Minor Revisions in January 1969

,l.j-z 4Z~L13

THIS IS A PRE-PRINT - SUBJECT TO CORRECTION

RADIUS-OF-DIUINAGE EQUATION

FOR

PRESSURE BUILD-UP

by

Anis A. Ishteiwy Oasis Oil Company cf Libya, Inc6

H. K. van Poollen Marathon Oil Company

Prepared for presentation at the Libyan Association of Petroleum Technologists’

Annual Meeting, January 25 - 26, 1967, at Tripoli, United Kingdom of Libya.

This paper is an excerpt from an MS Thesis in Petroleum Engineering presented

at the Colorado School of Mines in June, 1966.

ABSTRACT
.—— ——— —-

For two boundary conditions, radius-of-drainage relationships are

derived from pressure build-up analysis for a single well. The first boundary

condition considers a single well near an extensive linear barrier in an

otherwise homogeneous and infinite reservoir. In the second case, a single

well is assumed near a linear and extensive pressure source in an otherwise

homogeneous and infinite reservoir.

 -2-

Theoretical pressure build-up curves are plotted, the points at which

the effect of the boundary was felt were graphically determined, and then

correlation was made to arrive at equations for the radius of drainage. The

radius of drainage determined by these equations represents the distance

within which no discontinuity exists. It may be considered the distance to

which connected pore space has been proven.

 -3-

RADIUS-OF-DRAINAGE EOUATION

FOR

PRESSURE BUILD-UP

Introduction

When a single well is placed on production, a radius of drainage which

increases with time results. Variations of pressure transients with time

show that for some flow time the pressure drop is very small, and it approaches

zero at some distance from the well bore. Even though it appea?s that for

a specific flow time the pressure change has a value of zero at some distance

away from the well bore, this is only relative. According to the diffusivity

equation, pressure change occurs everywhere in the reservoir as soon as

pressure changes in the well bore. It was this apparent constancy of the

pressure at some point out in the reservoir that led to the concept of “radius

of drainage.”

Depending on their assumptions, many authors in the field of transient

. ~. c.- -+ ==d;l}~-of-drainage
equations. Among these
fluid behavior derived Ulf.=..e,..
.uu&=ti

are: Tek et al (1957), Hurst et al (1961), Jones (1962), Hutchinson and

Kern (1964) and van Poollen (1964). The equations developed by these authors

are applicable to a pressure build-up test only after a steady-state flow

condition has been reached. Horner (1961), Davis and Hawkins (1963),

Standing (1964), am? Gray (1965) Wesented

L equations for finding the distance

to a linear barrier. In Horner’s method, which was further developed by

the other authors, a considerably long shut-in time is required to establish

the second straight-line segment with considerable accuracy.

 -4-

Purpose and Scope of Investigation

It is often desirable to determine the radius of drainage under unsteady-

state conditions with short shut-in time. A drill-stem test, where the sum

of flow time and shut-in time is usually in the order of half an hour to two

hours, is a good example of that situation. It is the purpose cf this paper

to develop a relationship for finding the radius of drainage from pressure

build-up analysis under unsteady-state conditions.

In Case 1, the presence of an infinite linear barrier that completely

Gbstructs fluid flow in an otherwise infinite and homogeneous reservoir is

assumed. An actual field analogy would be a well near an extensive linear

fault. Another approximation for this idealized case would be a well in a

relatively thin reservoir near a high-contrast fluid-fluid contact, or a

reservoir exhibiting a sudden change in formation properties such as thickness,

porosity? or permeability.

In Case 2, a single well is assumed near a linear and extensive pressure

source in an otherwise infinite and homogeneous reservoir. A field approxi-

mation would be a well near a pressure-maintenance system whereby the pressure

at that boundary is essentially constant with time.

Equations Used

In a true infinite system, which is the situation considered in this

paper, the solution to the diffusivity equation which is frequently used has

been referred to as the point-source or the Lord Kelvin solution. This

solution is:

(1)

Where Ei is the exponential integral function.

 -5-

A. Case 1
are applied
When the principle of superposition and the method of images
equation
to the above solution for Case 2, the following pressure build-up

resultS:

(2)

I .
When the dimensionless time and dimensionless pressure droQ are Introduced,

equation (2) becomes

2 2
.- -r -r
w ‘1 +Ei ———— 1
(AP)D= -Eic.
-a2 (t+ O)D~ ! a2 ;D ~

(3)
- Ei [ (t-+%-l + ‘%$]
L- ‘IJJ L

Where

4k (t + ~) 4kt
(t + O)D = , and t:)= —
$ LICa2 f/IPC a2

B. Case 2
,m-— *L.
WFl~Jl --...,-.
.nnmnantl
LIl~halllc
c.,,,,* is fnllowed for Case ?, the followin,c,
.Ju.-. pressure

build-un equation IY?SUltS:

 -6-

(AP)D= -Ei~
~a2 ‘r:
(t + @)D ) “i[-l ‘E+=-d
(4)

Method of Iiivestifgatim

The method followed in this investigation is somewhat similar to that

presentedby Hurst, Haynie, and Walker (1961). Using equations (3) and (4)

for Case 1 and Case 2 respectively, theoretical pressure build-up curves are

plotted (see Figs. 1, 2, 3 & 4). For different dimensionless flow times

and an arbitrary distance to the boundary, a set of pressure build-up curves

is plotted whereby the ordinate is dimensionless pressure drop (AP)D and the

[)
t+e
abscissa is the logarithm of the dimensionless time ratio ~ . A

total of 26 different values is assumed for the dimensionless flow time. Flow

times range from 10 to 10,000. The distance to the boundary is then varied

and another set of build-up curves is obtained. --f-~~-+k:=

Ci Lii.&” range the
Siii~eW.lLII-l.

curves are essentially identical for each boundary condition, only sample

curves are shown in this paper. These curves have been photographically

reduced from a 15-by-20-in. grid to the present size. In all calculations

the value of rw is assumed to be 0.5 units.

The calculations were programmed in Algol for a Burroughs B5500 digital

computer, and the curves were plotted by use of a Calcomp plotter.

Analvsis of Results

For the boundary conditions considered in this study, the shape of

the curves (Figs. 1, 2, 3 &4) is characteristic of pressure buiid-up

behavior. It is seen that for early shut-in times (for large values of
 -7-

t+o
0 1the curves follow the logarithmic approximation to the point source
solution and form straight lines. As soon as the effect of the outer boundary

is felt by the well, the curves start deviating from the straight-line form.
It+(y
When the ratio of — approaches one, i.e., when shut-in times are very
\ G
)
large compared to flow times, all the curves converge towards a zero pressure

drop.

For each type of boundary, and within the range of distances considered

in this study, it is seen that the curves virtually coincide for corresponding

flow times regardless of the distance to the boundary. A very slight deviation
t+o
occurs at very large values of ~ . This deviation is noticeable only
()
when curves for small distances are superimposed on curves for very large

distances. Such small deviation can be seen, for instance, when Figures 1 and

2 on the original 15-by-20-in. grid are superposed and put on a light table.

Since in each type of bo~dary all curves for corresponding dimensionless

flow times essentially coincide, it seems reasonable to conclude that within

the range of distances investigated, the points of deviation from a straight

line, for corresponding dimensionless flow times, occur at the same values
Oft+o
Due to the inherent inaccuracy in determining such points
--%---”
()
graphically, readings were taken from all the curves (see sample curves in

Figs. 1, 2, 3, & 4). For each case of boundary condition, the readings were

averaged numerically and tabulated in Tables 1 and 2. For the sake of

t+o
simplicity, the values of I— at which the curves deviate from bein~
1° )
straight lines were referred to as the “points of deviation” or simply

Especially in the case of a linear barrier, some difficulty was en-

countered in determining the points of deviation for large dimensionless

 -8-

flow times. This difficulty is mainly due to the scale of the drawing.

Therefore, these curves were ignored.

Correlation of Results

In trying to arrive at a radius-of-drainage relationship for each type

!t+o
of boundary condition, the average values of — o at the deviation points
1)
were plotted vs. corresponding dimensionless flow times on a logarithmic

scale (see Fig. 5). It can be seen that the points fall approximately on a

straight line. An average straight line is drawn through these points.

An equation is fitted to the straight line to read as follows:

‘t + ~\ deviatiOn
= 1.13 ~ (5)
‘D 1/

4kt
~ by definition (6)
‘D = $pca

Equating (5) and (6) and solving for a, it is found that

(7)

Examples:

Two examples

equations.

Data:
-4 -1
k= 590 md, u = 0.60 Cp, c = 7.35 x 10 atm. ,

o = 0.20 t= 8.3 hours h

The dkt~il~~:

a= (actual distance to discontinuity) = 2000 cm or tD = 200.

 -9-

Case 1: Presence of a linear barrier.

t+e
From Fig. 2 for tD = 200 — dev. = 155
Q
()

f 1/2
kt
a = 1.88; t+o
i+ lJc(~) 1

1/2
“(0.59) (8.3) (3600)
= 1.88 = 2130 cm.
~70.2) (0.60) (7.35 x lo+) (155
d

cnllvc.e
---- - ,

tt@
()
~ev

From Fig. 4 for tD ‘ 200 . = 190

(3

1/2

. -11/2
= ~ .88[(0.~9) (8s3) (3600J I a A--,
1097 em.
“...
~(0.2) (0.6) (7.35 X 10-4) (190) 1

Conclusion
l-.
~Qr the two beundary conditions considered in this paper, approximate

radius-of-drainage equations have been derived from pressure build-up before

a steady-state condition is rsached. A

n role.+
p~”. of ~~~~~lure build-UD
r
VS. the

kgaritlm! 14 gives a straight line until the effects of a discontinuity

\u/
are felt by the test. When the curve deviates from being a straight line,

equation (7) can be used to find the radius of drainage.

