You are on page 1of 12

SPE

Society of PetroIeoJn Engineers of "'IME

SPE 13628

Pressure Behavior of Layered Reservoirs With Crossflow


by D. Bourdet, Flopetro/ Johnston
SPE Member

Copyright 1985, Society of Petroleum Engineers

This paper was presented at the SPE 1985 California Regional Meeting, held in Bakersfield, California, March 27-29, 1985, The material is subject to
correction by the author. Permission to copy is restricted to an abstract of not more than 300 words. Write SPE, P.O. Box 833836, Richardson, Texas
75083·3836. Telex: 730989 SPE DAL.

that reservoir heterogeneities are very frequently observed. For


ABSTRACT example, it is not exceptional that long build up tests never reach
Experience with a large number of well tests has shown that the h'omogeneous infinite acting radial flow behavior, traditionally
the analysis of some field tests using currently available interpreta- assumed when semi-log straight line analysis is attempted.
tion models is not satisfactory. Marked departure from homoge- Taking advantage of the computing facilities now available,
neous behavior is evident and none of the available heterogeneous even at the weBsite, a new approach as been recently proposed for
models, such as the double porosity models, yields a convincing well test interpretation. It considers not only the pressure response
interpretation. This is particularly apparent when the pressure but also the derivative of the pressure 1 ,2. It provides much more
derivative behavior is examined. characteristic response curves for analysis and in particular ampli-
In this paper, a new analytical solution is presented that des- fies the small variations produced on pressure behavior by reservoir
cribes the pressure response of a well intercepting a layered reser-
heterogeneities. The use of pressure derivative confirms frequently
voir with crossflow. The model is designed to describe a system
of alternating beds of relatively high permeability contrast. The the need for heterogeneous interpretation models but, on the other
interpreation models currently in use (homogeneous reservoir, two hand, the high sensitivity of the derivative curves shows also the
layers without crossflow, double porosity reservoir) are shown to limits of the presently available analytical solutions. It is found, in
be limiting forms of the new solution which is therefore a more some cases, that none of the usual heterogeneous models provide a
general description of transient pressure responses. go~d match on the derivative data plot even when conventional
The layered reservoir behavior is discussed in terms of the pressure curve analysis seems to reproduce reasonably well the
pressure and the derivative of pressure. The limits of applicability recorded pressure response. .,
of the homogeneous and the double porosity solutions for analysis In this paper, we will first review briefly several double porosity
of tests in layered formations are clearly defined. models used for analysis of heterogeneous formation responses, we
The new model is used to analyze an actual field case and will discuss the different hypotheses introduced in their mathemati-
a good match is obtained when no other solution is found to be cal derivations and will illustrate the difference in shape of the
applicable. In addition to the usual well and reservoir parameters resulting curves. Then, we will propose a new analytical solution
(total permeability thickness product, skin and well bore storage for a two layers system with cross-flow, present typical curves and
constant), analysis yields quantitative information on the storativity compare them to the classical heterogeneous responses. The ap-
plicability of the new model to actual reservoir configurations will
and permeability of the contrasting layers. be discussed and an example analysis of actual pressure data will
be presented.
... ...
INTRODUCTION DOUBLE POROSITY MODELS:
Well pressure responses, recorded during tests, are interpreted Several double porosity solutions have been proposed to des-
to provide a description of the well condition and various informa- cribe fissured formation responses. In all cases, two media are con-
tions on the reservoir. The analysis is carried out on pressure versus sidered: one, the fissures system, is of high permeability but does
time plots, either by identification of straight lines on portions of not store much of the tluid in place. The other, the matrix system,
the data (and determination of parameters of interest from slope has a very low permeability but contains most of the oil. One of the
and intercept values), or by curve matching on the global response. basic assumptions of the double porosity models is that all the oil is
This last approach considers the pressure data as a whole, includ- produced to the well via the fissures only, hence the matrix system
ing the curvatures present in the plots, and therefore it ensures a having no effective connection to the well. In the mathematics, this
higher degree of contldence in the results of interpretation. Curve assumption is implemented by making the permeability to the well
matching technique is now currently used to analyse a large variety for the matrix system zero.
of well and reservoir contlgurations.
As a result of the progress recently observed in pressure gauge
technology, pressure data reveal much more information than in
Pseudo steadllstate interporositll flow:
the past. Practice of curve analysis in well test interpretation shows
Va.rious types of matrix to fissure flow ha.ve been investigated.
References and illustrations at end of paper One, considered by Warren and Root 3 assumes the pseudo steady

