A Simplified Approach to Water Influx

Calculations-Finite Aquifer Systems

M. J. Fetkovich, SPE·AIME, Phillips Petroleum Co.

Introduction
All gas and oil reservoirs are associated to varying
extents with formation waters. The inclusion of the
effects of expansion or invasion of this water into oil
and gas reservoirs has ta.ken many forms, from recog-
nizing the effects of the expansion of the connate
water1 within the gas or oil reservoir itself, to calcu-
lating water influx or effiux across a boundary (with
the boundary usually being that of an oil or gas
reservoir).
There are four currently popular methods used for
calculating water influx into reservoirs. They are:
1. Schilthuis, steady state 1 - 3
2. Hurst Simplified, unsteady state1 • 2
3. Resistance or Influence Function, unsteady
state•-5
4. van Everdingen-Hurst Radial, unsteady state7
The first three methods have proved useful for pre-
dicting water drive performance after sufficient his-
torical data have been obtained to fix the necessary
influx constants. With what some consider to be dis-
appointing results,1· 8 the van Everdingen-Hurst Ra-
dial method is often used with geological and core
data when little or no performance history is available.
It has also been used to predict reservoir perfonnance
after enough historical data have been accumulated
to develop values of the influx constants, to and C.
In an attempt to include geometries other than
radial, derivations for both limited and infinite sys-
tems have been made to cover linear, 7 • 9 • 10 spherical, 11
elliptical, 12 thick-sand, 13 and wedge-shaped 18 reser-
voir-aquifer models.
The many rigorous geometrical representations that
have been developed cannot readily handle the effect
of interference between reservoirs. Electric analyzer
studies of the Smackover Limestone aquifer in Ar-
kansas by Bruce,14 of the Woodbine aquifer in East
Texas by Rumble et al., 15 and of the Ellenberger in
West Texas by Moore and Truby16 have shown that
reservoirs sharing a common aquifer can severely
interfere with each other, and that, for individual
reservoirs in a common aquifer, water drive perfonn-
ance calculations that do not consider interference
can be greatly in error.
Mortada11 developed a mathematical method with
which to handle interference in a basically infinite
radial aquifer system. The method has been applied
to field cases. 18 • 19 Coats concluded from his own study
that, "In predicting the pressure-volume behavior of
gas reservoirs situated on the common aquifer the
effect of interference from other reservoirs on the
common aquifer must be accounted for."
Another aquifer problem more recently presented
in the literature 20 is that of flank water injection for
pressure maintenance, either to initiate or to supple-
ment edge-water influx. A case history 21 shows that
we need to be able to study the effects of injecting
water into the aquifer instead of merely including
it in the hydrocarbon material balance equation.

This approach to water influx calculations offers a useful and flexible method of
forecasting and analyzing the performance of water drive reservoirs. The separation of
the water influx problem into a rate equatbn and a material balance equation, not
requiring superposition, makes the concepts and calculations quite simple and easy
to apply.

814 JOURNAL OF PETROLEUM TECHNOLOGY

 Little wonder that efforts11 • 22• 28 have been made to
simplify the water drive performance prediction meth-
ods, even to the point frequently of using the infinite
solution without trying to define fairly clearly the
limits and characteristics of the aquifer.
If we are to predict realistically the performance
of water drive reservoirs, then, a simple method must
be developed that can readily handle all the basic
geometries, interference from other reservoirs, and
water injection and production from the aquifer; the
method should also be flexible enough that it can be
further improved or added to as a problem requires.
We shall present here an approach that utilizes the
"stabilized", or pseudosteady-state aquifer produc-
tivity index and an aquifer material balance to repre-
sent the finite compressible system. Much of this has
been treated in the literature in the form of solutions
to individual well problems and reservoir material
balance derivations. For some reason - possibly a
concern for the early transient effects - any earlier
efforts to extend this avaiiable technolog-f to aquifer
or water drive problems have not been reported.
We hope to develop the idea that this simplified
approach is accurate enough for engineering purposes,
especially for field production forecasting of times
involving some 10 to 20 years, by comparing the PI-
Aquifer Material Balance solution with the van Ever-
dingen-Hurst solution through the use of example
problems. Solutions mainly involve finding a reason-
able rate equation for the problem, and considering
the aquifer encroachable water volume represented
in the material balance equation as being independent
of geometry only to the extent that basic mensuration
equations can be applied.
Basic Equations
The generalized rate equation for an aquifer without
regard to geometry or defining a specific type of
flow is:
(1)
with n usually being represented as unity (1) when
the flow obeys Darcy's law and is at pseudosteady
state or steady state. l,c is defined as the productivity
index (PI) of the aquifer and is analogous to the PI
of an oil well or the gas well backpressure curve co-
efficient.
The aquifer material balance for a constant com-
pressibility can be written in its simplest form as
-P = - ( w:,
p· ) We + P• , (2)
where p is the average aquifer pressure (shut-in), W ••
is the initial encroachable water in place at initial
pressure p;, and w. is the cumulative water effiux
from the aquifer or influx into a reservoir.
By combining Eqs. 1 and 2 (see Appendices A and
B for complete derivation), we can obtain the equation
expressing the instantaneous rate of water influx as
a function of time, and the inner boundary pres-
sure Pt0t·
JULY, 1971
(3)
(qlD;)max is defined as the initial open-flow potential
of the aquifer, again analogous to the open-flow
potential of an oil well or of a gas well. Fig. 1 is a
graphical representation of the generalized rate equa-
tion expressed as Eq. 1 and the aquifer open-flow
potential described above. Note that if we let W,;
become large, Eq. 3 reduces to the Schilthuis steady-
state equation
e,o = /fl) (p, - Pt01) • (4)
The final form of the cumulative water influx equa-
tion (given also in Appendix B)
w. = p:·
W· (p; - Pt0J) {l - e-[(q.,),,.as/W.1]f}
(5)
is not useful by itself because it cannot handle a
changing inner boundary pressure Pw! while repre-
senting the aquifer pressure always at its initial value.
Hurst2 • and others have handled this problem by the
method of superposition.
We can rewrite the equation to represent the cumu-
lative water influx over an interval of time At, then
start the problem again after every time interval (as
can be done for any material balance problem). With
the aid of the aquifer material balance equation, we
can redetermine a new aquifer shut-in pressure p,.,
then solve over a new time interval At. This re-
evaluation of the aquifer shut-in pressure each time
eliminates the need for superposition.
A significant point here is that we need not always
go back to the initial pressure to start a water influx
calculation. We can conveniently start it at any time
provided we can obtain a value to represent the
aquifer shut-in pressure.
The interval equation is
AW,,. =
w•• [p<,.-ll
---p;- - -
- Ptl!tnl
• {1 - e-C<q.,) • .,./w.,]Afn) (6)
The ratios W,i/p, and (qw.>max/W,, can be further
simplified to eliminate p, from the expressions, which
then do not need to be initiated agai.ii to new aquifer
shut-in pressures. These forms are retained so as to
keep their physical meanings.
The time interval is determined by
(7)
and the average pressure
-p _ Ptl!f<n-ll + P"'f<,.1
ID/ft - 2 (8)
represents the constant pressure used at the reservoir-
aquifer boundary during the time interval Atn. Fig. 2
depicts this pressure-time relationship and the step
curve that attempts to approximate it. This method
of representing the average pressure, pf/)1,., is appli-
cable to both past and future performance predictions.
To start the calculation again for the aquifer shut-in
81S
pressure p, we will make use of the general aquifer
material balance equation derived in Appendix B.
_P -_- [w· + ~ w.; +W cw,,
.
- w.> B.,,]
Pi +Pi,
• e1
where w. = l:" We.. , the total cumulative infiux (to
1 I
time tn) into the reservoir of interest. The term!.
2
is the total cumulative influx into other reservoirs
within the common aquifer and is farJier discussed
under Aquifer Interference. All other terms have
the conventional definition or have previously been
defined.
The realistic water i...'lflux rate and cumulative water
infiux relationship during an interval of time At is
depicted in Fig. 3 along with that which results from
using a step-function constant pressure as an approxi-
mation in any water infiux instantaneous rate equation.
Step-Function Solutions
It now appears that the simplification of the water
infiux problem is still none too simple. In reality,
though, we have reduced the problem so that we
can recognize that a simple time-incremented step-
function solution using the rate equation q'° = ''°
cP - p.,,1) to establish a constant rate over a time
interval; and the aauifer material balance equation
(9) P= - ( &:i ) W. : Pi to evaluate the shut- aquife~
in pressure after effiux from the aquifer, will give the
analytical solutions to the problem when At is allowed
W~; to become small. A At of a month in a normal reser-
voir problem does reproduce these analytical solu-
tions. (Co!l..stant rate steps over a At of 1 year for all
cases of r ..!r r ~ 5 reported in this study gave results
identical with those obtained using Eq. 6.) Fig. 4
illustrates this straightfoiward step-function approach.
For a time interval Atn, from t<11-1> tot.. , the work-
ing equation for the rate equation would be
(10)
The cumulative effiux during the time interval Atn
would be
AW,11 = Atn(q.,,), (11)

