Inflow Control Devices—Raising Profiles

Tor Ellis Maximizing reserves recovery using horizontal wells requires management
Marathon Petroleum Company (Norway) LLC
Stavanger, Norway of fluid flow through the reservoir. One increasingly popular approach is to use
inflow control devices that slow water and gas encroachment and reduce the
Alpay Erkal
Houston, Texas, USA amount of bypassed reserves.

Gordon Goh
Kuala Lumpur, Malaysia

Timo Jokela
Svein Kvernstuen
Edmund Leung
Terje Moen
Francisco Porturas
Torger Skillingstad
Paul B. Vorkinn
Stavanger, Norway
Heel Toe
Anne Gerd Raffn
Abingdon, England

Oilfield Review Winter 2009/2010: 21, no. 4.

Copyright © 2010 Schlumberger.
For help in preparation of this article, thanks to Ewen Connell,
Rosharon, Texas; and Mary Jo Caliandro, Sugar Land, Texas.
ECLIPSE, PeriScope, Petrel, ResFlow and ResInject are
marks of Schlumberger.
1. Al-Khelaiwi FT, Birchenko VM, Konopczynski MR and
Davies DR: “Advanced Wells: A Comprehensive
Approach to the Selection Between Passive and Active
Inflow Control Completions,” paper IPTC 12145,
presented at the International Petroleum Technology
Conference, Kuala Lumpur, December 3–5, 2008.
2. Raffn AG, Zeybek M, Moen T, Lauritzen JE, Sunbul AH,
Hembling DE and Majdpour A: “Case Histories of
Improved Horizontal Well Cleanup and Sweep Efficiency
with Nozzle-Based Inflow Control Devices in Sandstone
and Carbonate Reservoirs,” paper OTC 19172, presented
at the Offshore Technology Conference, Houston,
May 5–8, 2008.
3. Wellhead penetrations are holes in the wellhead through
which power cables and hydraulic lines must pass to
reach a device downhole. The number that can be drilled
is limited by surface area and by the amount of material
that can be removed from the wellhead without
compromising its integrity.
4. Jokela T: “Significance of Inflow Control Device (ICD)
Technology in Horizontal Sand Screen Completions,”
Bachelor thesis, Det Teknisk-Naturvitenskapelige
Fakultet, Stavanger, May 30, 2008.
> Heel-toe effect. Pressure losses along a horizontal wellbore in a homogeneous formation cause the
flowing tubing pressure to be lower at the well’s heel than at the toe. In time, and long before oil (green) from
sections near the toe arrives at the wellbore, water (blue) or gas (red) is drawn to the heel (top), resulting
in an early end to the well’s productive life. Inflow control devices inside sand screen assemblies equalize
the pressure drop along the entire length of the wellbore, promoting uniform flow of oil and gas through the
formation (bottom) so that the arrivals of water and gas are delayed and simultaneous.

