increases the strength of crossflow also increases.
Active Well in Layer 1. Observation Well in Layer 2: Fig. 13
The reversal in crossflow direction can not happen when both shows pressure transient behavior at an observation well for
layers or only the high permeability layer are fully open to three values of rD and for four values of k r 2/kr1. The
flow, as shown byjrevious studies with semi-permeable wall corresponding pressure derivatives are given in Fig. 14.
models. 1l ,24-2 In fact when both layers are fully These figures are generated by Eq. 24, but the dimensionless
perforated, the crossflow in the formation would reach a time and pressure are expressed in terms of average reservoir
steady-state condition at late times and would propagate as an parameters. For comparison purposes, the line-source
unchanged wave in the formation. l1 When one of the layers exponential integral solutions have also been added to the
is fully open to flow, the region of crossflow is stationary and pressure graph. Although the dimensionless time group of
does not change with time. 2S Note that the semi-permeable toIro2 would correlate the line-source solutions into a single
wall models impose an impedance to crossflow between layers set, it would not collapse the two-layer observation well
and assume that the crossflow is governed by pseudo-steady- curves.
state fluid transfer mechanism, thus flux is proportional to
pressure difference between the two layers. Our model Fig. 13 shows that only when rp is large and the contrast in
considers unimpeded crossnow in the formation, mainly kr2/krl is small, the two layer system behaves as a
controlled by vertical permeabilities of layers. homogeneous reservoir with average reservoir properties.
When kr2/kr1 becomes large, the layering effects are sensed at
A plot of dimensionless cumulative crossflow rate, area under the observation well. It is interesting to note that the
the curves of Fig. 10 as defined by Eq. 22, is shown in Fig. responses at rp=lO and rp=l00 are similar for all values of
11. This figure clearly shows the phenomenon of reversal in k r 2/krl, indicating that pressure distribution in layer 2 is
flow direction, with the location of flow reversal essentially uniform, at least up to a distance of rp=100. This
corresponding to the bottom of troughs. The magnitude of
effect is caused by a combination of limited-entry at the active
qeD at these troughs represents cumulative crossflow rate
well and vertical crossflow due to differing layer properties.
from layer 2 to 1, and the. final plateau corresponds to net The similarity of transient responses at ro=10 and rp=100 for
cumulative crossflow rate. Hence, the difference between the
two values is equal to total crossflow rate from layer 1 to kr2fkrl =1 illustrates the impact of limited-entry. Had the
limited-entry effects been unimportant the two curves for
layer 2. For example, for tp=105 and kr2/krl=I000, the total rp=lO and 100 would have correlated on the line-source
crossflow rate from layer 2 to 1 is equal to 0.513qw and the
solution based on tolro2. Prats27 has defined a heterogeneity
total crossflow rate from layer 1 to 2 is equal to (0.513-
parameter which identifies conditions where a two layer
0.167)qw=0.346qw. Therefore, 16.7% of the well's
reservoir behaves like a homogeneous reservoirs for fully
production rate from reservoir is contributed by layer 2 and penetrating wells. Our investigations indicate that partial
83.3% is contributed by fluids in layer 1, with 42% penetration effects tend to enhance the layering effects, and
(34.6/83.3) flowing via layer 2 and the rest directly flowing
that the criteria presented by Prats may not be rigorous for
through layer 1. This simple analysis demonstrates the
these cases.
importance of a high permeability layer in providing a conduit
for production of oil from a perforated low permeability zone. Active Well and Observation Well in Layer 1: Eq. 23 is used
to generate pressure and pressure derivative responses of Fig.
Crossflow flux and crossflow rates when the well is situated
15 when both active and observation wells are completed in
in the middle of the high permeability layer and perforated
layer 1 at the same horizon. Pressure curves are shown as
only through half of the interval are shown in Fig. 12. For
thin lines and derivatives are indicated by thick lines. The
this case the crossflow is always from the low permeability
dimensionless time and pressure are defined in terms of
layer to the high permeability layer, with the final qeD value
representing contribution of layer 1 to well's production rate. properties of layer 1, similar to drawdown curves of Fig. 4.
This behavior is similar to naturally fractured reservoirs where Again, contrary to homogeneous systems, the tD/rri group
the matrix system depletes into fractures for flow towards does not correlate various pressure responses to a common
producers. curve, indicating that reservoir layering and limited-entries are
affecting transients at the observation well. At any radial
Interference Test Behavior location, pressure curves are essentially unaffected by layering
for some period of time. after which the contrast in layer
Pressure transient behavior at an observation well in an properties becomes important. During these early periods the
interference test setting is function of not only its radial limited-entry effects are present to different extent at different
distance from an active well but also its vertical location in ro locations, which separates their corresponding pressure
the reservoir. Whether observation and active wells are is curves.
completed in the same layer or in different layers will have a
marked influence on test results. In addition, limited-entry Comparison of Figs. 13 and 15 shows that pressure transients
effects at both wells are important. To study these effects we are more affected by reservoir layering when the observation
552
M. D. ABBASZADEH, N. TOMITA AND T. TANAKA SPE 25664

