Performance Analysis of Wells With

Downhole Water Loop Installation

for Water Coning Control
L. Jin, A.K. Wojtanowicz, Louisiana State University

reservoirs with bottomwater drive. Hydrodynamically, water

Abstract coning is caused by pressure gradients resulting from well pro-
Downhole water sink (DWS) technology controls bottomwater duction from the pay zone. Although the theory is not difficult to
coning by draining water with a second completion placed under understand, water coning is difficult to control [Shirman(2) and
the oil-water contact. The technology has been studied theoreti- Abou-Sayed et al.(3)]. For many years, reservoir engineers have
cally with analytical, numerical and physical models, showing looked for some effective ways to control water coning, but it was
an increased rate of oil production and recovery. Also, improved not until the late 1990s that successful DWS technique field appli-
wells’ productivity with DWS has also been demonstrated in cations were reported [Swisher and Wojtanowicz(4), Bowlin et al.(5)
several field implementations. However, downhole drainage in and Ramos et al.(6)].
DWS wells requires independent lifting of considerable volumes Reportedly, DWS technology is a novel and effective method
of water that necessitates using either two tubing strings (for because it could not only prevent water coning, but also reverse
water and for oil) or one tubing and the tubing/casing annulus – the water cone after the water breakthrough [Bowlin et al.(5) and
for water and oil, respectively. Also, extensive water drainage in Shirman(2)]. For example, the results reported by Bowlin et al.(5)
the systems with weak bottomwater drive may cause a reservoir showed that the DWS installation significantly restored the oil pro-
pressure drop and the need to return the produced water from duction rate in a nearly watered-out well. By now, DWS comple-
the surface to the aquifer for pressure maintenance using desig- tions have been field-tested in numerous reservoirs all over the
nated injection wells. DWS well completion with the downhole world with good results [Ramos et al.(6), Ju et al.(7) and Davies
water loop (DWL) offers the benefit of re-injecting the drainage et al.(8)].
water back to the same aquifer in the same well without lifting Most of the theoretical work on DWS has been conducted at
the water to the surface. This could be achieved by designing a Louisiana State University. In addition, a considerable body
DWS well with three completions: the top (oil) completion, the of DWS research has been added by other authors [Gunning et
middle completion for water drainage and the bottom completion al.(9), Ould-amer et al.(10), Siemek and Stopa(11), Johns et al.(12) and
for water injection. Despite mechanical complexity of the triple Utama(13)]. Although all the theoretical evidence confirms effi-
well completion, there are two limitations of the DWL system ciency of the DWS technique, it also demonstrates its downside –
– the drained-and-injected water must be free from oil and the large amounts of water produced to the surface, pressure depletion
pressure interference between the two water completions must
of reservoirs with weak water drive and additional costs of water
be minimized.
handling and disposal.
In this work, a well performance (nodal) analysis model has
been developed for a DWL well completed in an oil reservoir
underlain by water layer of known thickness. In the model, the Development of DWL Technology
positions (depths) of the three well completions and the rates of In 1992, Wojtanowicz and Xu(14) proposed a DWL method to cut
production and drainage/injection are design parameters, while
back the volume of formation water produced by an oil well from a
all other properties are reservoir system properties. The model
hydrocarbon reservoir underlain by a water zone. The method em-
has been used to find the operational range of DWL for a given
ployed dual completion of the well inside the water zone, below
reservoir system and to compare DWL wells with conventional
the oil-water contact (OWC) to install the water loop equipment
wells, single-completed at the top of the oil layer.
(separated by a packer), in addition to the conventional completion
The results show that for each DWL system, there is such a
combination of the top production rate, bottom drainage-injec- in the oil zone (above the OWC). The water loop installation in-
tion rate and drainage-injection distance (D/I spacing) that would cluded a submersible pump, the upper (water sink) perforations and
result in water-free oil production. There is a minimum value the lower (water source) perforations. A submersible pump would
of D/I spacing above which the detrimental effect of pressure drain the formation water around the well from the water sink, and
interference between the two water completions is practically then would reinject the same water back to the water zone through
eliminated and the beneficial effect of water drainage on well the water source perforations. A simulation study was conducted
performance is strong – a two-fold increase of water drainage to investigate hydrodynamic performance of the method to restrain
rate would increase the critical oil rate by 80%. Also, because the water movement towards oil-producing perforations. The DWL
minimum D/I spacing is relatively small, DWL wells may be in- was mathematically modelled by computing flow potential distri-
stalled in reservoirs with thin layers of bottomwater. bution, generated by two constant-rate sinks (oil and water) and
one constant-rate source (water), located between the three linear
boundaries and the constant-pressure outer radial boundary. The
Introduction study revealed that the shape of the dynamic OWC in the well’s
vicinity could be effectively controlled by the method; therefore,
In the presence of water coning, water inflow to petroleum the oil production rates could be two to four times higher than
wells hampers production of oil or gas leading to early shut- the critical rates obtained when using conventional completion.
downs of the wells without sufficient recovery of hydrocarbons Also, the method had the advantage of becoming a solution to the
in place [Hernandez(1)]. This is a serious problem in maturing environmental compliance problem associated with disposal of
38 Journal of Canadian Petroleum Technology
 FIGURE 2: Flow streamlines from uniform source and point source to
point sink in 2D system.

