Chilingarian G.V. et al. Surface Operations in Petroleum Production, II
Подождите немного. Документ загружается.
195
A
I
I
M
r-------i
J
N
G
K
0
D
r-------i
H
L
P
Fig. 6-4. Diagrammatical representation
of
particle arrangement and packing density in clean gravels
(A-N)
and in
those
with a matrix
(0-P).
Loose packing is illustrated in C,
E,
F,
G,
and
I.
(After
Hails,
1976, in: Chilingarian and
Wolf,
1976, p. 448; courtesy
of
Elsevier Science Publishers.)
created at the top of the pack after gravel settles. In addition, a tight pack can
utilize a larger gravel in order to achieve the same degree of sand control as
a
loose
pack. Various placement techniques are available that are designed to obtain
a
tight
pack, one being vibration
of
the pack. These techniques are discussed later in this
chapter.
Due to the varying grain size distributions of gravels and formation sands, and
the random packing arrangements
of
gravel, it is necessary to rely on some
empirical means by which to determine the optimum gravel size. Two
of
the more
popular methods that are in use today are those proposed by Schwartz (1969) and
Saucier (1974). Both of these methods use sand sieve analysis data in sizing the
gravel, the same as used earlier by Hill (1941).
In referring to sieve analysis data, the ten-percentile point sand size
(Dl0)
is
defined as that point on the sand size distribution curve which corresponds to the
size mesh opening through which 90% by weight
of
the sand can pass through (see
Fig.
6-5).
The size of
10%
by weight
of
the sand is equal to or is larger than the
196
SIZE,
in
Fig.
6-5.
Examples
of
sand
sieve analysis.
ten-percentile sand size. In a similar manner, the forty-, fifty-, and ninety-percentile
sizes can be determined.
According to Hill (1941), one of the earliest works on gravel/sand size ratios had
success with a 5 to a maximum
of
8
(ratio of the ten-percentile sand size to the
ten-percentile gravel size). In over
4000
wells in Venezuela there was a good sand
control using an average of 6 times the ten-percentile sand size (Landreth, 1969).
According to Schwartz (1969), the gravel/sand size ratio is dependent upon the
uniformity coefficient of the sand and the expected flowrate of the well. Schwartz
defined the uniformity coefficient
(c)
as the ratio of the forty-percentile sand size
(Om)
to the ninety-percentile sand size
(D90):
=
(O40
)/(
O90)
(6-1)
If the uniformity coefficient is less than
5
(uniform sand) and the flow velocity is
less than
0.05
ft/sec, then the following gravel/sand size ratio should be used:
D,,
gravel/D,, sand
=
6
(6-2)
If the uniformity coefficient is greater than 5 (non-uniform sand) and/or the
flow
velocity is greater than
0.05
ft/sec, then the following gravel/sand size ratio
should be used:
Dm
gravel/D4, sand
=
6
(6-3)
In calculating the flow velocity, it should be assumed that only half of the
perforations are open to flow.
According to Saucier (1975), a gravel/sand size ratio of between
5
and 6 should
be used with the ratio being based upon the respective fifty-percentile
(
D50)
values,
regardless of uniformity of sand grain size or flowrate.
197
Any method used should also take into consideration previous experience in a
particular area and formation. Many factors such as tightness
of
pack, pack
thickness, flow rate, accuracy of core data, shape
of
gravel, and proper gravel
screening will all have an effect on the degree to which sand production
is
controlled.
Gravel quality
Maintaining good quality control of the gravel that is used
for
gravel packing is
an important aspect in obtaining an optimum completion. Proper gravel screening,
grain shape, gravel solubility, and gravel strength should all be considered when
choosing a gravel. The A.P.I. has established standard specifications for gravel size
and shape.
Screening
As pointed out by Boulet (1979, p.
165),
gravels are generally poorly sized at the
time that they arrive at the well location. In
51
gravel samples that were tested in
Boulet’s study, an average of
11.5
weight percent was undersized, with
10
samples
having greater than 20 weight percent
of
the gravel being undersized. The reason for
gravel undersizing is
(1)
the initial weakness of gravel and its breaking during
transport (weakness could be a result of natural fractures in grains
or
conglomerate
80
I
I
0
10
20
30
40
50
60
70
00
90
100
PERCENT OUT
OF
GAUGE MATERIAL
Fig.
6-6.
Permeability variations with out-of-gauge gravels. (After Boulet,
1979,
p.