well is located in a layer different from that of the active well. FIELD EXAMPLE
It is also seen that the characteristics of pressure transient at
the two observations wells are distinctly different, depending A buildup test from a limited-entry well in a multilayer
on whether the observation well is completed in a low or a reservoir is analyzed to illustrate the applicability of the
high permeability layer. This difference in transient behavior model proposed in this paper. Geologically, the reservoir
is important from well testing point of view. If interference consists of three main layers with a high permeability layer of
tests are conducted in a two-layer reservoir with two small thickness situated in the middle. The two low
selectively perforated intervals, one in a low and the other in a permeability layers are thick and contain sub-layer
high permeability zone, a combined interpretation of tests can heterogeneities. Both core data and log analysis indicate
provide properties of both layers. More research is needed to presence of a thin and very low permeability streak between
fully explore this feasibility. the high permeability layer 2 and the perforated interval. A
limited-entry well is completed in the bottom layer with
penetration ratio being equal to 0.2.
Dual Completions
A two-week buildup test with a downhole shut-in tool was
We consider the same two layer reservoir where half of the conducted on this well. Because of operational limitations the
layers are perforated in their center. The permeability first few hours of buildup data, including the final flowing
anisotropy ratios in the layers are equal to 0.1. pressure, could not be recorded. Therefore, reliable record of
Dimensionless pressure derivatives for this dually completed L1p measurements can not be made. This, however, does not
limited-entry reservoir as a function of layer permeability ratio affect the pressure derivative data. Type curves with a leaky
k r 2/kr1 are shown in Fig. 16. The dimensionless time and barrier similar to Fig. 5 were used to interpret the test. Fig.
pressure terms in this figure are defined in terms of 1:kr h and 18 shows the final analysis with a three-layer reservoir model
1:<I>h, where 1:kr h=k r 1h 1+k r 2 h Z and 1:<I>h=<I> 1h 1+<I>2 h 2, in consisting of a high permeability layer, a leaky barrier zone,
order for all responses to converge to the common level of and a low permeability perforated layer. The interpretation
0.5. The case of kr 2/krl=1 corresponds to a homogeneous showed that the ratio of permeability of the barrier zone and
reservoir with two limited-entry wells with penetration ratios low permeability layer 1 was about 0.0001 and the ratio of
of 0.045 and 0.455. The corresponding derivative is similar the high permeability layer 2 to the low permeability layer 1
to that of a single limited-entry well in a single layer was about 500. These numbers are in agreement with the
reservoir. As the kr2/krl ratio increases, derivatives become geology, core analysis and general understanding of the
similar to those from a naturally fractured reservoir with reservoir flow behavior. The permeability of the low
permeability was computed from the pressure match of Fig.
transient interporosity flow between matrix and fissures. 23
18. Permeability of the top third layer could not be calculated
The early-time behavior are influenced by restricted entry
from this test because its contribution could not be
effects, causing deviation from the 0.5 level compared to the
decomposed from that of layer 2. Another test for a well
dual porosity models
completed in the top third layer would be required to establish
its permeability.
Dimensionless transient flow rates from individual layers as a
function of permeability ratio kr 2/kr 1 is shown in Fig. 17.
This example illustrates that a two-layer reservoir can be
For all cases transient behavior mainly occurs at early times characterized from pressure transient data alone if one of the
and layer rates eventually become proportional to flow layers is perforated and there is a high contrast in the
capacity and apparent skin factor of their corresponding layers. properties of layers. It is foreseeable to expect that a
This is different from the behavior of a fully penetrating well multilayer reservoirs could be evaluated in a similar fashion if
in a stratified reservoir without true layer skin factors, where the layers are selectively opened and tested. Communication
individual layer rates are oniy function of flow capacity of between layers introduces distinct characteristics in pressure
layers, qDi=krihi/1:krh. The differing apparent skin effects transient data that increases the possibility of estimating layer
created for differing penetration ratios and penetration properties from bottomhole pressure data.
geometry in layers alters the kh proportionality relationship
for flow rates.· This is clearly seen for the case ofkr2/krl=10
for which krlhl=kr2h2. The corresponding flow rate of layer CONCLUSIONS
2, however, is higher than that of layer 1 because there is a
small apparent skin factor associated with the completion in 1. Comprehensive analytical solutions for transient behavior
layer 2 compared to layer 1, as estimated from the numerical in multilayer reservoirs with limited-entry wells of single
work ofYeh and Reynolds. 16 or dual completions have been derived. The solutions
consider crossflow in the formation and allow for differing
If transient layer flow rates are measured along with permeability anisotropy in layers.
boltomhole pressure data they can be used to estimate
individual layer properties by established techniques of 2. Sensitivity studies to the length and geometry of open
simultaneous flowrate and pressure analysis. 28 However, interval indicate that location of perforations has a large
there seems to be enough character in pressure derivatives of effect on transient tests, especially if the open interval is
Fig. 17 for high contrast in layer properties to allow for adjacent to a high permeability layer. The length of the
calculation of individual layer permeabilities from pressure perforated interval for a fixed completion geometry affects
data alone. Such possibility could be investigated through an the early-time data and is insignificant for late-time
appropriate sensitivity study prior to conducting a well test to behavior.
establish the adequacy of bottomhole pressure data or the need
for measurement of layer tmnsients.
553
8 TRANSIENT TESTING OF LIMITEP-ENTRY WELLS IN COMMUNICATING LAYEREP RESERVOIRS SPE 25664

