Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
the pressure gradients in the Howing Huid. causing How, and for horizontal How the
Assume that a two-dimensional radial How pressure P may be used for the value of rp.
exists. The equation for the rate of How However, for nonhorizontal How the veloc-
involves the quantity rp, which is the ity potential is used to express the force
velocity potential, a measure of the relative causing flow, because it applies a necessary
force at a point. The gradient of velocity correction to the pressure at various points
potential is a measure of the force gradient in the system. This correction amounts to
186 FINGERING AND CONING OF WATER AND GAS IN HOMOGENEOUS OIL SAND
the portion of the difference in pressure sufficient shut-in time has occurred to
between points that is due to hydrostatic allow equilibrium to be attained, the cor-
gradient, the effect of gravity on the mass of responding value of r. is the maximum
fluid between points of different elevation. radial distance from the well that influence
The rate of flow may be expressed by the extends in affecting pressure and drainage
following equation:* of fluid. Although usually this must be
arbitrarily assumed it should be noted that
[I] throughout the practical range of the values
where in the relationship, large differences in r.
et> = P - "!ogz cause relatively small differences in the
results, because the logarithm of the ratio of
In order for a water finger extending up- radii is employed. Where values of the
dip toward a producing well to be stable it is other factors are known, the equations
necessary that the pressure gradient at the developed may afford a means of measuring
top of the water surface equal the differen- an effective value of the drainage radius in
tial gravitational force between the water a uniform sand.
and the oil. This stable relationship is To illustrate applications of Eqs. 12 and
expressed by IS, consider the example defined by the
/let> = AP -/In (t - ~; r) [4] following data:
Differential weight between water and oil
Combination of these two equations in = /l,,!g = 0.141b. per sq. in. per ft. of depth.
the manner developed in the appendix
gives Eqs. 12 and IS. The minimum Dip of sand = ¥r = sin 8 = 0.1
distance that a finger can approach without Well radius = r .. = H ft.
being drawn into the well by an unstable
condition is given by For selected values of drainage radius of
600 and 6000 ft. and for the given condi-
tions, the relationship between distance of
the well above the static water table t and
the drawdown pressure /lP is presented in
The minimum vertical distance a well Fig. 2.
can be located above the static water table The asymptotic nature of the curves
without fingering is results from the fact that water cannot be
drawn toward the well from a location
beyond the drainage radius of the well.
WATER AND GAS CONING
Consider the coning of gas or water into a
The term r. in the foregoing equations well that partially penetrates a homoge-
represents the drainage radius, which is neous oil sand, as illustrated in Fig. 3.
the distance from the center of the well at Theoretical development of this problem
which the static or shut-in pressure of the has been presented by Muskat,l who has
well exists during flowing conditions, and shown that the following equation repre-
the difference between this static pressure sents the relationship of the factors involved
and the flowing pressure at the well con- in a homogeneous sand:
stitutes the drawdown pressure /lP. For a
value of static pressure measured after
/let> = et> .. - et>.
/let>. et> .. - et>.
= I _gIl,,!
/lP
h (I _ ~)
h
.• Nomenclature at end of text. [16]
M. G. ARTHUR
200
190 I I
6P
180 f - - rmin =
6~·g·Sin a'ln2
110 f - -
rw /J
160 f - - ( r,.
/ /
tmin=--
61"9 \
AP 1+lnr.-:
m,"
In~
)
/
V j
150 r - -
r",
/ Becom•• A:~rnptoti"
140 r- 6P" Well .ta~ic pr..... - flow pre .... lb •. /in.· / b ~.600·.t _
A "f.g • Differential wei,ht betw.en "'ater f oil V
130 f- a 0.14 Ib./in·/ft. of depth /
~~ .. dip of .and· 0.1 /
120 I- r.. • W.II radius' Y3 ft. I /
~
..:
110 I-
t • Req"ired hei,hf of ",ell aboy .. water /
t.bl. fOI"
well. ft .
no flow of water ;"V
... 100
90
;(I . 6000'
/
80 J
70 /
60 / ,.- -
50
V /'
40 ! /",re =600'
30
VI ~
20 II
10 f
I
10 20 30 so 60 70 110 90 100 110 110 130
6 P (Ib • ./i"~)
FIG. 2.-MINIKUK VEllTICAL DISTANCE OF AN EXAMPLE WELL ABOVE THE STABLE WATER TABLE
FOR WATER FINGEll NOT TO ENTER WELL.
