SPE

SPE 22931

Water Coning Calculations for Vertical and Horizontal Wells

Weiping Yang and A.A. Wattenbarger, * Texas A&M U.
'SPE Member

Copyright t 991. Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the 66th Annual Technical Conference and Exhibition of the Society of Petroleum Engineers held in Dallas. TX, October 6-9, 1991.
This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the authDr(s). Contents of the paper,
as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society
of Petroleum Engineers. Permission to copy is restricted to an abstract of not more than 300 words. illustrations may not be copied. The abstract should contain conspicuous acknowledgment
of where and by whom the paper is presented. Write Publications Manager, SPE, P.O. Box 833836, Richardson, TX 75083-3836 U.S.A. Telex, 730989 SPEDAL.

ABSTRACT

INTRODUCTION

Most authors have concentrated on correlations for critical rate

and breakthrough time in vertical and horizontal wells. WOR (wateroil ratio) has also been addressed in vertical wells. However, WOR
perfonnance in horizontal wells has not received much attention. The
purpose of this work is to develop a method suitable for either hand
calculation or simulation to predict (1) critical rate, (2) breakthrough
time, and (3) WOR after breakthrough in both vertical and horizontal
wells.

Many wells produce from oil zones underlain by "bottom

water". When the well is produced, water moves up toward the
wellbore in a cone shape. At certain conditions, water breaks through
into the well and concurrent oil and water production begins. This
phenomena is referred to as water coning.

An extensive sensitivity analysis of water coning was perfonned using numerical simulation. From this analysis, an empirical
coning correlation was developed based on the basic flow equations and
regression analysis. The fonnat of the correlation is similar to
Addington's gas-coning correlation. It predicts critical rate, breakthrough time and WOR after breakthrough.

WOR perfonnance at variable rate production conditions has

also been evaluated in this work. It was found that WOR has hysteresis, (i.e., WOR not only depends on the current production rate, but
also the previous production history). However, given sufficient time
after rate changes, hysteresis disappears. At such conditions, the
correlations can also give a good estimation of WOR for variable rate
cases.
This correlation provides a hand calculation method of coning
prediction for both vertical and horizontal wells. It can also be used as
a coning function for 3-D coarse grid reservoir simulation. The
correlation was tested and found to be reliable and accurate in predicting WOR, as well as critical rate and breakthrough time when water-oil
mobility ratio is smaller than 5 or viscous forces are not dominating.
References and illustrations at end of paper.

459

Many authors have addressed the coning problem in tenns of

critical rate (the maximum production rate without producing water),
water breakthrough time, and water-oil ratio (WOR) after water
breakthrough. Many methods have been developed for predicting
these behaviors.
Critical rate is probably the topic which has been discussed
the most. Since the first paper from Muskat and Wyckoff in 1935,
a number of correlations was developed for predicting critical rate. In
general, these correlations can be divided into two categories.
The first category detennines critical rate analytically based
on the equilibrium conditions of viscous forces and gravity forces. It
started by developing an oil potential function and then solved for the
critical rate by letting viscous forces equal the gravity forces.
However, the methods of calculating oil potential are various. For
example, Muskat and Wyckoff solved a Laplace equation for single
phase flow, while Chaney et a1. 4 and Chierici et a1. 5 used potentiometric models. Wheatly's15 method also falls into this category, but, he
took into account the influence of cone shape on the oil potential,
which others had not done before. Chaperonl6 and Giger l ? extended
this method to horizontal wells.
The second category is empirical correlations. Schols 8
developed a correlation from his lab experiment, while Hayland et
al. 19 developed their correlation from computer simulation runs.

WATER CONING CALCULATIONS FOR VERTICAL AND HORIZONTAL WELLS

Addington l2 also discussed critical rate calculation. However,

his concept of critical rate was different from others. Addington was
solving a closed outer boundary problem that never reaches steady-state
conditions, while others were dealing with open outer boundary
problems at steady-state conditions. Furthermore, Addington's critical
rate is decreasing with time or cumulative oil production, while others
had a constant critical rate.

SPE 22931

against the average oil column height above perforations after gas
breakthrough on a semi-log scale.
Based on this observation, Addington performed an extensive
parameter sensitivity analysis, from which slope and intercept of the
straight line was correlated with various reservoir and fluid properties
affecting coning performance. From this correlation, not only the
GOR can be predicted, but also the critical rate can be calculated.