405
2 PRESSURE BEHAVIOR OF LAYERED RESERVOIRS WITH CROSSFLOW SPE 13628

state interporosity flow condition. The effect of wellbore storage The first fissure flow being short lived with transient interporosi ty
and skin was introduced on this solution by Mavor and Cinco" and flow models, the (CDe(2S»), curves are not seen in practice and
a pressure type curve was presented by Bourdet and Gringarten 5 • therefore have not been included in the derivative components.
More recently, the derivative type curve for a well with storage and The dual porosity derivative response starts on the derivative of
skin in a double porosity system, pseudo steady state interporosity a p' transition curve, then follows a late transition curve labelled
flow, was introduced 6 as illustrated on fig-I. )..(CD )/+m/(1 - W)2 until it reaches the total system regime on the
Two parameters characterize the heterogeneity. The storativty 0.5 line.
ratio w gives an indication of the percentage of oil stored in the Two different models are illustrated on fig-2: one assumes the
fissure system: matrix blocks are spheres, the other corresponds to slab matrix
w = ..,----,:.(<P,-C-,-th-:,).L'---,-_ (1)
blocks. The two different geometries produce slightly different sha-
(<pCt h), + (<pCth)m pes on derivative curves as illustrated fig-3 8 •
The p' transition curves have the same shape as the homoge-
The interporosity flow parameter, ", shows how long it takes to neous responses. In fact, they are simply CDe(2S) curves displaced
reach the transition regime, i.e. the start of the matrix contribution by a factor two along pressure and time axes 6 • As a result, pres-
to the production: sure curves tend to produce, during transition regime, a semi-log
,,= 2km
arw -
k,
(2) straight line, of slope half of the true radial flow semi-log slope. On
derivative curves, the p' transition components stabilizes at 0.25
Two examples of double porosity behavior are illustrated on fig-I. level (instead of 0.5 for the CDe(2S) curves). In fact, actual double
The pressure response is described by three component curves5 : the porosity responses do not always follow perfectly the component
curve starts on a (CDe(tS»), during the fissure flow, then reaches curves, which are only limiting forms, and the semi-log behavior is
the transition regime described by a )..e- 2S and finally follows the not always present during transition 8 •
total system behavior on a (CD e(2S)),+m (respectively for example The presence of a sealing fault near the well produces very
A: (CDe(2S)), = 1, "e- 2S = 3.10-4, (CDe(2S»)'+m = 10- 1 and similar pressure curves. On fig-3, the responses for sphere and slab
105 , 10-7 , 10" for B). geometry are compared to a curve generated for a homogeneous
The derivative plots match on four component curves 8 : fissure reservoir with a sealing fault, with pressure change and elapsed
flow on a (CD e(2S»), derivative, then the transition regime, ex- time divided by tw0 8 • This example illustrates the non unicity of
pressed by a depression, is described first by a early transition the solution frequently encountered in welltest interpretation. This
curve, ["(CD)l+m/w(I - w)] (3.33.10- 4 for Aand 1.11.10- 2 for B), will be discussed in more detail at the end of paper.
and after by a late transition, ["(CD) I+m/ 1 - w], (with 3.33. 10- 5
and 1.11 . 10-3 ). Total system flow is reached on the 0.5 line. Dooble porosity derivative shapes:
Omega defines the displacement between the two homogeneous
behaviors of the pressure response (for small w values the two To summarize derivative responses, transition behaviors are
(C De(2S»), and (CD e(2S»)l+m curves are very different) and" fixes very different between the two types of double porosity model.
the pressure level during transition (with small "e(-2S), the transi- With the transient interporosity flow solutions, the transition starts
tion occurs at high PD). from early time and does not drop to a very low level. Frequently,
On the derivative response, the influence of the heterogeneous it does not even reach the 0.25 lines. With pseudo steady state
parameters can be explained as follows: w defines the shape of the interporosity flow, the transition starts later and the shape of the
derivative curve during transition regime, i.e. the duration and the depression is much more pronounced. There is no lower limit for
depth of the depression. Duration and depth of the transition are the depth of the depression, it is only dependent on the duration
linked, a small storativity ratio produces a long and therefore deep of this regime, between the end of fissure flow on (CDe(2S)), and
transition. The second heterogenous parameter, the interporosity the start of total system behavior on (CDe(2S»)'+m .
flow coefficient).. defines the position on the time axis of the transi- Layered systems with crossflow are known to produce in some
tion regime. A decrease of " value (all other parameters being un- cases a characteristic heterogeneous response but it is not always
changed) moves the depression to the right side of the plot, along possible to find a good match with any of the two usual double
the 0.5 line. porosity models, in particular when the derivative of pressure is
Transient interporosity flow : used. The shape of the transition is something between the response
produced by the two different interporosity flow hypothesis: it goes
A second type of double porosity model assumes transient in- below the 0.25 line of the transient interporosity flow model but the
terporosity flow. Incorporating wellbore and skin in the solution depression is not as sharp as a pseudo steady state inter porosity
proposed by de Swaan 7 , Bourdet and Gringarten presented a pres- flow transition.
sure type curve 5 on which the derivative curves were later added 8 In other cases, layered reservoirs with crossflow behave like
as illustrated on fig-2. a homogeneous system. The assumption that layered reservoirs
The type curve is built in the same way as for the pseudo behavior is the same as fissured formations is only valid in par-
steady state case of fig-I. Theoretically, pressure behavior is defined ticular cases. The intermediate behaviors between the homogeneous
by three component curves, (CD e(2S»), , p' and (CD e(2S))/+m response and the double porosity response have also to be con-
The p' dimensionless group is expressed as 6 : sidered. The two layers with crossflow solution presented here will
be shown to produce all the intermediate responses between the
homogeneous solution and double porosity, pseudo steady state
(3) solution.
*
with, for spheric&!. blocks,
LAYERED SYSTEMS:
Layered configurations, very frequently encountered in actual
6' = 1.0508 (4) reservoirs, have been the subject of many studies. Layered systems
without crossflow, when the production is commingled at the well,
and, for slab matrix blocks, or crossflow in the reservoir have all been considered~ to 1G •
In the present work, we consider an infinite system made of
6' = 1.81114 (5) only two communicating layers. It is an extension of the Warren
406
SPE 13628 D. BOURDET 3

