Sunteți pe pagina 1din 11

Calculation of the Production

Rate of a Thermally Stimulated Well


T. C. BOBERG
R. B. LANTZ
JUNIOR MEMBERS AIME

ABSTRACT
This paper presents a method for calculating the producing rate of a well as a function of time following steam
stimulation. The calculations have proved valuable in both
selecting wells for stimulation alld ill determining optimum
treatment sizes.
The heat transfer model accounts for cooling of the oil
sand by both vertical alld radial conduction. Heat losses
for any number of productive sands separated by unproductive rock are calculated for the injection. shut-in and
production phases of the cycle. The oil rate increase caused
by viscosity reduction due to heating is calculated by
steady-state radial flow equations. The response of Sllccessive cycles of steam injection call also he calculated with
this method.
Excellent agreement is shown betweell calculated and
actual field results. Also included are the results of several
reservoir and process variable studies. The method is best
suited for wells producing from a multiplicity of thin sands
where the bulk of the stimulated production comes from
the unheated reservoir. The flow equations used neglect
gravity drainage and sall/ration changes within the heated
region.

INTRODUCTION
This paper presents a calculation method which can be
used to predict the field performance of the cyclic steam
stimulation process. The calculation method enables the
engineer to select reservoirs that have favorable characteristics for steam stimulation and permits him to determine how much steam must be injected to achieve favorable stimulation. While the calculation represents a considerable simplification of physical reality and the results
are subject to numerous assumptions which must be made
about the reservoir, it has been found that realistic calculations can be made of individual well performance following steam injection.
The duration of the stimulation effect will depend primarily on the rate at which the heated oil sand cools which,
in turn, is determined by the rate at which energy is removed from the formation with the produced fluids and
conducted from the heated oil sand to unproductive rock.
A complete mathematical solution to this problem is a
formidable task, and finite difference techniques would undoubtedly have to be used. The calculation method preOriginal manuscript received in Society of Petroleum Engineers office
July 8, 1966. Revised manuscript received Oct. 31, .1966. Pa~er (SPE
1578) was presented at SPE 41st Annual Fall Meetmg held m Dallas.
Tex., Oct. 2-5, 1966. Copyright 1966 American Institute of Mining,
Metallurgical, and Petroleum Engineers, Inc.
DECEMBER, 1966

ESSO PRODUCTION RESEARCH CO.


HOUSTON, TEX.

sented here utilizes analytic solutions of simple related


heat transfer and fluid flow problems. The method is sufficiently simplified that it can be used as a hand calculation, although the calculations are somewhat lengthy and
laborious. For that reason, the analysis was programmed
for an IBM 7044 digital computer.
Well responses observed at the Quiriquire field in eastern Venezuela" have been matched using this program after
making suitable approximations for reservoir and well bore
conditions. One of the most valuable uses of this calculation method is to assess the effect of reservoir and proccess variables on the stimulation response. This paper contains results of several studies made of key reservoir and
process parameters. Among the most important of these is
the influence of prior well bore permeability damage. If a
well is severely damaged prior to stimulation, a higher
stimulation response will be observed than if it is undamaged. If a portion of this damage is removed. a permanent
rate improvement will occur.
THEORY
DESCRIPTION OF CALCULATION METHOD

The process of cyclic steam stimulation is essentially one


of reducing oil viscosity around the well bore by heating
for a limited distance out into the formation through the
injection of steam. Suitable modifications of the calculation
technique presented here can be made so that stimulation of wells by hot gas injection or in situ combustion can
also be calculated.
A schematic drawing of the heat transfer and fluid flow
considerations included in the calculation method is shown
in Fig. 1. In brief, the calculation assumes that the oil sand
is uniformly and radially invaded by injected steam. For
wells producing from several sands, each sand is assumed
to be invaded to the same distance radially. In calculating
the radius heated rh , energy losses from the wellbore and
conduction to impermeable rock adjacent to the producing
sands are taken into account. After steam injection is
stopped, heat conduction continues and oil sands with r
< r" cool as previously unheated shale and oil sand at
r> rh begin to warm. The effect of warming of oil sand
out beyond 1'" has little effect on the oil production rate
compared to the effect of cooling of the oil sand nearer the
wellbore than rho Thus, in computing the oil production
rate, an idealized step function temperature distribution in
the reservoir is assumed where the original temperature
exists for r > rh and where an average elevated temperature exists for r < rho The average temperature in the oil
sand for the region r < 1'" is computed as a function of
- - - - - - - -_._-

9References given at end of paper.


1613

time following termination of steam injection by an energy


balance. From the average temperature, the oil viscosity in
this region is determined. The oil production rate is calculated by a steady-state radial flow approximation which
accounts for the reduced oil viscosity in this region.

casing radius; for insulated tubing. 1', may be roughly approximated as the inside tubing radius.
The average down-hole steam quality X i for the entire
steam injection period is then

X,

SIZE OF THE HEATED REGIO'.;

= X""'f -

_ 2r.DKr,

Q" -

.( T,-T +2
aD) I
r

--------~--------~-

(1)

where I is read from Fig. 2 as a function of dimensionless


time at, 11',". For uninsulated tubing or the case where steam
is in direct contact with the casing wall, r, is the inside

-./VVVv

H EAT
CONDUCTION
OIL SAND
SHALE
OIL SAND

SHALE
OIL SAND

(2)

Heated Radills

During steam injection the oil sand near the well bore
is at condensing steam temperature L. the temperature
of saturated steam at the sand face injection pressure.
Pressure fall-off away from the well during injection is
neglected in this analysis. and T, is assumed to exist out
to a distance rio where the temperature falls sharply to
Tr, the original reservoir temperature. In reality, the temperature falls more gradually to reservoir temperature because of the presence of the hot water bank ahead of the
steam, but this is neglected to simplify the calculation.
The heated zone radius is calculated by the equation of
Marx and Langenheim: In the case of multisand reservoirs, if it is assumed that each sand is invaded uniformly
as though all sands had the same thickness h and were invaded by equal amounts of steam:

r.'
_ _ FLOW OF OIL,
WATER & GAS

, If,!