Methods are also available for predicting water breakthrough

time. Sobocinski and Cornelius6 , based on their experimental and
computer simulation results, developed a dimensionless plot which
traces the rise of cone apex from its build-up to breakthrough. Cone
breakthrough time and critical rate can be determined from the plot.
Bournazel and Jeanson7 evaluated this plot and developed a simple
analytical expression to fit the plot. Papatzacos et a1. 20 investigated the
cone breakthrough time in horizontal wells. Both single-cone and
simultaneous two-cone cases were considered. The solutions were
derived by a moving boundary method with constant pressure or gravity
equilibrium assumed on the moving boundary. Their solution only
applies for infmite acting reservoirs.
WOR after breakthrough in vertical wells was also addressed
by some authors. Bournazel and J eanson7 presented a method assuming
that water is separated from oil, the oil-water interface rises and stays
at some point of perforation interval. By calculating the length of the
perforation interval in the water, WOR can be predicted.
Byrne and Morse9 , Mungan10 , Blades and Strightll investigated the effects of various reservoir and well parameters on WOR
performance using numerical simulation. However, they had not come
up with a general predictive method.
Addington12 developed a set of gas-coning correlations for 3-D
coarse grid simulation. The correlation can be used to predict critical
coning rate and gas-oil ratio (GOR) after coning has been achieved.
Even though the correlations were developed for specific data in Prudoe
Bay field, the technique can be of use in water-coning evaluation.
Kuo and DesBrisay13 investigated the sensitivity of water
coning performance to various reservoir parameters using numerical
simulation. A correlation of predicting water cut performance was
developed from the sensitivity analysis.
Kabir14 studied water coning into gas wells using simulation.
The sensitivity of reservoir and fluid properties on water-gas ratio was
discussed.

We followed the same procedure as Addington did and

developed a water-coning correlation. As the first step of developing
the coning correlation, a one well model was simulated at a constant
total production rate.
The one well model was run on a
two-dimensional simulator. For vertical wells, a r-z radial model was
used and for a horizontal well, a 2-D x-z model was used. The well
was simulated with a wide range of properties.
Fig. 1 shows a sketch of a reservoir with a bottom aquifer
and a well perforated above the aquifer. As production begins, water
cones up toward the wellbore. If assuming that water is displacing oil
in a piston-like manner, then an imaginary current water-oil contact
can be defmed. Fig. 1 shows this contact by a dashed line. The oil
column height between the current contact and the bottom of the
perforation is defmed as the average oil column height below perforation, denoted by hbp . It can be calculated by writing a material
balance equation. The calculation is discussed in the Appendix.
As production proceeds, hbp decreases. At some point of
time, water breaks into the wellbore, the average oil column height
below perforation at this time is termed average oil column height
below perforation at breakthrough, denoted by hwb . After water
breaks into the well, WOR increases as t"p decreases.
After simulating a one well model at different properties for
both vertical and horizontal wells, we found that Addington's correlation form, with a slight modification, applies to water coning. That is,
the plot of WOR plus a constant, c, as a function of t"p is a straight
line after water breakthrough on a semi-log scale, as shown by Fig. 2.
The straight line relationship can be described mathematically as:
WOR
Log(WOR +c)

=0
= m(h bp -h wb) +Log(c)

hbp >h wb . . . . . . (1)

hbp S; hWb

c is a constant, depending on whether it is a vertical or a horizontal

well. Therefore, if the breakthrough height hwb' slope of the straight
line m and constant c can be determined, then, the whole process of
coning can be predicted.

This paper presents a water-coning correlation to predict

critical rate, water breakthrough time and WOR after breakthrough for
both vertical and horizontal wells. The correlation was developed
following the same procedure as Addington's. It can be used either as
a hand calculation method or a coning function for 3-D coarse grid
reservoir simulation.
APPROACH TO THE PROBLEM
Addington12 , when studying gas coning into an oil well,
observed that a straight line results when gas-oil ratio (GOR) is plotted

460

As we have mentioned in the Appendix, for a tank reservoir,

is linearly related to the cumulative oil production Np' the WOR
c vs. t"p plot can be easily converted to a WOR + c vs. Np plot.

~p

The method of determining hwb' m and c was developed from a

stepwise procedure. First, a number of simulation runs was made to
investigate the coning performance at different reservoir and fluid
properties both for vertical and horizontal wells. Then, for each
simulation run, WOR + c was plotted against t"p on a semi-log scale,
from which m and hwb were determined. Once the hwb and m data

SPE 22931

WEIPING YANG AND R.A. WATIENBARGER

was obtained for all the simulation runs, regression analysis was then
used to defme the relationship between m, hWb and various reservoir
and fluid properties.

1 + 39.0633 X 10-4
. ... (4)

We followed this procedure and developed a coning correlation

for both vertical and horizontal wells, respectively, the results will be
discussed in the following sections.

(5)

VERTICAL WELLS

[~]

The water-coning performance at different reservoir and fluid

properties was investigated using a 2-D r-z numerical simulator. Fig.
3 sketches the reservoir geometry, grid size and boundary conditions.
Following assumptions were made during the simulation:
1.
2.
3.
4.
5.
6.

m =0.015 [1 +485.7757 [_1_] 0.5

0.5
1
(1-45)(1-)..)]
rne
qn
1+M o.03
h1.7

No flow across the outer boundary.

Formation is underlain by a recharged bottom aquifer.
Only one perforation interval.
Reservoir is homogeneous but anisotropic.
Only water and oil are present at reservoir conditions.
Capillary pressure can be ignored.

Parameter Sensitivity Analysis

The parameter sensitivity analysis was made to provide data for
developing a predictive correlation of calculating breakthrough height
hwb and slope m.

(6)

To begin the parameter sensitivity analysis, a base case was set

up first and all the simulation runs were conducted by varying base case
data. Eleven parameters were varied to establish the 48 simulation
cases. The relative permeability data is tabulated in Table 1. The input
data for base case and all other runs are summarized in Table 2.

The parameters were grouped together based on the basic flow

equations and the grouping was confirmed by regression analysis. Eq.
4 guarantees that hwb can never go beyond h - ~ - hap.