188 FINGERING AND CONING OF WATER AND GAS IN HOMOGENEOUS OiL SAND
Eq. 16 relates the potential forces set up (b) Correct the value of 1/>' to the value of
II
in the fh!id flowing in the sand to the dif- 1/>., corresponding to the radius parameter
ferential gravitational forces between. the of the well by use of Fig. 5.
t ~~ t/Gas-Oillnferface
-t--------~I ~ l-----!>tt::!~::)':
r~ r FIG. 3.-WATEl1.-CONING AND GAS-CONING CONDITIONS.
oa-Water
Interface
n
gency represents the maximum height y of
GRAPHICAL SOLUTION OF CONING
a stable cone. (i = I -
For the case of water or gas coning to a
(b) Intersection of the tangent to the
well in which the penetration of the well
extends from one boundary of the oil sand curve with the value 1= 0 gives the value
to a point some distance from the static expressed by the following form of Eq. 16:
position of the water or the gas, the solution
may be obtained by the following procedure: (tt)i =0 = I _ g~;h
I. Determine the values of the potential
3. From the graphically determined
• t.1/> f .
ratIO t.1/>. or vanous values of from the
I
value of (tt)i -0 solve for llP, the maxi-
charts presented herein and plot ::. vs. ~. mum pressure differential without coning.
(a) I/>~ is given by Fig. 4 for the par- llP - gt.-yh [21J
ticular penetration of the problem and for a - (t.I/»
well-radius parameter p. of 0.001. 1- t.1/>. i- o
M. G. ARTHUR
An additional step is required in the dure outlined above. The basis for location
solution of the problem of water and gas of the horizontal dividing plane in the sand
coning in a well in which the interval open is that the potential over the well surface
17
16
V /,
v:-
15
14 /' I
V /1
"" ~,1
13 --
~),- V J
12
I ~1
l 1I
,/ 1/1
/
,
II" /1
JI
V I!
I I I!
8 / 1/ i J
7
V~ I/
1/ IfA
...J
c(
~
Z 6
~ ~ 0/
W
~
o 5
a..
4
~'\.S ~ ~ l>~
V
~z / /. ~ ~
v::
3
l /' ~ ,.....~ ~
L ~~~
V
I /~
jii P"'"
0
o .1 .2 .3 .4 .5 .6 .9 1.0
WELL PENETRATION
FIG. 4.-VELOCITY POTENTIAL DISTRIBUTION AT WELL RADIUS.
to production is located within the sand cf> .. must be the same for both the upper and
body, above the static water level and lower portions. The division is located by a
below the static gas-oil interface. In such a trial and error choice of values from Fig. 4
case the sand should be divided into two corrected by Fig. 5. Solution is greatly
parts and each part treated by the proce- facilitated by a plot of the corresponding
1.2
...
."-
~~~~ ~1\
~ \ \ '8
::i. \
1.1
0
~
~~
~
0
0
;t cP 1 14, ( Pot jential ~ ;yond ..t1'4 tnity of ell
1.0
~
a:
~ \ ~10 w. I Per ~t,.. on
...J t- = 0 10
4:
i= 0.9
\~~~~
Z
w I
~
0Q.
,Q
0 0.8 ~1I.
)\ \
~ ~f" ~ ~
~
t-
~
...J
4: ~ ~
'" "
~ phe ric. I low
i=
Z ~ P'= I~
r' \ ~
~
t'- ~
0.7
~
.~
W
\ 1\
w
t- \
~
O
\ '"1\
Q. \ I' I"-
\ r--
""
0., \ 1\ & f'-
0
0.6
..
\ " ~ "f'
""
i=
c(
a:
O.S
.ooot .0001 .0005 .000'1 ,oOt
\.OOC .005 .004
f\. .001 .006
I'-.
.008 .QI
000
900
800
'\
\. \.