• D/I spacing: distance between the water sink and injection

FIGURE 1: Downhole water loop (DWL) well completion.
completions in the aquifer.
produced water. From the standpoint of the reservoir engineering It is possible to use the flow potential theory to develop ex-
theory presented in the 1992 study, formation water could be kept pressions for the streamlines and isopotential lines for a number
away from oil-production perforations so that the oil recovery per of cases of 2D fluid-flow in the DWL well system, as showed in
well could also be improved. Figure 2. In a radial system, the change of pressure with respect
In another DWL study, Wojtanowicz and Shirman(14) deter- to radius dp/dr is inversely proportional to the radius 1/r [Smith
mined the effect of hydraulic communication (cement leak) be- and Pirson(16)]. This means that at a large distance from the well-
tween the water drainage and injection zones. The leak would bore, the pressure gradient is small. Thus, reinjecting the drained
reduce the size of the water drainage zone under the oil-producing water into the same aquifer at a vertical distance from the drainage
perforations and make the system inefficient. In that study, the completion would reduce hydraulic communication between the
downhole drainage injection was mathematically modelled as a two completions, while maintaining pressure (water drive energy)
system of three sinks operating under steady-state flow conditions in the aquifer. Recovery benefits of water replacement for aquifer
in a multi-layered porous medium. The results of simulation runs pressure maintenance are well documented [Kjos et al.(17), Paige et
revealed principal relationships between the reservoir engineering al.(18), Detienne et al.(19), Singh(20) and Abou-Sayed et al.(21)].
factors (fluid mobilities, configuration of geological strata and the
degree of zonal isolation) and the production design factors (the DWL System Experiment Design
position of well completions and the oil production and water in-
jection rates). Also, the study showed that for determined geolog- In the simulation model shown in Figure 3, a radial-cylindrical
ical conditions and well completion geometry, there was a unique reservoir is used, similar to that suggested by Hernandez(1). In
relationship between the water injection and oil production rates, order to simulate the DWL system, especially the water loop com-
which ensured stability of OWC, resulting in the continuous pro- ponent, three fictitious wells are placed at the same R-location and
duction of oil with minimum amount of water. perforated at different Z-locations. The top completion is located at
The objective of this work is to study the DWL systems fur- the top of the oil zone to lift the oil by submersible pump. The two
ther, and to: other completions for water drainage and re-injection are located
• Build a well performance (nodal analysis) model of DWL. in the aquifer – below the OWC and deeper in the aquifer, respec-
tively. The completions are separated by a packer that isolates their
• Model the pressure interference between the three
flowing pressures. Thus, a “water loop” is formed. With a high-
completions.
speed reservoir simulator, Eclipse 2007TM, the rating of a pump
• Understand how the key parameters, such as the D/I spacing, can be automatically adjusted by setting the top and bottom rates.
production/drainage-injection rates and clean water drainage (In the simulations, we used constant rates at all three completions
limit, would affect DWL performance. over the production time).
The simulation model comprises 31 grids in the radial and
20 grids in the vertical directions, with fine timesteps (approxi-
DWL Well Completion Method and mately half a day at early time of production and no more than
1 month for late time). The grid size in radial direction increases
Mechanisms logarithmically so that the radial locations of the block centres are
In water coning, the dynamic OWC deforms upward because of
the pressure drop caused by oil production. Thus, an equal pres-
sure drop in the water zone would keep the water from rising
[Wojtanowicz and Shirman(15)]. Figure 1 depicts the mechanisms of
controlling the DWL well system: a well is dually completed in
the oil and water zones and the two completions are separated
by a packer set inside the well at the depth of the oil-water con-
tact. Both the oil and water zone completions include a submers-
ible pump. The top submersible pump lifts the oil to the surface,
while the bottom pump drains the formation water around the well
and injects the water deeper into the same aquifer, thus preventing
the water cone from breaking through the oil column into oil-pro-
ducing perforations. The following definitions have been used in
this paper to describe the behaviours of DWL wells:
• Top rate: the production rate of the top completion in the oil
zone.
• Bottom rate: the water drainage rate at the water sink comple-
tion equal to the water injection rate, as shown in Figure 1. FIGURE 3: Coaxial simulation model of DWL well-reservoir system.
June 2010, Volume 49, No. 6 39
 1.0 7000
0.9

0.8 6000

0.7
5000
Water Cut

0.6

Pwftop, psi
0.5 4000
Qtop = 200 bpd
0.4
Qtop = 400 bpd
0.3 3000
Qtop = 2000 bpd
0.2 Qtop = 3000 bpd Qtop = 200 bpd Qtop = 2000 bpd
2000
Qtop = 5000 bpd Qtop = 400 bpd Qtop = 3000 bpd
0.1
Qtop = 5000 bpd
0.0 1000
0 200 400 600 800 1000 1200 0 100 200 300 400 500 600

Time, days Time, days

FIGURE 4: Water coning stabilization process with equilibrium water FIGURE 5: Stabilization of bottomhole flowing pressure.
cut.
show that the model may reach equilibrium condition after a con-
roughly in geometric progression. There are 10 layers in both the siderably long time period. We believe that the model properly
oil zone and aquifer in a vertical direction; the grid size is equally describes pressure interference between the three completions –
distributed in this direction. To simulate steady-state flow condi- the effect that entirely controls DWL system performance.
tions, the produced fluids from both completions are reinjected at
the outermost gridblock into the oil and water zones, respectively The reservoir and well data and fluid properties used in this
[Inikori(22)]. The reservoir model would maintain constant volume study are shown in Tables 1, 2 and 3 [Arslan(24)]. The data reveals
of fluid in place andv constant reservoir pressure, thus simulating a combination of high porosity and low-permeability – the prop-
a stabilized well production. The aquifer model is a coaxial cyl- erties of an oilfield in China [Song et al.(25)] where the porosity in
inder of large size connected to the bottom of the oil pay zone. the reservoir is as high as 0.3, while the permeability is only 80
After an initial transient period of developing water cone, the mD. High porosity with low-permeability is caused by the high
system attains steady state [Inikori(22), Arslan et al.(23), Arslan(24) clay content in the sandstone. Hu and Dietrich(26) also reported a
and Hernandez(1)]. The duration of the transient period depends similar field case in Canada. Neasham(27) gave a concrete analysis
on system properties (permeability, size, etc.) and the set limit for on the dispersed clay in sandstone reservoirs and its effect on fluid-
flowing pressure stability (convergence tolerance). Figures 4 and 5 flow properties. It is not uncommon to have reservoirs with high

Table 1: Reservoir and well data.