167;
courtesy of the
Society
of
Petroleum Engineers.)
198
I000
500
m
250-
J
m
4
w
I
IY
W
-
a
100-
50
grains), and
(2)
lack of quality control by the gravel suppliers. Tests indicate that
out-of-gauge gravel causes a reduction in gravel-pack permeability (see Fig. 6-6).
To minimize the occurrence of out-of-gauge gravel, the following preventative
steps should be taken: (1) Strict quality control by the gravel suppliers, and
(2)
rescreening the gravel at the wellsite.
I
-
-
SUB-ROUNDEC
A
ANGULAR
-
0
40/60
Gravel
shape
The shape of the gravel grains also has an effect upon the productivity of a
gravel-packed well as well as the ability of gravel to control the production of sand.
Suman (1975, p. 31) noted a minor effect of gravel angularity on permeability
(reduction) (see Fig. 6-7). It was speculated, however, that the higher permeability in
the case of spherical grains may have been due to the smoothness of the glass bead
surfaces that were used in these experiments.
According to Gulati and Maly (1975), angular gravel is not as prone to sand
invasion, because bridges are more likely to form; however, spherical gravel forms a
more consistent pack. Inasmuch as current gravel-packing theory is based upon
absolute stoppage
of
sand, sand bridges are not generally required. A consistent
gravel pack is, on the other hand, desirable and consequently spherical gravel is
preferable. In addition, an angular gravel has a much greater tendency to become
wedged into perforation slots and cause plugging. Krumbein and Sloss (1951) have
35L
LI
0010
0020
00400060
010
MEDIAN GRAIN DIAMETER,
in
Fig. 6-7. Permeability of gravel packs as a function of grain size and angularity. Angularity has a minor
effect. Higher permeabilities of
UCAR
Pac
(glass
beads) is likely due to smoothness
of
particle surfaces.
(After Suman, 1975, p.
30;
courtesy
of
Gulf Publishing
Co.)
199
0.9
0.7
0.5
.-
2
0.3
0
a,
r
Cl
[/)
._
L
0.1
0.3
0.5
0.7
0.9
Roundness
Fig. 6-8. Visual chart for estimating roundness and sphericity of sand grains. (After Krumbein and
Sloss,
1951, p. 81; also in Krumbein and
Sloss,
1963, figs. 4-10, p. 111; courtesy
of
W.H.
Freeman and
Co.)
developed a chart to classify grains on the basis of sphericity and roundness (see
Fig. 6-8).
Gravel
solubility
It is extremely important that the material used in gravel packing is stable under
the conditions to which it is exposed downhole. Wellbore conditions can be
extremely harsh and an initially effective gravel pack may be ruined due to gravel
dissolution.
As determined by Cheung (1985), most materials that are commonly used for
gravel paclung are not significantly soluble in HC1 acid. The solubility
of
these
materials ranged from less than
1%
for Ottawa and Heart
of
Texas sands to 1.8% for
sintered bauxite. All of the tested materials, however, did exhlbit higher solubility in
HC1-HF acid. The gravel solubility of these materials increased with hgher HF
concentrations and longer exposure times (see Figs. 6-9 and 6-10). Twenty-four-hour
dissolution rates in 7.5% HC1-1.5% HF ranged from
4%
for Ottawa sand to
20%
for
sintered bauxite and 96% for low-density bauxite. Due to their high solubilities in
HCI-HF acid, bauxitic materials should not be used in wells that are expected to be
treated with this acid.
200
2o
r
Acid Contact
Time
(Hour)
Fig.
6-9.
Gravel solubility in
7.5%
HCl-1.5% HF acid at
150
OF.
(After Cheung,
1985,
p.
870;
courtesy
of
the Society of Petroleum Engineers.)
As
pointed out by Reed (1980), Underdown and Kamalendu Das (1983), and
Watkins et
al.
(1985), siliceous gravel has been found to be soluble in steam.
According to Reed (1980), gravel solubility increases with increasing temperature
and pH.
As
can be seen from Table
6-1,
at pH levels above
11
rapid increases in
gravel solubility occur. Reed noted that generator effluent has typically
high
pH due
to
the presence
of
bicarbonate ions in the generator feedwater. The bicarbonate ions
decompose into
CO,
(vapor) and
OH-
anions (liquid)
and,
therefore, raise the pH
of
effluent.
TABLE
6-1
Quartz dissolution at various temperatures and
pH
during flow
of
water; pH was adjusted with NaOH
(After Reed,
1980,
p.