and Root model 3 to the case where both media are in contact with -CDe(2S) defines the well condition
the producing well. In other words, the matrix system permeability -It and w define the contrast between the two media
is not neglected and a radial flow rate occurs within the matrix -A defines the interaction between the two media.
system as well as from the matrix to the well. A dual system will be On figures 4 to 7, two families of double permeability curves
shown to be representative of complex conflgurations made of more are presented on semi-log and log-log scales. For all of them, CDe(2S)
than two layers. This type of modelization has been referenced as has been fixed to 1 and the crossflow parameter ~e-2S to 4·
double permeability behavior in the literature. 10- 4 • Five different values of It are used: 0.6, 0.9, 0.99, 0.999 and
1. On figures 4 and 6, the curves are generated for w = 10- 3
* and on figures 5 and 7 for w = 10-1 •
Three different regimes occur in a double permeability response
DOUBLE PERMEABILITY MODEL: (appendix C):
The equations describing fluid flow in medium 1 and 2, can be
-at early time, before the crossflow is established, the response
expressed as follows g :
is the same as for a two layers without crossflow.
-at late time, the system reaches a homogeneous behavior
characterizing the total producing system (total permeability
thickness, total storativity).
-at intermediate time, a transition behavior establishes.
(6)
On the semi-log plots (fig 4, 5), the dotted curves represent the
the assumptions for the horizontal flow are same as for the usual early and late time behaviors, namely the two layers no cross flow
analytical solutions developed for petroleum engineering. Vertical solutions (A = 0, appendix C) and the total homogeneous response
cross flow from layer 2 to layer 1 is assumed to be in pseudo steady (CDe(2S) = 1). The solid lines are the global double permeability
state regime. The constant a is introduced for consistency with responses. On the log-log plots (fig 6, 7), the double permeability
Warren and Root symbols and it will be discussed in more details curves are presented in terms of pressure response and derivative
at end of paper. of pressure. Figures 4 to 7 show that each parameter can modify
the. shape of the responses:
The following dimensionless terms are used, the values of ap ,
It: kappa affects the shape of the transition. For large It (close to
at and ac depending upon the system of units used (table 1).
1), the transition tends to flatten on the pressure plots, there-
klhl + k2h2 fore producing a very characteristic heterogeneous response.
PlD,2D = B
apq I'
(Pl,2 - p) (7) For low It (0.6), the "8" shape of the response is less pro-
nounced. On the derivative responses, the depth of the tran-
tD = at(k1 hl + k2h2)t sition increases with It, the limit being the double porosity
(8) response generated with It = 1.
[(¢Cth)l + (¢Ct h )21Ww 2
w: for small omega, the two layers no crossflow early time
r curve is very different from the late time curve "equivalent
rD=- (9)
rw homogeneous system", the transition is long (and therefore
klhl deep on derivative curves). For large w values, the global
It= (10) response tends to the homogeneous behavior curve for the
klhl + k2h2
total system.
~ Ae2S : for a given Ae 2S value, all the curves merge at the end
= arw 2 klhlk2h2
+ k2h2
(11)
of the transition: Ae- 2S defines the pressure of the start of
(¢Cthh the total system behavior.
W= (¢Cthh + (¢Cth)J
(12) The influence of the three heterogeneity parameters can be
summarized as illustrated on figures 8 and ll: the curves are gene-
CD = acC rated for several storativity ratios w (10- 1 , 10- 2 , 10- 3 , 0), all other
(13)
[(¢Cthh + (¢cth)21rw 2 parameters being fixed (CDe 2S = 1, It = 0.91l, Ae- 2S = 4 ·10-'.
The curves merge during the late transition, before the total system
Kappa is the ratio of permeability thickness for layer 1 to the total
regime is reached:
permeability thickness (traditionally not considered in the double
porosity hypothesis where It = 1). -At early time, the response is defined by four parameters:
(C D e2S h+2, It, ~e-2S and w.
Equations 6, expressed in dimensionless terms, describe a sys-
tem similar to the one proposed by Wijesinghe and Culhamu : -w is not acting any more at late transition time.
-The time of start of the total system flow is independent of
ItV2PlD = W
OPlD
8tD
- ~(P2D - PlD)
It.
-At late time, the only governing group left is CDe2S .
In conclusion, the heterogeneous characteristics of the respon-

i (1 -1t)V 2 PlD = (1 - w) 8P2D


otD
+ ~(P2D -
After transforming the above system in the Laplace space, it is
PlD) (14) ses are much more prononced for large permeability thickness ratios
It and small storativity ratios w (high diffusivity contrast). This
situation occurs for example in reservoir made of tight layers (such
solved analytically according to the method proposed by Kutlja- as silt) separated by thin fissured limestone breaks: the fissures
rov u as demonstrated in appendix A. In appendix E, wellbore provide most of the permeability whereas the tight layers store
storage effect and two skins 8 1 an 8 2 are introduced and the solu- most of the oil.
tion is simplifled to one single skin effect for both layers in appendix This condition of a high diffusivity contrast validates the use
C. Only this solution (Eq C.l) is considered in the following. of homogeneous interpretation model for many layered reservoirs:
when the diffusivities between layers are of the same order of mag-
Double permeability behavior: nitude, the responses do not show any trace of heterogeneity.
On the other hand, when the contrast is high beween several
When double permeability responses are plotted with the di- layers, complex systems cal). be modeled by using only two elements
mensionless pressure PD versus the dimensionless time group tD/CD, (with three or more media, the high contrast condition would not
the curves are defined with four parameters (appendix C):
407
4 PRESSURE BEHAVIOR OF LAYERED RESERVOIRS WITH CROSSFLOW SPE 13628