Wellbore Heat Losses

To calculate the size of the region heated by steam, it


is necessary to estimate the quantity of heat actually injected after well bore heat losses are taken into account.
Various methods are available for estimating wellbore heat
losses. ]0,11 A simple method which assumes a constant, average temperature of the injected steam and an average
initial geothermal gradient computes the cumulative energy
lost during injection Q" as:

MQ'J"

= hM.

(XiII", + h, .. - /z".) 'f, .


4Kr.(T, - T,.)t,N,

(3)

The function l. is plotted as a function of dimensionless time T = 4Kt,! h'(pC)] in Fig. 3. The use ofthis equation for multisand reservoirs also assumes that injection
times are sufficiently short and that interbedded shale is
sufficiently thick that no heating occurs at the mid-plane
of the shale during the injection period. Eq. 3 further assumes that the value of average density times heat capacity pC for the barren strata is the same as that for the
oil sands [(pC),h" Ie = (pC),].
For multisand reservoirs, a heated radius for each sand
r"i could be calculated assuming equal steam injection per
foot of net sand. An average heated radius rh could then
be computed from rio' = !h,r,,,'/!h,. Practically speaking.
the h ' for reservoirs consisting of more than three sands
of reasonably uniform thickness can be approximated by
using Eq. 3.

TEMPERATURE HISTORY OF THE


HEATED REGION (r",
r
r.)

< <

The average temperature Tav. of the heated region (or


FIG. I-SCHEMATIC REPRESE]'o;TATlOl'i OF HEAT TRANSFER Al'iD
FLUID FLOW CALCULATED BY MATHEMATICAL MODEL.

IOr-----------------------------------------~

t
~

'"~-10 F

====:::;...-:

0.01 ';;;-----~--_;:_:--------__:__=_-----~~--------~
0.01
0.1
1.0
10
FIG.
1614

2-1

FACTOR FOR WELLBORE HEAT

Loss

DETERMINATION.

FIG. 3--DIMENSIONLESS PUSITlOl'i OF STEAMI:DOUT REGION.


JOtTR'iAL OF PETROLEUM TECHl"OLO(;Y

regions in the case of a multisand reservoir) after termination of steam injection is calculated from an approximate
energy balance around the region r" < I' < r h :

T,,"

+ (T,

T,

- T..) [v,~, (1 - 8) - 0], OF

(4)

In Eq. 4, ~. and ~ are unit solutions of component conduction problems in the radial and vertical directions, respectively. 8 is a correction term which accounts for the
energy removed from the oil sand by the produced oil,
gas and water. If little energy is removed by the produced
fluids (the case of a low rate well), 8 can be neglected and
Eq. 4 reduces to a product solution for the average temperature of a series of right circular cylinders of radius
r" conducting three-dimensionally to initially unheated
rock. The development of Eq. 4 is given in the Appendix.
Values for V,' and ;;-, for the case of production from a
single sand may be read from Fig. 4. Relations from which
values of y, and Y, may be calculated are developed in
the Appendix (Eqs. A-8 and A-16).
ENERGY REMo\ED \\ITH
THE PRODUCED FLums

The quantity 8 in Eq. 4 accounts for the energy removed from the formation with the produced fluids and
is defined by:

If
I

8= 2

Hfdx
..
ZTrr,,'(pC),(L _ T..)' dlll1enslOnless .

(5)

I,

Eq. 4 must be solved in a stepwise manner since H" the


energy removal rate, is a function of T""g. Because approximations are used in deriving Eq. 4, as 8 approaches
unity 1',,," calculated by Eq. 4 can hecome less than T, ..
This is physically impossible. When it happens, T,," should
be taken as equal to T,.
H" the heat removal rate with the produced fluids, is
given by:

Hf

q""(H..,,

H,,), Btu/D

(6)

where
H"" = [5.61(pC)"

R"c"](T,," - T,.), Btu/STB oil

(7)
H" = 5.61 p,,[R,,(/lr - h r,.)

R",. h f .,], Btu/STB oil


(8)

wv

0000

1356

(P"")R ". bblliquid water at .60F ,


bbl stock-tank 011

I ... - P ... ,

(9)
when P" > p ... ,. and R".1' < R ... , R ... ,. = R" when p"", > p,,;
if R"., calculated by the above formula is greater than R,,,
then R"., = R"."
1.0

;:::-

0,8

r--.

~III ~

.........,

0.6

"'-

rh

"-

"

0.2

Vr

I
~

IIIII

I I III

_ _ 41l(t-t j )

Vz :8=---2-

IIIII I ~111 r
VZ -II~I~GL~ ~AI~~)

'"

0.1

Steady-State A pproxill1ation for the Productivity Index

For a given stage of reservoir depletion, the ratio of the


oil productivity index (q,J 6..P) in the stimulated condition
i" to that in the unstimulated condition J e can be estimated by an equation of the form:
- _ i" _
1
J-----J,
{t""

p.,,,

DEeEMIlER, 1966

c,

, dimensionless

(10)

+ c,

The geometric factors c, and c, include the pattern geometry and the skin factor of the well. For preliminary engineering calculations if no change in permeability occurs,
suitable values can be calculated from the formulas given
in Table I, which assume that r" r, and r,r r,.
The effect of prior well bore permeability damage is accounted for by using the effective well radius r" in the relations (Table 1). The effective well radius is related to the
and skin factor S by the equation
actual well radius

r"

(11)
Implicit in the assumptions used to derive Eq. 10 is the
assumption that heating and fluid injection have a negligible effect on the permeability to oil. To adequately predict changes in the permeability to oil during the injection and production phases, prediction of two-dimensional,
three-phase behavior is required. The complexity of adding
these equations would undoubtedly necessitate the use of
finite difference techniques to arrive at a solution and are
beyond the scope of the simplified model proposed in this
paper. When swelling clays are present, steam might reduce the permeability, although cases of plugging attributable to steam injection in permeable unconsolidated sands
are apparently rare.
Influence and Significance of the Skin Factor

The skin factor S has a large effect on the stimulation


response. The greater the skin factor, the greater c, will be
relative to c,; hence, the greater will be the influence of
the heated to cold viscosity ratio {tu'./P-., on the ratio of
stimulated to unstimulated productivity indices I. This can
be seen by referring to Eqs. 10 and 11 together with the
defining relations for c, and c, given in Table 1.