INITIAL AQUIFER PRESSURE, P;

• J.£H.u.1J
IOOO •(t) • [(4•~•11] I - REALISTIC RATE OF WATER INFWX
1 11 --- SlEP CONSTANT PRESSURE,Pwf

c
;;
.
IL
1w,0 • j '••wdl, CUMULATIVE WATER INFLUX
IOO
~

~ '<•·•>
...'
10

...
AQUIFER OPEN·FLOW POTENTIAL , .
~.;)

11 10 IOO IOOO
q.,- BBL/DAY
IOOOO 0 .,, '• .,.
TIME
'"
Fl1. 1-Aquifer "backpressure curve" with F11. 3--Calculated rate of water influx using a
open·flow potential, log-log plot. step-constant pressure at the aquifer
inner boundary compared with a
realistic representation.

TIME
Fl1. 2-Pressure-time relationship at aquifer inner F11. 4--A constant rate step-function approximation to
boundary as a step-function approximation. water influx over short time intervals.
816 JOURNAL OF PETROLEUM TECHNOLOGY
and the total cumulative efflux to time tn would be
Then to update the aquifer average pressure for
the next time interval,
-Pn = - ( w:i
p· ) W.,n
--
R ....t..
"' ....,.,. .......
J;' n•u•tinnc
,.... ..,..
In all derivation methods that attempt to predict water
influx and that assume a constant compressibility, it
is necessary to start with the same volume of initial
encroachable water in place for a given set of vari-
ables. Therefore, to predict water influx accurately
with the PI-Aquifer Material Balance approach, we
need only find a suitable rate equation.
Aquifer Productivity Index J.,
The aquifer productivity index J,., values used in this
study were calculated from a stabilized backpressure
equation for finite radial ftow conditions (8 360°).
The early transient period was neglected. For the
finite, slightly compressible, radial aquifers studied,
we used the "stabilized" pseudosteady-state rate
equation:
Table 4 summarizes the rate equations from which
PI can be calc-.ilated for finite radial and linear sys-
. (12) tems for pseudostady-state and steady-state condi-
tions. Also included in this table are the unsteady-
state equations for radial and linear transient ftow
that can be used for a system that does not reach
pseudosteady state or steady state during the period
+ Pi (13) of interest (see Fig. 20). Note that the infinite radial
ftow equation given in Table 4 is nothing more than
the Hurst Simplified water influx equation defined.
As in individual well problems, we could also intro-
duce the concept of skin into the equations to allow
theory to fit the observed data. Where changing the
internal aquifer radius, r r, would also cause a change
in the aquifer volume, Wei, the concept of skin would
allow us to vary PI without changing r r. This would
take on special significance if we attempt to match
historical inftux data from a best combination of
J,.,-W.,, while trying to conform with the existing
geometry of the system.
As a guide to the ti.mes at which pseudosteady state
and steady state are reached in a radial system, we
= can use this equation1 for pseudosteady state:
,...., 0.02 µCt tf:> T4 2 •
tP l = k , . (18)
and this equation1 for steady state:
(14)
,...., 0.04 µCt tf:> Tu'
t• = k (19)

We have then a productivity index for radial "sta-

The equations for a linear system could be derived
bilized" flow
like those for the radial system. All units in the above
J = 7.08 kh equations are in terms of days, centipoises, psi-i,
(15)
"' µ [In ( ;: )- !] . feet, and darcies.
In estimating times, we must remember to con-
sider the drainage boundaries that are established
The initial aquifer potential, (!/wi)mu; then is when there is interference from other reservoirs in
. (16) the same aquifer.

The initial encroachable water in place, W .,; , for Selection of Rate Equations
radial geometry (8 = 360°) is determined by Fig. 21 lists some possible types of aquifer ftow sys-
tems that could be used as a guide in selecting ap-
W.,i = S.~l (r..2 - r, 2) 8h Ct Pi . (17) propriate rate equations. Many problems can be
expressed in terms of essentially linear or.radial ftow.
Fig. 21 a describes a ftow system that is obviously
linear but whose distances between sealing faults de-
I scribe the cross-sectional area to be used with the
...... I aquifer rate equation. In water influx calculations we
~ I
~ k:- P;, INITIAL AQUIFER PRESSURE . are trying to describe the fiow in the aquifer itself. The
cross-sectional area at the aquifer-reservoir boundary
-CL , ,
!
-p . p, .
•-(~)We+
Wei
-
P·
I
is not neccssa...'ily applicable, especially after pseudo-
steady state or steady state has been established.
~ SLOPE 2_ • _.,!i. •-.,.,,..__.!..,,,,.._
i:i . 'dW1 W1 ; (Cw~Cf)WBw 1 Fig. 21 b describes ftow in a long, narrow reservoir.
That this type of fiow could be classed as linear has
~ FOR RADIAL GEOMETRY ·-,ew·coicrl-•fl•h been demonstrated by Havlena and Odeh11 from an
..
5
0 analysis of a gas reservoir 11 miles long and 1.5 miles
wide. Their analysis, using the material balance as an
INITIAL ENCROACHABLE
WATER-IN-PLACE equation of a straight line, indicated that the influx
w,i rate was proportional to the square root of time.
W,,CUMULATIVE WATER INFLUX Fig. 2 lc is an extension of the concept developed
Fig. 5--Graphical representation of the aquifer
by Fig. 21 b but in an additional dimension. Bottom-
material balance equation. water drive in a long, narrow field could be better
JULY, 1971 817
appro1-in1ated by radial fiow in the vertical direction, . (.t.) = !"'" cP - PwfA)
with h being the length of the reservoir. W etA - -