" \
\ \ \
700
\ \ \ \ \
"""
1\ \ \ \ \ \
"A \. \ \ \ \ \ \
'A, \ \ \ \ \ \ \ \
200 1\ 1\ 1\ \ 1\ \ 1\ 1\ 1\
; \
oJ
a::
II
100
90
i\l\ \ ,,\ ,\ \ \ ~~ ~
'\
"i\' " ""
y:; I.,,!I.
\
~
'\ \dI
LLI eo \~-
70
\ \ ~
I-
60
\ \ I~. , :- \ i\ \
".'
LLI
'\ \ \ \
~
~ 50
"\
«
a:: '<l \ \ \ \ 1\ \
40
«
Q.
30
K\ \ \ 1\ \ \\
~ 1\\\ 1\\ '
\
en
::>
25 20
\
~ ~ ~\\
«
a:: \
LLI
'z«"' 10
9
,
8
1\ \.
,'\.'\. \ \ \
\.\. \ 1\ \ \ \ \
;(
a:: 6
\.\ ,\\ \ \ \
o b. Well ,ft. ~\\\lill
5
~~\\\
h • Sa"d Thickness, ft.
4
~
~
3
l
-10 -9 -8 -7
POTENTIAL AT DRAINAGE
-6 -5 -4 ..,3 -2
RADIUS ~.
-I
1ll
0 +1 +2
values of well potential for each portion of Location of the dividing plane is shown
the sand against the distance of the divid- by Fig. 7. The values of cf>wT (well potential
ing plane from an arbitrary datum. in top portion) and cf>"B (well potential in
12.SO
~
I I I
12.40
~ 0 Potentia' above dividing plane
A Potential below clividinCj p'an. ___._
"
12.30
~ ')
+
.
l
12.20 ~ /
~
"li
."
~
J
• '2.10
V
"
/' k '\
~
12.00
V
/ TW T
11.90 / /
11.80
121 122. 123 124 125 126 12.1
Oi$tanc.. of Dividinq P'ane from Gas Oi' Interface ft.
FIG. 7.-S0LUTION OF LOCATION OF DIVIDING PLANE AND VELOCITY POTENTIAL OVER WELL SURl"ACE
FOR CASE OF SIMULTANEOUS CONING OF WATER AND GAS.
..OO,----"""T'----r------,---"""T'"----r---_
0.9
O.B
':1
-eo 0.6
mal(.imum equililt,.ium h.i9ht of: cone
o 1000'
I-
<:(
a: O.S
.J
<:(
I-
Z
ILl
I- 0.4
oQ.
0.3
Per-foro.tiO". +0
tap of oil 10ft_
0.2
0.1
o O.~4------·~~~5--~~~~------~0.=7-------0~.~B-------O~9~~........I~.O
PENETRATION ~
FIG. S.-EXAMPLE OF GRAPIDCAL SOLUTION OF EQUILIBRIUM HEIGHT OF WATER CONE, AND KAXDlUK
PRESSURE DRAWDOWN WITHOUT WATER CONING, FOR SELECTED VALUES OF DKAlNAGE llADIUS.
194 FINGERING AND CONING OF WATER AND GAS IN HOMOGENEOUS OIL SAND
,
= 29.75 per cent 12.14
1.000
.900 .rh ~
~ .
{/
~~ ~
1MlI'
~ .BOO ~~ ~ ..nno:
.,,,M/l
,
~ / 7~ i ~ 6+e = I - ~h(
M... 6P·I-.!..)
h
~
Value of h - Z at point of t.n9.n~y
9ives ma .. imum .'1uilibriuM depth of cone .
.500
/
.400
o ~ ~ ~ U ~ MUM M ~
PENETRATION %
FIG. 9.-EXAMPLE OF GRAPmCAL SOLUTION OF EQUlLI1IRlUM DEPTH OF GAS CONE AND MAXIMUM
PRESSURE DRAWDOWN WITHOUT GAS CONING, FOR SELECTED VALUES OF DRAINAGE RADIUS.
q,•• = Cq,'•• = 14.S4 X 0.836 = 12.1S determined for various values of drainage
M. G. ARTHUR
radius and are presented in Fig. 8. The radius for the given example. The low
maximum allowable pressure drawdown values of drawdown pressure that would
without water coning is calculated from result in water coning are caused by the
1.100
..:
vi
a:
\
.;.
c 1.600
C
--.. -
0
U
.
10
L
~
1.500 \ I00.