3. The behavior of inter-layer crossflow in the formation has TO =r/rw


been investigated. The results show that when a low r w = wellbore radius
diffusivity layer is partially completed, reversal in the Si =skin factor of layer i
direction of formation crossflow can occur. Near the s =Laplace variable
wellbore crossflow is from high diffusivity layer and away
to = O.OOO264kr lt/ch J.1Ctr w2
from the well it is from the low diffusivity layer. A high
z = vertical distance
permeability layer provides an effective conduit for fluid
production when a low permeability layer is opened to
za = z coordinates of limited-entry active well
z'a = z coordinates of limited-entry active well
flow.
zD =(z/rw)~(kr/kz) n
4. A study of interference test behavior with an observation
well located in high or low permeability layers has been zi = z coordinate of Layer i
conducted. Results show that pressure transient z'a = flfSt z coordinates of limited-entry observation well
characteristics are fundamentally different whether the z'b = second z coordinates of limited-entry observation well
observation well is completed in a low or in a high
permeability layer. Reservoir layering and limited-entry Symbols
influence transients significantly when the observation
well is located in a layer different from that of the active ~ = defined by Eqs. A-25 and 26
well. (J = Hankel transform variable
q» =porosity
5. Only when the distance between active and observation J,1 = viscosity, cp
wells is large and the contrast in layer permeabilities is A. = kz/~, vertical mobility, md/cp
small, a two-layer system behaves as a homogeneous lq = Tli / TIn , diffusivity ratio
reservoir with average reservoir properties, regardless of TI = kr / q»J.1C
partial penetration effects. Vi = defmed by Eq. Eq. A-17
£i = defined by Eq. A-27
6. A downhole shut-in buildup test from a stratified reservoir
has been analyzed by a three-layer reservoir model, with
the middle thin layer acting as a semi-permeable barrier.
ACKNOWLEDGMENT
The pressure derivative data exhibit mfluence of both low
and high permeability layers. The field example
We thank the management of Japan National Oil Corporation,
illustrates that a two-layer reservoir can be characterized
Technology Research Center, (JNOCffRC) for support and
by pressure transient data alone when one of the layers is permission to publish this work.
perforated and there is a large contrast in layer properties.

REFERENCES
NOMENCLATURE
I. Muskat, M.: Physical Principles of Oil Production,
B = formation volume factor, RB/STB McGraw-Hill :Book Co., Inc. New York (1949) 264-267.
b = penetration ratio, (Zb-za)1h1
bl = penetration ratio of layer I, (zbl-Zat>lhl 2. Nisle, R.G.: "The Effect of Partial Penetration on
bz = penetration ratio of layer 2, (zb2-zaZ)lh2 Pressure Buildup in Oil Wells", Trans. AIME (1958)
Ct = total compressibility, psi-I 213,85-90.
=
Co dimensionless wellbore storage constant
GL = Green's function for a fmite-Iength source well 3. Hantush, M.: "Drawdown Around a Partially Penetrating
Gd = Green's function for layers without source Well, Proc., ASCE (July 1961) 87, NO. HY4, 83-98
G s = Green's function for point source in infmite reservoir
4. Bilhartz, H.L. Jr. and Ramey, H. J. Jr: "The Combined
hi = thickness of layer i, ft
Effect of Storage, Skin and Partial Penetration on Well
J o = Bessel function of zero order Test Analysis," paper SPE 6453 presented at the SPE
kB = permeability of barrier zone, md, 52nd Annual Technical Conference and Exhibition,
kr = permeability in horizontal direction, md Denver, Ocl 9-12,1977.
kz =permeability in vertical direction, md
N = total number of layers 5. Streltsova-Adams, T.: "Pressure Drawdown in a Well
n = layer with the limited-entry well with Limited Entry," J. Pel Tech. (Nov. 1979) 1469-
p = pressure, psi 1476.
Po = initial reservoir pressure, psi
PO = klhl(Pi - p) / 141.2qwB J,1 6. CinCO-Leg, H, Miller, F.G. and Ramey, RUr.:
Qi = flow rate from layer 1, STB/D "Unsteady-State Pressure Distribution Created by a
qw = flow rate from well, STB/D Directionally Drilled Well, J. Pet. Tech. (November
QiI = dimensionless flow rate from layer i, Qi/qw 1975) 1392-1400; Trans. AIME, 259.
qgn = strength of point source located in layer n
7. Abbaszadeh, M. and Hegeman, P.S.: "Pressure-Transient
r = radial distance, ft
Analysis for a Slanted WeD in a Reservoir With Vertical
554
M. D. ABBASZADEH. N. TOMITA AND T. TANAKA SPE 25664