Data U.S. Field Unit Values Conversion Factor SI Unit

Datum depth (D) ft 15000 × 3.048 E−01 m
Thickness of oil zone (Ho) ft 50 × 3.048 E−01 m
Depth of WOC (static) (DOWC) ft 15000 × 3.048 E−01 m
Thickness of aquifer (Hw) ft 500 × 3.048 E−01 m
Reservoir pressure at datum depth (Pref) psi 6000 × 6.894 757 E+00 kpa
Position of top completion from formation top (Dtop) ft 0 × 3.048 E−01 m
Thickness of top completion (Htop) ft 20 × 3.048 E−01 m
Position of bottom completion from WOC (Dbot1) ft 10 × 3.048 E−01 m
Thickness of bottom completion (Hbot1) ft 10 × 3.048 E−01 m
Horizontal permeability in oil zone (absolute) (Kho) md 80 × 9.869 233 E−04 µm2
Vertical permeability in oil zone (absolute) (Kvo) md 48 × 9.869 233 E−04 µm2
Horizontal permeability in water zone (absolute) (Khw) md 80 × 9.869 233 E−04 µm2
Vertical permeability in water zone (absolute) (Kvw) md 48 × 9.869 233 E−04 µm2
Porosity in oil zone (Φo) fraction 0.3 × 1.000 E+00 fraction
Porosity in water zone (Φw) fraction 0.3 × 1.000 E+00 fraction
Well radius (rw) ft 0.292 × 3.048 E−01 m
Outer radius (re) ft 510 × 3.048 E−01 m
Depth to reservoir top (De) ft 14950 × 3.048 E−01 m
Total (rock+water) compressibility of the aquifer (Ct) 1/psi 7.00E-06 × 1.450 377 E-01 1/kpa
Distance between drainage and injection completions (D/I) ft 480 × 3.048 E−01 m
Thickness of re-injection completion (Hbot2) ft 10 × 3.048 E−01 m
Max pressure drawdown at top completion (∆Ptop) psi 800 × 6.894 757 E+00 kpa
Max pressure drawdown at bottom completions (∆Pbot) psi 1500 × 6.894 757 E+00 kpa
Tubing inner diameter (dt) in 1.25 × 2.54 E-02 m
Tubing head pressure (Pt) psi 50 × 6.894 757 E+00 kpa
Top completion skin value (estimated) (Stop) 0 × 1.000 E+00
Water sink completion skin value (estimated) (Sbot1) 0 × 1.000 E+00
Re-injection completion skin value (estimated) (Sbot2) 0 × 1.000 E+00
Tubing head temperature (Tt) oF 100 × (oF +459.67)/1.8 E+00 K
Bottomhole temperature (Twf) oF 180 × (oF +459.67)/1.8 E+00 K
Water injection limit (Qinj) bpd 8000 × 1.589 873 E−01 m3/d
Min bottom completion IPW* limit (Qbot min) bpd 0 × 1.589 873 E−01 m3/d
Max bottom completion IPW limit (Qbot max) bpd 8000 × 1.589 873 E−01 m3/d
Min top completion IPW limit (Qtop min) bpd 0 × 1.589 873 E−01 m3/d
Max top completion IPW limit (Qtop max) bpd 5000 × 1.589 873 E−01 m3/d

* IPW = Inflow Performance Window

40 Journal of Canadian Petroleum Technology

Table 2: Fluid properties’ data.

U.S. Conversion SI
Data Field Unit Water Oil Gas Rock Factor Unit
Pref psi 6000 6000 6000 6000 × 6.894 757 E+00 kpa
B rb/stb 1.02 1.26     × 1.000 E+00 rm3/sm3
C 1/psi 0.000003 0.000015   0.000004 × 1.450 377 E-01 1/kpa
µ cp 0.4 0.8     × 1.000 E−03 Pa·s
ρsc lbm/ ft3 68 48 0.0702    × 1.601 846 E+01 kg/m3
Pb psi     1000   × 6.894 757 E+00 kpa
Rs at Pb Mscf/stb     0.379   × 1.801 175 2 E+0 m3/m3