"'<
) WATER CoNING
0
-£ vi
j n:
c:
J
'" I'---..... .,.
-- -
0 1.400 ........... 90. .~
.
."
J
\ r-- r--- o
0
...
L
.
U
•
\ ~
'xof
E
\ GAS CONING ~
1..
::E 1.200 10.
"- ~ - 'C---
o
..
..."
.............
I-- r--- 60.
l'
-
1.100 -f-
Q.
E
3
...E
•
~
1.000 so.
o z ;5 5 6 7 8 9 10
Drainage Radius, r., Thou.ands of Feet
FIG. lo.-EXAMPLE Ol!' MAXIMUM PRESSURE DRAWDOWN WITHOUT CONING Ol!' WATER AND GASAS
l!'UNCTION Ol!' DRAINAGE RADIUS.
with the corresponding value, if the per- geneous sand, coning is greatly restricted
forations extended up to the gas-oil inter- by small lenses of relatively low permeabil-
face. Although the constants in the evalua- ity directly below the bottom of the well,
tion are entirely different, as may be seen since these small lenses greatly distort the
in Fig. 8, it would seem logical that the potential gradients in the sand.
values of drawdown without water coning
would be of the same order of magnitude for ACKNOWLEDGMENTS
each case. For a drainage radius of 1000 ft.,
the maximum drawdown is I.21b. per sq. in. The author appreciates the permission of
for the case of perforations extending to the the Union Oil Company of California' to
top of the productive oil sand, and 1.5 for publish this paper and especially thanks E.
the actual conditions of the well chosen as C. Babson for many suggestions and con-
an example. tributions. Thanks are also due to Jan Law
Gas Coning for his review and criticism.
The equation of the surface of the water Equating this to the general equation for
cusp is expressed by: y (Eq. 8), and solving for t gives
The limit of approach of the water finger The example cited in the body of the
to the well without exceeding the condi- report illustrates application of water
tion of stability can be determined. Fingers fingering in a homogeneous sand that has
of water approaching closer than this limit the geometric conditions shown by Fig. I.
will be drawn into the well. The potential In coning of water or gas into a well that
gradient set up from fluid flow may be partially penetrates a homogeneous oil
represented by sand as illustrated by Fig. 3, the relation of
the potential forces set up from the fluid
flowing in the sand to the differential
[9]
gravitational forces between the fluids
involved is shown by Muskat 1 to be
Substitution of the relationship expressed
by Eq. 6 gives A</> = </>.. - </>. = I _gA'Y h (I _:)
A</>. </>.. - </>. tli' h
[16]
[10]
Equations expressing the velocity poten-
tials deserve some explanation. Muskat
and the potential gradient set up by the has derived the following expression for the
differential gravitational forces between value of the potential over a sand in which
th,:, oil and water is given by there is a partially penetrating producing
well. The expression is valid for small
d</> dz • values of the radius parameter,
fir = !:J.-yg dr = !:J.-yg sm (J [II]
</> = {_ In reI + +
w X)r(I - w x) +
Simultanet)lls soluti(lfi (If the two pre- q reI - w - X)r(I w - x) +
ceding equations for r gives the minimum + In w + x + [p9 + (w + X)2]~
distance a finger can approach without W - X [p9+ +
(w - x)2]~i
being drawn into the well by an unstable
condition:
-i p'[f(2, I - W - x) - t(2, I - W+ x)
+ t(2, I -I- w - x) - rc2, i +. 10 + x)]
tli'
rmiD = [12] + 0(p4) \ [17]
A-yg sin (} In !!
r ..
The last term drops out because the
From geometric considerations, coefficient is zero. The term involving p2
is negligible for the range of values encoun-
y =t- r sin (} [13] tered. Eq. 17 is derived for a well of in-
DISCUSSION 199
finitely small radius. However, it does not Ko is the Hankel function or Bessel
conform to the requirements that the function of zero order and modified second
potential is uniform over the well surface. kind (sometimes called the third kind).