.......

1.0

TABLE I-EXPRESSIONS FOR c, AND c, USED IN EQ. 10 FOR CASE OF


NO ALTERATION OF PERMEABILITY BY STIMULATION

"-.....

;;r

c,

System

10

10C

Radial-Pe Constant

8 - DIMENSIONLESS TIME
FIG. 4--S0LUTION FOR

OF THE OIL RATE

For conventional heavy oil reservoirs which have sufficient reservoir energy to produce oil under cold conditions, the use of steady-state radial flow equations appears
to be adequate for predicting the oil production rate response to steam injection. These equations are inadequate
for tar sands and pressure-depleted reservoirs where the
bulk of the stimulated production must come from oil sand
which is actually heated rather than from the portion of
the reservoir that is still cold. The equations outlined in
the following section for this latter case will predict an unrealistically low rate response until the heated region itself
has become depleted or reduced in oil saturation.

",-

0.01

tj

Vr: 8=--2-

0.4

Il(!- l)

CALCULATlO~

AN[);;z (SINGLE SA;,\[).

I"(rhl rw J/ In(r,,/r Ie)


In ( -

'I, )

Radial-pi' Declining

U'

'I,'
2r,--

--"

I"(rhl TtIJ )/ln(re/ru:)

(")

In -

Th

","
V, +-.
2 :J

re

In("I,")-'h
1615

Another way to understand the effect of S is to consider


its relationship to k,j and rd, the permeability and radius
of the region around the well bore where permeability
differs from the formation permeability k:

r"
k - J ) In-=-.
S = ( -k
d
rtl
If kd is small relative to k and r"
large positive value.

TABLE 2-STEAM STIMULATION TEST DATA FOR WelL Q-594


Reservoir Characteristics

Net sand thickness, ft


Number of sands
Oil viscosity. cp

(12)

> r".,

4,050
470
183
16
133

Depth. It
Section thickness, ft

Pre-stimulation

then S will have a

Alteration oj Skin Factor hy Heating

Oil rate, BID


WOR, bbl/bbl
GOR, scflbbl

135
0.83
985

Stimulation
First Cycle

In some wells asphaltene precipitation may be a cause of


high skin factors. If this and similar damage can be alleviated by heating, suitable modification of c, and c, can be
made.
For the constant p, case where the skin factor S is reduced to S, following stimulation and r" > rd:
c,

+ In rh/r".
= S,
S + In relr",

1n r,/rh
; c:, = -----'---

In rJf",

Determination of the Oil RaIl'

To obtain the oil rate as a function of time. It IS necessary to know the unstimulated productivity index J, and
the static reservoir pressure p, as a function of the cumulative fluid withdrawals. Then the stimulated oil rate q"h
can be calculated:

J J,.

:'P. STB/D

(14)

where T is determined from Eq. 10; J, is obtained from


an extrapolated plot of the productivity index history of
the well prior to stimulation; and :'P the pressure drawdown is Pe - p", during the period of stimulated production. The static pressure p, can be estimated from an extrapolation of a plot of p, vs cumulative oil produced prior
to stimulation for those cases where the pressure maintenance benefits of the injected fluids can be neglected.
ESTIMATION OF PERFORMANCE
FOR SUCCEEDING CYCLES

To calculate the performance of cycles following the


first, account must be made for the residual heat left in
the reservoir during preceding cycles. The energy remaining in the oil sand after preceding cycles can be approximated by:
Heat remaining

= u,,'(pC)J1N, (T,,," -

T,.)

Injection time, including shut-in, days


Durat,ion of cycle including injection, days

Stimulated producing time. days


Actual oil production, bbl
Calculated oil production, bbl
Estimated cold production, bbl

(13)

A similar form can be written for the case where p, is declining.

qoh =

Wellhead iniection pressure, psig

(15)

An approximate method of taking this energy into account is to add it to that injected during the succeeding
cycle while assuming that injection takes place into a reservoir that is at original temperature T,. This additional
energy will result in new cycles having a larger heated radius for the same injection time and steam injection rate.
The major assumption involved in this approximation is
that interbedded shale and overburden and underburden
are at initial reservoir temperature at the beginning of
each cycle. This is a conservative assumption since calculated heat losses for cycles after the first will be higher
than those actually observed. A more optimistic analysis
would assume that all of the energy remaining, including
that in the shale, is added to the energy injected on the succeeding cycle.

Second Cycle

19.2
800
55
354
288
47,813
53,700
30,400

18.1
770
46
487
378
80,803
79,140
47,600

Steam injected, MM Ib

RESULTS
MATCH OF FIELD TEST PERFORMANCE
WITH CALCULATIONS

A stimulation field test conducted in the Quiriquire field


of eastern Venezuela provides an excellent check of the
utility of the calculation method. Comparisons of actual to
calculated results are presented for the first two cycles of
steam stimulation of Well Q-594 which produces from 16
sands in a 470-ft interval (Table 2)." A comparison between calculated and actual oil production rates after
stimulation for Well Q-594 is given in Fig. 5. It should be
noted that the calculated curve is a match of observed data
rather than a true prediction. The observed pressure drawdown. down time and water-oil ratio were used as input
data for the calculations.
The only skin effect used in the match of the production
history of the well was an effective skin due to the perforated casing. The effective well radius used in the calculation was 0.00176 ft compared to the actual casing radius
of 0.292 ft to account for a perforation density of four
holes per foot. A plot of the effective radius for various
perforation densities based on Muska!' is illustrated in Fig.
6. The external drainage radius r, was estimated as 570 ft.
The estimated decline in static pressure p, and cold productivity index J, as a function of cumulative oil produced
following steam injection is given in Table 3. The water
cut for this well following stimulation is given in Ref. 9.
The agreement of the calculated and observed rate history for the first cycle is excellent. Calculated cumulative
oil produced at the end of the first cycle differs by less than
-ACTUAL
---CALCULATED
DATA: SEE TABLE 2

400,----------------------------------,
>-

~ 300

........