Fig. 21d would be represented by most engineers

]UJA (p - PUJtA) + JUJB (p - PUJtB>
as a radial flow system of 180°, using a radius to flow •[Wei - WeA(t)] , (20)
equivalent to rr. However, by redefining the system where
to consider the dashed lines to be the new boundaries
and by treating the volume of water between the fault WeiA (t) = encroachable water in place available to
and the actual reservoir boundary as a part of the Field A at time t.
reservoir (so that the expansion of this portion of the If for simplicity we are assuming equal inner bound-
aquifer would take place with no resistance to flow) ary pressures and equal PI values for Fields A and B
we can readily see that it is, for practical purposes, a (equal influx rates),
linear flow situation. This approach should give an
optimistic answer, but not so optimistic as it would be . (t) =
W e•A Wei (t) (21)
2 ,
ii the probiem were treated as a radial flow system.
Fig. 21 e illustrates a reservoir located between two after Field B starts production and reaches pseudo-
parallel sealing faults that terminate in a large aquifer. steady state. The transient time will now be shorter
Flow into the reservoir would be linear, and there than the transient period of Field A producing alone.
would be an essentially constant pressure at the outer In our aquifer material balance equation, the inter-
boundary. This would require a steady-state approxi- ference term for other reservoirs with respect to a
mation With the productivity index, J"', being a func- t
tion of the length of the sealing faults and the distance given resevoir is given by the summation term l: We;,
2
between. which represents the sum of the cumulative influx into
Fig. 21f depicts a wedge sand. Solutions to this all other reservoirs in the common aquifer. This results
problem have been reported in the literature'· 18 in in additional depletion, or decline in the average pres-
terms of an extension of linear flow. Turned on end, sure of the common aquifer as a result of these fields'
it can also be treated as radial flow, with an angle 8 also having water influx.
and the width represented by the distance h. The expanded expression is more easily visualized
These illustrations are given only as a guide to show in the time-incremented step-function approach for a
that many reservoir-aquifer systems can be defined in time interval ll.t. The cumulative influx into all reser-
terms of radial or linear flow. Both the simplified voirs from Fields 2 to j (the field of interest is Field
method and the van Everdingen-Hurst solutions are 1) is
applicable if we view the problems in terms of finding
the proper representation of a rate equation. However, ll.We (ll.t) = JUJ<2> cP-PUJt<2>] ll.t+JtD<a> [p-pUJt<si] ll.t
the simplified method allows us to use different dimen-
sions or geometries when defining the aquifer produc- + ... + JUJ<;> [p - PUJ1<;i] t:.t. . (22)
tivity index and the aquifer volume for a given Also, when handling the problem from a time-
problem. incremented standpoint, we could even, for complete-
ness, include to some extent the change in compres-
Aquifer Interference sibility of the total system by allowing each field,
By separating the water influx problem as we have including all reservoirs within the common aquifer,
into a rate equation and a material balance equation, to contribute to the total compressibility:
we can examine each individually as to its effect on
interference. Consider an aquifer of radius r contain- Ct = S 0 Co + 511 C11 + SUJ CUJ + Ct. (23)
ing two similar fields, A and B (they need nofbe H we include all except the reservoir of interest (Reser-
similar when applying the simplified method). We voir 1) this becomes
assume Field A has been producing long enough to

= !~ /\ NB,,c +v,, GB11 C11 I\ J .

reach steady state. Let the productivity index of Field 0 ..L (' ,. ..L •
A be J"' = f(r). When Field B begins producing, the Ct I .... UJ ... UJ I C1 '
productivity index of Field A will increase, becoming
(24)
Ju,= f(r/2). From the sta11dpoint of the rate equation,
the deliverability of the aquifer for Field A will be where VP is the total pore volume of the aquifer and
increased after Field B begins producing. As pointed nonproducing fields. Muskat 27 points out that the
out by Bruce 14 in his study of the Smackover aquifer, indicated abnormally high compressibility, Ct = 36 X
the interference effect is totally one of "competition 10-e psi- 1 of the East Texas Woodbine aquifer could
among pools for the common water supply". be due to gas fields or gas caps of oil fields distributed
From the aquifer material balance standpoint, Field in the aquifer.
A would initially have an aquifer volume of Wei bbl If we do not wish to include the compressibility of
available to it for water influx. However, after Field the other reservoirs within the aquifer, Eq. 24 reduces
B begins producing, the aquifer drainage volume to the simple expression
available to Field A is reduced. It can be approxi. .
mated by the basic relationship given by Matthews Ct= CUJ +Ct. (25)
et a/. 26 - "at (pseudo) steady state the drainage vol-
umes in a bounded reservoir are proportional to the Water Injection into the AquUer
rates of withdrawal from each drainage volume." The usual method of treating water injection for study-

818 ;oURt'TAL OF PETROLEUltf TECHNOLOGY

ing pressure maintenance is to include a wa,ter injec- future performance predictions (reservoir pressures
tion term in the hydrocarbon material balance equa- and producing rates) the two treatments wiii accom-
tion. A form of the material balance equation for a plish the same thing.
gas reservoir is With the simplified procedure, where the problem
has been separated into its two basic components -
G 11B 9 = G(Bu - B 9 ;) + We + B,,, (W; W,,). productivity index and aquifer material balance -
(26) we can approach the problem the way we would
approach it to determine the resistance or influence
The basic assumption here with respect to water function. For a gas reservoir:
injection is that all water injected is instantly avail- 1. We can fix a best estimate of gas in place, G.
able to the reservoir, which would be realistic if the 2. Using the incremental form of the reservoir ma-
water was injected uniformly throughout the reser- terial balance equation for a time interval Atn and
voirs as in pattern waterflooding. However, when the two historical reservoir pressures, Pwtn and Pwr tn-1»
purpose is to maintain pressure, we generally use a we can solve for a water influx volume
flank water injection, with the injection wells located AWe,. = A(G,,B 0 ) - GAB 0 + A(W,,Bw). (27)
in the aquifer.
A more realistic approach is to include a water in- Then the average influx rate during the time interval is
jection term in the aquifer material balance equation
-e (Atn) = _
AW_en_
so as to incorporate the effects of the resistance to (28)
w Atn ' .
flow across the reservoir-aquifer boundary. For high-
permeability boundaries, the results would be essen- • h is
whtc · represented at time
. t,. + ln-1 (29)
tially the same. However, where the permeability at 2
the boundary is low, over a realistic time period little
or no water may enter the reservoir. The option should 3. We can plot the average influx rate ew
(Atn) as a
be available, at least, to study it both ways or in com- function of time.
bination. Eq. 9 includes water injection into the 4. We can calculate water influx rates as functions
aquifer in such a manner that the total water influx, of time, using various combined values of aquifer pro-
we is also a function of the water injected, w. = ductivity index and encroachable water in place. These
f(W;). rates of water influx are plotted with those calculated
In an interesting case history 21 of the Pegasus Ellen- using the material balance equation.
burger reservoir we are told of an attempt to maintain 5. We can select the best combination of J,p - W.;
pressure by using flank water injection to supplement to fit the problem. Although a statistical approach
edgewater influx. The peripheral project failed to
maintain pressure, resulting in very high pressures
around the injection wells. Injection into the central TABLE 1-HYPOTHETICAL GAS RESERVOIR AND
producing area was required to halt the pressure AQUIFER PROPERTIES.
decline. The water influx constants from the edge- Gas Reservoir Properties
water drive were established before water injection Initial gas reservoir pressure, psia 2,000
was begun. The Pl-Aquifer Material Balance ap- Porosity, fraction 0.20
Pay thickness, ft 100
proach would have been more successful in predicting
Water saturation, fraction PY 0.20
the final outcome. Initial formation volume factor for gas,
scf /reservoir cf 154.26
Historical Data Reservoir radius, ft 10,000
There are two differing treatments of historical data Gas gravity (to air) 0.700
from reservoirs subject to water drive. They are Pseudo criticai temperature for gas, 0 R 392
usuaiiy referred to as Pseudo critical pressure for gas, psia 668
1. The Material Balance as an Equation of a Reservoir depth, ft 7,000
Reservoir temperature, °F 130
Straight Line, 25 and
Initial gas-law deviation factor 0.780
2. The Resistance or Influence Function.'· 6 Initial gas in place, Bcf 776
• They differ mainly in their primary objectives. The Rate of take (1 MMcf/D to 8.59
straight-line approach attempts to determine the Bcf gas in place), Mscf/D 90,338
original gas or oil in place using the historical data, Total field wellhead potential, Mscf/D 250,000
whereas the resistance or influence-function approach Initial wellhead shut-in pressure, psia 1,600
fixes a best estimate of gas or oil in place and then Slope of wellhead backpressure curve 0.700
attempts to determine a best fit of the data to arrive Line pressure, psia 200
at a resistance or influence function F(t) with which Aquifer Properties
to predict future performance. Initial pressure in aquifer, psia 2,000
When the objective is to determine recoverable Permeability, md 10, 50, 100, 1,000
reserves, a precise value for oil or gas originally in r./r, 3,5,7, 10
place may not be justified because of the inaccuracies r. (using r, = 10,000 ft), thousands of ft 30, 50, 70, 100
involved in arriving at reliable values for residual gas Porosity, fraction 0.20
Aquifer thickness, ft 100
or oil saturation and sweep efficiency. If, however, in 6 x 10·6
Total compressibility for aquifer, 1/psi
determining original in-place values the resulting Viscosity of water, cp 0.50
influx coefficients, C and tn, are to be used to make
JULY, 1971 819
 -·
WATU-llllM 1'£11f~