In order to satisfy this condition for a A detailed discussion of Bessel functions
well of finite radius, the flux density q is given by Whittaker and Watson. 2 For
is not taken to be uniform over the well large values of the argument X, the follow-
surface. It is adjusted to values over finite ing asymptotic expansion is given for
increments of length of the well that will Ko by Dwight:'
enable a uniform potential to be obtained
over the surface of the well. Fig. 80 of Ko(X) ~C~)~ e-x [I - J!~2X+ 2~~8;;2
Muskat1 shows the adjusted and unad-
12 • 32 • 52 ]
justed values given by Eq. 17. Fig. 4 of - 3f(8XP + ... [191
this paper is a cross plot of the adjusted
values presented graphically by Muskat. Since the value of the argument X = 2mrp
Muskat indicates that the actual well is large, the value of the function Ko is
potential may be taken to be proportional neglibibly small for practical values of
to the value obtained from Eq. 17, using drainage radius. For p equal to or greater
a uniform flux density q and setting than one, Eq. 18 may be written
w= %x.
Fig. 4 is based on a particular well-radius 2
cP = 4qxln-p
parameter p. = 0.001, or rIO = 0.002h.
In order to correct the values given by
Eq. 20 is the basis for the values of cpo
Fig. 4 for potential at the well-bore radius
presented in Fig. 6.
to values of well-bore radius other than
p.. = 0.001, Eq. 17 was evaluated for
DISCUSSION
different values of the variables and with
w = %x. It was assumed that the variation N. VAN WINGEN,* Los Angeles, Calif.-This
paper presents an interesting application of
of the adjusted well potential with well
mathematical concepts developed by M.
radius is proportional to the variation of
Muskat. Because of the complexity of the
the unadjusted values of potential. The problem, the mathematics are of necessity
variation of CPo with P.. was found to be involved and although treatment is facilitated
negligible except for points very close to by means of Mr. Arthur's charts, a considerable
the well. Well-bore correction factors to knowledge of advanced mathematics continues
be applied to the value of well potential to be necessary in order to follow through
CP ..' from Fig. 4 are presented in Fig. 5. adequately the various steps of the analysis.
Variation of CPo is also included in Fig. 5 As such, the paper presents a formidable
for the extreme case of 90 per cent well obstacle to the average engineer.
penetration and at a point 10 per cent into In regard to the practical value of the results
to be obtained by this method, it is apparent
the oil sand beyond the extremity of the
that the simplifying assumptions, which, of
well.
necessity, have to be made as yet for the
For large values of the radius parameter, analytical treatment of a problem of this
the value of the potential in the sand is nature, may in certain areas severely limit the
given by Muskat to be quantitative usefulness of the data. For
n- '" example, it seems doubtful whether any pro-
cP = 4q [!7r n=1
k !ff Ko(2n7rp) cos (2mru') ductive zone in California could meet the
required prerequisite of homogeneity and yet
be sufficiently massive to permit the selective period in which the cone is formed, the situa-
perforation of but a portion of the productive tion is somewhat different.
sand to yield a commercially productive well. The rate of advancement of the cone is a
Likewise, inability to discern small lenses of function of the relationship between the vis-
low-permeability sand directly below the bot- cosities of the fluids and relative permeability.
tom of the well in an otherwise homogeneous During the time the cone is formed, oil is
sand body would result, as the author points produced from the well, causing a general rise
out, in the well being perforated unnecessarily of the water table, and conditions may prevent
high. This will not only cause a loss in well stable conditions from ever being reached. It
potential but may well result also in the obtain- would be interesting to make some investiga-
ment of a lesser ultimate recovery due to tions for different known viscosities and relative
by-passing of the encroaching water. One other permeabilities of the effect of the delay in the
point suggests· itself, and that is that the per- formation of the water cone.
missible well drawdown to prevent fingering In a water drive, the oil is produced as a
and coning will decrease continuously as water result of the replacement of oil by water. For a
encroaches with the depletion of the oil reserve, maximum recovery of oil, it is essential that
or in the event of an expanding gas cap. This the washing action in the area, invaded by the
again may mean that well rates may soon water, is as complete as possible, and that the
become uneconomic. volume of the reservoir, invaded by the water,
These comments are not offered to distract is a maximum at the time the well reaches its
in any manner from the value of this excellent economic limit. The mathematical develop-
paper, but merely to emphasize again that ment in the paper deals only with static condi-
extreme caution has to be exercised in applying tions, and it would be interesting to see an
analytical methods to actual oil-field problems. analysis of the effect of the rate of production
It is to be hoped that work of this nature will on the efficiency of replacement and the size
continue to be pursued intensively, so that and shape of the reservoir invaded by water at
ultimately we may perhaps learn to evaluate the time the well is abandoned.