....

III

~200

......
: 100....

(5

OLi__

~~_LilWLll

200

__

400

~~

__________

600

800

__

1000

TIME SINCE START OF FIRST STEAM INJECTION - DAYS


FIG. 5-C0Il1PARISOl'\ OF CALCULATED AND ACTUAL OIL RATES
AFTEft STEAII1 INJECTlOl'\ FOR QUlRIQUlRE WELL Q-594.

JOliR:\AL ,oFPETIIOLEUM TECHl\Ol.OGY

3 per cent from the cumulative oil actually produced from


the well. Temperature profiles obtained some time after
production had been initiated indicated reasonably uniform
heating of all 16 sands. Temperatures of highest and lowest
sands were cooler than the middle sands; however, the
average temperature across the producing interval agreed
closely with the value of T,,," calculated by this method.
The difference between the observed and calculated oil
rate for the second cycle of the Quiriquire well is greater.
Observed cumulative oil at the end of the second cycle
differs by about 10 per cent from that calculated. However, agreement between calculated and observed performance is still adequate for engineering estimates.
PROCESS

YARIABLE

STUDIES

The calculation method described here has been used to


study the effects of several reservoir and process variables
on the steam stimulation process. Studies of this type provide preliminary estimates of the energy required to
achieve desired stimulation levels and additional insight
into which variables have the greatest influence on the
production response of a steam-stimulated well (Figs .. 7
through 10). The incremental oil-steam ratio has been selected as the primary dependent variable for these studies
since it can be directly related to the economics of the
process. The incremental oil-steam ratio is defined as the
ratio of the stimulated less primary oil production to the
amount of steam injected expressed as barrels of condensate. The values of this ratio are maximum values which
occur when the computed oil production rate has returned
to the extrapolated value of the cold rate. The influences of
_.

_.-

~-.

--

---+~~--+-

./

...

I-

...

VI

oct
ar::
-'

'/
[7

~-+-

::::>

IO-J

-'
w

>
~

......
w

1'1

Static

Index

(M bbl)

Pressure (psio)

(B/D/psi)

0
100
200
300
400

490
410
330
250
170

0.312
0.281
0.256
0.231
0.206

skin damage, sand-shale ratio, stabilized water-oil ratio,


pre-stimulation oil production rate, gross sand thickness,
steam injection rate, total heat input and back-pressuring
of the well early in the production phase on the maximum
incremental oil-steam ratio are discussed in the following
sections.
Effect of Skin Damage

The amount of skin damage present in a well prior to


stimulation can have a tremendous effect on the production response of the well when it is steam stimulated. This
is true even if no well bore cleanup (i.e., damage removal)
is obtained although the stimulation benefits are greater
when well bore cleanup is achieved.
The effect of damage became evident when stimulated
production histories for some California wells became
available which showed better than 20-fold increases in
productivity following steam stimulation."'" This compared with three- to four-fold calculated increases which
would be predicted by steady-state flow formulas assuming no permeability damage. As an example, the effect of
different assumed skin factors on the stimulation response
of a hypothetical well (Well A) is shown in Fig. 7. Reservoir and fluid properties used in the example are shown in
Table 4. For all wells in Table 4, a temporary increase in
the water-oil ratio following steam injection was assumed.
This increase is shown in Fig. 8 as a function of the cumulative water produced divided by the cumulative water injected. This plot represents the average of actual data
from two Quiriquire wells. The calculated curves in Fig. 7
assume that the wells in question have no change in the
skin factor following steaming; thus, production returns
to the pre-stimulation rate. A permanent improvement in
productivity would result if post-stimulation skin factors
were reduced.

Effect of Oil Viscosity

For a given temperature rise, the viscosity reduction of


a low viscosity oil is much less pronounced than for a high
viscosity crude. Thus, peak stimulated oil rates following
steam injection will be smaller, the lower the original oil

,
,

o lOOOr---,---,---,----,---,---,---,----,--,

T rw= 0.25 FT
I

Cold
Productivity

Cumulative
Ojl Production

--

TABLE 3-ESTIMATED DECLINE IN STATIC PRESSURE AND COLD


PRODUCTIVITY WITH CUMULATIVE OIL PRODUCED FOLLOWING
STEAM INJECTION fOR WELL Q594

I
PERF. RADIUS = 0.25 IN.

~ 800r-----+---+---+----~- -+--+---.+--,
~

~ 600r----+--1-+-"<--+--~---+----L--~.. - ,

-_.

._-_._,

,
I

'--"'---

Ti

i
I,

400f---+--++-~""'

DATA GIVEN IN TABLE 4

;:)

'"
....
~

i
10

200r----+-~+---~--~--~~~---

STEAM
INJ

""'~;;;;;;;;;;;;;;;;:;:;:::::d
i

(5

NO. OF HOlES/FT
FIG. 6--EFFECTIVE \VELL RADIUS FOR PERFORATED CASED HOLE
(AFTER MUSKAT').
DECEMIIEII.1966

FIG. 7-EFFECT OF SKIN DAMAGE ON STEAM STIMULATION


RESPONSE.
1617

TABLE 4-STEAM STIMULATION TEST AND CALCULATION DATA

Reservoir Characteristics
Depth, It
Section thickness, ft
Net sand thickness, ft

Well A
(Fig. 7)

Well B
(Fig. 9)

867
546

3,000
200
67

335
7

Number of sands

Well C
(Fig. 10)

Effect of Sand-Shale Ratio

Well D
(Fig. 11)

3,740
200 to 1 ,088
234
18

4,000
1,400
467
36

Reservoir temperature,

of

Oil viscosity, cp:


at TI"

at 300F
Skin factor
Effective well radius,

ft'

97

100

120

125

900
13
20 to 60

4010 1,000
1 to 6

133
8

70
3

0.45

0.00176

0.00176

0.25

30

300

99

230

0.128
0.35
63

1.0
1.0
1,000

0.3
0.57
600

0.5
0-1.0

6.9

40

16.6

42 to 126

Pre-stimulation

Oil role, BID


Oil productivity index,
B/D/psi
WOR, bbl/bbl
GOR, scf/bbl

soo

Stimulation
Steam injected, MM Ib
Wellhead injection

surface conditions
Pressure, psig
440
780
770
Temperature, of
450
520
518
Steam quality,
dimensionless
1.0
0.95
0.95
Injection time, days
21
80
55
Shut-in time, days
13
4
5
*Includes effect of well completions, perforations, etc., jf any.