,.-'1/'ru•.1E¥£1111111GEN-MURST
OPl - - U llATL. IAL.
-
2100
noo;::-
could be used to make a selection, an engineer's
analysis based on intimate knowledge of each data
point and field history would be preferable.
A good starting point for I., and W Iii should be
... 1700 2400;
·~tlOO 22001
based on the basic geometry of the reservoir-aquifer
I::
-·
IOO IOOO system being studied. For a strictly radial geometry

.1:-
e•
,,_
r,• -~ the productivity index of the aquifer, J.,, the water in
•• .MO place, W ••, and the original gas in place, G, are all
i i•t00 functions of common variables in that I., = f (ln Tr,
.....
too
Ill too
ln T11); W.,, = f (Tl, T112); and G = f (Tr2). If aquifer
interference occurs at a later time, I., will change as
Ill 700 a result of a change in drainage radius Ta, but only as
i IO i MO the ln Ta(t), whereas the water in place W.,,(t) will
0 ... c : l 4 s : 1 c : = •1 u tl t4
TlllE-T£All$
ts 1; n ;; w20° cha.1ge as the square of r11(t), with ti'ie gas in place
remaining the same. Therefore it is possible to have
Fl1. 6
more than one value of water in place as a solution
during the producing life of a field.
Dt.!..n.ng t.'1e early times, before pseudoste.11tiy state
is established, the aquifer productivity index, J.,(t)
plotted vs the ln t and \ft for radial and linear flow,
respectiveiy, shouid be straight iines.'· 6 No fixed vaiue
of lw and w.,(t) exists during the early transient
period.
Discussion of Results
The method chosen with which to illustrate a com-
parison between the PI-Aquifer Material Balance
approach and the more rigorous solutions of van
Everdirulen-Hurst is a hvoothetical e:as reservoir sur-
rounded'"'by a futlte -radi8i aquifer. Using a gas reser-
voir does not require the introduction of variables
such as k,/k 0 relationships that may later be suspected

c tlOO
f
1"00
tlOO

'-: tlOO
llATEll-llllM
, . _ llO
'•/'r •7
1'£11f~

- u E¥£11DlllGEN-MURST
o Pl - AllUtFEll llATL. IAL.
-
noo
of contributing to some of the basic responses shown
by the water drive performance.
The properties used for the gas reservoir and
aquifer are listed in Table 1. So that the effect of the
early transient period could be investigated, we chose
a range of permeabilities and external radii of the
AtlO
IOO
I-
r,. MOO aquifer. In each case, the aquifer inner boundary
1.140 pressure was represented by the average pressure de-
...!« if IZllO termined from the solution of the gas reservoir ma-
• IOO
I to .....
1100

too
GAS RATE
terial balance. Values used for water viscosity and the
total compressibility are typical of those often used
Ii '° i
40
IO
Ill 100
Ill 700
100 zoo
in the literature for water inftux calculations.
A typical gas withdrawal rate of take of 1 MMcf/D
7 t t IO 11 tZ U 14 IS tt 17 • tt ZOO to 8.59 Bcf in place (1 MMcf/D to 7 .3 Bcf recover-
TllE-T£AllS
able with an 85-percent recovery factor) was used so
F11. 8 as to obtain realistic water influx values. A more rapid
gas wi.thdra.wai rate would result in less water influx
for the same reservoir and aquifer properties used in
this study. No attempt has been made to determine
recoverable reserves at abandonment based on re-
sidual gas and sweep efficiences. This could be han-
dled, however, by the methods suggested by Agarwal
et al. 28 All forecasts are carried out for a full period
of 20 years, that time defined by a 1-to-7.3 rate of
take. A constant field wellhead potential for the gas
reservoir was used for all cases.
Figs. 6 through 10 illustrate the water drive per-
formance for an aquifer with a permeability of 1,000
md at four different external aquifer radii - 30,000,
50,000, 70,000 and 100,000 ft (19 miles). In all cases,
the PI-Aquifer Material Balance solutions match
identically the gas producing rates, reservoir pressure,

820 JOURNAL OF PETROLEUM TECHNOLOGY

and cumulative water influx determined using the van
Everdingen-Hurst solutions.
The simplified approach does not utilize superpo-
sition, whereas the van Everdingen-Hurst solution
does. To investigate the effects of superposition when
producing rates are varied severely, a variable pro-
ducing rate situation was studied (Fig. 10). This was
done for t.'le largest aquifer radius. Excellent agree-
ment was obtained for this 1,000-md permeability
case.
Figs. 11 through 14 illustrate the water drive per-
formance when the aquifer permeability is changed
to 100 md. In these cases, the departure from the
van Everdingen-Hurst solutions is quite small with
respect to reservoir pressure and cumulative water
influx and is well within engineering accuracy. The
gas producing rates are identical.
Fig. 12 includes the additional points representing
results using the van Everdingen-Hurst radial infinite
solution. After early times, their solution departs from
the r4 /r, = 10 case about as much above the line as
llOO
f llOO
..<1100
MT£ll-lllllV£ l'£llfllllllAlC
~=

-
tOO llO
'1/'r =iO
£Y£11DlllGOMUIST
O 'l - MIWER llATL. llAI..
a Mii EVEll-N·lllllST,
--
mo~

t200 a
·~-
lll'lllTE
the simplified does below. What is interesting here is
that within the limits of field data, it would be difficult
to determine the actual extent of the aquifer. That is,
••
•::
e tao
200

...; t40
f tJoo
11"1200
tllOl
tooO
IOO
IOO
700 •
~
l
we could easily maintain that the performance data ~ t20 i ttoO GAS llATE i

-·
IOO
indicates an infinite radial aquifer. There would be
enough room to adjust the internal boundary pres-
.. tOO
I ao r: 500 ~
4GO !C

sures to force a fit to an infinite solution.

Fig. 14, showing the performance of the aquifer,
I:
. 3
:I 20
Ill-
IE 100
IOO
200 ....
tOO
illustrates why the cumulative water influx as calcu- 5000:-'i~.+-~~~7~t~t~W.,.._,.tl~Q~U-t~4-l~S~lt-t~7~M~"~2:
TllC-'IUllS
lated by the Pl-Aquifer Material Balance method
departs constantly from that calculated by the van Fig. 12
Everdingen-Hurst method. The departure results, not
unexpectedly, from a difference in influx rates during
the early transient period. After this period, the influx
rates agree quite well.
Figs. 15 and 16 illustrate the water drive perform-
ance for an aquifer of 50-md permeability. The de-
parture of the cumulative water influx from that
derived by the van Everdingen-Hurst solution is
most pronounced for the aquifer-to-reservoir ratio of
ra./r, = 10, an aquifer external radius of 100,000 ft.
The constant departure indicates that the difference
occurs as a result of the early transient period, as
shown in Fig. 14. Still, the reservoir pressure and gas
producing rate agree quite well.
Figs. 17 through 20 give the water drive perform-
Fie. 13

IOO •
500 0
0 I Z l 4 5 I 1 I t•tttztlt41Sttt7"1t20
TllE-'IUlll

Fig. 10 F11.14

JULY, 1971 821

ance using a 10-md aquifer permeability. In Fig. 17, index representing the fixed dimensions of 100,000
f 4 /fr = 3, the cumulative water inftux for the PI- ft for f,. of the aquifer, could not become established
Aquifer Material Balance solution is always greater, durin2 the 20-vear period of the forecast. Therefore,
indicating that the van Everdingen-Hurst solution was the Hmst-Siinplified-(Defined) equation given in Table
dominated by linear flow, resulting in a lower inftux 4 was used. The results obtained using this equation
rate than the radial flow detenrJnation. are quite good (sec Fig. 20). As in individual well
Fig. 19, ra.Irr = 7, shows a continuously increasing forecasts, the Hurst-Simplified (Defined) equation
departure of the cumulative water influx as a result could be used until pseudosteady state is established;
of transient flow etfects throughout. In these cases as then, after applying the material balance equation to
in all previous cases, the gas producing rates agree. determine the aquifer shut-in pressure, we could use
In all the 10-md aquifer permeability cases, the gas the pseudosteady-state rate equation for the rest of
reservoir is behaving essentially as a volumetric the forecast.
reservoir. Because the results presented in this study were
From a check of the time it takes to establish based on finite aquifer systems, it would be appropri-
pseudosteady state, it was found that the productivity ate to discuss briefly the terms "finite;; and "infinite"

ftMO
-
tlOO
llATCR-CllllYE 1'£11f'ORllNICC
••501111
'1/rr •3
- u lYUDlllGt:ll-IUIST
o Pl - AQUIFER llATL. IAL.
llATER-DlllYE fl[llf'GlllMllCE
a•to llO
'a/'r • ~
- u EVEllDlllG[ll•lllM$T
o '1 - AQUIFER llATL. IAL.