specifically the effects of the various complicat- These investigations would seem necessary
ing factors that now have to be ruled out from for an interpretation of the mathematical
the consideration as being too complex to lend developments of the paper. The numerical
themselves to mathematical treatment. example shows that for a development with
600 ft. radius (± 30 acres per well) the allow-
W. T. LIETZ,· Los Angeles, Calif.-I con-
sider Mr. Arthur's paper a valuable contribu- able drawdown is very small, so that, unless
tion to the theoretical background of water and permeabilities are extremely high, the rate of
gas coning. In the application of the results of production would be so small that in some
this theoretical analysis, the following points cases the well would reach its economic limit
may need some further investigation. It before the cone has been formed, owing to the
appears from the mathematical development decline in productivity index observed in many
that the maximum height to which a water reservoirs.
cone can be drawn up without entering the These considerations apply also to the case
well is independent of the permeability of the of water and gas coning discussed in the second
rock, viscosities of the fluids and relative per- part of the paper.
meabilities of the rock after replacement of one M. G. ARTHUR (author's reply).-Since it
fluid by another. was realized that the mathematics involved in
The only factors appearing in the formulas a development of the problem are a formidable
are Slope of formation, distance of well from obstacle to many engineers, the paper pre-
edge water and difference in density of fluids. sented attempted to indicate the relative effect
This result would seem logical, since the of the various variables to those who would
"thor considers the time that the cone has accept the validity of the mathematical
ieached its maximum height ~nd is therefore development.
stationary. However, when one considers the Some California oil fields having high perme-
• Shell on Company. ability and low standard deviations of perme-
DISCUSSION 201
ability· are sufficiently homogeneous for coning In primary and secondary water drive the
to occur. However, most California fields have objective is not to prevent water encroachment
relatively low effective permeability at right but to prevent undue fingering of the ad-
angles to the bedding planes and the fingering vancing water front. :For edge water advancing
type of water encroachment is commoner than at a relatively constant rate in a homogeneous
coning. sand, the relationships presented in the paper
In fields where there is evidence that coning will determine the lead that a finger or cone
may occur, intervals of low permeability such can have on the advancing front. If the general
as shale lenses may be located by electric logs encroachment is so rapid that stable conditions
or by cores, and if such a break is at a con- cannot be attained, fingering and coning will
venient point the well may be plugged to this be of less magnitude than shown by the paper.
point before the liner is run, or cemented blank The effect of rate of production on efficiency
liner may be perforated above this point. If of replacement or washing action is a dynamic
vertical permeability is such that coning can problem of a nature different from coning and
occur, little oil will be lost ultimately because fingering. It must be attacked from an entirely
the level of the bottom water probably will different approach, which would involve many
rise, flushing oil out of the sand below the more variables than those considered in this
shale lens at which the well was bottomed. paper.
If the water table fails to rise, the interval The solution of the problem of preventing
below the shale lens may be perforated as a unreasonably low drawdowns without water
salvage operation late in the life of the field, coning is to stay some distance away from the
at which time drawdowns will be low and static edge-water position. The allowable pres-
adequate dehydration facilities probably will sure drawdowns increase exponentially with
have been constructed to handle the production. increasing distance of the well from static
Where water encroaches the oil reservoir edge-water position. In the numerical example
with depletion, it is unavoidable that produc- of water coning presented, uneconomical rates
tion rates without coning will eventually be- of production would be required in order to
come uneconomic. The solution would be to prevent water coning. The alternatives would
accept the production of water with the oil or be to accept production of water with the oil
to prevent it by plugging back the wells. It is or to plug back the well to a greater distance
hoped that consideration of the relationships from the bottom water.
presented in the paper will aid in determining
In regard to well density, the spacing is not
the extent of these operations.
a good measure of the drainage radius. Drain-
• J. Law: A Statistical Approach to Inter- age radius very often is greater than the dis-
stitial Heterogeneity of Sand Reservoirs. This
volume, page 202. tance between wells.