The effect of the ratio of sand to shale thickness on the


incremental oil-steam ratio for a single cycle is depicted in
Fig. 10 for the conditions assumed for Well C in Table
4. The net sand thickness and number of sands were held
constant in these calculations. and the gross section thickness was varied to vary the sand-shale ratio. The decline
in the incremental oil-steam ratio as the sand-shale ratio
decreases (increasing gross section thickness) is the result
of increased heat losses to the interbedded shales. These
greater heat losses result in an accelerated temperature decline; hence, an increase in the rate of decline of productivity occurs. The above conclusions about sand-shale ratio hold only as long as the average individual sand thick
ness is roughly constant. Thicker sands will cool more
slowly and the stimulation response will last longer for the
same sand-shale ratio than will be the case for thin sands.
Effect of Post-Stimulation Water-Oil Ratio
And Gross Sand Thickness

1,500
600
0.95
18 to 54

viscosity. However, it is not obvious how much less incremental oil will be produced over the entire cycle length
since the lower heat removal rate with the produced fluids
will cause stimulated production to last longer for the lighter oil. If there were no heat losses from the heated portions
of the oil sands, the increased cycle length for the lower
viscosity oil should give the same incremental oil recovery
as for the heavier oil although at a reduced stimulated rate.
Consequently, heat losses to unproductive rock are important in evaluating the effect of oil viscosity on incremental oil recovery.
The effect of initial oil viscosity is shown in Fig. 9 for
Well B of Table 4. These results indicate more than a 50
per cent increase in incremental oil recovered for a 1,000cp oil over a 40-cp oil.

The effects of post-stimulation water-oil ratio and level


of steam input per foot of gross thickness are shown in Fig.
11 for a well having the conditions indicated for Well D
in Table 4.
Fig. 11 indicates that the incremental oil-steam ratio depends markedly on the quantity of steam injected per foot

iii 1.8
c:a

'.....
~

1.6

1.4

CI:

:E 1.2

t;;

'..... 1.0

o
4.

0.8

...:E
...

CI:

~<:
Ow
CI:::l

u..:c(

CI:>

Oz 0.8
~O
z~

-c(
W-A

C)::l

z~

1.6
1.2

0.4

\
\'

:I:ti
CI:
0..

0.0

I
i
i
1

1.

i
,

1000

-T--t--'-

~
0.5
..."'al

I
II
I

J.

.1

OIL VISCOSITY - CP

.1

100

FIG. 9-EFFECT OF VISCOSITY 0:\ bCREMENTAL OIL/STEA~1 RATIO.

Ii

0.0

~ r-- 1

-0.4

0.4
40

0.6

CI:
U

c(-

uJJ

BASIC DATA GIVEN TABLE 4

2.0

-'

~~0.4
Oal

~ ~ 0.3

-+-- L

__~,_;~

-"11F---+--t----+-~____t__-~-~

"'0

~
i= 02
~
.
... co:
co:
u
0.1 f--i-----j

I
I

DATA:!m

~ABl~ 4

0.2

0.4

0.6

0.8

1.0

1.2

CUM. WATER PROD.jCUM. WATER INJ.


FIG. 8--EFFLCT OF STEAM STIMULATION ON WATER-OIL RATIO.
16111

FIG. 100CALCULATEU EFFECT OF SANU/SHALE RATIO 0:\


INCREMENTAL OIL/STEAM RATIO.

JOllIC\AL OF I'ETItOLEl!\t TECHNOLOG\'

of gross section and on the stabilized water-oil ratio. The


detrimental effect of a high pre-stimulation water-oil ratio
is a consequence of the high heat capacity of the produced
water. Since the heat capacity of water is approximately
twice that of the crude oil. a high water-oil ratio results
in a high rate of energy removal as fluids are produced
from the formation.
High producing gas-oil ratios will also have a detrimental
effect on stimulation benefit. Both of these effects can be
seen from Eqs. 8 and 9. Under high temperature conditions a high gas-oil ratio can remove a large amount of
water in the vapor phase accompanied by a high latent
heat of vaporization. Consequently, both high water-oil
ratios and gas-oil ratios cause a more rapid decline in reservoir temperature.
Thicker sections will require a proportionately greater
input of steam to achieve a given stimulation response, all
other variables being equal. The curves plotted in Fig. 11
apply specifically only for the conditions of Table 4 (i.e.,
an undamaged well having the indicated cold productivity
with negligible cold rate decline expected over the cycle
life, and a very thick section where heat loss rates for the
many individual sands can be assumed essentially equal).
Process COlltrol Variables

Three variables relating to process control are the rate


of steam injection, the total heat input and the degree of
back-pressuring of a well early in the production phase of
a stimulation cycle.
Within the limits of injection pressure considerations, it
appears to be beneficial to inject the steam at as high a rate
as possible. The high injection rates provide two benefits:
(1) well bore heat losses as a percentage of total heat injected are reduced and (2) a given amount of energy can
be injected in a shorter period of time, thus reducing the
lost production while the well is being steamed.
It should be noted in Fig. 11 that as the cumulative
steam input is increased, the incremental oil-steam ratio
curves pass through a maximum and begin to decline at
higher steam input. Thus, there appears to be an optimum
level of steam input for a given set of operating conditions,
and it should be possible to optimize the length of steam
injection to maximize the incremental oil-steam ratio. The
-' 2.4
CD
CD

ms/hn = STEAM RATE (LB/HR) /NET FT


COLD OIL RATE 0.5 (STB/D)/FT

.......