porosity and low-permeability. For example, this effect is widely Table 3: Relative permeability and capillary pressure.
spread across the Norphlet formation in the southern coast of the
U.S. [Dixon et al.(28)]. The reason for choosing the data in this Case Sw krw kro Pc (psi) Pc (kpa)
study is because water coning is more severe in these kinds of res- 0.20 0.0000 1.0000 1.428 9.846
ervoirs [Joshi(29)]. However, we believe the choice of data should 0.30 0.0002 0.8374 1.250 8.618
not significantly affect findings of the study. 0.40 0.0039 0.6328 1.071 7.384
A 0.55 0.0366 0.3337 0.803 5.536
Approach 0.70 0.1526 0.1187 0.536 3.696
0.80 0.3164 0.0391 0.357 2.461
Our main objective was to develop an understanding of the ef- 1.00 1.0000 0.0000 0.000 0.000
fects of DWL well completion variables on well system perfor-
mance. Hence, three operational parameters have been selected for 0.20 0.0000 1.0000 1.428 9.846
the simulation study: top (production) rate, drainage-injection rate 0.30 0.0029 0.6944 1.250 8.618
(bottom rate) and the D/I spacing. In order to see individual effects, B 0.40 0.0128 0.4444 1.071 7.384
we would set the value of one parameter and vary the other two pa- 0.55 0.0605 0.1736 0.803 5.536
0.70 0.1898 0.0278 0.536 3.696
rameters to observe their relative effects. During each simulation
0.80 0.3547 0.0000 0.357 2.461
run, we would bring the system to equilibrium by stabilizing the 1.00 1.0000 0.0000 0.000 0.000
flow so both the water cut (WC) and flowing pressure (pwf) would
become constant. We also assumed zero skin damage and a com-
plete integrity of the cement sheath [Arslan(24)]. damaged. (Even minimal oil content in injection water would de-
With the high-speed reservoir simulator, Eclipse 2007TM, which posit a residual oil-saturated skin zone around the injection com-
was used in the study, a typical simulation run would be as follows: pletion, thus reducing permeability to water and injectivity of the
fix the drainage rate and change the top rate and distance. (For ex- completion.) Therefore, it is important to avoid oil in the injection
ample, for drainage rate 3200 B/D, we would change the top rate water [Pang and Sharma(30) and Abou-Sayed et al.(3)].
to 480 B/D, 640 B/D and 800 B/D, with D/I spacings for each top Let’s consider the oil content limit in the injection water. Figure
rate at 0 ft, 50 ft, 100 ft, 200 ft, 300 ft, 400 ft and 480 ft. The total 9 shows that oil content in injection water changes with D/I spacing
number of runs for this scenario is 1×3×7=21.) Typical simulation and injection rate. When the top rate is high and the bottom rate is
scenarios used in this study are shown in Table 4. low or the D/I spacing is short, there will be little oil in the injec-
tion water. These results correspond to the drainage completion
located 10 ft below the OWC. Also note that if the drainage com-
pletion is placed 20 ft below the OWC, there will be no oil in the
Results and Observations injection water.
In conventional wells, operated at a constant liquid produc- To produce oil at a higher rate, while avoiding oil in injec-
tion rate, the increased water cut leads to less oil produced – a tion water, an Inflow Performance Window (IPW) can be de-
common problem in production practices. In contrast, all simula- veloped for different combinations of the top and bottom rates,
tions of DWL wells consistently show lower water cut and higher as shown in Figure 10. In the figure, there are three plots of Qbot
oil rates, as shown in Figures 6 and 7. The results can be summa- vs. Qtop, each of them being an “envelope” representing different
rized as follows: D/I spacing: 10 ft, 100 ft and 480 ft. The top boundary of each
• Larger spacing between water drainage and injection com- envelope represents the maximum oil-free water drainage rate,
pletions (D/I spacing) dramatically improves well perfor- while the lower boundary corresponds to the maximum top rate
mance by reducing water cut and increasing oil rate. of oil with no water. (Points inside the envelopes correspond to a
• There is an optimum value of D/I spacing (≈200 ft in this
case), above which there is no further improvement. Table 4: Simulation scenarios.
• Critical rate increases with increased D/I spacing.
Bottom Rate Top Rate D/I Spacing
• At optimum D/I spacing, the effect of water drainage on well
(bpd) (bpd) (ft)
performance is strong; a 60% and 100% increase of water
drainage would increase oil production by 55% and 80%, 480 0,50,100,200,300,400,480
respectively. 800 640 0,50,100,200,300,400,480
• For each DWL system, there is a combination of the top 800 0,50,100,200,300,400,480
(production) rate, bottom (drainage-injection) rate and D/I
spacing that would result in water-free oil production. The 480 0,50,100,200,300,400,480
observation is consistent with previous studies of DWS sys- 2000 640 0,50,100,200,300,400,480
800 0,50,100,200,300,400,480
tems [Shirman(2), Siemek and Stopa(11), Arslan(24), Johns et
al.(12), Hernandez(1) and Utama(13)].
480 0,50,100,200,300,400,480
• For a given D/I spacing, an effective increase of oil pro- 640 0,50,100,200,300,400,480
3200
duction requires synchronized increases in production and 800 0,50,100,200,300,400,480
drainage rates, as shown in Figure 8. A sole increasing of the
top production rate is not effective. 480 0,50,100,200,300,400,480
Oil-free water looping is an important condition for DWL. If 4000 640 0,50,100,200,300,400,480
the drainage-injection water contains some oil, not only will the 800 0,50,100,200,300,400,480
aquifer be polluted, but also the injection completion will be
June 2010, Volume 49, No. 6 41
 1.0 1.0
1.0 Qtop = 480 bpd Qtop = 480 bpd Qtop = 480 bpd
0.9 0.9
0.9 Qtop = 640 bpd Qtop = 640 bpd Qtop = 640 bpd
0.8 0.8
0.8 Qtop = 800 bpd Qtop = 800 bpd Qtop = 800 bpd
0.7 0.7
0.7 (Qbot = 2000 bpd) (Qbot = 3200 bpd) (Qbot = 4000 bpd)
Water Cut

Water Cut

Water Cut
0.6 0.6
0.6
0.5 0.5
0.5
0.4 0.4
0.4
0.3 0.3
0.3
0.2 0.2
0.2
0.1 0.1
0.1
0.0 0.0
0.0
0 100 200 300 400 500 600 0 100 200 300 400 500 600 0 100 200 300 400 500 600

D/I Spacing, feet D/I Spacing, feet D/I Spacing, feet

FIGURE 6: Water cut changes with top rate and D/I spacing at various bottom rate.

combination of the rates that the oil is water-free and drainage water A typical nodal analysis involves solving the system of two
contains no oil – i.e., segregated production.) It is evident that with the equations, TPR and IPR, for the optimum values of flowing
injection point approaching the water drainage completion, the pressure and production rate. Golan and Whitson(34) and Beggs(35)
performance window of the DWL system moves into the area of presented a wide selection of available calculation models and ap-
larger water rates and smaller oil production. plied the nodal analysis method to production oil and gas wells, as
It is important to see that the DWL water re-injection tech- well as modifying the method for injection wells.
nology would work even for a close distance between drainage Arslan et al.(23) and Arslan(24) applied the nodal analysis method
and injection completions (small D/I spacing). This means that the
to DWS wells. They pointed out that oil wells would be generally
DWL system does not require deep drilling or deviated wells to in-
produced at their highest possible rate to maximize the cash flow.
ject produced water. Moreover, the system would not reduce the
water-drive ability of the aquifer (reservoir pressure) while giving However, the maximum allowable pressure drawdown at the two
a low-water-cut oil production. From the previously described re- completions of the DWS well was a limiting factor caused by prac-
sults we can see that there exists a critical (minimum) D/I spacing tical considerations, such as well integrity, sand control, gas lib-
in the DWL system. At this distance, the oil rate is at the maximum eration, etc. The nodal analysis for DWS wells was based on the
and there is no oil in the injection water. following operational principles:
• Produce at maximum possible top rate (economic goal).
• Keep the pressure drawdown at or below the maximum al-
System (Nodal) Analysis of DWL Well lowable pressure drawdown for both completions (comple-
tion limit).
Nodal analysis is a well-known method for assessing perfor-
mance of petroleum wells as production systems – from the outer 1.0
(Qbot = 2000 bpd)
boundary of the reservoir to the sand face, across the perforations 0.9
and completion section to the tubing intake, up the tubing string 0.8
with its restrictions and down-hole safety valves, past the surface 0.7
choke, the flow lines and separators. The name “nodal” was pro-
Water Cut

0.6 D/I = 50 ft
posed by Gilbert(31), who proposed considering an oil well as a se-
0.5 D/I = 100 ft
ries of hydraulic components bracketed by appropriately placed
0.4 D/I = 200 ft
nodes. Later, Brown and Beggs(32) and Mach et al.(33) systemati- D/I = 300 ft
cally studied Gilbert’s idea and further improved the system anal- 0.3
D/I = 400 ft
ysis method. They clearly defined the concept by reducing the 0.2
D/I = 480 ft
complex production system to three basic correlations describing: 0.1

