OTC 22963

Optimal Inflow Control Devices Configurations for Oil Rim Reservoirs

Ali Mojaddam Zadeh, Per Arne Slotte, Robert Aasheim, Atle Johnsen Gyllensten, Kristine rland, Statoil ASA

Copyright 2012, Offshore Technology Conference

This paper was prepared for presentation at the Offshore Technology Conference held in Houston, Texas, USA, 30 April3 May 2012.
This paper was selected for presentation by an OTC program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Offshore Technology Conference and are subject to correction by the author(s). The material does not necessarily reflect any position of the Offshore Technology Conference, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Offshore Technology Conference is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of OTC copyright.

Abstract
Judicious selection of intelligent completions will increase oil recovery and constrain unwanted water and gas production
from the reservoir.
Field experience, and results from extensive simulation studies, shows that the water, oil and gas production are highly
dependent on the type of inflow control devices (ICDs) and the ICD configuration. Therefore, choosing the best ICD
configuration is a key point to maximize the total oil production. In a long horizontal well the primary objective is to
maximize oil production from the whole completion interval and avoiding early breakthrough of gas or water in parts of the
well. In particular there is a need for compensating for the friction induced pressure drop in the well which unchecked will
result in significantly higher production and early break through from the heel sections than from the toe. The secondary
objective is to limit flow from sections that, due to for instance heterogeneities, still suffer from early break through. Devices
with autonomous valves should be especially suited for meeting this secondary objective.
Traditionally, optimal ICD configurations have been selected using a trial and error approach simulating different
configurations. This time consuming approach may result in non-optimal configurations since it is impossible to test all of the
possibilities.
We have developed a semi-analytical mathematical model for calculating the optimal ICD strengths for long horizontal wells
in high permeable oil rim reservoirs. The mathematical model is utilized in a computer program which finds the optimal ICD
configuration selecting ICDs from a pre-defined set of off-the shelf devices. The optimal configuration can be found in a
fraction of a second, and no flow simulations are needed.
Optimized ICDs with RCP valves (autonomous control devises) is compared with optimized spiral control devices, and
benefits and restrictions are presented.
Introduction
The nature of oil rim reservoirs makes long horizontal wells an attractive option for increasing well-reservoir contact and
reducing drawdown. However, increasing the horizontal wellbore length leads to some production challenges. In a long
horizontal well with open hole completion, the drawdown in the heel section of a well is much higher than the drawdown in
the toe section. This is because of the higher cumulative frictional pressure loss in the heel section than the toe. Thus, higher
production in the heel than the toe section is expected. Consequently, the inflow from the reservoir to the well and water/gas
breakthrough are non-uniform. This phenomenon gives partial water or gas breakthrough and lower oil recovery and sweep
efficiency.
Advanced completions in horizontal wells have been applied in recent years to avoid mentioned problems. Utilizing
advanced completion gives the chance of controlling unwanted fluid and optimizing of the oil production from a long
horizontal well. Inflow Control Devices (ICDs) as a type of advanced completion have been used in producing wells to
enhance oil production and restrict unwanted fluid from the reservoir to well. ICDs are passive control valves which are not
adjustable or retrievable. Therefore, it is crucial to choose the best type and configuration for a single well before starting the
production.
In this study we have used channel type Spiral Inflow Control Devices (SICDs) and Autonomous Inflow Control Devices
(AICDs) to study the effect of advanced completions on production performance In the channel type SICDs, a set of
chambers with specific length and diameter is wrapped up around a pipe cover by screen. By flowing of reservoir fluid
through the chambers the fluid direction will change numerous times and creates pressure drop (Preston Fernandes,et al.

OTC 22963

2009 ). The pressure drop over the SICD is determined by the nominal strength of the SICD, which depends on the length
and diameter of the channel. A single horizontal well can be completed with SICDs of different strengths and optional
spacing between the valves. This makes it possible to design the most efficient SICD well completion setup by selecting
different combination of strengths and number of SICDs for each segment.
Autonomous valves are passive inflow control devices which are designed to automatically choke back fluids with lower
viscosity, thus partially closing sections of the well with high gas or water fraction. In the special type of the AICDs which is
used in this study, there is a floating disk bound between the inlet and outlet of the valve which can move both toward
upstream or downstream of the valve based on viscosity of the produced fluid (Vidar Mathiesen, 2011). When a less viscos
fluid enters the valves inlet the floating disk will move towards the upstream and choke back. AICDs have different types
with different tuning parameters. Selection between different types of AICDs for implementation depends on rock and fluid
properties in a reservoir.
In this paper, we are going to introduce a semi-analytical algorithm which can be used as a pre-processing tool for selection
of most efficient and reliable ICD configuration before running reservoir simulation. The aim is to find a unique ICD setting
which results in equalized inflow for a long horizontal without needing detailed information about the rock property
variations along the wellbore in the reservoir.
Pressure drops
There are three types of pressure drop in a conventional reservoir with a long horizontal well completed by ICDs:
Drawdown pressure
Pressure drop in the wellbore
Pressure drop in ICD
Drawdown pressure: The differential pressure in the sand body of the reservoir which drives the fluid to the wellbore.
Drawdown pressure is linearly proportional to total flow rate of single well and inversely proportional to absolute
permeability of the reservoir (L.P.Lake, 1977), as can be seen in the following equation:

P DD

r
q
ln( e ) ,... (1)
2 k rw

Eq. 1 defines the pressure drawdown in a cylindrical reservoir where the oil drainage radius re, is the distance between the
well and gas-oil contact, k is the vertical permeability and is viscosity. The reservoir produces from a well with the radius
of rw with inflow rate of q per meter. In reservoirs with high average permeability, the drawdown pressure will be negligibly
small compared with other pressure drops going to be discussed.
Pressure drop in the wellbore: By flowing of the reservoir fluid through the tubing in a long horizontal well, the fluid
pressure will be decreased due to friction and accelerations. The acceleration pressure drop is caused by variation in flow
area, flow direction or density of the fluid. We have assumed oil inflow is passing through a long horizontal pipe (tubing)
with no variation in shape and flow area. Also, the initial inflow is oil with no gas in the wellbore so that the fluid density is
almost constant before gas breakthrough. Therefore, the pressure drop because of acceleration is neglected in this study.
However, pressure drop caused by frictional pressure should be considered. The pressure drop in the tubing is given in the
Eq. 2 (Victor L. Streeter, 1975):

dPf

dx
where,

f Q 2
, .. (2)
D5

is fluid density, D is tubing diameter, and Q is volumetric flow rate of fluid in the tubing. The fanning friction

factor f (F.S. Manning et al, 1991) depends on the Reynolds number (Eq. 3).

Re

Q
D

, ..... (3)

For laminar flow (Re< 2000) fanning friction factor can be calculated using Eq. 4a (Robert L. Daugherty, et al, 1985)

16
, ... (4a)
Re

For Re> 4000, following correlation in Eq. 4b (Haaland, S.E., 1983) can be used to obtain fanning friction factor:

OTC 22963

6.9 e 1.11

3.6log

Re 3.7 D

, . (4b)

where e is the absolute roughness of the tubing in the same unit as D. In the transition between laminar and turbulent flow
(2000<Re<4000) linear interpolation in the Reynolds number can be used. Note that f is almost constant for high flow rates
which give a frictional pressure drop proportional to Q2.
Pressure drop along ICDs: by introducing the ICDs in the well completion we should consider an extra pressure drop for the
fluid passing through the devices. The amount of pressure drop is determined by fluid flow rate through the ICD, mixture
density and mixture viscosity of the fluid passing through the valves. Empirical equations using experimental data is used to
define the relation between the pressure drop through the valves, fluid properties and fluid flow rate. For SICDs the pressure
drop over a given well segment is given by (Bourgoyne, et al, 1986):

PSICD
where

mix

cal . mix
mix cal

0.25

mix
Q
.aSICD .
cal
n

, (5)

cal and cal are the density and viscosity of a known fluid respectively which is used for calibration of the devices.
and mix are mixture density and viscosity respectively. aSICD is an empirical constant , based on the measurements of

calibration fluid flow through a particular type of SICD, Q is the total inflow through the SICDs and n is the number of the
SICD valves in the well segment.
The equalized-flow ICD configuration
One of the main purposes of the ICDs is to create a more equal inflow along the length of the wellbore by compensating
for the increased drawdown created by the wellbore friction. In the following we will assume an equal inflow and derive the
corresponding ICD settings for a homogenous reservoir.
Fig. 1 schematically shows a simple horizontal well completed with ICDs in a homogenous reservoir.

Fig. 1Schematic drawing of a simple horizontal well completed with ICDs in a homogenous reservoir.