-'
CD
CD

2.0

-~-.--------.--

----~

IX

---

-------

m s/h n =1500

i=
<I:

1.6

....<I:
~

on

.......
-'
(5
-'

1.2

m s/h n =1500
ms/h n = 500

.S

<I:

....~

CONCLUSIONS
I. The simplified calculation method presented herein
can match history for conventional heavy oil wells and
can be used with confidence to make preliminary field selection and process variable studies for these cases. Steadystate radial flow equations, while adequate for predicting
steam stimulation response for most conventional heavy
oil wells, should not be used for tar sands and depleted reservoirs. For these cases the bulk of the oil production
comes from the region actually heated rather than from the
unheated region.
2. Process variables studies using the calculation show
that (a) wells having a high skin factor prior to stimulation
will respond most favorably to steam stimulation; a permanent rate improvement results if heating removes a portion
of the skin; (b) low produced gas-oil and water-oil ratios,
high steam injection rate, high sand-shale ratio, thick sands
and high original oil viscosity benefit the stimulation effect;
(c) thick sections require a proportionately greater energy
input to achieve a given incremental oil-steam ratio; (d)
flashing of produced water which causes rapid cooling
and deterioration of the stimulation response can be avoided by back-pressuring the well early in the producing
cycle. This will permit more incremental oil to be produced
than would be attained if drawdowns were maximized
throughout the entire cycle.
'::This is true while the heated radius Th is still much smaller than
the drainage radius reo As rh becomes nearly equal to rtJ, productivity
increases rapidly with increasing Tit again.

Z
....

maximum in the incremental oil-steam ratio is the result


primarily of four factors: (I) increased heat losses associated with the larger heated radius and longer cycle times
which resulted from higher energy inputs. (2) increased
lost production as the length of the injection period is increased, (3) a lower rate of increase of the heated radius
with energy injected as the heated radius becomes large.
and (4) a diminishing incremental benefit to the productivity index by further increasing the heated radius. *
The back-pressuring of a well early in the production
phase of a stimulation cycle can result in substantial increases in the calculated cumulative oil produced at cycle
end. Back-pressuring prevents or minimizes the flashing
of produced water to steam which removes large quantities
of energy which would otherwise be available to prolong
the stimulation effect. A pumping well can be back-pressured by one of two methods. First. the annulus pressure
can be controlled manually while the well is pumped off.
Second, the well can be back-pressured more or less automatically by the column of liquid that will exist above the
pump when pump capacities are rate limiting. Table 5 illustrates the effect of back-pressuring by the pump limiting
method on the computed results for Well Q-594.
Probably the optimum program of back-pressuring a
well would be one in which the flowing bottom-hole pressure is maintained slightly above the saturation pressure
for steam at the existing bottom-hole temperature. This
would provide the maximum drawdown possible without
flashing a large fraction of the produced water to steam.

IX

0
0

200
150
250
100
50
mst;/h - M LB STEAM/FT OF GROSS INTERVAL

FIG. II-THEORETICAL PREDICTlO'i OF Il'iCRDIE'iTAI. OlJjSTEA~I


RATIO \"5 STEA!\OI }l'iJECTED.
UE(;E'IIIIER. 1966

TABLE 5-EFFECT OF BACK PRESSURE DURING THE EARLY PART OF


PRODUCT ION PHASE

300

Calculated Cumulative Inc. Oil


ot Cycle End. STB

Case

Method of BackPressuring

(Well Q594)

Pump limited
No bock-pressure

35,000
17,000
1619

NOMENCLATURE

r,

well radius used in Eq. 1 for well bore heat


loss calculations, ft

rd

r,

=
=
=

radius of the region of damaged permeability


k d , ft
drainage radius of well, ft

a = geothermal gradient, F/ft


b

square root of dimensionless time for radial


temperature decay (Eq. A-5, dimensionless)
B, = constant appearing in Eq. A-14, ft
c"

C,

C"

=
=

constants appearing in Eq. 10, dimensionless


(Table I)
average specific heat of gas, over temperature
interval T,. to T",", Btul scf, F
0

c,,,

C.'

average specific heats of oil and water over


the temperature interval T" to T .. ,.,., Btu/lb
of
"
,

depth of the producing formation, ft


average thickness of the individual sand mem_
bers, ft

hi

= enthalpy

of liquid water at T,,," above 32F.

r.
r,

specific enthalpy of vaporization of water at


T
Btu/lb

rw

actual well bore radius, ft

R"

total produced gas-oil ratio. scflbbl at stocktank conditions

Rw

total produced water-oil ratio, bbll bbl at stocktank conditions

.lR w = Rw - Roo, bbllbbl at stock-tank conditions


R"o

R"" =

DV" ,

hiT = specific enthalpy of liquid water at T" Btu/lb


hIs = specific enthalpy of liquid water at T" Btu/lb
h, = individual sand thicknesses, ft

h, = artificially increased sand thickness used in Eq.


A-14, ft
HI = rate at which energy is removed from the formation with the produced fluids at time I,
Btu/D

Ho. = [5.61(pC)" + R,Ic',l(T,,," - T,.), Btu/STB oil


Hon = 5.61 p,,[R ... (h I - hI) + R ... ,/zIgl, Btu/STB oil
I = dimensionless factor read from Fig. 2 as a
function of aI, I r,,'
J = ratio of stimulated to unstimulated productivity
indexes. dimensionless
J., J" = stimulated (hot) and unstimulated (cold) productivity indexes. respectively, STBI D/psi
J" = zero order Bessel function of the first kind
J, = first order Bessel function of the first kind
k = formation permeability, darcy
kd = damaged permeability, darcy
k" = oil permeability, darcy
K = formation thermal conductivity. Btu/ft/D/oF
I, = thickness of an individual interbedded shale.
ft
IJ = artificially reduced shale thickness used in Eq.
A-14. ft
M. = total mass of steam plus condensate injected
in the current cycle, Ib
N, = number of individual sand members
P. = static formation pressure existing at a distance
r, from the wellbore. psia
P .. = producing bottom-hole pressure, psi a
Pm, = saturated vapor pressure of water at T",", psia
q. = oil production rate, STB/D
qQ' = oil production rate during stimulated production, STB/D
Q", = cumulative energy lost from wellbore during
steam injection, Btu
r = radial distance from the well bore, ft
1620

inside tubing radius, ft

r w = effective well bore radius, ft

Btu/lb

hi.