• Flow through porous medium to the wellbore – known as in- 0.0

flow performance relationship (IPR). 0 500 1000 1500 2000

• Flow through vertical or directional conduit – known as Qtop, bpd

tubing performance relationship (TPR).
FIGURE 7: Water cut vs. top rate at bottom rate, 2000 bpd, for
• Flow through horizontal pipe in the surface. different D/I spacing.
400 500
(Qbot = 800 bpd) (Qbot = 2000 bpd)
350 450

400
300
Top Oil Rate, bpd

Top Oil Rate, bpd

350
250
300

200 250

150 200

150
100
D/I = 10 ft D/I = 200 ft D/I = 10 ft D/I = 200 ft
100
D/I = 300 ft D/I = 480 ft D/I = 480 ft
50 D/I = 50 ft D/I = 50 ft D/I = 300 ft
50
D/I = 100 ft D/I = 400 ft D/I = 100 ft D/I = 400 ft
0 0
0 500 1000 1500 2000 0 500 1000 1500 2000

Top Liquid Rate, bpd Top Liquid Rate, bpd

FIGURE 8: Top completion oil rate changes with top completion liquid rate at various D/I spacing and bottom rate.
42 Journal of Canadian Petroleum Technology
 2500 2500 2500

(Qbot = 1200 bpd) (Qbot = 1400 bpd) (Qbot = 1600 bpd)

2000 2000

Bottom Oil Rate, bpd

Bottom Oil Rate, bpd

Bottom Oil Rate, bpd
2000
Qbot = 200 bpd Qbot = 200 bpd Qbot = 200 bpd
Qbot = 400 bpd Qbot = 400 bpd Qbot = 400 bpd
Qbot = 800 bpd Qbot = 800 bpd 1500 Qbot = 800 bpd
1500 1500
Qbot = 1200 bpd Qbot = 1200 bpd Qbot = 1200 bpd
Qbot = 1800 bpd Qbot = 1800 bpd Qbot = 1800 bpd
Qbot = 2400 bpd Qbot = 2400 bpd Qbot = 2400 bpd
1000 1000 Qbot = 3200 bpd 1000 Qbot = 3200 bpd
Qbot = 3200 bpd
Qbot = 4000 bpd Qbot = 4000 bpd Qbot = 4000 bpd

500 500 500

0 0 0
0 100 200 300 400 500 600 0 100 200 300 400 500 600 0 100 200 300 400 500 600
D/I Spacing, feet D/I Spacing, feet D/I Spacing, feet
FIGURE 9: Oil content in injection water depends upon D/I spacing and the combination of production and drainage rates.

• Control the water drainage rate, such that the operational the DWS top completion. A considerable improvement in water
point (top and bottom rates) is below the “flip-flop” line to control and well productivity could be achieved by using DWS in
avoid reverse coning (environmental limit). a heavy oil reservoir.
• Select submersible pumps’ rating, such that that the both Because of the similarity between DWS and DWL wells, the
completions inflow could be lifted to surface (TPR limit). DWS nodal analysis method could be extended to DWL wells by
As the nodal analysis only gives well performance at a given evaluating the pressure vs. flow rate relationship for the top com-
time, the calculated production rates change with time. For a se- pletion for a range of bottom rates and D/I spacings. As an addi-
ries of timesteps, the model was used to predict the DWS well’s tional constraint, oil should be eliminated from the water drainage
recovery over the well’s life. At each timestep, oil depletion (incre- so the injected water is oil-free. This constraint is added to the
mental recovery) was estimated and the new oil pay thickness was DWL well performance plots, shown in Figures 11 and 12 for
updated from material balance. Future rates were predicted succes- drainage-injection rates 4000 B/D and 7200 BWPD, respectively.
sively in this manner. By using this method, it was concluded that In the plots, several upward lines represent TPR with variable
DWS technology could increase oil production rates more than water cut – each line for a different D/I spacing. The numbered in-
two-fold when compared with conventional wells. tercepts (1, 2, etc.) indicate the rate/pressure conditions of natural
Qin and Wojtanowicz(36) applied the DWS nodal analysis flow at the top completion. It is clear that DWL increases produc-
method to heavy oil reservoirs with bottomwater. They considered tion rate and reduces water cut compared to a conventional well.
a variable rate-dependent water cut at stabilized well production The results also show that larger D/I spacing improves DWL per-
and various ways to reduce the water cut with DWS. They found formance until D/I optimum value is reached.
out that the operational range of production rates with variable In Figures 11 and 12, lines AB and A’B’ represent the limit of
water cut was limited for heavy oil. Substantial value of pressure oil-free water drainage. For each D/I spacing and drainage rate,
drawdown needed for heavy oil inflow was diminished after the there is a minimum production rate (Qtop) needed to prevent oil
water breakthrough. Therefore, DWS bottomwater drainage was drainage. Figure 11 shows a case when the oil-free drainage limit
instrumental in stimulating inflow of heavy oil without water to is met for all values of D/I spacing. When the bottom rate (Qbot)
is 4000 B/D and the oil-free injection limitation line AB is below
2000 the IPR line, this means that reservoir energy is enough to sustain
1800 natural flow rates. In this case (the 850 B/D rate with 0.75 water
1600 cut), the DWL well’s production rate is 1050 B/D with 0.56 water
1400
cut for the 50 ft D/I spacing. Moreover, the DWL production rate
would increase to 1280 B/D with water cut reduced to 0.31 for the
1200
480 ft D/I spacing. For the two values of D/I spacing, the oil rates
Qbot, bpd