943;
courtesy
of
the Society of Petroleum Engineers)
Temperature Si concentration in effluent
a
(mg Si/l)
(“C)
(OF)
PH
7
PH
9
pH
10
pH
11
pH
12
23 73
<
1.0
<
1.0
<
1.0
<
1.0
<
1.0
93 200
<
1.0
<
1.0
<
1.0
4.3 19.7
177 350 6.0
3.4
4.9 42.7
1120
260
5
00
53 69 65 189 1920
a
Flowed at
3
ml/min through
2.5-cm
diameter by
7.6-cm
long pack of
10-20
mesh “Heart of Texas”
quartz sand.
201
100-
ao
-
p
60-
2
>
-
._
0
Y
C
Y
2
40-
0
4
8
12
16 20
24
Acid
Contact
Time
(Hour)
Fig.
6-10.
Gravel solubility
in
12%
HC1-3%
HF
acid at
15O0F.
(After Cheung,
1985,
p.
870;
courtesy of
the Society
of
Petroleum Engineers.)
Underdown and Kamalendu Das (1983) have compared the steam solubility of
siliceous gravel with that of sintered bauxite and resin-coated sand. When exposed
to
500
O
F
water for 72-192
hrs,
siliceous gravel samples had significant weight loss
(>
30%)
at all pH values tested (see Table 6-11). In comparison, sintered bauxite was
stable, with less than
4%
weight loss, when exposed to 560-600
O
F
water with a pH
of
11
for 72
hrs
(see Table 6-111). Resin-coated sand was found to be stable at pH
levels of
9
and below; however, at a pH of
11
significant weight loss
(>
25%)
occurred (see Table 6-IV).
Some operators have been using sintered bauxite to gravel pack thermal wells.
This man-made material, which has good thermal stability, typically has
a
minimum
of 0.8 sphericity and roundness and does not generate fines during pumping or
transporting (Carborundum, 1983). Sintered bauxite is, however, quite expensive
and rather soluble in HC1-HF acid. Another disadvantage lies in the fact that if
high
pH steam is injected into a well that is completed with a sintered bauxite gravel
pack, it is likely that the siliceous formation sand will be dissolved by the steam and
the sintered bauxite will be pushed into the voids that are created by dissolution.
202
TABLE
6-11
Weight
loss
of siliceous material (After Underdown and Kamalendu Das,
1983,
p.
202)
Sample Temperature Time PH Weight loss
(OF)
(W
(8,
wt/wt)
Ottawa sand
540-580 192 7 31.9
16/20
U.S.
mesh
500-540 72 9 38.5
500-540 72
11
46.1
Venezuelan sand
12/18
U.S.
mesh
530-570 72
11
56.0
TABLE
6-111
Weight loss of bauxitic material (After Underdown and Kamalendu
Das,
1983,
p.
202)
Sample Temperature Time PH Weight loss
(OF)
(hr)
(%.
wt/wt)
Sin tered bauxite
560-600 72 7 0.7
20/40
US. mesh
560-600 72 9 1.3
560-600 72
11
3.5
High
alumina beads
560-600 72 7 2.3
20/40
US. mesh
560-600 72 9 2.4
560-600 12
11
3.7
In an effort to reduce gravel dissolution by steam, Watkins et al.
(1985)
suggested
adding ammonium salt to the generator feedwater. The ammonium ions decompose
during steam generation and give off hydrogen ions. The hydrogen ions reduce the
alkalinity of the effluent. The addition
of
ammonium salts to generator feedwater
reduced the effluent pH from
11
to values between
9
and
10.
The silica dissolution
was
found to be reduced by
94%.
TABLE
6-IV
Weight
loss
of
resin-coated sand (After Underdown and Kamalendu Das,
1983,
p.
202)
Temperature Time PH Weight
loss
a
(OF)
(hr)
(W,
wt/wt)
480-500 72 7 0.5
470-500 12
9
1.6
470-500 72
11
25.8
520-570 72
11
24.3
a
20/40
U.S. mesh
203
Gravel strength
During the gravel packing operation, the gravel can be subjected to many
crushing forces as it passes through the pumps and crossover tool ports and
impinges against the casing. The hardest non-fractured gravel available, therefore,
should be used. It is extremely important to take all preventive steps to insure that
the gravel does not breakup under these conditions, because the grain size and shape
will be altered and fines will be generated.