HORIZONTAL WELLS

From these simulation runs, it was found that the constant, c,

for vertical wells is 0.02. Therefore, the WOR changes can be
described by the following equation:

The same procedure of developing correlations for vertical

wells was followed here for horizontal wells. First, the WOR
behavior at different reservoir and fluid properties was investigated by
numerical simulation, then the breakthrough height hwb and slope m
were determined, fmally, the regression analysis was used to correlate
hwb and m with various reservoir and fluid properties.

WOR =0
hbp > hwb
(2)
Log(WOR+0.02) =m(hbp-h wb) + Log(0.02) h bp S h wb . . . . .
The WOR from each simulation run was least square fitted by the above
equation, from which the height hWb and slope m was determined. The
last two columns in Table 2 list the m and hwb for each run.

Generalized Correlations
Parameter sensitivity analysis shows that height hWb and slope
m are functions of the various reservoir and fluid properties. These
functions were defmed using the regression analysis.
As Table 2 shows, hwb increases with production rate qt and
oil viscosity, etc. However, the increase of hWb is limited by a natural
constraint:
hWb S h - hp - hap. . . . . . . . . . . . . . . . . . .. (3)
With this in mind, we came up with the following results:

461

A 2-D x-z model was used in the simulation. Fig. 4 sketches

the reservoir geometry, grid and boundary conditions. In addition to
the assumptions made for vertical wells, it was further assumed that
the horizontal well is long and fully penetrated so that a 2-D x-z
geometry can be used.

Parameter Sensitivity Analysis

The sensitivity of various reservoir and fluid properties on the
coning behavior in horizontal wells was investigated extensively by
varying the base case data. Eleven parameters were varied and
evaluated by 47 simulation runs. The input data for base case and all
these runs are summarized in Table 3. The relative permeability curve
is the same as in vertical weU

WATER CONING CALCULATIONS FOR VERTICAL AND HORIZONTAL WELLS

From the sensitivity analysis, it was found that the best way of
presenting WOR data is to plot WOR + 0.25 as a function of average
oil column height below perforation hbp ' The resulted plot is a straight
line on a semi-log scale, which can be described mathematically by the
following equation:

The WOR results from each simulation run was curve fitted by the
above equation, from which the breakthrough height hWb and slope m
were determined. The last two columns of Table 3 list the hwb and m
for each run.

Generalized Correlations
Parameter sensitivity analysis shows that the breakthrough
height hWb and slope m are functions of reservoir and fluid properties.
The height hwb increases with production rate, oil viscosity, etc.
However, the same argument for hwb in vertical wells still applies here,
that is, the increase in hwb is limited by a natural constraint:
hWb

h - hap' . . . . . . . . . . . . . . . . . . . . . . . (8)

SPE 22931

HOW TO CALCULATE CRITICAL RATE

The correlation for hwb can be used as a critical rate correlation. Assuming that a well is produced at a rate of qt> then, right at the
height hwb' water breaks into the well. To see this process other way
around, assuming that the height is at hwb, then, if the production rate
is above 'It, the well produces water; if rate is below 'It, the well does
not produce water. Therefore, the rate solved from Eq. 4 or 10 is
actually the critical rate at height hwb'
To demonstrate that this is the case, we made five simulation
runs, the input data for these runs are the same as in base case except
production rate. Fig. 5 shows the five production schedules and the
corresponding WOR performance. Schedule A, Band C have a
constant production rate of 1000, 2500 and 4500 RBID, respectively.
Schedule D and E have a variable rate which starts at 2500 RB/D,
then, when hbp drops to 65.12, production rate is increased to 4500
RB/D in schedule E; decreased to 1000 RB/D in schedule D. The
figure depicts that at the height of 65.12, when rate is higher than
2500 RB/D, the well is coning water; when rate is below 2500 RB/D,
the well is not coning water. Therefore, the critical rate at the height
of 65.12 is 2500 RB/D. Of course, at different height, the critical
coning rate is different, which can be solved from Eq. 4:
(13)

thus:
h-h ap
hWb

(9)

With this in mind, we came up with the following results:

h-h ap ]2=1 +
[ hWb

4.7921X1O-4X~.32[_1 ]0.65 [..:...] [_1_]

(10)
4

Xa0.18 [
m=0.004 1+2.7496
[
h

xD

X~

04 [ ]0.5
q~

qD

1 +MO.

](11)

(I+Mo. 25 )(I-A)0.3

k k' h2~'Y
h co

qeD' . . . . . . . . . . . . . . . . . . .. (14)

JL o

for vertical wells.

Similarly, for horizontal wells, critical rate can be solved from Eq.
10 as:
=

4.7921X1O-4x~.32 [_1_] 0.65

xD

M k:oLh~'Y q
JLo

eD

h~p

1
I+M.4 (h-h ap)2_h2bp

(15)

. . . . . . . . . . . . . . . . . . (16)

These equations show that critical rate is decreasing with

height ~P' thus, critical rate is decreasing with time or cumulative oil
production.
(12)

Again, the parameters were grouped together based on the basic flow
equations and the grouping was confirmed by regression analysis.