30 Oilfield Review
Extended-reach and multilateral horizontal drill-
ing techniques significantly increase wellbore/
reservoir contact. This augmented contact allows
operators to use less drawdown pressure to achieve
production rates equal to those of conventional
vertical or deviated wells. The ability to optimize
results from these standard configurations through
better reservoir fluids management has been
greatly enhanced by the development of remotely
operated inflow control valves and chokes. These
devices enable engineers to adjust flow from indi-
vidual zones that are over- or underpressured or
from those producing water or gas that may be det-
rimental to overall well productivity.
Long sections drilled horizontally through a
single reservoir, however, present a different set
of challenges. In homogeneous formations, sig-
nificant pressure drops occur within the open-
hole interval as fluids flow from TD toward the
heel of the well. The result may be significantly
higher drawdown pressures at the heel than at
the toe. Known as the heel-toe effect, this differ-
ential causes unequal inflow along the well path
and leads to water or gas coning at the heel
(previous page). A possible consequence of this
condition is an early end to the well’s productive
life and substantial reserves left unrecovered in
the lower section of the well.
Water or gas breakthrough anywhere along the
length of the wellbore can also result from reser-
voir heterogeneity or from differences in distances
between the wellbore and fluid contacts. Pressure
variations within the reservoir caused by reservoir
compartmentalization or by interference from
production- and injection-well flow can also
lead to early breakthrough.1 Carbonate reservoirs,
because they tend to have a high degree of fractur-
ing and permeability variation, are especially vul-
nerable to uneven inflow profiles and accelerated
water and gas breakthroughs.2
Many completions designed for long-reach
wells include sand control systems. If these com-
pletions do not have isolation devices such as
packers, annular flow can lead to severe erosion
and plugging of sand screens. In the past such
annular flow effects were countered with gravel
packs or expandable sand screens. But gravel
packs often reduce near-wellbore productivity.
Expandable sand screens require complex instal-
lation procedures and are prone to collapse later
in the well’s life.
In traditional completions the solution to an
increase in water or gas cut is to reduce the
choke setting at the wellhead. This lowers draw-
down pressure, resulting in lower production
Winter 2009/2010
Nozzle-Type ICD
Helical-Channel ICD
> Leading ICD types. Fluid from the formation (red arrows) flows through multiple
screen layers mounted on an inner jacket, and along the annulus between the
solid basepipe and the screens. It then enters the production tubing through
a restriction in the case of nozzle- and orifice-based tools (top), or through a
tortuous pathway in the case of helical- and tube-based devices (bottom).
rates but higher cumulative oil recovery. However, ated wells and in several types of reservoirs.4
this simple solution generally does not work in These devices are usually part of openhole com-
wells drilled at high angles. pletions that also include sand screens. In addi-
In wells completed with “intelligent” technol- tion, ICD completions often use packers to
ogy, operators may shut off or reduce flow from segment the wellbore at points of large permea-
offending zones using remotely actuated down- bility contrast. This strategy combats water con-
hole valves. But horizontal wells designed to opti- ing or gas cresting through fractured zones, halts
mize reservoir exposure are often poor candidates annular flow between compartments and allows
for such strategies. Extremely long wells often for isolation of potential wet zones.
have many zones. The limit on the number of ICDs are also effective in reservoirs where
wellhead penetrations available may render it their ability to regulate inflow rates creates a suf-
impossible to deploy enough downhole control ficient pressure drop at the toe of the wellbore
valves to be effective.3 Additionally, such comple- for the reservoir fluid to flow or lift filtercake and
tions are expensive, complex and fraught with other solids to the surface.
risk when installed in long, high-angle sections. This article describes various ICD designs
As a consequence, operators often choose to and how they are modeled to suit particular
produce these multiple-zone wells using isolating applications. Case histories from Asia, the North
devices such as swellable packers. To mitigate Sea and the Middle East illustrate how these pas-
OSWIN09/10—Rick, story #2—Figure 03
crossflow and to promote uniform flow through the sive devices enable operators to increase well life
reservoir, they have turned to passive inflow control and ultimate recovery.
devices (ICDs) in combination with swellable pack-
ers. By restraining, or normalizing, flow through Velocity Control
high-rate sections, ICDs create higher drawdown Inflow control devices are included in the hard-
pressures and thus higher flow rates along the bore- ware placed at the formation/borehole interface.
hole sections that are more resistant to flow. This They use a variety of flow-through configurations
corrects uneven flow caused by the heel-toe effect including nozzles, tubes and labyrinth helical
and heterogeneous permeability. channels (above). These devices are intended to
Whether intended for injection or production, balance the well’s inflow profile and minimize
ICDs have applications in horizontal and devi- annular flow at the cost of a limited, additional
31
Production rate per length, bbl/d/ft 12 All ICDs are permanent well components and
10 are rated by their flow resistance. Essentially, the
rating signifies the total amount of pressure drop
8 created across the device with a reference fluid
property and flow rate. Nozzle- and orifice-type
6
devices enjoy an advantage over channel ICDs: The
4 nozzle size and therefore the ICD rating can be
adjusted easily at the wellsite, before deployment,
2
in response to real-time drilling information. The
0 designs of ICDs are typically based on predrilling