Pr is the reservoir pressure at drainage radius of the well. By following the green arrow loop from (x=0 (at toe), Y=drainage
radius of the well) coordinate, to the same place and writing the pressure relation we get:

Pr P DD ( x 0) P ICD ( x 0) P Friction x P ICD x P DD x Pr ,.... (6)

Where P is the drawdown pressure, P
the pressure drop along ICDs and P
is the frictional pressure drop
through tubing. Assuming equal inflow from the sand phase into the well and constant permeability in all directions would
give constant drawdown if the distance from the well to the drainage radius does not change. Also, if we consider that the
permeability is high enough then the pressure drop in the sand phase is negligible compare to the pressure drop in the well.
Thus, the drawdown pressure will be canceled out in Eq. 6 and the expected pressure drop along the ICDs at x is:
DD

ICD

Friction

OTC 22963

P ICD x P ICD ( x 0) P Friction x ,... (7)

As can be seen from the Eq. 2, the frictional pressure loss is function of distance from the reference point (in this case x=0 at
toe) and cumulative flow rate passing through the pipe. By integration over the distance from the toe, the expected pressure
drop for ICD at specific distance from the well can be obtained:

P ICD x P ICD ( x 0)

f ( x, Q). .Q( x) 2
dx ,.........
D5

(8)
The relation between the fanning friction factor f ( x, Q ) and total flow rate is not simple (Eqs. 2 and 4) and the integration
in Eq. 8 cannot be solved analytically. Numerical integration with mid-point rectangle method was used to obtain the
frictional pressure drop along the tubing.
On the other hand, Eq. 5 shows a relation between the pressure drop through ICDs and the fluid property. Knowing the equal
inflow along the horizontal well and calculating the expected ICD pressure drop at each section of the well by using Eq. 8,
we can find a unique

aSICD
for each segment of the well.
n2

In the channel type ICDs, aSICD determines the amount of pressure drop across one valve with specific fluid conditions (rate,
density and viscosity). The nominal bar rating in Table 1 refers to pressure drop created by when a standard 12meter SICD
joint, is exposed to 26 Sm3/day flow rate (Henriksen et. al, 2006).
Table 1 Different nominal pressure drop for channel type ICDs..
Nominal bar rating
0.2
0.4
0.8
aSICD bar/(Rm3/day)2
0.00076
0.0015
0.0026

1.6
0.0044

3.2
0.0086

The second variable that should be defined to be able to specify a unique configuration is the number of SICDs per section of
the well (n). A well section is a user defined length of a horizontal well which can have a range from one SICD to full
coverage of SICDs, in addition the section may be open hole completed (zero pressure drop). Maximum number of SICD in
each segment is the section length divided one single SICD joint length. The higher number of ICDs in each segment gives
less resistance against the flow from the sand phase to the segment which results in lower pressure drop.
There are a finite number of available strengths, and the number of SICDs per section is an integer. Thus, the best possible
aSICD and n couple for each section must be found. This best selection does not exactly give the wanted pressure drop, and
Fig. 2 shows the calculated pressure drop (y axis) from heel to the toe (x axis), which is shown by the blue line, versus
expected pressure drop made by SICDs at each section of the well numbering from heel to the toe of the well(red line). Each
section on the red line is obtained by a single (aSICD,n) couple.

Fig. 2 Pressure drop in horizontal well. Blue line: Frictional pressure drop in the well calculated by Ed. 8 and zero pressure drop
at toe, Red Line: pressure drop made by SICDs using the Eq. 9 and selecting the best (aSICD,n) combination for each section of the
well. X axis: Section numbers, numbering from heel to toe; Y-axis: pressure drop in bar.

OTC 22963

To study the effect of obtained ICD configuration with the algorithm, a simple reservoir model with following specification
mentioned in the Table 2 has been made. For simplicity, the reservoir model has a constant permeability, and pore volume
multiplier is applied to the upper most and lower most layers for the pressure support. Also, only oil and gas phases are
considered as the mobile phase in the in the first set of simulations. In additions, there are grid refinements in lateral and
vertical directions toward the well location to be able to capture both gas and water coning effects.
Table 2Simple generic model reservoir parameters
Parameter
Value
Reservoir thickness
30 m
Reservoir dimension
498m X 2925m
Length of horizontal well
2858.4 m
Well bore diameter
0.156 m
Horizontal permeability
2000 mD
Vertical Permeability
1000 mD
Porosity
20%
Initial reservoir pressure
157.5 bar
Oil density
805 kg/m3
Oil viscosity
1.75 cP

Parameter
Oil formation volume factor
Gas-oil ratio (GOR)
Water Density
Water Viscosity
Water formation volume factor
Gas density
Gas viscosity
Gas formation volume factor

Value
1.17 rm3/sm3
60 m3/m3
1045 kg/m3
0.45 cP
1.07 rm3/sm3
0.8 kg/m3
0.017 cP
0.0068 rm3/sm3

The simulation is controlled by constant gas rate of 144 000 Sm3/day in prediction mode. The SICD setting derived from
semi-analytical algorithm is introduced to the reservoir model to study the oil inflow equalization. The results were compared
with same reservoir conditions but the well with openhole completions and also a well with 100% coverage of 3.2 bar SICD
along the well (K.H. Henriksen, et al, 2006). Fig.3 and Fig.4 show the oil and gas inflow profile respectively at each
segment along the horizontal well completed with the SICD setting with the purpose of equal oil inflow versus different time
during nine years of production.