be met). In other words, layered systems behave like two layers POSSIBLE EXTENSIONS:
with crossflow: layer 1 is the sum of the high permeability fissured
breaks, layer 2 is the sum of the low permeability zones. With the development of computerized interpretation techni-
ques, it is now possible to analyse well test data faster 20 and to
Limits 0/ the double porositll solution : obtain a much more detailed reservoir description 2l . New analyti-
cal solutions are frequently proposed. For example, Cinco-Ley has
Figures 6 and 7 illustrate that the double permeability solu- recently considered in the double porosity solution the infiuence of a
tion can produce all intermediate behavior between standard ho- skin effect at the surface of the matrix blocks 22 . A similar approach
mogeneous responses and the double porosity solution that assumes for layered systems has been discussed by Gao Cheng-TailS with
pseudo steady state interporosity fiow. For low permeability thick-
the semipermeable wall model. This has to be further investigated
ness ratio (I\': = 0.6), the response is very close to the homogeneous
for a better understanding of the inter layer flow behavior and the
curve corresponding to the total system. For I\': = 0.9, the devia-
related parameters, hand 0'. The double permeability solution can
tion from the homogeneous response is already significant but the
double permeability curve is still very different from the double also be extended to interference testing between wells (appendix D)
porosity curve (I\': = 1). A I\': of 0.99 is clearly not enough and even or to vertical interference testing 23 .
when the ratio of the permeability thickness of the two media is
one thousandth, (I\': = 0.999), the response does not fit exactly with
the double porosity solution. As a result, application of the double
porosity model to layered configurations must be performed with *
caution. Even if both models provide a good match, the use of
double porosity solution on layered reservoir responses may result
in an incorrect estimate of the heterogeneous parameters.
This is illustrated on the semi-log and log-log plots of figures
CONCLUSION
10 and 11: a double porosity curve (CDe 2S )/+m = 100, W =
0.222, he- 2S = 10-& is shown to be very similar to the double A new analytical solution has been proposed to describe the
permeability curve generated for C D e2S = 100, W = 0.1, I\': = pressure response of layered reservoirs with crossflow. The solution
O.g,andll.e- 2S = 7 . 10- 7 • By using the double porosity model on
double permeability data,the storativity ratio W is overestimated. is shown to be general and includes the traditional homogeneolls
Pressure responses of the two models can be very close when reservoir solution, the two layers without crossflow solution and
the storativity ratio is large and therefore the transition is of short the double porosity pseudo steady state interporosity flow solution
duration: the early time no crossfiow regime produce curves very as limiting forms. This double permeability model ca.n produce all
similar to homogeneous reservoir responses and can be matched on intermediate behavior between the homogeneous type response and
a CDe 2S curve as a fissure regime. In addition, double permeability that of double porosity. As a result, the new model demonstrate
responses are found, in some cases,. to exhibit two parallel semi- that the applicability of double porosity model for layered reser-
log straight lines, usually considered characteristic of the double voirs is very restricted.
porosity model.
When the storativity ratio is small, the transition is longer and The double permeability responses exhibit three characteristic
its shape is well defined. In general, the solution is unique and only flow regimes: first a two layers without crossfiow behavior, then
one model matches the data. a transition and finally a homogeneous behavior representing the
total system.
* The use of the model was illustrated with an actual build-up
EXAMPLE OF APPLICATION that no other published solution was found to describe accurately.
The use of the double permeability solution is demonstrated on
a 13 hour long build-up, performed after one month of drawdown Experience has shown that, unlike the example presented, in-
(data table 2). A log-log plot of the pressure data shows, especially terpretation is frequently not unique: several types of reservoir
in terms of derivative, a characteristic heterogeneous shape: after configurations can produce similar responses that match the data
the usual hump corresponding to well bore storage, the derivative equally well.
follows a depression before it stabilizes at a constant level.
Several matches were attempted with the usual heterogeneous The development of high accuracy pressure measurements com-
interpretation models and none was found to provide a satisfactory bined with the introduction of computing facilities at the wellsite
solution: double porosity pseudo stady state transitions are too reveal the need for a more extended set of theoretical models for
sharp, transient model curves are too smooth in transition regime. well test interpretation in order to make full use of the recorded
The introduction of one or two sealing faults cannot reproduce any data. Small variations in pressure response are meaningful and can
better the derivative shape of the build-up data. Finally, a good be &nalysed to provide a detailed description of reservoir dynamic
match was obtained with the double permeability model as shown behavior.
on figure 12 (well and reservoir data and result of analysis are listed
in table 3). On figure 12, the thin derivative curve represents the
Well testing provides more than a gross description of the
corresponding double porosity curve (I\': = 1) for comparison with
the actual solution for I\': = 0.975. producing system but on the other hand, the interpretation cannot
This illustrates the matching procedure with the new model: in be concluded in many cases by a diagnosis performed only on
general the diagnosis of double permeability behavior is not straight pressure data. Other sources of inform&tion, notably geological
forward and often it can be recognised only after other solutions features, are required to reduce the number of alternative solutions.
have been unsuccesfully attempted. In the example, the match has
been refined by introducing the parameter K, on a first tentative
match with the double porosity pseudo steady state solution.