Heel Measured depth Toe reservoir models, and changing the rating of
> Reducing the influence of high-flow-rate areas. In a heterogeneous model ICDs reduced the fluid channel- or tube-type ICDs is more difficult, time-
inflow rate (blue) at the heel (within orange circle) to half that predicted for a screen-only completion consuming and not easily done on location.
(red). However, they increased the inflow rate from the lower two-thirds of the well (within green oval)
including the toe.
Modeling: Static and Dynamic
Historically, nozzle-type ICDs have been designed
using a ratio of the pressure drop at the device
pressure drop between formation and wellbore Helical devices force the fluid to flow through inlet as calculated by Bernoulli’s equation to the
(above).5 They accomplish this by changing the channels that have a preset diameter and length. average formation drawdown pressure derived
flow regime from the Darcy radial flow within the The differential pressure provided by these from Darcy’s equation. When this ratio is close
reservoir to a pressure-drop flow within the ICD. devices is determined by friction against the to unity, the ICDs are self-regulating.
Each of the basic types of ICD uses a different channel surface and is a function of the flow rate The designs based on these assumptions are
operating principle to achieve this backpressure. and fluid properties.6 simple and effective in horizontal wells with rela-
The pressure drop within a nozzle-type ICD is This viscosity sensitivity may result in ineffi- tively high productivity indexes (PIs) and mini-
a function of flow rate as the fluid passes through ciencies, however, when backpressure at break- mal flow restrictions. The same number and size
restricting ports inserted into the basepipe or through streaks is not significantly greater than of ICD nozzles are assigned to each joint of tubing
into the housing outside the basepipe. As that in areas producing oil that has lower viscos- from toe to heel. This approach usually improves
Bernoulli’s principle states, the pressure drop ity because of entrained water and gas. flow uniformity through the reservoir, counters
through a port increases as the square of the Orifice ICDs are similar to nozzle-based much of the heel-toe effect and balances flow
fluid-flow velocity, which increases as the port devices. Backpressure is generated by adjusting from heterogeneous zones.
opening diameter decreases. the number of orifices of known diameter and But these goals may be achieved at the cost of
Nozzle-based ICDs are self-regulating com- flow characteristics in each tool. The orifices are overly restricting the flow from high-permeabil-
pletion components. That is, given the uncer- inserted in a jacket around a basepipe. Another ity, high-rate oil zones. Additionally, this method
tainty of permeability variations along the option consists of an annular chamber on a stan- eliminates flexibility for zonal control and does
horizontal section of the wellbore, each ICD joint dard oilfield tubular. The reservoir fluid is pro- not include the effects of variation in zonal poros-
will behave independently of the local heteroge- duced through a sand screen into a flow chamber ity thickness, saturation and oil/water contacts.
neity and fluid type, which may change over time. from which it then flows through parallel tubes to For more-precise designs, engineers can turn
The former may occur because of compaction or the production string. Like helical-channel ver- to modeling using tools such as the Schlumberger
subsidence around the wellbore, and the latter as sions, these tubular ICDs also rely on friction to ICD Advisor software. Using steady-state systems,
OSWIN09/10—Rick, story #2—Figure 04A
a result of the inevitable influx of water or gas. create a pressure drop that is determined by the the experts model wellbore hydraulics to deter-
As fluids more mobile than oil, such as water tube’s length and inside diameter. Some recently mine tubing and annulus flow, flow direction and
or gas, flow into the wellbore at higher velocities introduced ICDs are best described as tube- completion-specific flow correlations. Reservoir
than that of oil, backpressure at the point of channel and orifice-nozzle combinations. flow is determined through PI models.
ingress increases. This slows the flow of forma- Some wells may benefit from a recent ICD Incorporating data from offset wells, LWD
tion fluids through high-permeability intervals or innovation with a valve that reacts to an upstream tools, geology and other sources, engineers opti-
streaks, preventing water or gas from reaching or downstream change in pressure. The autono- mize well designs by determining near-wellbore
the wellbore ahead of reserves in less permeable mous ICD adjusts the flow area when the pres- performance at a specific time. They test various
sections of the formation. sure differential across it changes. scenarios and completion designs to balance flow,
decrease water cut, control gas/oil ratios and, by
5. Alkhelaiwi FT and Davies DR: “Inflow Control Devices: Petroleum Technology Conference, Kuala Lumpur,
Application and Value Quantification of a Developing December 3–5, 2008. varying the number of isolation packers per well
Technology,” paper SPE 108700, presented at the 7. Maggs D, Raffn AG, Porturas F, Murison J, Tay F, section, verify the effects of annular compart-
International Oil Conference and Exhibition in Mexico, Suwarlan W, Samsudin NB, Yusmar WZA, Yusof BW,
Veracruz, Mexico, June 27–30, 2007. mentalization (next page, top right). In doing so,
Imran TNOM, Abdullah NA and Mat Reffin MZB:
6. Al Arfi SA, Salem SEA, Keshka AAS, Al-Bakr S, Amiri AH, “Production Optimization for Second Stage Field they are determining the impact of packer den-
El-Barbary AY, Elasmar M and Mohamed OY: “Inflow Development Using ICD and Advanced Well Placement sity on production in the presence of ICDs.
Control Device an Innovative Completion Solution from Technology,” paper SPE 113577, presented at the SPE
‘Extended Wellbore to Extended Well Life Cycle’,” Europec/EAGE Annual Conference and Exhibition, Rome, Finally, they determine the number and sizes of
paper IPTC 12486, presented at the International June 9–12, 2008. nozzles to be deployed in each compartment.