Fl1.15 fll. 18
llATER-DlllYE 1'£11f'a.MMICE

-
•• 50 llD
r,/rr •tO t5llO
c llOO -u EYERDlllGlll·lllM$T
f llOO O '1 • AQUIFER llATL. IAL. -;:t

200
---I:
-i 1700

:-i
IZOD·a
tlOO\

••MA
1;;; no =
GAS llATE
IOO i
...
1·::~-- -----0-----------0..::~ SOD ~
44111 !C

0
Ill HO
lit 700
i 100 .
:I
500 o:-Jtl"'!,Z2~l~4-5!;--'.!:l-=-7~l~l~IO~ll~Q~U~t~4~15L....1.ll-t~7~to-ft.&....JI:
TlllE·YEAllS

Fis. 11 F11. 19

-a -
llATU-DlllYE l'UFa.MMICE

-
a• IOllD
'e/'r • 3 t5llO t5llO
-u EVEU-11·111M$T tlOO
O 111 - - U llATL. IAL. -;:t ftlOO

1:·-
ln=
-
!!90
11001

~
i
IOO
IOO •
1!111to

i i 1100

r::
GAS llATE

I:
1. I
Ill HO
lit TOO
IOO
0 0

Fis. 17 fll. 20
822 JOURNAL OF PETROLEUM TECHNOLOGY
 TABLE 2-RADIAL FLOW WATER INFLUX VARIABLES USED FOR THE Pl·AQUIFER MATERIAL BALANCE
SOLUTIONS FOR A 20-YEAR FORECAST OF THE HYPOTHE"flCAL GAS RESERYIOR
k == 0.010 darcy k - 0.050 darcy
Initial (pseudosteady state) (pseudostaady state)
Encroach·
Radius able Stabili· stabiii·
of Water in Initial zation Initial zation
Aquifer Place J .. Potential Time J .. Potential Time
r./r, (ft) (108 bbl) (8/D/psi) (8/D) (years) (8/D/psi) (8/D) (years)
3 30,000 107.52 40.620 81,240 3.o 203.10 .. ,.,,."""
4UO,.l:UV 0.6
5 50,000 322.56 16.477 32,954 8.2 82.385 164,770 1.6
7 70,000 645.12 11.840 23,680 16.1 59.200 118,400 3.2
10 100,000 1,330.56 9.1202 18,240 32.9* 45.601 91,202 6.6

k = 0.100 darcy k - 1.000 darcy

(pseudostaady state) (pseudosteady state)
Stabili- Stabili·
Initial zation Initial zation
J .. Potential Time J., Potential Time
(8/D/psi) (8/D) (years) (8/D/psi) (8/D) (years)
3 30,000 107.52 406.20
---
812,400 0.3 4,062.0 8,124,000 0.03
5 50,000 322.56 164.77 329,540 0.8 1,647.7 3,295,400 0.08
7 70,000 645.12 118.40 236,800 1.6 1,184.0 2,368,000 0.16
10 100,000 1,330.56 91.202 182,400 3.3 912.02 1,824,000 0.33
•Stabilization time for this value of J • exceeds duration of forecast.

as applied to water influx problems. "Finite", as used 2. The deliverability or productivity index, J,,,, is
in this study, indicates only that finite dimensions very large (infinite). As a special case of water influx,
were used for defining the aquifer productivity index, an infinite productivity index is always assumed when
J,,,, and the aquifer volume, We;. "Infinite", when ap- the expansion of the water within the hydrocarbon
plied to an aquifer, can take on at least three different reservoir itself is included in the reservoir material
meanings. balance equation by the addition of a water compres-
1. The aquifer volume Wei is very large (infinite). sibility term.
This can result in a Schilthuis steady-state aquifer 3. Transient flow exists during the entire period of
behavior. interest, with the result that an infinite solution is
applicable.
For the studies involving the largest aquifer radius
used- 100,000 ft - the 10-md aquifer permeability
case was the only one that could be classed as in-
finite - and then only because the infinite solution
could be applied. Its volu.T.e of water influx was so
insignificant as to cause the gas reservoir to behave
like a volumetric reservoir. The 100-md case response
as a finite aquifer (even with no transients being con-
sidered for the simplified solution) was such that it
appeared to behave like an infinite aquifer solution
a-linear flow d-linear flow (see Fig. 12). The term "infinite" when applied to
water influx problems should always be qualified as
LARGE
'I ,
' CONSTANT PRESSURE
AQUIFER-Pe)/
__ , ~
~ ,.
to which of the above definitions is meant.
In review, the good results obtained with the PI-
Aquifer Material Balance approach are suprising
~
F.fvt.~ when we consider that the additional flow contribu-
tions from the early transient period have been omit-
ted and that there exists the condition rr < < r", im-
b-linear flow e-linear steady state posed in the derivation of the pseudosteady state
radial flow equation. Variations of the constant in
the term [In (r"/rr) - *]were studied by using -Y.2
and - 1, as well as some of the other suggested meth-
ods of expressing the inner boundary pressure, PtDt.
In all cases, the results obtained were significantly
poorer than those reported in this study.
Certainly in many cases the additional capabilities
f-radial flow, of the PI-Aquifer Material Balance, when properly
c-radial flow,
bottom water wedge sand included instead of omitted, can far outweigh any
early transient effects omitted. In many cases where
Fig. 21-Types of fiow systems for rate equation. the tr->..nsient is of long duration; as for the 10-md
JULY, 1971 823
TABLE 3-RADIAL FLOW WATER INFLUX VARIABLES B 0 = oil formation volume factor, reservoir
USED FOR THE VAN EVERDINGEN·HURST SOLUTIONS bbl/surface bbl
FOR A 20-YEAR FORECAST OF THE
HYPOTHETICAL GAS RESERVOIR B.,, = water formation volume factor, reservoir
Ratio of bbl/surface bbl
Dimensionless c1 = formation {rock) compressibility, psi- 1
Aquifer Time to Water Influx cu = gas compressibility, psi - 1
Permeability Real Time Constant
(darcies) (l/year) (cu ft/psi) Co = oil compressibility, psi- 1
ct = total or effective aquifer compressibility,
0.010 0.38483 75,396 psi-1
0.050 1.9242 75,396
0.100 3.8483 75,396 c.,, = water compressibility (includes the effect
1.000 38.483 75,396 of dissolved gas), psi- 1
Cu = gas well backpressure curve coefficient
(gas well productivity index)
cases given in this study, it makes little difference e = natural logarithm base 2. 71828
whether the transient effects are included or not. e.,, = water influx or effiux rate, reservoirbbl/D
Example Calculations
e,,, (~In) = average influx or effiux rate during time
interval (atn), reservoir bbl/D
An example of a water drive perfonnance prediction r:!.. -
v -
;";.,;~1 nae! 1n
J..l..UW.Uol. 6~ £u.
nlQir,:a. Rcrf
y .........."", - -...
for a gas reservoir using the PI-Aquifer Material Bal- G, =cumulative gas production, Bscf
ance approach is given in detail in Table 5. The cal- h = aquifer thickness, ft
culations were performed on a desk calculator using i = subscript to denote initial value or condi-
the simple trial-and-error procedure of iterative sub- tions (except for cumulative water in-
stitution. The iterative calculations are shown only jected, W,)
for Years 1 and 20. During the period of constant J.,, = aquifer productivity index, reservoir bbl/
producing rate, the second trial was always within 1 D/psi
psi of the final answer. When the producing rate was k = aquifer permeability, darcies
limited by the backpressure curve, an additional itera- L = length, ft
tion was required. n = exponent of backpressure curve, also
used as a subscript to denote the end
Conclusions of an interval
The PI-Aquifer Material Balance approach to water
influx calculations offers a very useful and flexible
p= average aquifer pressure (shut-in pres-
sure), psia
method for forecasting and analyzing the performance
Pcn-1) = average aquifer pressure (shut-in pres-
of water drive reserv0trs. The separation of the water
sure) at the beginning of an interval,
influx problem into a rate equation and a material
psia
balance equation, not requiring the use of superposi-
Pe = external boundary pressure, psia
tion, makes the concepts and calculations quite simple
p; = initial aquifer pressure, psia
and easy to apply.
Pii= gas reservoir average pressure (shut-in
Acknowledgments pressure), psia
I wish to thank Phillips Petroleum Co. for permission Ptt = wellhead tubing flowing pressure, psia
to publish this paper. The assistance of M. W. Canon, Pta = wellhead shut-in pressure, psia
who programmed the calculations and helped with all p.,,, = inner aquifer boundary pressure, psia
the computer runs, is gratefully acknowledged. Pwt = a constant inner boundary pressure for a
time interval (Atn) (see Eq. 8), psia
Nomenclature PI = productivity index, reservoir bbl/D/psi
b =width, ft qu = gas flow rate, Mscf/D
Bu =gas formation volume factor, reservoir Qu = average gas· flow rate during an interval,
bbl/scf Mscf/D