Pressure Support," SPE Formation Evaluation, 22. van Everdingen, A.F.: "The Skin Effect and Its Influence
(September 1990), 277-284. on the Productive Capacity of a Well," J. Pet tech. (June
1953) 171-176; Trans. AIME 198.
8. Hantush, M.S. and Jacob, C.E.: "Steady Three-
Dimensionl Flow to a Well in a Two-Layered Aquifer," 23. Gringarten, A.C.: "Interpretation of Tests Fissured and
Eos Trans. AGU, 36 (1955), 286-292. Multilayered Reservoirs with Double-Porosity Behavior;
Theory and Practice", J. Pet. Tech. (April 1984) 549-564.
9. Russell, D.G. and Prats, M.: "Performance of Layered
Reservoirs with Crossflow, Single Compressible-Fluid 24. Bremer, R.E., Winston, H. and Vela, "An Analytical
Case," Trans. AIME, 215 (1962) 53-67. Model for Vertical Interference Tests Across Low-
Permeability Zones," SPEJ (June 1985) 407-418.
to. Russell, D.G. and Prats, M.: "The Practical Aspects of
Interlayer Crossflow," J. Pet. Tech. (June 1962) 589-894. 25. Ehlig-Economides, C. and Ayoub, J. A. : "Vertical
Interference Testing Across a Low-Permeability Zone,"
11. Deans, H.A, and Gao Cheng-tai: "Pressure Transients and SPE Formation Evaluation (October 1986) 497-510.
Crossflow in a Multilayer Reservoir: Single-phase
Flow," paper SPE 11966 presented at the 1983 Annual 26. Bourdet, D.: "Pressure Behavior of Layered Reservoirs
Technical Conference and Exhibition, San Francisco, With Crossflow," paper SPE 13628 presented at the 1985
September 5-8, 1983. SPE California Regional Meeting, Bakersfield (March
27~29).
12. Gao Cheng-Tai: "Single-phase Fluid Flow in a Stratified
Porous Medium With Crossflow," J. Pet. Tech. 27. Prats, M.: "Interpretation of Pulse Tests in Reservoirs
(February 1984) 97-106. With Crossflow Between Contiguous Layers," SPE
Formation Evaluation, (October 1986) 511-520.
13. Kazemi, H. and Seth, M.S.: "Effect of Anisotropy and
Stratification on Pressure Transient Analysis of Wells 28. Kuchuk, F., Karakas, M. and Ayestaran, L.: "Well
with Restricted Flow Entry," J. Pet. Tech. (1969), 639- Testing and Analysis Techniques for Layered Reservoirs,"
647. SPE Formation Evaluation (August 1986) 342-354.

14. Javandel, I. and Witherspoon, P.A.: "Analytical Solution 29. Carslaw, H.S. and Jaeger, J.C.: Conduction of Heat in
of a Partially Penetrating Well in a Two-layer Aquifer," Solids, Oxford University Press, 2nd edition 1959.
Water Res. Res. (April 1983), Vol. 19, No.2, 567-578.
30. Gringarten, A.C. and Ramey, H.J., Jr. : "Use of Source
15. Yeh, N. and Reynolds, A.C.: "Analysis of Pressure Data and Green's Functions in Solving Unsteady-State
From a Restricted-Entry Well in a Multilayer Reservoir," Problems in Reservoirs," Soc. Pet. Eng. Jour. (OCtober
SPE Formation Evaluation (March 1989),81-89. 1973), 285-296.

16. Yeh, N. and Reynolds, A.C.: "Computation of the APPENDIX· A


Pseudo skin Factor Caused by a Restricted-Entry Well
Completed in a Multilayer Reservoir," SPE Formation Transient Pressure Solution for a Limited-Enby Well in a
Evaluation, (June 1989),253-263. Layered Reservoir

17. Kuchuk, F.: "Pressure Behavior of Horizontal Wells in We consider an N layer reservoir with a limited-entry well
Multilayer Reservoirs with Crossflow," SPE 22731 located in layer n, as shown in Fig. 1. There are n-l layers
Presented at the 66th Annual Technical Conference and below and N-n layers above the perforated layer. Because of
Exhibition of the Society of Petroleum Engineers, the symmetry about the z axis, only the r and z components
Dallas, Texas Oct. 6-9, 1991. of flow need to be considered.
18. Wilkinson, D. J. and Hammond, P.: "A Perturbation The method of source and Green's functions is used to solve
Theorem for Mixed Boundary Value Pressure Transient for pressure distribution in the reservoir. Variations of this
Testing," Transport in Porous Media, vol. 5, (1990) 609- method has been used in the past to solve many pressure
659. transient problems. 2,6,7,17,29,30 Briefly, ftrSt a solution
for an instantaneous point source (Le., Green's function) is
19. Churchill, R.V.: Operational Mathematics, McGraw. obtained for the configuration of Fig. 1. Next, the point
Hill Book Co. Inc., New York City (1992) 420-439. source solution is integrated along the perforated interval to
yield the instantaneous solution for a fmite-length line-source
20. Stehfest, H.: "Numerical Inversion of Laplace well. The continuous solution for a well producing at a
Transforms," Algorithm 368, Communications of the constant rate is generated by integration of the instantaneous
ACM, Vol. 13, No.1 (January 1970) 47-49. line-source solution with respect to time.
21. Agarwal, R.G., AI-Hussainy, R. and Ramey, H.J. Jr.: Instantaneous Point-Source Solution
"An Investigation of Wellbore Storage and Skin Effects
in Unsteady Liquid Flow: I. Analytical Treatment," Soc.
The instantaneous point-source solution, Gp(r;r',z;z',H), for a
Pet. Eng. Jour. (September 1970), 270-290; Trans.
AIME249. point source located at r' and z' satisfies the following
diffusivity equation and boundary conditions:
555
10 lRANSIENT JESTING OF LIMITED-ENlRY WELLS IN COMMUNICATING LAYERED RESERVOIRS SPE 25664