462

The critical rate calculated in this manner is different from the

rate calculated from the classic steady-state methods. The reasons are
that classic methods are associated with the open outer boundaries
under steady-state conditions. And the critical rate is the rate below
which there is no water production at any time. This method is for a
closed boundary problem, which never reaches steady-state conditions.
Critical rate is the rate below which there is no water production at a
particular time.

SPE 22931

WEIPING YANG AND R.A. WATIENBARGER

HOW TO CALCULATE BREAKTHROUGH TIME

DISCUSSION OF RESULTS

Fo.r a tank reservo.ir, the average o.il co.lumn height belo.w perfo.ratio.n hbp is linearly related to the cumulative o.il pro.ductio.n Np. Then,
the cumulative o.il pro.ductio.n at breakthro.ugh can be calculated fro.m
the breakthro.ugh height hwb:

The co.rrelatio.n can also. be used to' predict WOR fo.r variable
rate cases. The predictio.n is based o.n the assumptio.n that WOR has
no. hysteresis, i.e., WOR is o.nlya functio.n o.f current height hbp and
current pro.ductio.n rate, previo.us pro.ductio.n histo.ry has no. influence
o.n the current WOR. Under such an assumptio.n, the co.rrelatio.ns are
valid fo.r variable rate case, o.nly hwb and slo.pe m have to. be recalculated each time when rate changes.

hWb

=h

(Np)bt
Atp(1-s wc -sor)

- hap - h p ' . . . . . . . .. (17)

A sample calculatio.n fo.r a vertical well is sho.wn in Fig. 10,

where so.lid line represents the WOR calculated fro.m co.rrelatio.n while
circle represents simulatio.n WOR. The pro.ductio.n rate starts at 2500
RBID, decreased to. 1000 RBID at height o.f 42 ft, then increased to.
4500 RBID at height o.f 21.6 ft.

so.lve fo.r (Np)bt' we have:

the breakthrough time can be predicted as:

tbt

(Np)bt . . . . . . . . . . . . . . . . . . . . . .
qt

(19)

This pro.cedure applies to. bo.th vertical and ho.rizo.ntal wells.

The breakthro.ugh time calculated fro.m this co.rrelatio.n is co.mpared
with o.ther metho.ds and simulatio.n results. Fig. 6 sho.ws the co.mpariso.n fo.r a vertical well, in which co.rrelatio.n breakthro.ugh time is
co.mpared with So.bo.cinski's metho.d and simulatio.n results, (the
breakthrough time fro.m simulatio.n was taken as the time when water
cut equals 0.01). Fo.r this case, the co.rrelatio.n gives a very go.o.d
appro.ximatio.n to. the simulatio.n results. But, So.bo.cinski's metho.d is
o.bvio.usly to.o. high. The reaso.n eQuId be that So.bo.cinski's co.rrelatio.n
is o.nly fo.r o.pen bo.undary problems.
Fig. 7 sho.ws the co.mpariso.n fo.r a ho.rizo.ntal well, where co.rrelatio.n is co.mpared with Papatzaco.s's metho.d and simulatio.n. Again,o.ur
co.rrelatio.n result matches the simulatio.n result. Ho.wever, Papatzaco.s's
breakthro.ugh time is to.o. high, the reaso.n co.uld be that his metho.d o.nly
applies to' infinite acting reservo.irs.

HOW TO CALCULATE WOR AFTER WATER BREAKTHROUGH

To. fmd the WOR at height h bp fo.r a given pro.ductio.n rate, first,
calculate the breakthro.ugh height hwb fro.m Eq. 4 o.r 10 and slo.pe m
fro.m Eq. 5 o.r 11, then use the fo.llo.wing equatio.n to' fmd WOR fo.r a
vertical well:
WOR = 0
h bp > hWb (20)
Lo.g(WOR+0.02) = m (hbp-hwb ) + Lo.g(0.02) hbp:S;;h wb
and use the fo.llo.wing fo.r a ho.rizo.ntal well:

A sample calculatio.n fo.r a co.nstant rate case was made fo.r a

vertical and ho.rizo.ntal well respectively. The results were co.mpared
with the simulatio.n results. The co.mpariso.ns are sho.wn in Figs. 8 and
9. The figures sho.w that co.rrelatio.n gives a go.o.d match to' the
simulatio.n results.

463

A similar sample calculatio.n was made fo.r a ho.rizo.ntal well.