1000 would increase by 54% and 76% compared to the conventional

800
fwtop = 0, D/I = 10 ft
well’s production rate, respectively.
600 fobot = 0, D/I = 10 ft Figure 12 shows a case when the D/I spacing is restricted by the
400
fwtop = 0, D/I = 100 ft oil-free drainage limit. When bottom rate is 7200 B/D, the oil-free
fobot = 0, D/I = 100 ft
fwtop = 0, D/I = 480 ft
injection limitation line A’B’ intercepts the IPR plot, which means
200
fobot = 0, D/I = 480 ft that if the D/I spacing is greater than 100 ft, the reservoir energy
0 is not enough to lift the oil to the surface. To maintain a higher
0 100 200 300 400 500
production rate, a submersible pump should be installed at the top
Qtop, bpd completion – otherwise the oil will be injected to the aquifer.
FIGURE 10: Space DWL performance window. 6000
6000 IPR
IPR
Qbot = 4000 bpd
5500 TPRConv
TPRConv
5500 TPRDWS
TPRDWS
5000 Qbot = 7200 bpd TPRDWLD/I = 10 ft
5000 TPRDWLD/I = 10 ft
Pwf, psi

TPRDWLD/I = 50 ft
Pwf, psi

TPRDWLD/I = 50 ft
4500 1 – WC = 0.75 TPRDWLD/I = 100 ft
4500 1 – WC = 0.75 TPRDWLD/I = 100 ft
2 – WC = 0.50
2 – WC = 0.60 TPRDWLD/I = 200 ft
TPRDWLD/I = 200 ft 3 – WC = 0.43
3 – WC = 0.56 4000
4000 4 – WC = 0.36 TPRDWLD/I = 300 ft
4 – WC = 0.47 TPRDWLD/I = 300 ft
5 – WC = 0.40 TPRDWLD/I = 400 ft