dGdD'
Tz;~-
dzo
(2(1 +-s)=GdDj =0
Kj
(A-I4)

where the Hankel transformation is defined as: 19


(A-I)

GdDj =j TO GdDjJO(C11'o)dro (A-IS)


at %=0 (A-2) o

The boundary conditions of Eqs. A-8 through A·ll remain


at % = %j (A-3)
unchanged. The solution to the second-order ordinary
differential Eq. A-I4 is:

(kJlz ) !!!.eL =(kz ) ~ at % =%j (A-4)


j i1z Jl Oz
j+1 (A-I6)

iJGpN =0 at % = %N (A-5) where.,


iJz
Gpj = Po at t =0 (A-6) vl =_1_ ((12 +.!....) (A-I7)
Tz; Kj
The solution to Eq. A-I may be expressed as the sum of two
Green's functions, Gdi and Gsi; where, Gdi satisfies the
diffusivity equation without the source term and Gsi is for a Determination of the Gsi Term
point source in an infinite reservoir to model the source term.
The Green's function for a point-source in an infinite
Determination of the GdLImn anisotropic reservoir is given by19,30 :

The diffusivity equation and the associated initial and


boundary conditions in terms of dimensionless parameters
defined by Eqs. 8 through 13 for the Green's function Gdi
become:
In terms of dimensionless parameters and for r'=O:

(A-7)
(A-I9)

(A-8)
The Laplace transform of Eq. A-I9 is;

(A-9)

(A-20)
(A-IO)

(A-H)
and, the zero-order Hankel transformation of Eq. A-20 is:
The Laplace transforms of Eq. A-7 is,
(A-21)
1 a lKJ
- ,
dD) 2~udD' s-a
- - T O - - ' +TZi~--GdDj=0 (A-12)
TO Oro ( dro iJz o Kj Therefore, the point-source solution for the layer containing
the source is given by;
where,
(A-22)
(A-13)
Defemination of Coefficients Ai and Bi=-
The zero-onler Hankel transformation ofEq. A-12 with respect
to I'D results in: Applying the boundary conditions of Eqs. A-8 through A-IO
to Eqs. A·I6 and A-22 results in a system of 2N equations
556
Mr Dr ABBASZADEH, Ne TOMITA AND Ie TANAKA SPE 25664

with 2N unknowns for coefficients Ai and Bi. Generally, the APPENDIX • B


resulting system of equations can be solved numerically by
any numerical matrix solver. However, the Guassian Transient Presswe Solution for Dual Limited-Entry Wells in
elimination/substitution method is used here to derive analytic a Two-Layer ReservQjr
expressions for Ai and Bi. In particular, the results for layer n
containing the source are: We consider a two-layer reservoir where each layer is partially
completed and is flowing into a common tubing, as shown in
Fig. 2. The method of source and Green's functions is also
(A-23) used to obtain pressure distribution in the system. Here, we
directly start with the instantaneous line-source solution and
its integration with time to obtain the pressure field.
(A-24)
(B-1)

where, the recursive relationships for pare:


whose Laplace and Hankel transformation gives :
fJi+l (Ei + 1)e-2Vi+l 1 Di+l
+ (Ei _1)e-2Vi+lIDi
f3i = f3 i+ 1(Ei - l)e-ivl+1'0;+1 + (Ei + l)e-ZV'+I'0; PD «(J, s) =GUi ( (J, s) + q (s)G Ls ( (J, s)
di (B-2)

for nSi < N (A- 25) The instantaneous finite-length line-source solutions, GLdi,
for layers are obtained from the approach presented in
fJi-l (I + Ei )e-2ViZDi-l + (1- Ei)e-2ViZDi Appendix A. Therefore, the point-source Green functions can
be taken from Eq. A-22, integrated and substituted in Eq. B-2:
f3i = f3i_l(I- Ei)e-2V;'0;-1 +(1+ Ei)e-ZV;'o;
for lSi < n (A-26)

and,
(B-4)
A.·V·
E· = & & (A-27)
& A.i+lVi+l
where,
Finite-Len~th Line-Source Solution
(B-5)
The Green's function for a finite-length line source well is
obtained by integrating the point-source solution, GPDn,
along'the perforated interval. The result is: (B-6)

-Un {a [a e-V,.(2Zo.-1-ZD) + e-V,.( 2Z o.+Z D)]


Gw = 2bv; bl and b2 are penetration ratios and A21 =A'lfA1. As before,
PII PII-l
boundary condition of Eqs. A-8 through A-I0 in terms of
(e V.'60 - eV.'''' )+(f3l1eV.'D - e-V.'D Xe-V.'60 _ e-v.,.o )} pressure apply. Using these boundary conditions along with
pressure Eqs. B-3 and B-4, expressions for coefficients AI,
(A-28) BI, A2, B2 are obtained.