Fig. 11 sho.ws the co.mpariso.n o.f co.rrelatio.n with simulatio.n results.
Again, so.lid line represents the WOR calculated fro.m co.rrelatio.n while
circle represents simulatio.n WOR. The pro.ductio.n rate starts at 2500
RBID, decreased to. 1000 RBID at height o.f 30.5 ft, then increased to.
4500 RBID at height o.f 12.3 ft.
The figures sho.w that every time when rate is changed,
co.rrelatio.n predicts a mo.re abrupt jump o.f WOR. Ho.wever, as time
go.es o.n after rate changes, co.rrelatio.n WOR gradually appro.aches
simulatio.n WOR. This trend is o.bserved in bo.th figures. The
deviatio.n o.f co.rrelatio.n fro.m simulatio.n WOR is the result o.f
hysteresis assumptio.n. Right after rate changes, previo.us pro.ductio.n
rate is still playing its ro.le, the WOR deviatio.n is mo.st severe, WOR
has hysteresis. But, given sufficient time after rate changes, the
influence fro.m previo.us pro.ductio.n histo.ry is diminishing, and
co.rrelatio.n WOR is appro.aching simulatio.n WOR, which implies that
WOR hysteresis disappears.
WOR hysteresis can also. be seen fro.m Fig. 5. After pro.ductio.n rate is decreased to. 1000 RBID in schedule D, WOR do.es no.t
fo.llo.w schedule A curve, indicating that pro.ductio.n histo.ry befo.re rate
change do.es have so.me influence o.n the WOR after rate change, i.e.,
WOR has hysteresis. Ho.wever, WOR difference between two.
schedules is really small, hysteresis is no.t severe here. The same
trend can also. be o.bserved by co.mparing schedule C and schedule E.
Since rate o.nly changes o.nce in schedule D and E, hysteresis is no.t
very important, co.nsequently, co.rrelatio.n can give a go.o.d appro.ximatio.n fo.r such cases.

CONCLUSIONS
This paper presents a water co.ning co.rrelatio.n to. predict
critical rate, breakthro.ugh time and WOR after breakthro.ugh fo.r bo.th
vertical and ho.rizo.ntal wells. The co.rrelatio.n was develo.ped based o.n
the basic flo.w equatio.ns and regressio.n analysis using the data fro.m
numerical simulatio.ns. The fo.rmat o.f the co.rrelatio.n is similar to.
Addingto.n's gas co.ning co.rrelatio.n and it can be used in a similar
way, i.e., either as a hand calculatio.n metho.d o.r a co.ning functio.n fo.r
a 3-D co.arse grid simulatio.n. Fro.m o.ur experience, the co.rrelatio.n
can give meaningful approximatio.n when water-oil mo.bility ratio. is
smaller than 5 o.r visco.us fo.rces are no.t do.minating. The accuracy

WATER CONING CALCULATIONS FOR VERTICAL AND HORIZONTAL WELLS

may become less for values outside this range. With this in mind and
recalling other assumptions made, we draw the following conclusions:
1.

As water cone moves up, critical rate gradually decreases.

Eqs. 14 and Eq. 16 predict this critical rate for vertical and
horizontal wells, respectively.

2.

For a tank reservoir, the ilwb correlation, Eqs. 4 and 10 can be

used to calculate water breakthrough time for vertical and
horizontal wells, respectively. The calculation procedure is
described by Eq. 19.

3.

For constant rate cases, WOR after breakthrough can be

predicted from Eq. 20 or Eq. 21 by calculating ilwb and m from
Eqs. 4 and 5 or Eqs. 10 and 11.

4.

This study found that WOR has hysteresis. That is, previous
rates or rate changes do have some effects on the current WOR.
But, given sufficient time, these effects disappear.

5.

If rate does not change very frequently, that is, there is enough
time for hysteresis to disappear, the method can be used to
predict WOR for variable rate cases. The prediction is only
approximate since it is based on the non-hysteresis assumption.
The approximation is more accurate at times long after the rate
changes occur.

NOMENCLATURE
A
Bo
h
hap
hbp
ho

cross sectional area, ft2

oil formation volume factor, stb/rb
initial oil formation thickness, ft
oil column height above perforations, ft
average oil column height below perforation, ft
current oil zone thickness, ft
~
perforation length, ft
ht
total formation thickness, ft
hw
current water zone thickness, ft
hWb
breakthrough height, ft
kh
horizontal permeability, md
ley
vertical permeability, md
ko
oil effective permeability, md
k ro '
oil relative permeability at Swe
~,
water relative permeability at 1-Sor
L
horizontal well length, ft
LOG LOG of base 10
m
slope
M
water oil mobility ratio
Np
cumulative oil production, stb
p
pressure, psi
PI
parameter groups
P2
parameter groups
~
critical coning rate, stb/D
qD
dimensionless production rate
qeD
dimen~ionless critical coning rate
qt
total fluid production rate, RB/D
rw
wellbore radius, ft
rDe
dimensionless drainage radius

SPE 22931

drainage radius, ft
connate water saturation
residual oil saturation
Sor
time, days
t
breakthrough time, days
~t
dimensionless time
tD
dimensionless breakthrough time
tDBT
WC water cut
WOR water-oil ratio
drainage width; ft
X.
dimensionless drainage width
xD
oil viscosity, cp
Jl.o
water
viscosity, cp
JI.w
oil gravity, psi/ft
'Yo
water gravity, psi/ft
'Yw
q,
porosity, fraction
water-oil gravity difference, psi/ft
Il:y
fraction of perforated interval
o
fraction of oil column height above perforation
A
re
Swe