GRAVEL-PACKING
FLUIDS
Choice
of
the proper gravel-packing fluid is an extremely important aspect in the
design of a gravel pack. As with any completion fluid, it is essential that the fluid be
clean and compatible with the formation rocks, minerals and fluids. All
completion-fluid system equipment should be thoroughly cleaned prior to gravel
packing. Additionally, the completion fluid should be filtered to remove all solids
that are larger than 2 pm. If it is necessary to add solids to increase viscosity and
fluid density, or to control fluid loss, then these solids must be capable of being
easily removed by either dissolving in the formation fluids or by being acidized. In
addition, if viscosifiers are added, a breaker should be added that will return the
viscosity to that of water within 24 hrs. Fluids not requiring breakers or solids are
the most desirable gravel packing fluids.
Borden et al. (1982) have classified water-base carrier fluids into four major
categories.
Low-viscosity carrier
fluids
Included in low-viscosity carrier fluids are water, brine, oil, diesel, and gas oil.
Small amounts
of
HEC or other polymers can be added to aqueous-base fluids to
increase fluid viscosity up to 100 cP. Low-viscosity fluids can carry up to five
pounds
of
gravel per gallon of fluid, depending on the type
of
surface pumping
equipment used. With standard surface blending equipment, the average is one
pound of gravel per gallon of fluid, because the gravel must go through the pump.
The lower the viscosity of the fluid, the better the compaction of the gravel. At high
viscosity, the grains will be separated temporarily. Eventually, the grains will settle
to expose the upper portion
of
the liner and
so
defeat the purpose of the pack. The
compaction percentage is directly proportional to the free-falling velocity of the
gravel in the carrier fluid. The average free falling velocity of
US.
12-16 gravel in
water is 32.4 ft/min, whereas with
U.S.
6-8
gravel the average
is
46.6 ft/min
(Wright and Solum, 1967).
High-viscosity carrier
fluids
The high-viscosity carrier fluids are composed of water solutions with added
polymer or cross-linked polymer. These fluids are capable of carrying up to 15 lb of
204
gravel per gal of fluid and, therefore, require much less fluid. The high viscosity
reduces fluid loss, and fines do not have a tendency to migrate towards the
perforations and cause plugging (Borden et al., 1982). These fluids do, however,
exhibit a significant amount of gravel settling or poor compaction after placement,
which can cause voids in the pack and a loss of sand control (Elson et al., 1984).
Medium- or intermediate-viscosity carrier fluids
Medium- or intermediate-viscosity carrier fluids typically have viscosities ranging
from 300-400 cP. In the case of these fluids, perforation plugging by fines usually
does not present problems (Borden et al., 1982). Tests indicate that medium-viscos-
ity fluids should be used in all high-angle gravel-packing operations (Elson et al.,
1984).
Foam or low-density carrier fluids
Elson and Anderson (1982) have investigated the use of foam as a carrier fluid
for gravel-packing operations. Foam has a low density,
high
apparent viscosity, high
lifting capacity, and a relatively low cost. These characteristics make foam a
desirable carrier fluid, especially in low-pressured zones that are prone to lose
circulation and in water-sensitive formations. Lost circulation is much less likely to
occur with foam due to the reduced hydrostatic pressure. Formation damage in
water-sensitive formations is greatly reduced, because the zone is exposed to much
lower volume of water. When foam is used as a carrier fluid, liner centralizers
should be used sparingly because they can induce packing irregularities.
SCREEN
OR
LINER SELECTION
A screen or slotted liner is placed inside the wellbore to retain the gravel between
them and the formation (or casing) and to allow the flow of fluids into the wellbore.
The width of the slots must be narrow enough to prevent the gravel from passing
through them. Coberly (1937) recommended that the slot width be twice the largest
grain diameter and, therefore, he relied upon the occurrence of bridging. Inasmuch
as sand bridges are not usually stable under typical downhole conditions, this sizing
method is not advisable. Absolute stoppage of gravel is necessary and the slots
should be sized smaller than the smallest gravel size. On an average, the difference
in size should be 0.015 in. to retain all the gravel in place. The overall average open
area should be approximately 2% of the surface area of the liner. This figure may
vary according to the production expected (Wright and Solum, 1967).
A gravel pack liner is simply a casing with slots that have been cut into it. The
slots are cut vertical to prevent any significant reduction in pipe strength. Slotted
liners are usually much less expensive than wire-wrapped screens and less prone to
damage when running into the wellbore. There is, however, a limitation as to how