(B-7)

(A-29)

Pressure distribution in a reservoir is the continuous form of


the instantaneous line-source solution. This is given by
integration of the instantaneous solution with respect to time,
which in Laplace and Hankel spaces is written as, Az =e-ZV2 '\ D{[r(all - a1Z )- (all + a1Z )e-zv2'2 ]qiD
== 1== +[ra3e-ZV2 '2 + (1 + r )a4e- V2 '\ )rZD} (B - 9)
PDII(U,ZD'S) =-GW(U,ZD'S) (A-30)
S

Bz = D{[r(all -alZ)-(all +a1z)e-v2' 2}rID


Therefore, the wellbore pressure is obtained from the
combination of Eqs. A-28 and A-30. The resulting expression
for the wellbore pressure, however, is a function of depth, z. +[ra3e-ZV2 '2 - (1- r)a4 ]i'1zD} (B-IO)
where,
557
12 lRANSJENf TESTING OF LIM1TED-ENlRY WELLS IN COMMUNICATING LAYERED RESERVOIRS SPE 25664

(B-ll)

(B-12)

(B-13)
Transient layer rates are obtained by requiring that pressure
from the individual layers to be equal at the wellbore, and that
a 4 = ~[e-V2(2'2-'U) _ e-V2(2.2-.02)] (B-14)
the sum of layer rates be equal to total production rate from
v2 the reservoir. This yields two equations with two unknowns
for layer flow rate whose solutions are,
D_ 1 (B-15)
1- (l_r)e-2V2'2 + (I + r}e-2V2" _ 1 V;-U2
and,
(B-29)
q1D = 2s qV ( V -U ) + b v A.:z1 (U - \-1 )
1 2 2 2 2 1

(B-16)
- 1 U; - ~ (B _ 30)
qlD = 2s qV1(V2 -U2)+b2v2A.:z1(U1 - \-1)
and £ is dermed by Eq. A-27 for i=1. Average dimensionless Wellbore pressure is obtained by substituting the above
wellbore pressures from layers are derived by integrating expressions in either Eq. B-17 or B-18,
pressure solutions along their perforated length. The results
after integration become,
(B-31)
(B-17)

(B-18) ADDmON OF LAYER SKIN FACTORS

where, Generally, the skin factor of layers are different from one
another. The above equations can be modified readily to
u; =~(sinhv1zb1-sinhv1z"d include the effect of layer skin factors. Using the van
Everdingen's22 skin formulation as an additional pressure
+"!"[I- cosh v1(Zb1 - Z"1)]+"4 (B-19) drop proportional to layer flowrate and open interval
V1 thickness, the wellbore pressure is related to layer pressure as,
U2 = "2 (sinh V1Zb1 - sinh v1z"d (B-20)
(B-32)
De
"1 =~;-----:--­
v1(Zb1 - Z"1) cosh V1z1 (B-33)
[(au + a 12 )e-2V2' 2 + (all - a 12 )e-2V2" ] (B-21)
Equating Eqs. B-32 and B-33, and using the condition that
sum of layer rates is equal to well's production rate, the layer
Uz = De [a 3e-2V2'\ +2a4e-V2 "] (B-22) flowrates are obtained as,
v1(Zbl - Z"1) cosh v1Z1
1 -v (. _.) 1
"4 = Zb1 - Z,,1 + -e \ ., 01 - -
V1 v1
(B-23)

and,

~= v (D ) (eau +a12)
2 zb2 - z,,2
[a4e-V2" + e-V2 ('02 +.,) _e-V2('u+',)] (B-24) Wellbore pressure is derived by substituting the above
expressions for layer rates in either Eq. B-32 or B-33.
- 1
V2=-:----~
V2(Zb2 - Z"2)
[w1(e V2 ' b2 _e V2 ' 02 )-w2(e- V2 ' U _e-v2 ' 02 )+2W3 ] (B-25)
558
SPE25664

T ABLE-1 Two-Layer Reservoir Parameters Used in


Comparitive Numerical Simulations

Parameter ~ ~ ~
hI, ft 400 400 400
h2,ft 40 40 40
za,ft 400 390 270
Zb,ft 440 400 350
kr2/krl 100 1000 1
+1/4P2 1 1 1
kz1/krl 0.1 0.1 0.1
kz2/ka 1 1 0.1

TABLE-2 Two-Layer Reservoir Parameters Used


in Interference Test Studies

Parameter casu ~
hI. ft 400 400
h2. ft 40 40
za. ft 270 270
Zb,ft 350 350
z'a,ft 410 270
z'b,ft 440 350
+1/4P2 1 1
kz1/kr1 0.1 0.1
kz2lka 0.1 0.1

- - - - - - - - - - - - - - - - - - - - - ZN
Layer N

Zn

I
;';';'."
:.:.:.:.
Layern (k rn • kzn • hn• lPn) :.'.;.:.