REFERENCES
1. Muskat, M. and Wyckoff, R.D.: "An Approximate Theory of
Water Coning in Oil Production," Trans. AlME (1935), 114, 144161.
2. Buckley, S.E. and Leverett, M.C.: "Mechanisms of Fluid Displacement in Sands," Trans. AlME (1942) 146, 107-116.
3. Meyer, H.I., and Garder, A.O.: "Mechanics of Two Immiscible
Fluids in Porous Media," Journal of Applied Physics, November
1954, Vol. 25, No. 11, p. 1400.
4. Chaney, P.E., Noble, M.D., Henson, W.L., and Rice, T.D.:
"How to Perforate Your Well to Prevent Water and Gas Coning,"
Oil &: Gas Journal, May 7, 1956, p. 108.
5. Chierici, G.L., Ciucci, G.M., and Pizzi, G.: "A Systematic Study
of Gas and Water Coning By Potentionmetric Models," JPT,
August 1964, pp.923-29.
6. Sobocinski, D.P., and Cornelius, A.I.: "A Correlation for
Predicting Water Coning time," JPT, May 1965, pp.594-600.
7. Bournazel, C. and Jeanson, B.: "Fast Water Coning Evaluation,"
Paper APE 3628 presented at the SPE 46th Annual Fall Meeting,
New Orleans, October 3-6, 1971.
8. Schols, R.S.: "An Empirical Formula for the. Critical Oil Production Rate," Erdoel Erdgas, Z., January 1972, Vol. 88, No.1, pp.
6-11.
9. Byrne, W.B. and Morse, R.A., "The Effects of Various Reservoir
and Well Parameters on Water Coning Performance," paper SPE
4287 presented at the SPE 3rd Numerical Simulation of Reservoir
Simulation of Reservoir Performance Symposium, Houston,
January 10-12, 1973.

464

SPE 22931

WEIPING YANG AND R.A. WAITENBARGER

10. Mungan, N.: "A Theoretical and Experimental Coning Study,"

Soc. Pet. Eng. J. (Iune 1975) 221-236.
11.

Addington, D.V.: "An Approach to Gas-Coning Correlations for

a Large Grid Cell Reservoir Simulator," JPT (November 1981)
2267-74.

13.

Kuo, M.C.T., and DesBrisay, C.L.: "A Simplified Method for

Water Coning Predictions," Paper SPE 12067, SPE 58th Annual
Fall Meeting, San Francisco, October 5-8, 1983.

14. Kabir, C.S.: "Predicting Gas Well Performance Coning Water in

Bottom-Water-Drive Reservoirs," SPE Paper 12068, presented at
the 58th Annual Fall Meeting, San Francisco, October 5-8, 1983.
15. Wheatly, M.I., "An Approximate Theory of Oil Water Coning,"
SPE Paper 14210, SPE 60th Annual Fall Meeting, Las Vegas,
NV, September 22-25, 1985.
16.

17.

With these assumptions, the oil material balance equation can be

written as:

Blades, D.N. and Stright, D.H., Ir., "Predicting High Volume

Lift Performance in Wells Coning Water," J. Can. Pet. Tech.
(October-December 1975) 62-70.

12.

Chaperon, I.: "Theoretical Study of Coning Toward Horizontal

and Vertical Wells in Anisotropic Formations: Sub critical and
Critical Rates," SPE Paper 15377, SPE 61st Annual Fall Meeting,
New Orleans, LA, October 5-8, 1986.

hts o

h~~so

21.

(hCh) 0.0 + (h-li)(I-s we) + Iis or . . . . . . (A-I)

= (h -li)A~(1-swc) + Iiso~~

. . . . . . . . . . (A-2)

the left-hand side equals the oil left in the reservoir, it should equal the
original oil in place minus the cumulative oil production Np ;

substitute this equation into Eq. (A-2), we have:

Solve for

ii, we have:

Ii

A~(1

NpB
. . . . . . . . . . . . . . . . . . . . (A-5)
swe sor)

And
hbp

19. Heyland, L.A., Papatzacos, P., Skjaeveland, S.M.: "Critical Rate

for Water Coning: Correlation and Analytical Solution," SPE
Reservoir Engineering, November 1989.
20.

multiplying both sides by the cross-sectional area A and the porosity,

we have:

Giger, F.M.: "Analytical 2-D Models of Water Cresting Before

Breakthrough for Horizontal Wells," SPE Paper 15378, SPE 61st
Annual Fall Meeting, New Orleans, LA, October 5-8, 1986.

18. Piper, L.D., Gonzalez, L.M.: "Calculation of the Critical Oil

Production Rate and Optimum Completion Interval," SPE paper
16206, presented at the SPE Production Operations Symposium
held in Oklahoma, March, 87-10, 1987.

= h -Ii -hap -h . . . . . . . . . . . . . . . . . . ..
p

TABLE 1. Relative permeability data

Papatzacos, P., Herring, T.R., Martinsen, R., Skjaeveland,

S.M.: "Cone Breakthrough Time for Horizontal Wells," Paper
SPE 19822, SPE 64th Annual Fall Meeting, San Antonio, TX,
October 8-11, 1989.

!w
0.1500
0.2000
0.2500
0.3000
0.3500
0.4000
0.4500
0.5000
0.5500
0.6000
0.6500
0.7000
0.7500
0.7750
0.7880
0.8000
1.000

Yang, W.: "Water Coning Calculations for Vertical and

Horizontal Wells," MS thesis, Texas A&M University, August
1990.