3500 6 – WC = 0.35 TPRDWLD/I = 400 ft 3500

QtopMin (no oil
7 – WC = 0.31 QtopMin (no oil in the bottom)
8 – WC = 0.25 in the bottom)
3000 3000
0 200 400 600 800 1000 1200 1400 1600 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Qtop, bpd Qtop, bpd
FIGURE 11: Nodal analysis for DWL wells for water drainage-injection FIGURE 12: Nodal analysis for DWL wells for water drainage-injection
rate 4000 bpd. rate 7200 bpd.
June 2010, Volume 49, No. 6 43
Conclusions control in petroleum wells. The authors would like to express their
appreciation to the DWSTI-JIP for supporting this work.
The paper considers a new well system, DWL, with combined
benefits of the DWS and water injection wells. Effects of DWL
operational parameters: D/I spacing and top and bottom rates have NOMENCLATURE
been studied for a selected well-reservoir system. Also studied B = formation volume factor, rb/stb
was the effect of the oil-free water drainage limit on DWL perfor- C = compressibility, 1/psi
mance design. A DWS nodal analysis method has been modified D/I = D/I spacing, ft
and used to analyze DWL well performance. The study leads to the fobot = oil fraction at the bottom completions, fraction
following conclusions:
fwtop = water fraction at the top completion, fraction
• Larger spacing between water drainage and injection com-
Ho = thickness of the oil (pay) zone, ft
pletions (D/I) improves well performance by reducing water
Hw = thickness of the aquifer, ft
cut and increasing the oil rate.
• There is an optimum value of D/I, above which there is no IPR = inflow performance curve
improvement. Pb = bubblepoint pressure, psi
• Critical rate (water-free oil) at the upper completion increases Pref = pressure at the reference depth, psi
with increased D/I spacing. Pwf = well bottomhole flowing pressure, psi
• At optimum D/I spacing, the effect of water drainage on well Qbot = liquid drainage/injection (bottom) rate at the bottom
performance is strong; a two-fold increase of water drainage completions, B/D
would increase oil production by 80%. Qtop = liquid production rate at the top completion, B/D
• For each DWL system, there is a combination of top produc- Ro = radius of the oil (pay) zone, ft
tion rate, bottom drainage-injection rate and D/I spacing that Rs = solution gas/oil ratio, Mscf/stb
would result in water-free oil production. The observation is Rw = radius of the aquifer, ft
consistent with previous studies of DWS systems. TPRConv = tubing performance curve for conventional well
• For a given D/I spacing, an effective increase of oil pro- TPRDWL = tubing performance curve for DWL well
duction requires synchronized increases in production and TPRDWS = tubing performance curve for DWS well
drainage rates. A sole increase of the top production rate is WC = water cut, fraction
not effective because it would result in higher water cut. µ = viscosity, cp
• A DWL well could work effectively for a small D/I spacing. ρsc = surface density, lbm/ ft3
This means the system may be feasible in reservoirs with rel-
atively thin bottomwater. The minimum water thickness re- SI Metric Conversion Factors
lates to the optimum value of D/I spacing.
• DWL wells could significantly increase oil production rates B/D × 1.589 873 E−01 = m3/d
in comparison to conventional wells. cp × 1.0 E−03 = Pa·s
• DWL wells are inferior compared to DWS wells because ft × 3.048 E−01 = m
oF × (oF +459.67)/1.8 E+00 = K
their production rates are limited by D/I spacing and usu-
ally smaller than that of DWS wells. However, they have the in × 2.54 E-02 = m
advantage of improved tubing performance with little or no lbm/ ft3 × 1.601 846 E+01 = kg/m3
water and cost savings of water lifting and handling. This as- md × 9.869 233 E−04 = µm2
pect of DWL-DWS comparison deserves a separate study. Mscf/stb × 1.801 175 E+02 = m3/m3
• Nodal analysis of DWL wells, presented here, with variable 1/psi × 1.450 377 E-01 = 1/kpa
water cut and oil-free water drainage limit has practical merit. psi × 6.894 757 E+00 = kpa
The method would help to design the D/I spacing and decide
if one or two downhole pumps were needed for the system.
References
1. Hernandez, J.C. 2007. Oil Bypassing by Water Invasion to Wells
Mechanisms and Remediation. PhD dissertation, Louisiana State
Future Work University and Agricultural and Mechanical College, Baton Rouge,
Long-term and cumulative effects of formation damage are im- Louisiana (09 July 2007).
portant to the DWL system. For example, factors such as total sus- 2. Shirman, E.I. 1998. Experimental and Theoretical Study of Dynamic
Water Control in Oil Wells. PhD dissertation, Louisiana State Univer-
pended solids and dispersed oil in the injection water would cause sity, Baton Rouge, Louisiana.
injectivity decline and reduce the half-life of the injection com- 3. Abou-Sayed, A.S., Zaki, K.S., Wang, G.G., and Sarfare, M.D. 2005. A
pletion. The addition of models describing injectivity decline at Mechanistic Model for Formation Damage and Fracture Propagation
the injection completion is needed to better describe DWL well During Water Injection. Paper SPE 94606 presented at the SPE Euro-
performance. pean Formation Damage Conference, Sheveningen, The Netherlands,
High-injection rate is required to make DWL work effectively, 25–27 May. doi: 10.2118/94606-MS.
which necessitates high-pressure at the injection completion. As 4. Swisher, M.D. and Wojtanowicz, A.K. 1996. In Situ-Segregated Pro-
a result, the risk of fracturing needs to be correlated with time- duction of Oil and Water—A Production Method With Environmental
Merit: Field Application. SPE Advanced Technology Series 4 (2): 51–
dependent skin damage at the injection completion. It seems that
58. SPE-29693-PA. doi: 10.2118/29693-PA.
fracturing potential is a design parameter of the DWL well. 5. Bowlin, K.R., Chea, C.K., Wheeler, S.S., and Waldo, L.A. 1997.
In order to evaluate the feasibility of DWL, a large number of Field Application of In-situ Gravity Segregation to Remediate Prior
different reservoir-aquifer systems need to be considered and an- Water Coning. Paper SPE 38296 presented at the SPE Western Re-
alyzed. The analysis requires dimensionless groups, rather than gional Meeting, Long Beach, California, USA, 25–27 June. doi:
case-by-case studies. Determination of dimensionless groups that 10.2118/38296-MS.
are specific for DWL is instrumental in carrying out the feasibility 6. Ramos, L., Luzardo, B., Rojas, L., Villavona, A., Guerrero, J., and
study. Hermes, G. 2003. Downhole Water Sink Completion Improves Pro-
duction Decline in Venezuelan Reservoirs with Strong Aquifer
Support. Paper SPE 84047 prepared for presentation at the SPE An-
nual Technical Conference and Exhibition, Denver, 5-8 October.
Acknowledgements 7. Ju, B., Dai, S., Fan, T., Wang, X., and Wu, H. 2005. An Effective
Method to Improve Recovery of Heavy Oil Reservoir with Bottom
This study is part of the LSU research program Downhole Water Water Drive. Paper IPTC 10521 presented at the International Petro-
Sink Technology Initiative (DWSTI) – Joint Industry Project (JIP) leum Technology Conference, Doha, Qatar, 21–23 November. doi:
aimed at research and development of DWS technology for water 10.2523/10521-MS.
44 Journal of Canadian Petroleum Technology
 8. Davies, D.R, Narayanasamy, R., Kristensen, B., and Somerville, 28. Dixon, S.A., Summers, D.M., and Surdam, R.C. 1989. Diagenesis
J.M. 2008. Analysis of Possible Applications of Dual ESPs—A Res- and preservation of porosity in Norphlet Formation (Upper Jurassic),
ervoir-Engineering Perspective. SPE Prod & Oper 23 (2): 257–267. southern Alabama. AAPG Bulletin 73 (6): 707–728. June, 1989.
SPE-99878-PA. doi: 10.2118/99878-PA. 29. Joshi, S.D. 1991. Horizontal Well Technology. Tulsa, Oklahoma: Pen-
9. Gunning, J., Paterson, L., and Poliak, B. 1999. Coning in dual nWell Publishing.
completed systems. J. Pet. Sci. Eng. 23 (1): 27–39. doi:10.1016/
30. Pang, S. and Sharma, M.M. 1997. A Model for Predicting Injectivity
S0920-4105(99)00006-6.
Decline in Water-Injection Wells. SPE Form Eval 12 (3): 194–201.
10. Ould-amer, Y., Chikh, S., and Naji, H. 2004. Attenuation of water