32 Oilfield Review
 The advantages of this steady-state modeling Oil Rate, Gas Rate, Water Rate, Water Cut, BHP,
are quick designs, high-resolution near-wellbore bbl/d Mcf/d bbl/d % psi
models and quantification of the upside potential Open hole
of oil production with decreased water and gas 7,759 698 2,411 23.7 3,794
cut. However, this approach delivers only a snap- 3 x 4 mm, second joint
shot in time and cannot predict or quantify the
8,821 798 1,263 12.5 3,752
value of delaying water or gas breakthrough. This
3 x 4 mm, joint
step requires the investment of considerably
more time and effort to perform dynamic simula- 9,290 837 762 7.6 3,740

tions, such as using the Petrel software reservoir
engi­neering workflow in conjunction with the
Multi-Segmented Well (MSW) model of the
ECLIPSE reservoir simulator.
This model treats the well as a series of seg-
ments and allows engineers to model indepen-
dently three-phase flow, liquid-gas holdup and
the implications of using ICDs and flow control
valves over the life of the well. Each modeled seg-
ment can be angled upward or downward and can
contain different fluids to account for an undulat-
ing well path.
Ideally, dynamic modeling is accomplished
using a full-field geologic model. But often this is
not practical, even with high-performance, paral-
lel computing hardware, because of the long com-
putational simulation necessary to complete the
runs. A more practical solution begins by extract-
ing a sector model from the ECLIPSE full-field
simulation model that can extract flux, pressure
or no-flux boundary conditions to reduce dynamic
simulation time while honoring the geologic het-
erogeneity and interference from nearby wells.
Reducing the number of geology grid cells
offers more-sensitive runs. Furthermore, the sec-
tor model can be combined with the full-field
model. The area of interest is then modified to
refine the grid and upscale from the geologic
model, and the well trajectory is loaded. The
segmented well with ICDs and packers is then
> Packer density impact. By isolating compartments within heterogeneous
formations, it is possible to reduce water cut and sand production
considerably while maintaining or, as in this case, increasing oil production.
Reservoir engineers first test the model for optimum packer density before
determining the number and sizes of ICDs needed for the completion. In this
example, installing three 4-mm-diameter nozzles per joint reduced water cut
to 7.6% compared with 23.7% in an openhole completion. At the same time
production increased from 7,760 to 9,290 bbl/d [1,233 to 1,476 m3/d] without
a significant increase in bottomhole pressure (BHP). When the same nozzle
configuration was used on every second joint, water cut was reduced to 12.5%.
The wells were part of a second-stage devel- [610-m] lateral and one 1,000-ft [305-m] lateral
opment of a mature field; challenges included a to be placed precisely with respect to fluid con-
stacked sand reservoir with uncertain dips and tacts and reservoir boundaries (below). This
unconsolidated sands. The company also sought option also required openhole sand screens and
to avoid formation damage during drilling, mini- passive ICDs to enable production contribution
mize drilling costs, and maximize production and from the entire length of the wellbore.
drainage of remaining reserves while minimizing Rotary steerable systems were used to drill
water cut.7 the wells as far from the water contact as possi-
While the horizontal well option was less ble to delay water production and as close to the
costly than an alternative plan that included overlying shale boundary as possible to capture
drilling three deviated wells, it was technically attic oil. An LWD assembly that included a deep
more challenging. It required one 2,000-ft azimuthal resistivity distance-to-boundary tool—
Gas zone Oil zone Water zone
000
–4,

created in the ECLIPSE simulation.

The Sweet Spot

The advantage gained by the ability to quickly –3,875
OSWIN09/10—Rick, story #2—Figure 06
incorporate new data into completions was dem-
onstrated in a field offshore Malaysia. Having
A
opted, for economic reasons, to drill two long
horizontal wellbores into a target characterized
by a thin oil rim with a gas cap and active aquifer, B –3
,75
0
the operator included ResFlow ICDs in the com-
0 m 750
pletion design. Because they are nozzle-type
devices, it is easy to adjust and optimize them on 0 ft 2,500
location in response to new LWD data without > Well placement. As part of an ongoing field expansion, this small area within a field offshore Malaysia
costing valuable rig time. was targeted for development using one 2,000-ft lateral (A) and one 1,000-ft lateral (B). The thin oil rim
(green) is bounded by a strong waterdrive (blue) and a gas cap (red).Depth contours are labeled in
feet. (Adapted from Maggs et al, reference 7.)