radius of region originally heated, ft

Sk

S =
S, =
I

t, =

1',,,"

T, =

T, =

=
=

v" v,

v" V, =
V =

normal (unstimulated> water-oil ratio, bbllbbl


at stock-tank conditions
water produced in the vapor state per stocktank bbl oil produced, bbl water vapor (as
condensed liquid at 60F)/STB
kth term in the series solution for v" dimensionless
skin factor of well, dimensionless
skin factor remaining after a stimulation treatment, dimensionless
time elapsed since start of injection for the current cycle, days
time of injection (current cycle), days
average temperature of the originally heated
oil sand at any time t, OF
original reservoir temperature, OF
condensing steam temperature at sand-face injection pressure. OF
temperature difference at r, Z and I above initial reservoir temperature, OF
T,,,," - T", of
unit solution for the component conduction
problems in the rand z direction, respectively. dimensionless
integrated average of v" v, for 0 < r < rio and
all h" respectively, dimensionless
T, - T,., OF

w = constant appearing in Eq. A-16 equal to 4a


(t - t,). sq ft

W, = constant appearing in Eq. A-16, ft

Xi =
X,,,,(

Yo =

=
z=

Y,

average downhole steam quality during mJeclion phase, Ib vapor/lb liquid plus vapor
wellhead steam quality, Ib vapor/lb liquid plus
vapor
hypothetical thickness used in Eq. A-14, ft
zero order Bessel function of the second kind
first order Bessel function of the second kind
vertical distance from bottom of lowest sand
in interval, ft
N.~

~
;=1

h,

a = overburden thermal diffusivity. sq ft/D

/30 =
8

oil formation volume factor, STB oill res bbI


quantity defined in Eq 5, dimensionless
JOUR:-'AL OF ,PETROLEUM TECJ\:\,OLOGY

7i =

dimensionle~s time
_
_
2
_
$, = eTerfc hiT) +--;=\I T -

The average unit solution forv, is obtained by solving


the one-dimensional heat conduction problem in the radial
direction:

v ....

~~',--(
r~)
r?r
(Jr

= volumetric

heat capacity of reservoir rock including interstitial fluids, Btu/ cu ft, OF


p", pw = stock-tank fluid density of oil and water, Ib/
cu ft
T = 4Kt,/ h'(pC)" dimensionless
(pC)1

ACKNOWLEDGMENTS
The authors wish to express their appreciation to Esso
Production Research Co, for permission to publish this
paper. Also, the comments and help of A. R. Hagedorn
and A. G. Spillette are gratefully acknowledged.

~,
ill

(A-3)

with boundary and initial conditions:


v" = 1

t = t,

0< r< r

v, = 0

t = t,

r> r

v,. = 0

t? t,

"

(A-4)

r __ "
Ct::

If the thermal properties do not vary with r, the temperature solution for r < r,. is'
if'

v, = .... '

REFERENCES

e-b'!fi,(y)J,,(ry/r,,)dy
y' [J, (y) Y n (y) - in (y) Y, (y)]'

(A-5)

1. Armstrong, Ted A.: "Steam Injection Has Quick Payout in


Wilmington Field", Oil & Gas Jour. (March 21, 1966) 78.
2. Carslaw. H. S. and Jaeger, J. c.: Condurtion of Heat in Sol
ids, 2nd Ed., Oxford U. Prf'ss, London, England (1959) 346.
3. Carslaw, H. S. and Jaeger, J. c.: ibid., 5.1.
4. Long, Robf'rt .T.: "Case History of Steam Soaking in Kern
River Field. California". JOUf. Pet. Tech. (Sept .. 19(5) 91\9
99.3.
5. Luk ... Y. L.: Intewals of Ressel Functions. McGraw Hill Book
Co.. N. Y. (1962) 314.
6. Marx, J. W. and Langenheim. R. H.: "Reservoir Heating by
Hot Fluid Injection", Tum"., AIME 11959) 2]6, 312314.
i. Muskat, M.: Physical Principles of Oil Produrtion. lst Ed ..
McGraw Hill Book Co., N. Y. (1949) 194. 215.
R Owens. \Y. D. and Suter. V. E.: "Steam Stimulation-Nf'wf'st
Form of Secondary Pf'troleulll Rf'co\'f'ry", Oil & Gas JOUf.
(April 26, 1965) 82.
9. Payne, R. \V. and Zambrano, G.: "Cyt:iie Steam Injedion
Helps Raise Venezuf'la Production". Oil tIC Gas Jllur. (May
24. 19(5) 78.
10. Ramey. H. J . Jr.: "Wellborf' Heat Transmission", Jour. Pel.
Tech. (April, 19(2) 427-435.
II. Squier, D. P., Smith. D. D. and Dougherty, E. L.: "Calclllatf'd
Telllperatnre Behavior of Hot \Vatpr InjPetion Wells", Jour.
Pet. Tech. (April, 1962) 436-440.

where b' = aU - t,)/r,,'.