APPENDIX
For a tank reservoir, there is no flow across the outer boundary.
The height hbp is uniquely related to the cumulative oil production. The
relationship can be derived from a material balance equation. As shown
by Fig. 1, three regions have to be included when writing a material
balance equation, the aquifer, water invaded region and the oil column
between top of the reservoir and current water oil contact. In the
aquifer, it is assumed that oil saturation is zero, the region between
initial water oil contact and the current water-oil contact is defined as
the water invaded region, in which oil saturation equals the residual oil
saturation. In the region above the current water-oil contact, it was
assumed that oil saturation is still at its initial level 1 - !we.

465

O.OOOOE+OO
4.0000E-03
1.0200E-02
1. 6600E-02
2.3200E-02
3.0500E-02
3.9200E-02
4.9700E-02
6.3000E-02
7.9800E-02
0.1000
0.1244
0.1525
0.1698
0.1784
0.1870
0.1870

~o

0.9500
0.7500
0.5876
0.4462
0.3325
0.2450
0.1770
0.1200
7. 2400E-02
3.7400E-02
1.6300E-02
5.6400E-03
7.7000E-04
3.8000E-04
1.9000E-04
O.OOOOE+OO
O.OOOOE+OO

(A-6)

SPE 22931

WATER CONING CALCULATIONS FOR VERTICAL AND HORIZONTAL WELLS

TABLE 2. Simulation inQut data and results - vertical wells

case

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

18
19
20
21
22
23
24
25
26
27
28
29
30

Ich

lev

r.

h"p

hp

4000
2000
3000
4000
6000

200

1300

160

3.75

16.25

iLo

1.5

iLw

fl:y

0.31

0.0996 0.207 2500

f/J

50
100
200
400
800
1000
1300
1600
1800
100
160
200
260
3.75
13.75
23.75
43.75
8.75
16.25
26.25
36.25
0.5
1.5
3.0
4.0
0.20
J.31
0.40
0.50
0.70

31
32
33
34
35

0.0779
0.0893
0.1102
0.1198

36
37
38
49

0.1
0.207
0.30
0.40

40
41
42
43

1000
1500
3500
4500

44
45
46
47

hWb

-0.0366
-0.0271
-0.0323
-0.0366
-0.0432

65.12
92.84
75.26
65.12
53.34

-0.0445
-0.0394
-0.0366
-0.0329
-0.0298

58.90
61.50
65.12
70.88
77.40

-0.0379
-0.0366
-0.0351
-0.0340

63.24
65.12
68.34
70.17

-0.0381
-0.0366
-0.0364
-0.0361

60.34
65.12
68.10
71.71

-0.0366
-0.0339
-0.0324
-0.0319

65.12
62.16
60.02
55.46

-0.0375
-0.0366
-0.0342
-0.0329

71.20
65.12
60.00
54.61

-0.0460
-0.0366
-0.0294
-0.0271

36.05
65.12
92.61
105.13

-0.0364
-0.0366
-0.0364
-0.0366
-0.0366

69.13
65.12
63.68
61.68
58.68

-0.0338
-0.0354
-0.0377
-0.0386

71.76
68.47
62.63
60.42

-0.0362
-0.0366
-0.0366
-0.0367

66.38
65.12
65.53
65.37

-0.0481
-0.0429
-0.0329
-0.0304

43.40
52.34
74.58
81.67

Note: a blank entry in the table indicates that the parameter has the same value as base case or case 1.

466

SPE 22931

WEIPING YANG AND R.A. WATIENBARGER

TABLE 3. Simulation

case

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30

kh

xa

4000
1000
2000
3000
6000

200

1151.5

160

hap

20

in~ut

data and results - horizontal wens

/L o

/L w

A-y

2303

1.5

0.31

0.0996

tP

0.207 2500

50
100
200
400
800
600
800
1300
1500
100
200
260
300
1
10
40
60
1200
1600
2600
3000
0.5
3.0
4.0
5.0

31
32
33
34
35

0.0779
0.0893
0.1102
0.1198

40
41
42
43

0.1
0.30
0.40
0.45

44
45
46
47

1000
1500
3500
4500

-0.0392
-0.0229
-0.0303
-0.0353
-0.0452

36.04
66.23
48.02
40.40
31.10

-0.0378
-0.0366
-0.0392
-0.0406
-0.0419

42.98
39.75
36.04
33.45
30.73

-0.0441
-0.0424
-0.0382
-0.0364

33.60
33.74
37.23
39.07

-0.0406
-0.0390
-0.0387
-0.0383

31.79
38.42
41.82
44.13

-0.0464
-0.0414
-0.0386
-0.0377

46.19
40.29
30.73
28.01

-0.0290
-0.0331
-0.0419
-0.0447

49.87
43.20
33.58
31.22

-0.0417
-0.0392
-0.0381
-0.0376
-0.0364

37.30
36.04
35.04
33.98
32.51

-0.0348
-0.0370
-0.0411
-0.0428

40.86
38.27
34.21
32.73

-0.0392
-0.0391
-0.0407
-0.0409

36.15
35.83
35.48
35.37

-0.0641 20.83
-0.0503 27.34
-0.0335 42.70
-0.0298 48.37

Note: a blank entry in the table indicates that the parameter has the same value as base case or case 1.