Winter 2009/2010 33
 Critical Components
Vilje
Besides their ability to enhance drainage effi-
ciency and boost cumulative oil recovery, ICDs
Production line offer the industry relatively inexpensive, low-risk
Gas lift line Alvheim FPSO East Kameleon components for technology-driven strategies. They
Water injection can be easily added to development programs that
and disposal line
Umbilical include sand control and horizontal wells.
East riser In the Norwegian sector of the North Sea, engi-
base
neers at Marathon Petroleum Company (Norway)
West riser LLC concluded that the recoverable reserves in
base Kneler A
South the relatively thin oil columns of the Alvheim and
Boa riser base
Volund fields were directly and consistently linked
to the amount of net pay exposed to the wellbore
Kneler B (left). To establish maximum contact, Marathon
therefore drilled single-, dual- and trilateral wells
with horizontal sections ranging in length from
1,082 to 2,332 m [3,550 to 7,651 ft].
The Marathon team realized that to fully
exploit the benefits of the correlation of recover-
Volund able reserves to net feet of reservoir contact, it was
important that the entire length of the comple-
tions contribute to production. Early in the project
> Layout of Alvheim and Volund fields in the Norwegian area of the North Sea. [Courtesy of Marathon
they decided to use both ResFlow nozzle-type ICDs
Petroleum Company (Norway) LLC.]
and helical-type ICDs in all production wells—a
total of 10 wells at Alvheim and 1 at Volund.
the PeriScope bed boundary mapper—was used Production from both wells compared favor- As a result of this technology-based approach
to steer a smooth wellbore trajectory. ably with that from other deviated wells in the and the favorable geology, Marathon has increased
The longer lateral came online without the area drilled conventionally through the stacked its booked reserves at Alvheim from 147 million
assistance of gas lift at 2,300 bbl/d [366 m3/d] of sands of the field. However, even including costs to 201 million bbl [23 million to 32  million m3]
oil and a water cut of about 10%. This level of of the additional technology—rotary steering sys- of oil and from 196 to 269 Bcf [5.5  billion to
water production was expected from mobile tem, LWD and ResFlow ICDs—the overall project 7.6 billion m3] of gas.
water in the oil rim and is not associated with cost was 15% less than it would have been using The fields have been in production less than
breakthrough from the water leg. The second traditional well construction methods. In addi- two years, and the completions include numerous
well, drilled updip from the first, required gas tion, increased sweep efficiency gained by well technologies, making it difficult to attribute
lift to clean up and initially produced about placement and ICDs has increased the asset value specific results to a single methodology. However,
1,900 bbl/d [302 m3/d] with 20% water cut. by an estimated 100,000 bbl [16,000 m3] of oil. overall water production at the Alvheim floating,
production, storage and offloading (FPSO) facility
is less than originally expected. A good example is
30,000 Actual oil production rate 80 the 24/6-B-1CH well, which has a 13-m [43-ft] oil
Predicted oil production rate
Actual water cut 70
column and an active aquifer. The well has been
25,000 Predicted water cut produced at higher rates than originally planned
60 without significant onset of, or increase in, water
Oil production rate, bbl/d

20,000 production (left). Both these outcomes, though

50
Water cut, %

their causes are inconclusive, suggest ICD success

15,000 40 in maintaining an even flow profile.
OSWIN09/10—Rick, story #2—Figure 07 When one completion planned as a single lat-
30
10,000 eral evolved to a trilateral, engineers also learned
20 a valuable lesson concerning planning for the use
5,000 of ICDs and multilateral installations. Because
10
the actual completion departed from the original
0 0 plan, the flow rate was different than predicted.
Jun 16, Aug 08, Sept 24, Nov 13, Jan 02, Feb 21, Apr 12, Jun 01, Jul 21, Sept 09, The ICDs chosen for these installations were of a
2008 2008 2008 2008 2009 2009 2009 2009 2009 2009
design type that could not be readily changed,
Date
and thus optimized, on location. The result was
> Production improvements. In Well 24/6-B-1CH of the Alvheim field, the 13-m oil column with an active
gas and water coning earlier than expected in
aquifer was produced at a higher drawdown than originally planned. As shown in the graph, the
resulting higher production volumes were achieved without significantly increasing water cut over both laterals.
predicted values, which is indicative, if not conclusive, that an even inflow profile was achieved.