TABLE 4-RADIAL AND LINEAR AQUIFER RATE EQUATIONSZ

Type of Boundary Radial Flow Linear Flow

q., 7.08 khcP - p .. 1) 3(1.127) kbh (p - P•1)

Finite--closed (no flow) at outer boundary =
q. - µ. L
p. lr''fl\' r.
r. )
- 4J
3 1

Finite--constant pressure at outer boundary q. =

7.08 kh (p. - P•1)
q. =
1.127 kbh <e• - P•1>
p. [In C: )) µ. L

Infinite q. 7.08 kh (p, -Pw1) kbh (p, - p.1)

= q .. =
p. [ In " 14.23
t/>p. Ct fr"!] p.
"6.33 let
--;;c.
{Hurst Slmpllfled (Defined)]

824 JOURNAL OF PETROLEUM TECHNOLOGY

 TABLE !5-EXAMPLE OF A WATER DRIVE PERRllRMANCE PREllllCTllON FOR A GAS RESERVCllR USING THE Pl-AQUIFER ll•ATERIAL BALANCE APPROACH
RESERVOIR MATERIAL BALANCE CAl.CULATIONS Pl - AOUIFEA MATERl ..L BALANCE CALCULATIONS IAOUIFEA K • 111Dmd;'lllr, • 101
.....
•'D
·-I
. Tiii( •t,, •
-t,__,, ....... -(~j· ;-w.
G - Go lr.
.... .... ...
•• 1.N
...
c,1.,.1 - "'11• ..... I+ ....
.....
111rw1.1 + ,... ,., ;..I li..1 - .....,
-~w;:}• ... .. w, " - '. ~
'
tn •Go Go B - Go G -11; ;r z1,.1 ..
'" ... Pwt --1-- --1--
I- I Eq.I
iii (i;} 11,

121.
-(11111-I
I0.331
•1151 13hl4hlt' :&151 71UID-lll 111.llDll-1121 171 +Ill ZM91•lll USING 110 1111•1111 1121 +1I?
1IZU1.

("'~:r
(14),._1•C14)n
l
1111.-1 • 1r11:1.
- -I - - 1241 .... 1 1171-1111 ·-I
--·
1111•1111
:&!Ill !Ziff l.114h IOI 1.5131hlo-'
11211
-.0-IZll

_._,_ .... _!!__

.... ... . .... ....
- ..... ....
-
DAYS DA,YS MICM l'UIA PllA A PllA PllA I'll RES Ill RESllL OSI PllA
_ _ISi
__ __111_ _ _ _111_ _ 1111 1111 1111 1111 IZll 1111
!.!L _J!!__ ~L _J!_ _!!!__ __!!!!._ __!!!__ -- IMI flll __!.!!__ 1171 11•

I
• •
I
I
• I 111.11111 71UID 1.- 111a D.711
• >•,331 lllUI
• I I I

-··
...
11
•
t
311 115 I0,331 12.171 :12.17ll
n.m1
U.17lt
743.127
743.127
741.1127
711.llDll
114.•
114.N
·-··-
1.151'1
'""
II•
11411
D.71l
ua
ua
llll
1lll
1121
• >11,331
ii
B
-.Ii
llllJI
1111.li
lllU
-I
--1
17.I
11.1
l-1 Utz.Ml
I.Ill.. .
1,217,llD
l.21Z,MI
1,111...
1.111.llD
1.111
1.011
1.111 I.I 1•.z

··- - ....
=
l
...
1'I I0,131 n.m
...,.
IO.MI 71l. . 711.111 UI• •• l71D 1IOI i.n
1«1 I
>•,331 i
!
1•11
1m.11
-.1
111l.I
111.Z
Ill.I ........ ,......
U11.1ll l,7J1.lll
...
1• 7.1
1U
1111.1
11M.1

-··
I 11,331 n.m '77.•1 717... l71D 1111 >•.131
,.,
--- - ~ •
4 1• 11,331 n.m m•z MUii 111... 1.14174 lltn I.JM 1721 1111 >•.111 1111.11 121.i 1,311.311 11.111.111 11.114 zu 1111.1

i
I 1121 II.DI nm 1M.IOI 111.1• 1tU2' ll1111 llll 1111 1111 >•.131 111111 1m.• Ill.I 1,115,111 11,m.IM 11111 •1 1•.1
l:i
I
1
11• I0,331
I0,331
:12.171
:12.111
111.111
Zit.Ill
Ill.Ill
141.111
141...
711...
1.17111
1.74111
lllM ua 1111

""
12•
IllI
>•.131
>•.131
a

i=
IUl.11
1111.11
mu
1•.1
11111
Ill.I
111.1
i ll.llll.~ 11......
ll.11U14 11.111.m
II.QI
11.111
II.I
!I.I
1M3.1
__ z
1121.5

-...
.
I II I0,331 nm llUI~ llUll lll.llZ l.71tM lllZ1 U11 1411 11M >•.DI 1121.I llU
8 11.111.111 a,144.- 14.111 M.I

··-
z
,,
·-
.... 11,331 n.m 81.ltlJ 411.IQ Ill.Ill Ulm lllll I.Ill IUI 1111 >•,331 1151.'1 1•.1 llU c ll,lll.1$1 71.111.111 11.111 111.1 1•1'

....
I

- ·-
1111.t
"' I0.131 n.m 1111.Jlll 44U11 IM.111 llMf ll2' 1. . >11,331
i..
1•.1 Ill.I
~
11.111.111 a.111.111 '1.lll 141.1 1111.I
" ll
• 111U llOU Ill.I 11.-.111 111.1•.111
••• 111.1 11:12.1
II
11
1111 11,331
I0,331
11.m
11.m
llZ.1111
•.1111
UU17
• .Ill
.....
m.114
111.111 llnll
'1111

........ ·- ,..... ·-........ ....