* *

408
SPE 13628 D. BOURDET 5
NOMENCLATURE:
5-Bourdet,D., and Gringarten,A.C.: "Determination of Fissure
a =Unit conversion factor (table 1) Volume and Block Size in Fractured Reservoirs by Type-
B =Formation volume factor Curve Analysis", paper SPE Q2Q3 presented at the 55th
Annual Fall Technical Conference and Exhibition of SPE of
Ct = Total reservoir compressibility
AlME, Dallas, (Sep. 21-24, 1980).
C = Wellbore storage constant 6-Bourdet, D., Ayoub, J.A., Whittle, T.M., Pirard, Y.M. and
h =ThickneBB 0/ the zone Kniazeff, V.: "Interpreting Well Tests in Fractured Reser-
I =Modified Bessel function voirs", World Oil (October lQ83) 77-87.
7-de Swaan, A.O.:"Analytic Solutions for Determining Natu-
k =Permeabilit1l rally Fractured Reservoir Properties by Well Testing", Soc.
K =M odified Bessel function Pet. Eng. J. (June, 1Q76) 117.
p =Pressure 8-Bourdet,D., AIagoa,A., Ayoub,J.A. and Pirard,Y.M.: "New
q =Production rate Type-Curves fol' Tests of Fissured Formations", World Oil
(April lQ84).
r =Radial distance Q-Barenblatt, G.L, ZeItov, Ju.P. and Kocina, LN., "Basic Con-
s =Laplace variable cepts in the Theory of Seepage of Homogeneous Liquids in
S = Wellbore skin factor Fissured Rocks (Strata)" Soviet J. App. Math. and Mech.,
1Q60, XXIV, N05, 1286-1303.
t =Time 10-Jacquard, P., "Etude Mathematique du Drainage d'un Res-
a =Interporosit1l shape factor Eqn. (2) and (11) ervoir Heterogene", Revue de 1'Institut Francais du Petrole
fl' =Interporosit1l transient transotion group defined inEqn. (3) (IQ60) XV, No.10.
ll-Lefkovits, H.C., Hazebroek, P.,AIIen, E.E., and Matthews,
6' =Shape coefficient Eqn. (4) and (5)
C.S.: "A Study of the Behavior of Bounded Reservoirs Com-
K =Permeabilit1l ratio defined in Eqn. (10) posed of Stratified Layers", Soc. Pet. Eng. J. (March 1Q61)
>. = Intreporosit1l flow parameter defined Eqn. (2) and(l1) 43-58; Trans., AlME, 222.
It = V iscosit1l 12-Duvaut, G., "Drainage des Systemes Heterogenes", Revue
de l'Institut Francais du Petrole (Oct. 1961) XVI No.10, 1164.
<p =Porosit1l 13-Katz, M.L. and Tek, M.R.," A Theoretical Study of Pressure
w =Storativity ratio defined Eqn. (1) and (12) Distribution and Fluid Flux in Bounded Stratified Porous
Systems with Crossllow" , Soc. Pet. Eng. J. (March lQ62) 68-
Subscripts: 82; Trans., AlME, 225.
D =dimensionlesB 14-Russell, D.G. and Prats, M., "The Practical Aspects of
/ =/iSS?J.Tf! medium Interlayer Crossllow", J. Pet. Tech. (June 1Q62) 589-5Q4.
i =initial 15-Russell, D.G. and Prats, M., "Performance of Layered Re-
m =matriz medium servoirs with Crossllow -Single-Compressible-Fluid Case",
Soc. Pet. Eng. J. (March lQ62) 53-67; Trans., AlME, 225.
M =ma!ch 16-Kutljarov, V.S., "Sur la Determination des Parametres des
w =well Couches Fissurees et Poreuses d'apres les Donnees d'Arrivee
de Liquide vers des Puits", lQ67, (Traduction LF.P., March
* lQ70, N0705).
ACKNOWLEDGEMENTS: 17-Tariq, S.M., and Ramey, H.J., "Drawdown Behavior of a
The author is grateful to the management of FLOPETROL Well with Storage and Skin Effect Communicating with Lay-
JOHNSTON SCHLUMBERGER for permission to publish this pa- ers of Different Radii and Other Characteristics" paper SPE
per. 7453 presented at the 53rd Annual Fall Technical Conference
The study presented here has evolved from many discussion and Exhibition of SPE of AlME, Houston, Oct. 1-3, 1978.
with co-workers at FLOPETROL JOHNSTON SCHLUMBERGER: 18-Gao Cheng-Tai, "Single-Phase Fluid Flow in a Stratified
Porous Medium with Crossflow", J. Pet. Tech. (Fev. 1984)
B. Deruyck, A. Douglas, A. Gringarten·, V. Kniazeff, P-A. Landel, 97-106.
S. Tariq·· and T. Whittle. Participation of A. Bocock in the analysis 11l-Wij esinghe, A.M. Culham, W.E., "Single-Well Pressure Tes-
of the field example is also acknowledged. ting solutions for Naturally fractured reservoirs with Ar-
(o!< now with S.S.-I.) bitrary fracture Connectivity", paper SPE 13055 presented
(U now with S.W.S.) at the 59th Annual Fall Technical Conference and Exhibition
* of SPE of AlME, Houston, Sept. 16-19, 1984.
20-Bourdet,D. and AIagoa,A.: "New Method Enhances Well
REFERENCES: Test Interpretation", World Oil (Sept 1984).
1-Bourdet,D., Whittle,T.M., Douglas,A.A. and Pirard Y.M.: 21-Clark,G. and Van Golft-Racht, T.D.:"Pressure Derivative
"A New Set of Type-Curves Simplifies Well Test Analysis", Approach to Transient Test Analysis: a High Permeability
World Oil (May 1983) 95-106. North Sea Reservoir Example" paper SPE 12Q59 presented
2-Bourdet,D., Ayoub,J.A., and Pirard,Y.M.: "Use of Pressure at the Ul84 European Petroleum Conference of SPE, London,
Derivative in Well Test Interpretation", paper SPE 12777 Oct. 25-28, lQ84.
presented at the 1984 California Regional Meeting, Long 22-Gringarten, A.C.:"lnterpretation of Tests in Fissured and
Beach, CA, (April 11-13, lQ84). Multilayered Reservoirs with Double-Porosity Behavior: The-
3-Warren, J.E. and Root, P.J.:"Behavior of Naturally FractUl'ed ory and Practice", J. Pet. Tech. (April 1984), 549-564.
Resrvoirs", Soc. Pet. Eng. J. (Sept. 1963) 245 23-EhIig-Economides, C.A., Ayoub, J.A.:"Vertical Interference
4-Mavor, M.J. and Cinco, H.:"Transient Pressure Behavior of Testing Across a Low-Permeability Zone", paper SPE 13251
Naturally Fractured Reservoirs", paper SPE 7Q77 presented presented at the 59th Annual Fall Technical Conference and
at the 1979 California Regional Meeting, Ventura, CA, (April • Exhibition of SPE of AlME, Houston, Sept. 16-19, lQ84.
18-20, 1979).
409
6 PRESSURE BEHAVIOR OF LAYERED RESERVOIRS WITH CROSSFLOW SPE 13628

24-Stehfest, H.:" Algorithm 368, Numerical Inversion of Laplace


Transforms", Communication of the ACM (Jan. 1070) 13, With
No.1, 47-40.