34 Oilfield Review
 Recently, a different operator expanded on
the application of ICD completions, not to coun-
Production, bbl/d
ter the effects of uneven inflow profiles but to
counter uneven pressure profiles. In one horizon-
tal well extending more than 5,200 ft [1,600 m]
through a high-permeability reservoir in a large
Middle East field, the pressure differential
between heel and toe was 200 psi [1.4 MPa] with
the higher pressure at the heel.8
An initial production log confirmed what was
expected given the pressure profile: A downward
crossflow of fluids from heel to toe was detected
during a shut-in logging pass. In addition, produc-
tion logging measurements acquired while the
well was flowing showed water moving downward
from the heel and oil flowing to the surface. Logs
also indicated production was coming from only
the first 10% of the lateral.9
Based on the results of static modeling, the
operator recompleted the well with 22 ResFlow
ICDs and, to segment the well, seven swellable
packers on the production string. Logs acquired
after recompletion indicated that crossflow had
been eliminated and production was coming
from the entire lateral. Water cut was reduced
from 30% to less than 10%, and the actual inflow
profile matched that predicted by the static ICD
model (above right).10
A Clean Start
Predictably, it has been observed that the differ-
ence in pressure drops between the heel and toe
caused by friction losses in an openhole horizon-
tal well increases with wellbore length. This
disparity can lead to the filtercake being prefer-
entially lifted from the wellbore wall at the heel
and to poor inflow performance caused by corre-
spondingly higher skin at the toe.
Studies have shown that in relatively high-
permeability environments, the best cleanup
results—removal of filtercake after drilling or
completion—are obtained through proper chem-
ical treatment and extended flowback with high
rates.11 In 2006 Saudi Aramco completed two test
wells equipped with ICD systems, one in a sand-
stone formation and the other in carbonate rock.
In the sandstone there were concerns over water
and gas coning through high-permeability
streaks, and the operator sought to decrease the
impact of the heel-toe effect to improve cleanup
and sweep efficiency. The 8½-in. openhole com-
pletion included 5½-in. screens with ResFlow
ICD nozzles on every joint of tubing. For compart-
mentalization and better inflow control, small,
swellable elastomer packers were placed on
every second joint. The horizontal section was
2,540 ft [775 m] long.
Winter 2009/2010
Simulated flow profile
ICD completion actual flow profile
Heel Measured depth Toe
1 ICD 1 ICD 2 ICDs 2 ICDs 3 ICDs 3 ICDs 4 ICDs 5 ICDs
ICD Swellable packer
> Inflow profile from production log measurements. After installation of ICDs and swellable packers,
production logging tools were run to acquire an inflow profile along the length of the well at low,
medium and high flow rates. The inflow profile shown was obtained with the well flowing at the
medium rate. Crossflow evident in earlier logs has been eliminated and flow contribution is evident
from the entire lateral. The actual inflow profile (green) was very close to the simulated one (red).
(Adapted from Krinis et al, reference 8.)
The well was produced at 6,000 to 7,000 bbl/d well was logged again. The new data indicated an
[953 to 1,113 m3/d] for 4 months. A production improved flow profile and the tool was able to
log was then acquired. The log data, as well as the travel an additional 350 ft [106 m]. Four hours
inability to get the tool within 650 ft [198 m] of later, the logging tool was run again, this time to
TD because of solids-laden mud filling the toe of within 50 ft [15 m] of TD (below). The rate was
the wellbore, indicated the well had not cleaned reduced to the original 6,000 to 7,000 bbl/d and
up despite the prolonged flow period. data from the final logging run indicated a
The flow rate was then increased to 9,000 to permanent change had been made to the
10,000 bbl/d [1,430 to 1,590 m3/d] for 4 h and the inflow profile.
10,000 Initial log at 6,000 bbl/d
Initial log at 9,000 bbl/d
9,000
Repeat log at 9,000 bbl/d
8,000 Repeat log at 6,000 bbl/d
7,000
Production rate, bbl/d
6,000
5,000
4,000
3,000
2,000 OSWIN09/10—Rick, story #2—Figure 09
1,000
0
9,000 9,300 9,600 9,900 10,200 10,500 10,800 11,100 11,400 11,700
Measured depth, ft
> Cleanup through higher rates. After the logging tool failed to reach TD and
log data indicated no production contribution from the toe after an initial flow
period of 4 months (red), the rate was increased to about 9,000 to 10,000 bbl/d
for 4 h and the production log was rerun. Log data yielded an improved flow
profile, and the tool was able to travel an additional 350 ft (gray). Four hours
later, the logging tool was run to within 50 ft of TD (green). The rate was
returned to about 6,000 to 7,000 bbl/d, and log data demonstrated a permanent
change had been made to the inflow profile (blue). (Adapted from Sunbul et al,
reference 12.)
8. Krinis D, Hembling D, Al-Dawood N, Al-Qatari S, 10. Krinis et al, reference 8.
Simonian S and Salerno G: “Optimizing Horizontal Well 11. Shahri AM, Kilany K, Hembling D, Lauritzen JE,
Performance in Nonuniform Pressure Environments Gottumukkala V, Ogunyemi O and Becerra Moreno O: “Best
Using Passive Inflow Control Devices,” paper OTC 20129, Cleanup Practices for an Offshore Sandstone Reservoir
presented at the Offshore Technology Conference, with ICD Completions in Horizontal Wells,” paper SPE
Houston, May 4–7, 2009. 120651, presented at the SPE Middle East Oil and Gas
9. Krinis et al, reference 8. Show and Conference, Bahrain, March 15–18, 2009.
35
 Standard ICD with ICD with Target Rate, into two segments; the injector and producer are
Screen Same Nozzle Different m3/d located in Segment 1 and the second oil producer
Size, 1.2 Configuration
cm/joint, in Segment 2.
in All Zones Nozzle Size The injector is a vertical well drilled through
Zone 1 5,800 4,604 3,570 3,500 the Not, Ile, Tilj and Åre 2 Formations and pro-
800 to 1,800 mD vides sweep and pressure support for two hori-
0.9 cm/joint
zontal producers. About 250 m [820 ft] deep, the
Zone 2 748 1,233 820 800