'1111 tm
1111
1111 I
>•.•
>•,331
ll .....
1111.1 11:12.1 Ill.I

.... i 11.111.111 1a.111.111 Ill.Ill Ill.I 1a.1

-.,.
i
.
1•.1 ..I 11.m.111 111.431..1 111.111 111'1
um
.....
13 1711 I0,331 111.1411 MUii I.um 1111 I.IQ llll 11141 >•• 121.1

·-..... . ·-... -· -·
Ill >11,331 111U 111'4 11,111..1 1ll.111,471 Ill.Ill llU 1743.1
"Ii 1111 i 11.m •1.122 Ill.Ill 1111 1111 8

... •.i
•m t.11111 1111 >11,331 llOU 1m.1 12.111.711 m•111 111.111 Ill.I 1111.1
147' 8 11.m

....
a1.•
....
- .__ Ill.I
" 11
II.Ill
lllll
II.Ill
a.m
llUI!
...t:tl
111.111
111.ln 111.111 Ill un
IQ
111
111'
111 1
11.311
ll.111
II.Ill
II.Ill Ill.I
1111.1
1111.1 llU
-1.511 111,111,741
24.111.111 241.111.111 ..,.
111.111 llU 1111.1
1a1.1
11 71.flt IU7J m... 141.111 IM 1111 11,111 au 1au llU 21..1... lll. . . . . n1m Ill.I 1111.1
II M,111 Zl.141 •.SID . . .I 111.111 1.11111 Ill 741 !III 11.m M,111 nu 114.1 11.111.111 lll.ZZZ.111 211111 111.J 1117.7

II•
•
t
•
- :llS 11.111
II.Ziii
51,111
17,241
.......
12.lll
a.111
Ill.I..
Ill.I'll
lll.lrll
lll.lill
Ill.Ill
111.IDZ
1•111
141.IA
Ill.Ill
111.llDll
•J.UI
Ill.Ill
t.nm

··- ....
•.-1
·-I
.,. ·-
.....
114

....
111
.,.
.. llZ
Ill
Ill
i

I
....
11,711

11.m
u.m
II.Ziii
17,IDI
17,lll
17.111
Ill.I
711.1
71U
111.1
1117.7
11111.l
1117.l
1117.l
111.7
SID.I
MU
Ml.7
·-1 11,111,1111
11.111.1111
11,111.llll
11.111.1111
m.111.111
111.•.•
111,llUll
111.151.911
Ill.Ill
Ill.IA
Ill.Ill
Ill.Ill 4'U mu

f:QUATICHIS:

{ -~w-1••·}
NOTH:

; GAi RHUVOIR MATERIAL IAlANCE EllllATllJll lllTM WATER IOIFLUX.

• W. •Ii,; li,,.11 - Jwjl I - l•• J1 : EQUATION I 111 PAPER.
"THE NlllED DUANTITY IS OITAINlO FIOM THE NlV10UI TRAll lllTHIN ATIME INTUVAL.
.. .... DEVIATION FACTOIS DITAINH HOil REFUEllCE It
"ii" ... Aft.OT DF Jo VI.,¥ FDI AGIVEN REIUVOlft IS USED FDI TMll CALCUlATIDN.

ALL &UH: OATA FOR THll PllOILEJI AMA8 1M TAil ES I ANO I.

.. • Co 1'111 - "111' ; TOT Al FlllD lllELLHEAD IACl-PRESSUH CURVHOUATION. j" • -(.;;~ Ji ; EQUATION 2 tfil PAPER .

llllERE •• 0.11117, DITZ

 qgc = constant gas ftow rate, Mscf/D
(q'°;)max = initial open-ftow potential of the aquifer,
reservoir bbl/D
Ta = external radius of aquifer, ft
Tr = internal radius of aquifer, ft
t = time, days
Atn = time inteiVal n
tp, = time to establish pseudosteady state, days
t, = time to establish steady state, days
V,, = pore volume
W = initial water in place, surface bbl
We = cumulative water influx into a reservoir
or eftlux from the aquifer, reservoir
bbl
.:lWe,. = cumulative water influx or efHux during
an interval, reservoir bbl
Wei = cumulative water influx into reservoir (J)
within the common aquifer, reservoir
bbl
We; = initial encroachable water in place at
pressure p;, reservoir bbl
We; (t) = encroachable water in place at time (t),
reservoir bbl
W; = cumulative water injected, surface bbl
W,, = cu.'Ilulative water produced, surface bbl
z = gas deviation factor
cp = porosity, fraction
µ. = viscosity of water, cp
References
1. Craft, B. C. and Hawkins, M. F., Jr.: Applied Petroleum
Reservoir Engineering, Prentice-Hall, Inc., Englewood
Cliffs, N. I. (1959) 126, 206, 224.
2. Frick, T. C.: Petroleum Production Handbook, McGraw-
Hill Book Co., Inc., New York (1962) ll, Chap. 32,
5-22.
3. Pirson, S. J.: Elements of Oil Reservoir Engineering,
2nd ed., McGraw-Hill Book Co., Inc., New York (1958)
608.
4. Hutchinson, T. S. and Sikora, V. J.: "A Generalir.ed
Water-Drive Analysis'', Trans., AIME (1959) 216, 169-
. 178.
5. Hicks, A. L., Weber, A. G. and Ledbetter, R. L.: '"Com·
puting Techniques for Water-Drive Reservoirs", Trans.,
AIME (1959) 2Ui, 400-402.
6. Coats, K. H., Rapoport, L. A., McCord, J. R. and
Drews, W. P.: "Determination of Aquifer Influence
Functions from Field Data", J. Pet. Tech. (Dec., 1964)
1417-1424.
7. van Everdingen, A. F. and Hurst, W.: ''The Application
of the Laplace Transformation to Flow Problems in
Reservoirs", Trans., AIME (1949) 116, 305-324.
8. Lowe, R. M.: "Performance Predictions of the Marg
Tex Oii Reservoir Using Unsteady-State Calculations'',
J. Pet. Tech. (May, 1967) 595-600.
9. Mueller, T. D.: "Transient Response of Nonhomogenous
Aquifers", Soc. Pet. Eng. J. (March, 1962) 33-43.
10. Nabor, G. W. and Barham, R. H.: "Linear Aquifer
Behavior", i. Pet. Tech. (May, 1964) .561-.563.
11. Chatas, A. T.: "Unsteady Spherical Flow in Petroleum
Reservoirs", Soc. Pet. Eng. J. (June, 1966) 102-114.
12. Coats, K. H., Tek, M. R. and Katz, D. L.: "Unsteady-
State Liquid Flow Through Porous Media Having Elliptic
Boundaries'', Trans., AIME (19.59) 216, 460-464.
13. Katz, D. L., Tek, M. R., Coats, K. H., Katz, M. L.,
Jones, S. C. and Miller, M. C.: Movement of Under·
ground Water in Contact with Natural Gas, Monograph
-Project N0-31, AGA, New York (1963).
14. Bruce, W. A.: "A Study of the Smackover Limestone
Formation and the Reservoir Behavior of Its Oil and
Condensate Poois", Trans., AIME (i944) 155, 88-119.
15. Rumble, R. C., Spain, H. H. and Stamm, H. E.: "Reser-
826
voir Analyzer Study of the Woodbine Basin", Trans.,
AIME (1951) 192, 331-340.
16. Moore, W. D. and Truby, L. G., Jr.: "Pressure Per-
formance of Five Fields Completed in a Common
Aquifer", Trans., AIME (1952) 195, 297-302.
17. Mortada, M.: "A Practical Method for Treating Oilfield
Interference in Water-Drive Reservoirs", Trans., AIME
(1955) 204, 217-226.
i8. Coats, K. H., Tek, :&i. R. and Katz, D. L.: HMethod for
Predicting the Behavior of Mutually Interfering Gas
Reservoirs Adjacent to a Common Aquifer", Trar..s.j
AIME (1959) 216, 247-251.
19. Robinson, M. P.: "Pressure Interference Correction to
the Material Balance Equation for Water-Drive Reser-
voirs Using a Digital Computer", Trans., AIME (1958)
213, 418-422.
20. Bonet, E. I. and Crawford, P. B.: "Aquifer Behavior
with Injection", J. Pet. Tech. (Sept., 1969) 1210-1216.
21. Cargile, L. L.: "A Case History of the Pegasus Ellen·
burger Reservoir", J. Pet. Tech. (Oct., 1969) 1330-
1336.
22. Carter, R. D. and Tracey, G. W.: "An Improved Method
for Calculating Water Intlux", Trans., AIME (1960)
219, 415-417.
23. Stanley, L. T.: "Curve-Fitting Cuts Material Balance
Calculations", Pet. Eng. (Aug., 1961) 90.
24. Hurst, W.: "Water lntlux into a Reservoir and Its Appli-
cation to the Equation of Volumetric Balance", Trans.,
AIME (1943) 151, 57-72.
25. Havlena, D. and Odeh, A. S.: "The Material Balance as
an Equation of a Straight Line - Part II, Field Cases",
J.. Pet. Tech. (July, 1964) 815·822=
26. Matthews, C. S., Brons, F. and Haz.ebroek, P.: "A
Method for Determination of Average Pressure in a
Bounded Reservoir", Trans., AIME (1954) 201, 182-191.
27. Muskat, M.: Physical Principles of Oil Production,
McGraw-Hill Book Co., Inc., New York (1949) 583.
28. Agarwal, R. G., Al-Hussainy, R. and Ramey, H. J., Jr.:
''The Importance of Water Influx in Gas Reservoirs",
J. Pet. Tech. (Nov., 1965) 1336-1342.
29. Russell, D. G. and Prats, M.: "Performance of Layered
Reservoirs with Crossflow - Single-Compressible-Fluid
Case", Soc. Pet. Eng. J. (March, 1962) 53-67.
30. Manual of Back-Pressure Testing of Gas Wells, I?~r­
state Oil Compact Commission, Oklahoma City, Oltla.
(1966).
APPENDIX A
Aquifer Material Balance
A material balance equation may be developed for
a finite aquifer system as follows.
Total Volume at Pressure p
[(w~J?~ts) (!'nd~i,-= )]
+
Pore Volume at Pressure p
Original Pore Volume Lost Pore Volume
r(l \ Volume of Initial)
Contents -
Lossof
( Pore ': olume
)l
atp, \ atp IJ
Total Voidage Volume at Pressure p
_r( Volume Eftluxed and Produced \ 1
-l ~~ n (A-1)
In the algebraic form using the standard AIME no-
menclature,
{[W B'°] + [W, B'°]} - {[W Bwd - [c,(p, - p)W Bw1U
= {[We]+ [W,, Bw]} . (A-2)
JOURNAL OF PETROLEUM TECHNOLOGY
Dividing through by B"'i, which can be represented in graphical form as shown
in Fig. 5.
{[ w ::J + [ W; :;; ]}- W ~ -[c 1 {p; - p) w]} Note that the term W., is not total water in place,
W B"'; (which represents the total aquifer pore vol-
= {[!::] + [ W, ::; ]}. . . . . (A-3)
ume). The aquifer will still be 100-percent saturated
with water when all the aquifer pressure is depleted;
that is, when P = 0. Note, too, that the detenrJna=
Substituting tion of W"', the initial encroachable water in place,
B~ . . , -, is not basically geometry-dependent except to the ex-
B
-. = l + C"'(.Pi - pJ, (A-4) tent that iundamentai mensuration ruies can be ap-
"'' plied. Isopachous planimetry would be the most rigor-
we obtain ous of all approaches.
W [1 + c"' (pi - p)] - W [1 - c1 (pi - p)]
APPENDIX B
1 Water Influx Equations
= -
8
. [W. + B"' (W, - W;)] , . . (A-5)
Aquifer Rate Equation
"''
or The aquifer rate equation independent of geometry is
W B"'i {[1 + c"' (pi - p)] - [1 - c, (pi - p)]} q'° = I'° (p - Pt01)1.o • • . . . . (B-1)
= W. + B"' (W, - Wi) ; . . . . (A-6) The aquifer rate equation when graphically de-
picted is analogous to the productivity index curve of
collecting terms,
the oil wells and to the backpressure curve of the gas
W B"'' [(c"' + c1) (p; - p)] = W. + (W, - W1) B"'. wells (see Fig. 1).
The rate-time relationship for water influx against
. . (A-7) an increasing t:..p is shown graphically in Figs. 3 and 4.
The cumulative influx into the reservoir or efllux
Rearranging Eq. A-7 we have from the aquifer is determined by
-P -_ - [W•(c"'+ +(W,c - WWi)B"']
B"';
+ Pi·
w. = i qtl/dt
t
. . . . . (B-2)
1)