* A= (
(1 - w)s +
1 - II':
" _ ws +
II':
,,)2 + 4,,2
K,(I - K,)
(A,12)
APPENDIX A
Substituting 0-1,2 2 into (A.S), it comes
Solution lor a linite radius well :
The following conditions are used to solve the system of equations PlD = a1B1Ko(0-1rD) + a2B2Ko(U2 TD)
{ (A.13)
6: Pw = B 1K o(0-1 TD) + B 2 K o(U2 TD)
Initially, the pressure is constant everywhere in the reservoir, i.e.
at initial time: with
a1 = 1 + (1/")[(1 - w)s - (1 - K,)0-1']
pdr) = PI(r) = Pi (A. 1) { a2 (A. 14)
= 1 + (1/")[(1 - w)s - (1 - 11':)0-,']
The reservoir is infinite acting:
From condition (A.3),
lim PI
"-+00
= r-+oo
lim P2 = Pi (A.2)
(1 - a2)Ko(uITD) B,
(A.1S)
(I- al)Ko(u1 rD)
At sand face, both layers have same pressure:
and with the constant flow rate condition (A.4),
(A. 3)

The downhole flowrate is expressed as:

(AA)
(A. IS)

After changing the system (Eq.6) in dimensionless terms (Eq.I4), With


it is transformed into the Laplace space.
b = s(1 - a1)(lI':a2 + 1 -1I':)U2KO(U1)K1(U2)
(A.17)
- ,(1 - a2)(K,a1 + 1 - K,)u1Ko(0-2)Kdul)
The pressure, for a finite radius well can be expressed as:
The solutions are the modified Bessel functions [0 (o-rD) and KO(o-rD)'
For an infinite system (condition A.2), only the Ko function applies. _
PD = 1
/ ( D(s)
K~(o-l) + K~(0-2)
E(S») (A.IS)
With S
Pw = AKO(o-rD)
{ Pw = BKo(o-rD) (A.6)
. where
A.S changes into: KO(o-) = Ko(u) (A.10)
1 UK1(0-)
K,0-2 Alio(o-rD) = wsAKo(o-rD) =

l
D(s) (a2 - 1)(K, a1 + 1 - K,) (A.20)
- >"[BKo(o-TD) - AKo(o-rD)] a2 - a1
(a1 - 1)(lI':a2 + 1 - 11':)
(1 - K,)0-2BKo(o-rD) = (1 - w)sBKo(o-rD) E ()
s = .:....:::-~--'---~ (A.21)
(A.7) al - a2
+ >"[BKo(o-TD) --AKO(o-rD)]
The Laplace space solutions are inverted by means of a numeri-
it writes cal inversion scheme presented by Stehfestll4 •
[/CO-2 - ws - >"]A+ >"B = 0
{
>..A + [(1 - K,)0-2 - (1 - w)s - >"]B =0 (A. 8) *
Solutions are possible when: APPENDIXB:

(A.O) solution lor a well with wellbore storage and two skins:

It gives: With skin effects, SI and S2, the condition (A.3) is changed into:

0-4_((1- w)s + >.. + ws + >..)~+ [ws + >"][(1 - w)s + "]- ,,2 = 0 (B.I)
1-1\': K, K,(I-K,)
(A.IO)
This polynomial has two u 2 solutions (both are positive numbers): (A.IS) becomes:

(A.ll) The wellbore storage effect is introduced in (A.4):

410
SPE 13628 D. BOURDET 7

qB = C dpw _ 211"rw(klhl {)Pl + kshs ()P2) (B.3) lim D(8) = 1-11: (C.5)
dt p {)r p ()r rw )._0

lim E(8)
).-0
= II: (C.6)
and (A.16) expresses now:
When W > 11:, (0'12 and D) and ((T2 2 and E) are exchanged.
(TIKI ((TI)B I = - (1/8 - 8CDPD)
[a2(K~(us) + 8d - (K~(U2) + 8,)]/c The limiting form of (C.l) is:
) (T,Kt!us)B, =(1/8 - 8CDPD)
(B.4)
[al(K~((TI) + 81) - (K~(uI) + 8,s)]1 c PD = 1/8(8CD + 1- + ___11:_ _ _ )
II: (C.7)
Where 8 + K~(J~=:8) 8 +K~(~)
c = (lI: a2 + 1 -1I:)[al(K~(ud + 8 1 ) - (K~(Ul) + 8 2 )] (B.5) Double poro8ity 8olution :
- (lI:al + 1 - lI:)[a2(K~(u2) + 8 1 ) - (K~(U2) + 82)]
The double porosity solution is the limit when II: -> 1;
The well pressure is:
· A = Ii m ((I-W)8+>' - w(l-w)s+>. 8+ (1
I1m""
It-I It-I 1 - II: (1 - W)8 + >.
-II:
.
)2)
•••

(c.8)
r 2 r ((I-W)S+>')
where It~Ul = It~ 1-11: (C.9)
= 00
lI:a2 + 1-11:
F( 8) = -a-2-'-'(K-~A.(:-U2-:-)-+-8="1-'-)-_-'(:-:CK"'~:-:-((T-2-:-)-+-8::-7 . 2 w(1 - W)8 + >.
2) (B.7)
I1m (T2 =
It-I (I-W)8+>'
S (C.I0)
lI:al + 1-11:
al(K~(Ul) + 81 ) - (K~(Ul) + 82) The term K~(ud cancels out and with

G( ) _ K~(U2) + 82 lim E(8)


It-I
= 1 (C.ll)
8 - a2(KH(Tt) + 8 1 ) - (K~(U2) + 8 2 )
(B.8)
KH(Td + 8 s the usual double porosity solution (pseudo steady state interporosity
flow) is restituted:

*
APPENDIX C: PD = 1/8(SCD + 1 ) (C.12)
8 + KO(
1
W(I-Wlo+). 8)
11-00+).
8olution lor a well with wellbore 8torage and skin:
Homogeneous 8olution :
When 8 1 = 8 2 , the general solution (B.6) simplifies:
When II: = w, (C.I) simplifies into the solution for a well with
_ / ( D(8) E(8)) wellbore storage and skin, in a homogeneous reservoir:
PD = 1 8 8CD + 8 + K~(Ul) + 8 + K~((T2) (C.l)

D(,) and E(,) are defined in (A.20) and (A.21). PD = 1/8(SCD + 1 ) (C.13)
8 +K~(v'B)
In the following, we will express various limiting forms of the general
solution (C. I). Double permeability model: early time behavior:
Two layers without cross/low: At early time when t -> 0, the Laplace variable 8 -+- 00. Then