200 to 500 mD injection well is an openhole completion with
0.7 cm/joint ResInject injection control devices, sand screens
Zone 3 961 1,677 3,128 3,200 and a pack of resin-coated gravel to prevent
100 to 2,000 mD
2.2 cm/joint annular flow.
3 Engineers from Reslink and Statoil designed
Total injection rate, m /d 7,509 7,514 7,518 7,500
the system. They modeled injection rates
> Optimizing injector ICD design. The injection rates used in the different expected for three zones using different comple-
completion scenarios of the Stær structure demonstrate that the injectors tion techniques: standard screens alone, ICDs of
can be optimized based on permeability and nozzle design to obtain the
the same nozzle size and number of nozzles per
desired rates in each zone. (Adapted from Raffn et al, reference 13.)
joint, and different numbers of ICDs per joint
(left). The team chose to use the same nozzle
configuration along the entire wellbore instead of
Engineers suspected that the higher rates designing injection projects must consider per- specific ICD nozzle sizes and numbers for each
required to clean up the entire production interval meability contrasts, heel-toe effect, formation zone. This choice reflected the fact that while dif-
had exacerbated the heel-toe effect in the tradi- damage, creation of thief zones and injectivity ferent designs in each zone achieved target injec-
tional openhole completions. Modelers matched changes at the wellbore.13 tion rates, simulations supported maximum
the production log data to a static reservoir simu- Just as they do with inflow control, ICDs injection rates in the upper zones.14
lation and replaced the ICD completion in the address these challenges by balancing fluid out- These simulations were run to evaluate the
simulation with a standard screen completion. flow along the entire length of the injection well- economics of using ICD injectors on Stær and to
They then increased the rate in the standard bore. If the well has a high-permeability streak, select the nozzle design. Two static near-wellbore
screen completion to 15,000 bbl/d [2,400 m3/d]. the ICD self-regulating feature prevents a signifi- simulations were used to compare water distribu-
That simulation indicated an extreme heel- cant increase of local injection rate. This ability tion: The first was based on injection into the
toe effect: The toe was contributing only 25% as to automatically control fluid mobility results in matrix, including its permeability variations,
much production as the heel. By contrast, simu- better water distribution and pressure support and the second considered injection into a frac-
lated ICD completions with 15,000-bbl/d rates and thus enhanced areal and vertical sweep of oil tured zone.
showed better balance of the inflow, including reserves in all zones. It also delays water break- In the first case, the upper, high-permeability
much higher contribution from the toe.12 through, and because ICDs can control injection zone received an uneven share of the injected
These findings are significant in that they pressure and rate, there is minimal risk of near- water. However, with ICDs, peak outflow was
show ICD completions allow extended well wellbore fracturing. reduced by 50% and zones with lower permeabil-
lengths in both these formations without compro- These capabilities matched the management ity received more water. For the second static
mising the balancing effect or cleanup efficiency goals of the Statoil team planning the 2004 devel- model, a 12-m [39-ft], 20-D layer was added to
in the lower sections of the wells. That result opment of the Urd field—a satellite producing to
12. Sunbul AH, Lauritzen JE, Hembling DE, Majdpour A,
allows the operator to contact more formation the Norne FPSO vessel in the North Sea. Placed on Raffn AG, Zeybek M and Moen T: “Case Histories of
with fewer wellbores without fear of sacrificing production in 2005, the Urd oil field contains two Improved Horizontal Well Cleanup and Sweep Efficiency
with Nozzle Based Inflow Control Devices in Sandstone
cumulative production. heterogeneous structures: Svale and Stær, which and Carbonate Reservoirs,” paper SPE 120795,
OSWIN09/10—Rick, story
are 4#2—Figure 12 and 5.6 mi], respectively, from
and 9 km [2.5 presented at the SPE Saudi Arabia Section Technical
Symposium, Alkhobar, Saudi Arabia, May 10–12, 2008.
Reversing Direction the main field. The field was developed using three
13. Raffn AG, Hundsnes S, Kvernstuen S and Moen T: “ICD
Though they are called inflow control devices, subsea templates and pipelines for oil production, Screen Technology Used to Optimize Waterflooding in
ICDs are also used to manage fluid outflow in water injection and gas lift. Management goals for Injector Well,” paper SPE 106018, presented at the SPE
Production and Operations Symposium, Oklahoma City,
injection wells. In some cases modeling reveals the ICD injection system included Oklahoma, USA, March 31–April 3, 2007.
that it is more effective to place ICDs in the injec- • optimizing pressure support and sweep effi- 14. Raffn et al, reference 13.
tor well than in the producer. And in many ciency for all zones 15. Tachet E, Alvestad J, Wat R and Keogh K: “Improve
Steam Distribution in Canadian Reservoirs During
instances installing the devices in both the injec- • delaying water breakthrough in high-permea- SAGD Operations Through Completion Solutions,”
tor and producer wells is the best option. bility connected zones paper 2009-332, presented at the World Heavy Oil
Congress, Porlamar, Venezuela, November 3–5, 2009.
Injector wells often penetrate and give pres- • avoiding fractures that may dominate water 16. Fram JH and Sims JC: “Addressing Horizontal Steam
sure support to several reservoir intervals with distribution. Injection Completions Challenges with Chevron’s
Horizontal Steam Test Facility,” paper 2009-398,
varying characteristics. To avoid water break- The Stær structure was completed with one presented at the World Heavy Oil Congress, Porlamar,
through at production wells, reservoir engineers injector containing ICDs and two horizontal oil Venezuela, November 3–5, 2009.
producers containing intelligent technology for 17. Fram and Sims, reference 16.