. . . . . . . . (A-8) Differentiating we have

To further generalize the equation to include inter-
ference effects of other reservoirs in a common aquifer,
' . . . . . . . . (B-3)
or

- [w· + ~ w.1 + (W, - Wi)B"'] (B-4)

P= - (c'° + c1) W B"'; + Pi '
Using the aquifer rate equation, we obtain
. . . . . . . . (A-9)
(B-1)
where W. represents the cumulative water influx for
the reservoir of interest, and W •1 represents cumula- then
tive water influx into reservoir (J) within the common dW. -
aquifer. The water compressibility can be considered -;ft=,'° (p - p"',). . . . . . . (B-5)
then as effective compressibility, which includes the
compressibility of the other nonproducing hydrocar- At initial conditions we can define the maximum
bon reservoirs. capacity or initiai open-ftow potential of the aquifer,
Eq. A-9 is the general equation, but to simplify the =
when p"'1 0, as
further derivation we will set the interference, water (qtlli)max = l'°(p.), • • • • • • • • (B-6)
production, and water injection terms to zero; that is,
:I W.1 = 0, W, = 0, and Wi = 0. or
We then have
• . • • . • . (B-7)

Therefore,
. . . . . . . . (A-10)
Defining [(c'° + c1) W B"'i] p; = W.;, as the initial q'° = (qtO.)mu (p - PVJ/) • • • • • (B-8)
Pi
encroachable water in place, we can write for the
aquifer material balance equation Then
dW• _ (q.,)mu: (p- _ P )
-P = - ( P· ) w:i w. + P• • . . . (A-11)
dt P•
1
"' '
(B-9)

JULY, 1971 827

or and will use any consistent set of units.
(n __ ,)__ -
dW, = '"2tD.,DIM (p - PVJt)dt . Cumuladve Water Inln Equation
P• Now we can derive the more useful cumulative water
From the material balance equation slope, influx equation. If we combine the equations

dp
dW,
= - (_EL)
W,, ' .
(B-11)
t
W, = [ ewdt . (B-2)

or and

-
dp = - (w:,
p·) dW,. (B-12) (B-23)

Combining Eqs. B-10 and B-12, then

dp = - J!.!_[(qwi)mu (p - P"'t)dt] (B-13) (B-24)

W,. P•
Simplifying and separating variables,
or
_ dp __ (qto.>mu dt. (B-14)
(p - Ptot) W,,
Then (B-25)
-
J
Pl
dp -
cP - Ptor) -
(qw.Jmax
w,, I' dt. (B-15)
Eq. B-25, when integrated between limits

Rearranging and changing limits on p, we obtain

(qto.Jmu [dt
' = J _-~·
dp • • (B-16)
(B-26)
w.. _
p
(p - Ptor) gives
Integrating between limits gives us

[ (qw.}~ax]
W,,
t= m[~ -Ptot] ,
P - Ptot .
(B-17)
(B-27)
which can be expressed as
but
P• - Ptot = eC<G'••>•u:1w..1t (B-18)
P - Ptot (qci)mu = Ito (p.) (B-6)
but Substituting and rearranging, we arrive at the final
form of the cumulative water influx equation.
(B-19)
or
q"' - (B-20)
le = (p - Pct) • (B-28)

Tnerefore, It is interesting to note that both the instantaneous

water influx rate-equation and the cumulative inftux
J"' (p, - Pwt) = eC<f••>au!W.<Jt (B-21) equation are identical in form with equations derived
q"' by Russell and Prats29 for predicting the performance
Now, defining ew = qw , of layered reservoirs. Their results and conclusions
should be directly applicable when the simplilied water
e _ Ito (p, - Ptor) influx approach is used. .JPT
to(f) - eC<G'.1)au/1V.,]1 ' (B-22)

which is the final fom expiessing the instantaneous
rate of water influx as a function of time and the
internal boundary pressure, (pto1). The equation is
quite general and totally independent of geometry,
O,.lnml m•nuscrlpt received In Society of Petroleum EnslnHrs
offlcm Aug. 5, 1969. Rmvised m•nuscript received Aprii 6, 1971.
Pilper (SPE 2603)w• Pf'ft•nted •t SPE 44th Annu•I F•ll MHtlne:,
held In Denver, Colo., Sept. 211-0ct. l, 1969. C Copyrlsht 1971
Amerlc.n Institute of Mlnlne:, Mmllurgic•I, •nd Petroleum Ene:i·
nHrS, Inc.

828 JOURNAL OF PETROLEUM TECHNOLOGY