Since J 1 (y) Y" (y) - J" (y) Y 1 (y)

=2...
Eq. A-5 reduces to:
....y

(f)

v,,

e-I>'II"J,(y)J,,(ry/r,.)dy

(A-6)

o
The average of I', from 0 to r" is given by:

OVERBURDEN

---;---

----f------1---

APPENDIX
DERIVATION OF EQUATIONS
NEGLIGIBLE ENERGY REMOVAL
WITH PRODUCED FLUIDS

The heat transfer model for conduction cooling of the


heated oil sands following termination of steam Injection
consists of approximating the sand-shale sequence by a
stack of variable thickness. equal radii sand cylinders.
These cylinders initially at constant temperature T, COII_
duct heat as a function of time to variable thickness interbedded shales and to semi-infinite overburden and underburden having thermal properties identical to the oil sands.
The geometric approximation of the sand-shale sequence
is illustrated in Fig. 12.
When conduction radially and vertically are the only
mechanisms of heat transfer. the temperature of any point
within the originally heated cylinder can be expressed as
the product:
v

V v,

v~

DECEMBER, 1966

--t I

---t---I

I
I

(A-l)

where v, and v~ are unit solutions of component conduction problems in the rand z directions, respectively. Similarly, an integrated average temperature for the heated
regions may be computed:

v= V v, ~~

12

(A-2)

UNDERBURDEN
FIG. 12-GEOMETRIf. ApPROXIMATION OF MULTIPLE
SAND,SHALE SEQUENCE.
1621

h, = thickness of sand j
I, = thickness of shale j
e- h'''''},(y)dy
Vr =

Luke' gives the following solution of the integral


A-7:
1

00

-b" "-

--

Vr

y = [M. (X, h'q + hI, - hl,V(-:rr,,'(pC), (L - T,)


NJ)- h (y is a hypothetical thickness
which. when added to the individual sand
thicknesses, accounts for all the energy injected including that lost to the shale during the injection phase.)

(A-7)

III

Eq.

The average temperature V, can then be found by integrating over all th~ in.dividual sands:
Bl+hl
F,

where s, is defined as:

S,

(-ITl),

(A-8)

Sf.

/:=0

1'(1.5

yr. k!

1'(2

k)T'( 1

k)

k) 1'(3

N.,
~
;=1

F,dz
/

k)

F,

IN.,
~

~;'
_

(A-lO)

and it follows that:

[( ITI)"

W, +
-- W, erl U.' , - W, ert-;=-

(2

k) (3

]
k)

(A-II)

.1',.

where

U', = B",

The average unit solution for


is obtained by solving
the one-dimensional heat conduction problem in the vertical direction:

W;

B"

If

4a( I -

(t

il'v,

ill',

ilz' =

Tt

with boundary conditions \',


itial condition is

( )_,0

all
~ I all

v. t"z .

'j

v,

(A-12l
=

0, t

>

t, as

z ~ :x. The in-

z outside the regions of thickness h


z within the regions of thickness II,.

j'

The solution of the conduction equation subject to the


initial and boundary conditions is'
C/J

('

2\/-:rl'it

(A-l3)

Carrying out the integration with the initial condition


gives:
I [
z
. h, - z
z - B,
= 2 erl-=+ erf-=+ erl------===-+ erl

2yat

+ ert z

2\/at

+ h" -

2y'at

- _B,

+ el-I

B.,

2v'at

+ ii., -

+ . . . . .]

2v'al

2v'at

= B,_, + h + i
= 0
h =h +y
i, =-~ I, - y (if i, < 0,
j _,

j _,

CASES \\HERE ENERGY REMOYAI.


WITH PRODUCED FLUIDS MUST
HE TAKEN INTO ACCOUNT

If significant energy is removed with the produced fluids, some account should be made for this energy removal.
Consider an energy balance taken on the sum of the thicknesses of the originally heated regions Z and radius rio:
t

Zr.r,,'(pC)l~

Z-:rr,,' (pC), V -

1622

H,dx -

H.., (A-17)

ti

where H, = the energy conducted to the shale and oil-sand


outside the originally heated region, Btu.
Eq. A-17 states that the energy contained within the
originally heated region at any time is equal to that contained immediately prior to production less that produced
with the fluids removed from the formation and less that
conducted to shale and oil sand outside of the originally
heated region. A rigorous evaluation of H, has not been
possible. However, a satisfactory approximation for H, is
as follows.
(A-IS)
where
t

the two sands of thickness


Iz j and h )+, plus the shale of thickness I, are
treated as a single sand. The y is then recalculated to account for one less sand).

(A-16)

For a single sand reservoir. \', is plotted vs 4(Y(t - l')lh,'


in Fig. 4.

1]

Bm - ji",

(A-14)
where B,
B

W.,-Iw) +

"

t,) .

l',(t"z') exp [ - (z - z')'/4l'it)dz'

-Cf)

~w
--=- [exp( -

\I

+ iim - B"
+ h" - Bm

W, = B"
IV, = B,., - B"

+ ii, -

yw

VW

exp( - W,'lw) -exp( - W,'lw) -exp(- W/lw)


(k+1.5)

v,

B,

TI'I! m=ln=l

Hence, Eq. A-8. together with Eqs. A-lO and A-II. form
the solution for V,. A plot of
\'s b' is given in Fig. 4.

VZ

[W
W.
W, ert, ' + w, erl ,-

N.

-----:--c-.--

yw

(A-l5)

m= 1

vz

II,

Substituting Eq. A-l4 into Eq. A-15 and integrating gives:

2
S,,=

~
;=1

B,

(A-9)

From Eq. A-9:

S'tl

N,

1,12

V= V -

Hldx

t,

Z-:rr,,' (pC),

(A-19)

Where conduction is the sole mechanism of temperature


JOt:R:'I'AL OF PETROLEIJM TECH:,\,OLOGY

Ii

decay, and the energy removed with the produced fluids is


negligible:

= V(I -

Combining Eq. A-20 with Eqs. A-I7 and A-IS. and dividing by Z.".r,,'(pCl, gives
= V - 2 V8 - V(l - 8) (I which simplifies to E4. 4.
Of several approximations for H, which were tried, the
one used here gave best agreement with field data. Note
that for the limiting case where conduction IS negligible
-7 1), this approximation gives:

v,

v,vol

and Eqs. A-I7, A-IS and A-20 reduce to Eq. A-2.17 may
be considered an effective driving force for conduction
which takes into account the effect of temperature reduction due to the removal of energy by the produced fluids.
The definition of 8 from Eq. 5 combined with Eqs. A17 through A-19 gives Eq. 4. By Eqs. A-19 and 5:

DECEMBER, 1 .. 66

(A-20)

8)

(v,v,

T,,," = T,.

(T, - T,)[l - 28]

(A-22)

From the definition of 8 this may be seen as giving the


correct value for Ttlvg in this limiting case.

***

)623

S-ar putea să vă placă și