control of three zones. The reservoir is divided

36 Oilfield Review
simulate a fracture. When ICDs were included in
the model, the fracture experienced a water-
injection rate increase of only about 10%; there
was a 10-fold jump when only a standard screen
was used in the same model.
A third evaluation used a full-field reservoir
model to estimate the effect of the improved
water distribution. This evaluation included an
injection well equipped with ICDs in scenarios
similar to those analyzed by the near-wellbore
simulator in the first two cases.
The simulations concluded that given a high-
permeability channel, the use of ICDs increased
cumulative oil production by 10% over that
achieved with use of a standard screen alone.
They also showed that with no high-permeability
zone present the ICDs would improve cumulative
oil production by 1% and that the most likely case
was somewhere between the two (right).
In 2008, based on the success of this water-
injection project, Statoil installed another
injection well equipped with ResInject ICDs in
the Svale structure. The well has performed
according to objectives.
Control of the Future
The success of ICDs is now drawing the attention
of producers concerned with inefficient flow
from long laterals. Among these are heavy oil pro-
ducers. For more than 15 years, steam-assisted
gravity drainage (SAGD) has been the process
of choice for development of fields producing
heavy oil. Despite this history, the process is not
well understood.15 It may be that the current
steam distribution in horizontal injection wells
designed to heat and drive oil to deeper produc-
tion wells is less than optimal, particularly in
heterogeneous reservoirs.
Besides the common difficulties associated
with creating uniform flow through any reservoir,
two-phase water systems (liquid and vapor) used
in SAGD wells add to the difficulty of control. In
addition to single-phase-flow concerns relating to
fluid-velocity profiles and pressure drops associ-
ated with piping configurations, many other fac-
tors including flow-regime effects, water holdup,
phase splitting, droplet size, slugging and other
variables are introduced in two-phase flow.16
Typically, SAGD injection liners are slotted
along the entire section—a configuration that
does little to optimize steam distribution. To fight
the heel-toe effect, many operators today use
dual steam conduits in horizontal steam injec-
tors—one landed near the heel of the well and a
second near the toe.
Winter 2009/2010
Production history
Modeled standard completion,
80 high-permeability channels
Modeled standard completion,
70
no high-permeability channels
60
50
Water cut, %
40
30
20
10
0
0 200,000 400,000 600,000 800,000 1,000,000
Cumulative oil production, m3
> Water-cut models. For the most part, the actual water cut in this well was
lower than predicted by either model. Though the field is in early production,
the improved numbers may reflect enhanced sweep achieved by the use of
ICDs. (Adapted from Raffn et al, reference 13.)
In an effort to better understand SAGD pro-
duction and find more efficient solutions to its
challenges, Chevron has constructed a surface
horizontal steam-injection facility at its Kern
River field near Bakersfield, California, USA.
Researchers there are focusing on evaluation and
deployment of equipment for accurate and reli-
able steam placement along laterals in horizontal
injection wells to improve recovery.17
Their proliferation in recent years is testi-
mony to the effectiveness of ICDs. Use of ICDs
has allowed operators to realize full value from
the ability to drill long laterals, thereby exposing
large volumes of the reservoir to the wellbore. In
fact it can be argued that inefficient drainage
owing to uneven flow through the reservoir
threatened to impose economic limits on well-
bore length that were far short of the technical
limits. Today’s lengths are measured in kilome-
OSWIN09/10—Rick, story #2—Figure 13
ters rather than in meters, as they were less than
a decade ago. —RvF
37