Nomenclature

a= radius of drainage in centimeters

t= producing time before shut-in, in seconds

k= permeability in darcies
= porosity; a dimensionless fraction
$
= viscosity in centipoise
B
-1
c= compressibility in atmospheres

q= production or injection rate in cubic centimeters per second

..*m+.fl*
at standard ~-.-.. Conditions
----

B= formation volume factor in reservoir units per standard

surface units

h= i%rmation +hi~~ness
s...-... in centimeters
I..

P= pressure drop at time t and radius r, in centimeters

P,t
APD = dimensionless pressure drop

dimensionless flow time

‘D =
r= well radius in centimeters
w

References
n-..:
lkivLa,
“ E. c
and Hawkins, M. F., 1963, Linear Fluid Detection by Well
“. ,
Pressure Measurements: Jour. Petroleum Technology, Oct., p. 1077-79.

Gray, K. E., 1965, Approximating Well-To-Fault Distance from Pressure Build-Up

Tests: Jour. Petroleum Technology, July 1965.

Horner, D. R., 1951, Pressure Build-Up in Wells: Proc. Third World Petroleum
Congress, the Hague, p. 503-21.

Hurst, William, Ha~ie, O. K., and Walker, R. N., 1’361,Some Problems of

Pressure Build-Up: Paper SPE-147 presented at 36th Annual Fall Meeting
of Sot. Petroleum Engineers in Dallas, Texas (Oct. 8-11, 1961).

Standing, M. B., 1964, Discussion of Linear Fluid-Barrier Detection by Well

Pressure Measurements: Jour. Petroleum Technology, March, p. 259-62.

Tek, M. R., Grove, M. L., and Poettmann, F. H., 1957, Method for Predicting
Back-Pressure Behavior of Low-Permeability Natural-Gas Well: Trans. AIME,
210 - 302, 1957.

van Poollen, H. K., 1964, A Hard Look at Radius-of-Drainage and Stabilization

Time Equations: Oil and Gas Jour., Nov. i, p. ii8-2S.
Acknowledgments

The authors wish to express their appreciation to the Graduate School

of the Colorado School of Mines for allowing publication of this study.
 I
‘~”-t– \ I I
I
I
I
I
I

.._. —.- ——

.1
&
g“
o
u
a
~1-
In
u
a
a-
\
I

FIGURE 1 THEORETICAL ~Ul~u-ur u.,,,.

FOR A WELL NEAR A LINEAR E..-...__
AT A a = 100 cm
.

)= ‘F
~—
..-.
1 M
10 1’00
.(I
,$.

,
 —-

cl)
+ <
t-
.’

/
,’

,
0

t“

t=!
4 G-1
0
I 1 I -“
II Zt c1 II ;1

S3Ud 37N01S 110

I
 1
/

/
. .

T. .

, 1 /

I
/

g
o
0
k+
m II
/

/
/

/ .

L
m
La
M
3
cl
H
L

I O’.z .
s~,J &NOISNs%J
— — &.
0 “o
-J-l
S“*
,., ,.:
8“0 I I

I
.

I
 -
--=s+~-zz====-
---
‘.’\-3~~s:?+
:7 -- -—— .==-==.” —
=.b_

i
~.
,

———-- - ——”
.-. ~ .,
..-2% ,—k —
‘,. \,
\.
\\’
I ‘L, x. ‘-l
\.
. .-—.
—-—-——.—— ++
= #—
..

— _. .. . . ...
w. J
-.--—
—..
—
————-_-_::—-_ n:--
“1 .-

-:--. _... _ _.—

.. ------———
r

. ..~
~. -. .—-
-:
,

I& O+–-. ___~

ml
K
0;

~:f~

:j
___~
m .+
U* I
gl
z
y

.~

m
~.

m
J FIGURE U THEoIWTICAL
FOR A WELL
pRESSURE SO

Iw
-. -.,,. L.. .-’.,- ..-’

a o
OIKNSIWESS TH RATIO

T+ e’”” “ ‘
i

Q
 10’

10

‘D
=1.13
() t+e
~ dev.
+

‘D

+ CASE I

o CASE2

Ic

FIGURE
5 FLCY1’
OF DIMENSIONL5S TTME
‘nm:;us
VALUES
OF $+2 ON
BUILD UP CURVTS ‘Oii?fiE
DEVIATION
FROM STRN(WT LTN E OCCUR.. .

II ‘4
2 103 I

(w de?”.