The two layers without cross flow solution is restituted by taking li m ",,=
A I'1m (( - w) + - -
1 --W- - s >. - - +
>. ( -1)2 ...) (C.14)
the limit>. ..... 0 in (C. 1): •_00 '-00 1- II: II: 1- II: II: S
When II: > W,
The limits for 0'1 2 and 0'2 2 are given in (C.3,C.4), and for D(8)
li m ((I-W W)+ - -
>. - - +
>. "
,2 ... ) and E(s) in (C.5, C.6). As a result, at early time, a two layers with
Ii m",,=
A
----s (C.2)
).-0 ).-0 1- II: II: 1- II: II: crossflow system behaves like two layers without crossflow.

and Double permeability model: late time behavior:


I·1m (Tl = I'1m (1 -
),-+0
2
).-0
- -W8
·1 - II:
>. +"
+ -1--
- II:
, 2 ... )
At late time, when t -+- 00, 8 -+- O. Then,
(C.3)
l-w
=--,
1-11:

ImA= Iim (>.
- - + ->.+ (1--- - -w) r.. -+ ;k8+82)
W ...
·
I 1m
),-+0
0'2 2 = ),-+0I'1m (w- K,
8 + ->.
K,
+" ...
,,2 )
• -+0 .-0 1 - II: K, 1- K, K, ~
(C.15)
(C.4)
W I-w -
= -8 • 2 >.

lim 0'1 = lIm - - + - +
>. (I-It}'
---L-
W
It'
1.
2
8 + 8 ...
)
(c.16)
II: .-+0 .-+0 ( 1 - K, K, I-It + It

It results,

411
8 PRESSURE BEHAVIOR OF LAYERED RESERVOIRS WITH CROSSFLOW SPE 13628

.-0 c-w +
• 2 • I-I< - K 2
w S + S ... )
_-0
lim 002 = hm -L-
l-IC
1.
K
(C.17)
=s
and,

~}211:(1-II:)e2 + ...)
._0
lim D(s) = lim ((1- IN
.~o 1-11:
_
II: 4)..2 (C.18)
TABLE 2
Build - up data
=0
elap.t:d time prunre ehange t:lap.t:d time pre.rare change

lim E(s) = lim ( I«t~I<~.+ s ... ) (C.19)


6.1, hr fl.p, p.i 6t, hr /j.p, p.i
.~o .~o 1«1-10)
1 5.83333E-03 139.65 76 0.65917 2019.4
=1 2 7.08333E-03 164.59 77 0.68400 2029.3
3 8.75000E-03 197.17 78 0.71250 2040.2
At late time, the response follows the homogeneous behavior cor- 4 1. 16667E-02 260.24 79 0.75000 2053.9
5 1.50000E-02 321. 08 80 0.82500 2082.7
responding to the total system, a.s expressed in (C.13). 6 1. 87500E-02 400.99 81 0.90450 2109.0
7 2.04167E-02 414.86 82 1. 0035 2142.1
8 2. 25000E-02 469.51 83 1.1040 2173.1
* 10
9 2.50000E-02
2.75000E-02
502.51
547.03
84
85
1. 2045
1. 3035
2201. 5
2227.8
11 3.08333E-02 601.67 86 1. 4040 2252.7
APPENDlXD 12 3.41667E-02 653.30 87 1. 5045 2276.4
13 3.83333E-02 713.47 88 1. 6035 2299.6
Line SOllrr.e. sQllltion 14 4.25000E-02 770.59 89 1. 7040 2321.9
15 4.75000E-02 832.06 90 1. 8045 2343.4
For a line source well, equation A.15 simplfies: 16 5. 16667E-02 880.70 91 1. 9035 2364.4
17 5.58333E-02 927.15 92 2.0040 2384.8
18 6.00000E-02 978.60 93 2.1045 2403.6
1- a2 19 6.41667E-02 1012.1 94 2.2035 2422.5
B 1 =---B2 (D.l) 20 6.83333E-02 1050.5 95 2.3040 2441.1
1- a1 21 7.25000E-02 1089.4 96 2.4045 2459.1
22 7.66667E-02 1122.7 97 2.5035 2476.6
The pressures in layer 1 and in layer 2 are: 23 8.08333E-02 1154.0 98 2.6040 2493.6
24 8.50000E-02 1183.7 99 2.7045 2510.5
25 8.91667E-02 1213.6 100 2.8035 2526.5
{ fFW = (1/ e)(al ;:=. a#.Ko(ulrD) + a2 a1.~a\ Ko(U2rD»
tJ
(D.2)
26
27
9.33333E-02
9.75000E-02
1238.1
1265.3
101
102
2.9040
3.0045
2542.2
2556.9
a
P2D = (l/s)U::::. a'. Ko (001 rD) + a1.-:....~!! Ko (U2rD)) 28
29
0.10167
0.10583
1289.0
1311.4
103
104
3.0540
3.1035
2564.4
2572 .1
30 0.11000 1332.6 105 3.2295 2590.1
31 0.11417 1352.8 106 3.3795 2611. 4
32 0.11833 1372.2 107 3.5295 2632.1
33 0.12250 1392.1 108 3.6795 2652.2
34 0.12667 1409.6 109 3.8295 2671.6
35 0.13417 1439.2 110 3.9795 2690.6
36 0.14042 1464.0 III 4.1295 2709.3
37 0.14458 1476.9 112 4.2795 2726.9
38 0.14875 1492.0 113 4.4295 2744.1
39 0.15292 1505.0 114 4.5795 2761. 2
40 0.16125 1529.6 115 4.7280 2777.8
41 0.16958 1553.1 116 4.8780 2793.9
TABLEt 42 0.17792 1574.1 117 5.0280 2809.6
43 0.18625 1594.6 118 5.1780 2824.7
!Lnite Qonv.ereions 44 0.19458 1612.9 119 5.3535 2842.0
45 0.20292 1631. 2 120 5.5290 2859.0
46 0.21125 1648.1 121 5.7030 2875.2
47 0.21958 1663.4 122 5.8785 2891. 3
48 0.22792 1678.6 123 6.1680 2916.8
SYMBOL OILFIELD METRIC DARCY 49 0.23625 1692.3 124 6.3990 2936.4
UNITS UNITS UNITS 50 0.24458 1705.4 125 6.4125 2937.4
51 0.25292 1718.3 126 7.6575 3032.2
52 0.26125 1730.0 127 7.8000 3041. 9
q STBID dm 3 /s em 3 /s 53 0.26958 1741.9 128 8.0535 3059.2
P. ep Pa-s ep 54 0.27792 1752.6 129 8.1780 3067.6
k rod p. m 2 darcy 55 0.29042 1768.6 130 8.4285 3083.8
56 0.30250 1782.9 131 8.6790 3099.5
r,h ft m em 57 0.31500 1795.7 132 8.9280 3114.7
P psi a kPa atm 58 0.32792 1808.4 133 9.3030 3136.9
59 0.34000 1820.4 134 9.5535 3151. 4
t hr hr S 60 0.35250 1831. 5 135 9.8040 3171.9
Ct psia- 1 kPa- 1 atm- 1 61 0.36500 1842.3 136 10.179 3185.2
62 0.37750 1852.9 137 10.554 3204.3
C bbl/psi dm3 /kPa. cm 3 /atm 63 0.39000 1862.7 138 10.803 3216.8
a" 141.2 1O- 8 /21r 1/21r 64 0.40250 1872 .3 139 11.054 3228.9
3.6.10- 8 65 0.41917 1884.4 140 11. 304 3240.7
al 0.000264 1 66 0.43583 1895.8 141 11. 553 3252.1
ae 0.8936 1.592.10- 4 0.1592 67 0.45250 1906.5 142 11. 804 3263.2
68 0.46917 1917.7 143 12.054 3274.1
69 0.48583 1930.2 144 12.303 3284.7
70 0.50250 1937.8 145 11. 054 3295.3
71 0.51917 1947.1 146 12.804 3305.7
72 0.55917 1969.5 147 13.053 3315.7
73 0.58417 1982.5 148 13.179 3320.6
74 0.60917 1995.1
75 0.63417 2007.8