::.:.:.: - , . -

Zn-1
Za Zb

Layer 2

Layer 1

FIG. 1 - Schematic of a mUltilayer reservoir with a limited-entry well in one of the layers

559
· ·:
I
kr2, kz 2, h2, <112
K~~~~_+-1_________
- - - - - - -__ n

~.-.;-.
•.•...
Layer 2
Z2

Z1
~
': k:::::::::::::::::::::: :::~~::::::::::::::::::::::: . .J
- I I :::['
~I~.
Zb21 Za21 1_ 4'

II
--
't

0.1
CaseC

Case A
<:l
.~ A,-WilllTrililgh Perm layer
k r 1, kZ 1, h1, <111 Layer 1 B: well In tow perm layer
0.01
c: well in Homogeneous res.
Case B
Une = Analytical
Za1 I Zb1 Symbol = Numerical

lr 0.001 Ii- ,
10'
""",I
10'
! " ,,"..

10'
! " ,,""

10'

to
'"

=
"",I
10'

0.000263"'t
4J,Jl1c,rw
' " "",I

2
10·
!" "",I
10'
' " "",I
10·
" " !!Ill!
10·

FIG. 2 - Schematic of a two-layer reservoir with dual completions


FIG. 3 - Comparison of analytical and numerical solution for
the two-layer reservoir with properties of Table 1

'"gJ
1
10 F I Iii i iiil ' i i i i liil iii i i ilil i i i I i ill! j' i i i iii! i i i i I lib
10 E' I

10·
kik,=1

~ ~ k2 /k l =10
:::[' ... -,~~~;_~._:;.~.~-;_ ..............
:::[' 10" kik ,=10
~I~. ~I~ ":;':::.~~,
4'
-
't
11- 10"
<:l
kik,=100
4'.

II
<:l
-
:::t 0.1 \, ",.
" ""
;.. \..............
" \
" \
" \
'" 100
..~:.:-.;'l.ft. __

-~ kik,=1000 -% k B / k l = 10'5 " \


.. \
- 3 layer case ... 0.01
kB / kl = 10-4 \. '- 1000
10" ...... 2 layer case '. ,
kB / k l = 10'3
kik,=10,OOO,000 kik,=10,000 - No barrier (2-layer)
";0
"~~3
, ! , " "~,I '" " ! ! , "",1 , , \ , " , •• 1 '"" ",*, 0.001 L , """'! " ,,,1111 "'III"! 1 I 'pili' "'ii!lll " """! ","'". , ! '11111'
10" b • ! , !! 10" 0 000263k,t
0
10· 10' 10· 1~ 1~ 1~ 1~ 1~ 1~ 1~ 1~ 1~
. 10' t _. 2
0.000263k,t
0- 4Jd.J.1C,rw to = 2
4Jil-L1C ,rw
FIG. 4 - Pressure derivative as a function of k2 /k l for FIG. 5 - Pressure derivative for the semi-permeable barrier model
two and three layer reservoirs of three layer reservoir
SPE256 64

~
!.
Start of Partial
Penetration Effect

-o::~~
~.

II
--
""'"
<:l
0.1

-~ 0.Q1
- za!h, =0
...... Za! h, =0.4
Za! hI =0.8
0.001 ~"""""""""'uL:-..J,.;..L...LU.wL:-'-.L.LL.LWL~""""'''''''''L:-''''''''LL1olJuL:-..J,.;..L...LU.wL:-'-.L.LL.w.IlI
101 10· 105 106 107 106
O.OOO263Ajt
tD = 2
4J..f.l.tc,rw
FIG. 6- Sensitivity of pressure derivative to location of perforated interval
as a function of contrast in k21k1 >1 for two layer reservoir

100

b=0.025
10

~
::(" k 2 !k1 =1

~~
:;
II
<:l
- 0.1
k 2 !k1

~!k1 =100
=10

-~
0.Q1

k2 ! k 1 =1000
0.001
10' 10' Hi' 10' 10' 10· 107 10· 10·
O.OOO263Ajt
tD = '\Jl\c,r;
FIG. 7 - Sensitivity of pressure derivative to the length of perforated interval
as a function of contrast in k2/k1 > 1 for two layer reservoir

561
- kz2/k r2 =1.0
........... kzi kr2 =10. 1
II ...... kzik r2 =10-2
C:l
0.01 .- _. kzikr2 =10.3
-~
._._.. kz2/k r2 =10. 5

0.001 ~~"""""'..I.I.I.-...,I.-,.,JI.o..oI-&.Iu.u.a..---'-""""a...&...I.U.L..-..L...L...""""""I.lolo- .........................u.u..-...a....a....u~---..I--'-I""""""