467

hWb

-0.0489 17.46
-0.0300 54.31
-0.0272 63.61
-0.0253 71.29
0.20
0.31
0.40
0.50
0.70

36
37
38
39

16

WATER CONING CALCULATIONS FOR VERTICAL AND HORIZONTAL WELLS

10

SPE 22931

WaR + c

hap
1 - swc

hp

I
I

hbp h
ht

0.1
h

sor

initial

wac

V V V V 'if
Fig. I-A sketch of well configurations for calculating 0.01 L-_----L_ _.L..-_--L_ _.l-.-_--L._ _..l...-_--'-_----.J
20
40
60
80
100
120
140
160
average oil column height below perforations
0
hbp (ft)

Fig. 2-WOR

+ c vs.

hbp plot from a simulation run

~ ~~~
.

/'/

...
HH~!Gl!

InH al "

VV

I\..

I~ ~
~

I ~

Fig. 3-Simulation grid for a vertical well

Fig. 4-Simulation grid for a horizontal well

468

WEIPING YANG AND R.A. WATTENBARGER

SPE 22931

10

W
.. ~O~R_+~0~.0~2~____________________________~

11

Breakthrough Time (days)

3000r-------~----~~~----------------_.

Schedule C

-+-

-A- Correlation

Schedule B

-*""
o

Schedule D

Schedule E

2500

Schedule A

2000

1000

1000
2500
4500
1000
4500

4500

D 2500

2500

Simulation

kh

.. 4000 md

Ie.,

= 200 md
= 1300 ft
= 100 ft

hop =Oft
hp = 20ft
II. = 1.5 cp
IIw = 0.31
.y = 0.0996 psi/ft
.. 0.207
~

1500

B 2500

Sobocinski

-e-

'.h

P,oduction Schedule

1000

500

0.01

L_...L.-_....l.-_---L_--'-_---ll-_-'--_...L----'

20

40

60

80
100
hbp (ft)

120

140

oL-_---L_ _L - _ - L_ _L -_ _- L_ _~_ _~

160

10

Production Rate (x1000 BBLs/D)

Fig. 5-Critical rate analysis

WOR at different production schedules

Fig. 6-Vertical well breakthrough time comparison

between correlation, simulation and Sobocinski's
method

Breakthrough time (days)

10000~~~-~----~~~----------------

WOR + 0.02

100~=-~~--------------------------~
Correlation

Simulation

10

1000

100

.. 1000 md
- 50md
.. 1151.5 ft
h
.. 160 ft
hop -Oft
II. - 1.5 cp
IIw .. 0.31
.y .. 0.0996 psi/ft
.. 0.207
~
.. 1500 ft
L

kt.

Ie.,

kh

'.

Ie.,

5000

6000

7000

.. 1300 ft

.. 160ft

= 0 ft

= 20 ft
= 1.5 cp
= 0.31

/j,.

Papalzcous

II.
IIw

Simulation

.y

= 0.0996 psi/ft

= 0.207
.. 6000 RBID

'It

10L-~--~----~--~----~--~----~~

4000

'.

hop

0,1

Correlation

= 2000md
= 100 md

8000

9000

10000

0.0 1 L ______---L________..L--_ _ _ _ _--'-_ _ _ _ _ _

Production rate (BBLs/D)

Fig. 7-Horizontal well breakthrough time comparison

between correlation, simulation and Papatzcous'
method

0.2
0.4
0.6
Recovery (% original oil in place)

Fig. 8-WOR comparison between correlation and

simulation for a vertical well

469

0.8

12

WATER CONING CALCULATIONS FOR VERTICAL AND HORIZONTAL WELLS

1YOR + 0.02

WOR + 0.25
10r-------------------------------------i

SPE 22931

10

Simulation

Simulation

Correlation

Correlation

= 1000 md

kh

kh

k.,
r.

'" 200 md
= 1151.5 It
= 160 It
h
hOI' '" 20 It
Po '" 1.5 cp
Pw '" 0.31
.. y = 0.0996 psi/lt
'" 0.207
~
= 2303 It
l
q. = 2500 RB/D

k.,
r.

0.1L----L--~----~--~----~---L----~--~

20

40

60

80
100
hop (ft)

120

140

qt.

60

40

20

Correlation

Simulation

'"
'"
'"
h
'"
hOI' '"
Po =
P w '"
.. y
'"
~
l
'"

k.,
r.

qt.

4000md
200md
1151.51t
160 It
20 It
1.5 cp
0.31
0.0996 psi/lt
0.207
2303 It

40

60

80

h",,>30.5
h",,>l2.3
h",,<12.3

{2500
1000
4500

L -__- L____L -__-L____L -_ _- L_ _ _ _

20

h",,>42.0
h"p>21.6
h",,<21.6

80
100
hbp (ft)

120

140

160

Fig. lO-WOR hysteresis analysis for a vertical well

WOR + 0.25
10r---------------------------------------,

{2S00
1000
4500

0.01L----L--~----~---L----~--~--~--~

160

Fig. 9-WOR comparison between correlation and

simulation for a horizontal well

0.1

1300 It
h
'" 160 It
hop '" 3.75 It
h. = 16.25 It
p. '" 1.5 cp
0.31
Pw
"Y = 0.0996 psi/lt
= 0.207
~

0.1

= 4000 md
= 200md

100

120

__

__

140

160

hbp (ft)

Fig. ll-WOR hysteresis analysis for a horizontal well

470