412
102

•••• example A Ae-2S Co e 2S


.... example B
__-------------------------------------------1o25
10-20

10
10-1
3.10-3

-- .. ~!-- .........

·;/~O:5, AC O
.;/ " w(1·w)

10_1~~------~~--------~~~------------r-~-L------_+--------~-L~~-/.-3-.1-0--5------~-'-'\~--------~
TABLES 10-1 10 102 103 104 105 106
Well and rel1ervoir parameter,
to/CD
q=3000BBLID
Fig. 1-Wellbore storage and skin type curves for a double· porosity reservoir (pseudosteady*state blocks to fissure flow)
B=!.2 (Ref. 6).
5 1
Ct = 6· 10- P8I-

h = 60 II
4>=0.1 10 2
p= !.6cp
rw = 0.29/1
Res.lt. o~alr.i.

( PD)
fl.p M
= 9.16.10-' p.i- 1

(
IDICE..) = 26 hr- 1
fl.1 M Coe2S,1010
(CD.'S)1+' = 0.58
W= 10- 3 10
".-'5 = 0.05 ~C
106
---------------------
~ = 0.976 Q.
kh = 698 mi./1
......
C
C = 6.3· 10-3 Bbllp.i 0......
CD = 1877 C
S= -4 ~

C
Q.

10-1
10-1 1 10 10 2 10 3 104

Fig. 2-Wellbore storage and skin type curves for a double-porosity reservoir (transient block to fissure flow) (Ref. 8).
6

2 layers
10 2 r---------+---------+---------+----------------------------; no crossflow

o
a.

o
2
a. 1

W,10-3

10-1~--------__--------__--------~--------~~--------+_--------4 o
10-1 10 103 104 10 5 10-1 10 10 2 103 104

Fig. 3-Three dIfferent Interpretation models produce very similar responses Fig. 4-Semilog plot of double-permeability curves

6 10

2 layers
no crossflow

4
.0
a.
0
U
0 .....
a.
~
0
a.
2 10-1

.(I)
W,10-3
-C
I'T1
0 10-2~-----------------------r----------__------------+_----------~
10-1 1 10 102 10 3 10 4 10-1 10 10 2 103 104
f- ....

\.>-'
Fig. 5-Semllog plot of double-permeability curves Fig. 6-Log-log plot of the double-permeability curves of Fig 4
c~

rv
oa
10 6

4
.0
0.
0
U
--. 0
0.
~
0
0.
10-1 2

W,10-1

10-2~--------------------------------~----------~-----------4 o
10-1 1 10 102 103 10 4 10-1 10 10 2 103 10 4

Fig. 7-Log·log plot of the double-permeability curves of Fig. 5 Fig. 8-Semilog plot of double-permeability curves

10

1
·0
0.
0
U
--.
~
0
0.

10-1

Ii, 0.99

10~~----------~--------------------~----------~----------~
10-1 1 10 10 2 103 10 4

Fig. 9-Log-log plot 01 double-permeability curves of Fig. 8.


8 102~----------~----------~----------~----------~

- - double permeability
••• double porosity - - double permeability
- double porosity

10
.0
homogeneous Q.
total system o
Coe2S, 102 ()
....
~
o
Q.

o 10-1~~--------~-----------+----------~r---------~
10-1 10 10 3 10 4 105 10-1 10 2 10 3

Fig. 10-Semilog plot of double-porosity and double-permeability curves. Fig. l1-Log-log plot of the double-porosity and double-permeability curves of Fig 10.

10
double permeability
double porosity

.0

-....
Q.

()
o

~
o
Q.
10-1

10-2~--------+---------~---------+--------~r---------T---------4
10-1 10 10 3 10 4 10 5
N
OJ Fig. 12-Log-log match of bUildup data

You might also like