102

FIG~ 8 - Sensitivity of pressure derivative to permeability anisotropy -


two layer reservoir

H__'-0",
~~~ "--__ ..-- Layer 2

~ 11 II \\ l~
'Ii!; L_aY_e_r_2

I
I +- Layer 1 ..-- Layer 1
I
I

FIG. 9 - Conceptualization of formation crossflow in a two layer reservoir

562
SPE256 64

0.2 0.1? i
4
0.0 <!:'c 0.0 to = 10
-0.2
............
~
a:co -0.1 '.
•••••••••••••••••••. L\
~ ~
-0.4 ~ \
g ~ -0.2 \... •••••••••.

>
c -0.6
to = 10
5
~ \ /
.9
-0.8 ~ oo~ .' \
.~
-
5
-1.0 .0.4 to = 10
-1.2 ~
-1.4x10·3
~".~'

8 -0.5

10 100 1000 10000


ro ..O..6h f , , ""-1' '" " ••• , '" """ '" "",' I "
1 10 100 1000 10000
ro
FIG. 10 - Formation crossflow flux in a two layer reservoir,
limited-entry well in low permeability FIG. 11 - Cummulative formation crossflow rate in a two layer
reservoir, limited-entry well in low permeability

01'
e>
c.>

0.2? r= 0.2
Line Source Solution, Po = 0.5Ei(ro2/4to)
10

O'0F::::::: ···--··········::~·······::::..::7··
-0.2~ "' ~ •••••••••, ~
J
0.0

-0.2 ~
at ro = 1000
100
10
;

-0.4 -0.4 . .<. JI~


.~~ 2

c v-:1~
>u -0.6 -0.6 rP
o ~
.9 +-- II
cS
~
-0.8 -0.8 ~

-1.0 -1.0 0.1

-1.2 -1.2

10· 105 106 107 108 109 10 10


10 100
ro O.OOO263L(kh) t
tOt L(4Jh)JLICt~
FIG. 12 - Formation crossflow flux and cummulative rate,
limited-entry well in high permeability FIG. 13 - Pressure transient behavior of partially penetrating observation well in
a two layer reservoir, active well in layer 1 and observation well in layer 2
1r--..-r"T"I'Tmr-rTTnm.......-r"T"I'Tmr-rTTnm.......-r"T"I'Tmr-rTTnm...-..-r.,.,..,....

- ro=1000
.••... ro=100
000 ro=10

0.OOO;l63})kh) I
ID, = ~(tjIh) #IC,?w

FIG. 14 - Pressure derivative of partially penetrating observation well in a two


layer reservoir, active well in layer 1 and observation well in layer 2

10
-
... ,
.. ---- --- ... --- .
,..... • I;;;.. ·~ ••••••••I ••••••••

•••••••••••••••••• •••••••..•••• kzlkl- 000

.."".............•.
..........
/*.,~ ••• _
~.,~
"4",-..
.100
.10
••••••' i0\'4'-:"
••, / 'S4~-:""II..,II:e~~~_~~~~~~
:/ .
'~Jl ...
i
:'
••- ro=1000
'..... ro=100
0.01 ro=10

10'

FIG. 15 -Transient behavior of partially penetrating observation well in a


two layer reservoir, both active and observation well in layer 1

564
SPE256ob,

~'1.;n':-'-"""'''''''''1iL:n':-'-'''''''''''''''''10l.:.-..a..'''''''''''''''''101.:.-..a.................0"1.:-.................0L.1:--'-"-Uu.wIOL.:-·- '-..............
1O.
1 1 1
O.000263I,(kh) t
tD, - I,(fth) JtlC,Z

FIG. 16 - Pressure derivative of dually completed two layer reservoir

1.0 ~-------------:-"'""::"::-::-:--:-=:-::-:=-=-="':":"=-=-:=-=-=--:-=
-------------------------------------------------2--
.................................................................................................
......................................
0.8

~ 2

~
8 0.4 =-------::--====::::::=======1
=------
0.2
.......................................... 2

O.OOO263I,(kh) t
tD, = I,(fth) JtlC,Z

FIG. 17 - Transient layer flow rates in a dually completed two layer reservoir

565
10 tot··..··················-·······························•...•.•....•...•...•...! - .

6
i:
~ 51
4
i
!
~

----------------I--~~~
3
"i: 2
Q)
C
~ 1
~
~
:J
,... O.
~
R~gm>z:p'rzu"~
i

~ ,:
0.. !
;,
m f/)
f/)
! i
~ 0.1 ......................................................······················1·······..···································································r········· .
c::
o
"00
c::
Q) Line =Analytical ! i
E Symbols =Field Data 1 i
C
0.01 H· · ~ ·r r..·r..rTrfi · ·..· ·r · r ·I
I f ..rlTrt
I· · r - -~ _nmrnrT--rll±l
6 8
10 10
Dimensionless Time

FIG. 18 - Interpretation of field pressure buildup test with three layered model

Вам также может понравиться