SPE-28489-PA. doi: 10.2118/28489-PA.
coning using dual completion technology. J. Pet. Sci. Eng. 45 (1–2):
109–122. doi:10.1016/j.petrol.2004.04.004. 31. Gilbert, W.E. 1954. Flowing and gas-lift well performance. API
11. Siemek, J. and Stopa, J. 2002. A Simplified Semi-Analytical Model for Drilling and Production Practice 20 (1954): 126-157.
Water-Coning Control in Oil Wells with Dual Completions System. 32. Brown, K.E. and Beggs, H.D. 1977. The Technology of Artificial Lift
Journal of Energy Resources Technology 124 (4): 246–252. Methods. Tulsa, Oklahoma: PennWell Books.
12. Johns, R.T., Lake, L.W., Ansari, R.Z., and Delliste, A.M. 2005. Pre- 33. Mach, J., Proano, E., and Brown, K.E. 1979. A Nodal Approach for
diction of Capillary Fluid Interfaces During Gas or Water Coning Applying Systems Analysis to the Flowing and Artificial Lift Oil or
in Vertical Wells. SPE J. 10 (4): 440–448. SPE-77772-PA. doi: Gas Well. Paper SPE 8025 available from SPE, Richardson, Texas.
10.2118/77772-PA.
13. Utama, F.A. 2008. An Analytical Model to Predict Segregated Flow 34. Golan, M. and Whitson, C.H. 1986. Well Performance. Boston, Mas-
in the Downhole Water Sink Completion and Anisotropic Reservoir. sachusetts: International Human Resources Development Corporation
Paper SPE 120196 presented as part of the student paper contest at (IHRDC).
the SPE Annual Technical Conference and Exhibition, Denver, 21–24 35. Beggs, H.D. 1991. Production Optimization Using NODAL Analysis.
September. Tulsa, Oklahoma: OGCI–Petroskills.
14. Wojtanowicz, A.K. and Xu, H. 1992. A New Method to Minimize 36. Qin, W. and Wojtanowicz, A.K. 2007. Well Performance Analysis for
Oilwell Production Water Cut Using A Downhole Water Loop. Paper Heavy Oil With Water Coning. Paper CIPC 2007-162 presented at the
CIM 92-13 presented at the 43rd Annual Technical Conference of the Canadian International Petroleum Conference, Calgary, 12–14 June.
Petroleum Society of CIM, Calgary, 7–10 June. doi: 10.2118/2007-162.
15. Wojtanowicz, A.K. and Shirman, E.I. 1996. An in-situ method for
downhole drainage-injection of formation brine in a single oil pro-
ducting well. In Deep Injection Disposal of Hazardous and Industrial
Waste, ed. J.A. Apps and C. Tsang, 403–420. New York: Academic This paper (2008-173) was accepted for presentation at the 9th Canadian
Press. International Petroleum Conference (the 59th Annual Technical Meeting
16. Smith, C.R. and Pirson, S.J. 1963. Water Coning Control in Oil Wells of the Petroleum Society), Calgary, 17-19 June, 2008, and revised for pub-
by Fluids Injection. SPE J. 3 (4): 314–326; Trans., AIME, 228. SPE- lication. Original manuscript received for review 11 June 2008. Revised
613-PA. doi: 10.2118/613-PA. paper received for review 1 April 2010. Paper peer approved 7 April 2010
17. Kjos, T., Sangesland, S., Michelet, J.F., and Kleppe, J. 1995. Down- as SPE Paper 138402.
Hole Water-Oil Separation and Water Reinjection Through Well
Branches. Paper SPE 30518 presented at the SPE Annual Conference
and Exhibition, Dallas, 22–25 October. doi: 10.2118/30518-MS.
�����������������������������������������������������������������������
18. Paige, R.W.,
������������������������������������������������������������
Murray, L.R., Martins, J.P., and Marsh, S.M. 1995. Op- Authors’ Biographies
timising Water Injection Performance. Paper SPE 29774 presented at
the Middle East Oil Show, Bahrain, 11–14 March. doi: 10.2118/29774- Lu Jin is a Ph.D. candidate in the Craft and
MS. Hawkins Petroleum Engineering Depart-
������������������������������������������������������������������������
19. Detienne, ����������������������������������������������������������
J-L., Danquigny, J., Lacourie, Y., and Espy, M. 2002. Pro- ment at the Louisiana State University. He
duced water re-injection on a low permeability carbonaceous reservoir. worked as a research assistant at the China
Paper SPE 78482 presented at the Abu Dhabi International Petroleum
University of Petroleum from 2005 – 2007.
Exhibition and Conference, Abu Dhabi, UAE, 13–16 October. doi:
10.2118/78482-MS. Jin holds a B.S. degree in petroleum engi-
20. Singh, K. 2002. Designing a Produced Water Re-Injection Program in neering from Daqing Petroleum Institute in
Bekapai Field. Paper SPE 101938 presented at the SPE Asia Pacific China and an M.S. degree in petroleum en-
Oil and Gas Conference and Exhibition, Melbourne, Australia, 8–12 gineering from Louisiana State University.
October. doi: 10.2118/101938-STU.
21. Abou-Sayed, A.S., Zaki, K.S., Wang, G., Sarfare, M.D., and Harris,
M.H. 2007. Produced Water Management Strategy and Water Injec- Andrew K. Wojtanowicz is Texaco envi-
tion Best Practices: Design, Performance, and Monitoring. SPE Prod ronmental chair and professor in the Craft
& Oper 22 (1): 59−68. SPE-108238-PA. doi: 10.2118/108238-PA.
and Hawkins Petroleum Engineering De-
22. Inikori, S.O. 2002. Numerical Study of Water Coning Control with
Downhole Water Sink (DWS) Well Completions in Vertical and Hor- partment at the Louisiana State University.
izontal Wells. PhD dissertation, Louisiana State University, Baton Wojtanowicz is a registered petroleum and
Rouge, Louisiana (19 March 2002). environmental engineer in Louisiana and
��������������������������������������������������������������������
23. Arslan, O.,
��������������������������������������������������������
White, C.D., and Wojtanowicz, A.K. 2004. Nodal Anal- petroleum engineer in Oklahoma. He has
ysis for Oil Wells With Downhole Water Sink Completions. Paper held faculty positions at the New Mexico
CIPC 2004-242 presented at the Canadian International Petroleum Institute of Mining and Technology and the
Conference, Calgary, 8–10 June. doi: 10.2118/2004-242. AGH University of Science and Technology
��������������������������������������������������������������������
24. Arslan, ��������������������������������������������������������
O. 2005. Optimal Operating Strategy for Wells with Down- in Krakow, Poland. A UN expert in drilling
hole Water Sink Completions to Control Water Production and
engineering, Wojtanowicz has also worked as a drilling engineer,
Improve Performance. PhD dissertation, Louisiana State University
and Agricultural and Mechanical College, Baton Rouge, Louisiana drilling supervisor and drilling fluids technologist in Europe and
(May 2005). Africa. In 1992 – 1993 he worked for Conoco Inc. as an envi-
25. Song, F., Hou, J., and Su, N. 2009. Model building for Chang-8 low ronmental research fellow developing the dewatering technology
permeability sandstone reservoir in the Yanchang formation of the for closed-loop drilling systems. In 2008, Wojtanowicz was a vis-
Xifeng oil field. Mining Science and Technology (China) 19 (2): 245– iting distinguished professor of petroleum engineering at the Petro-
251. doi:10.1016/S1674-5264(09)60047-8. leum Institute in Abu Dhabi, UAE. Since 1997, Wojtanowicz has
�������������������������������������������������������������������������
26. Hu, K. and Dietrich, J. 2009. Hydrocarbon Reservoir Potential in Car- directed a JIP concerning research, development and field imple-
boniferous Sandstones in the Maritimes Basin, Eastern Canada. Proc., mentation of DWS. He holds an M.S. degree in mining/petroleum
Frontiers + Innovation—2009 CSPG CSEG CWLS Convention, Cal-
engineering and Ph.D. and D.Sc. degrees in petroleum engineering
gary, 4–8 May, 746–749.
�����������������������������������������������������������������
27. Neasham, J.W. 1977. The Morphology of Dispersed Clay in Sand-
from the AGH University of Science and Technology in Krakow,
stone Reservoirs and Its Effect on Sandstone Shaliness Pore Space and Poland. He has also served as the 2003 – 2004 SPE distinguished
Fluid Flow Properties. Paper SPE 6858 presented at the SPE Annual lecturer of the DWS technology, received the SPE 2008 Distin-
Fall Technical Conference and Exhibition, Denver, 9–12 October. doi: guished Achievement Award for Petroleum Engineering Faculty
10.2118/6858-MS. and the SPE 2003 Health Environment and Safety Award.
June 2010, Volume 49, No. 6 45