Fig. 3 Oil inflow profile in homogeneous generic

reservoir model with two phase including SICD setting
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time (days)

Fig. 4 Gas inflow profile in homogeneous generic

reservoir model with two phase including SICD setting
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time (days)

Also, the same reservoir model is used with an open-hole long horizontal well and the inflow profiles are shown in Fig. 5 and
Fig. 6. The inflow profiles show early gas breakthrough at heel section of the well in openhole conditions. On the other
hand, in the model with equal inflow SICD settings, there is a good distribution of oil and gas inflow throughout the whole
length of the horizontal well (Fig. 3). This will give much later and simultaneous gas breakthrough in the reservoir (Fig. 4).

OTC 22963

Fig. 5 Oil inflow profile in homogeneous generic

reservoir model with two phase in openhole well
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time (days)

Fig. 6 Gas inflow profile in homogeneous generic

reservoir model with two phase in openhole well
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time (days)

Fig. 7 and Fig. 8 show the inflow profile of the well completed with current practice of full coverage of 3.2 bar SICD in the
same reservoir conditions. Comparing Fig. 5 with Fig. 7 show that the oil inflow profile for openhole completion will not be
stabilize both before and after breakthrough during nine years simulation time while the oil inflow for the well completed by
full coverage of 3.2 bar SICD, will be stabilized after the gas breakthrough time ( after 600th days Fig. 8).

Fig. 7 Oil inflow profile in homogeneous generic

reservoir model with two phase including full coverage
of 3.2 bar SICD
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time (days)

Fig. 8 Gas inflow profile in homogeneous generic

reservoir model with two phase including full
coverage of 3.2 bar SICD
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time (days)

The total oil production will be affected by the selection of SICD in the well completion. Fig. 9 shows the total oil production
in first 2000 days of simulation for the three cases, and there is about 15% increase in oil production for the SICD
configuration that gives equal inflow profile compared with the case with open hole conditions.
Comparing all three cases show that the equal inflow ICD settings is the best of the three alternatives, but the main effect is
obtained by simply having SICDs in the well completion. The ICDs will stabilize the oil inflow throughout the well, in
particular after the gas breakthrough where the oil production will be gravity driven as described by the GORM model (Arne
Mjaavatten, et al, 2006).
In the next simulations water was added as a mobile phase, and the oil-water contact were assumed to be 0.5 meter below the
well location depth to be able to study the effect of water on SICD performance. SICD setting with the purpose of even oil
inflow (derived from algorithm) and SICD setting with full coverage of 3.2 bar SICD was tested in the same reservoir model.

OTC 22963

Fig. 9Cumulative oil production for a reservoir simulation when only oil and gas are mobile phases: A well with open hole
completion (red curve); A well with full coverage of 3.2 bar SICD (green curve); and SICD suggested by mentioned algorithm (blue
curve)

Fig. 10 and Fig. 11 show oil and gas inflow profile respectively, for the well having even oil inflow SICD setting. Oil inflow
profile remains almost even throughout the well during 9 years simulation time-lapse both before and after gas breakthrough.

Fig. 10 Oil inflow profile in homogeneous generic

reservoir model with three phase including even
inflow SICD setting
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time(days)

Fig. 11 Gas inflow profile in homogeneous generic

Reservoir model with three phase including even
inflow SICD setting
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time(days)

The simulation was run for the same reservoir conditions but with the well completed by full coverage of 3.2 bar SICD. Oil
and gas inflow profile during simulation time lapse are shown in Fig. 12 and Fig. 13. Early gas breakthrough in very first
time steps is obvious in the Fig. 13, while the oil inflow will be stabilized and even throughout the well. However, early gas
breakthrough will give uneven inflow before the gas breakthrough. Cumulative oil production will show little difference
between the well completions with full coverage of 3.2 bar SICDs and even inflow SICD setting. The existence of water
phase seems to mask the benefits of even oil inflow throughout the well in this homogenous reservoir model.

OTC 22963

Fig. 12 Oil inflow profile in homogeneous generic

reservoir model with three phase flow including full
coverage of 3.2 bar SICD
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time(days)

Fig. 13 Gas inflow profile in homogeneous generic

reservoir model with three phase flow including full
coverage of 3.2 bar SICD
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time(days)

Robustness of ICD setting with the different flow rates

SICDs are passive valves, and the valve characteristics will not be changed during the well production life. Therefore, the
SICD selection with the purpose of equalizing the inflow should be robust with respect to rate changes. To be able to find
equalized inflow SICD configurations, we have assumed a fixed total oil inflow to calculate pressure drop in a long
horizontal well while the oil flow rate from a real well will change during the production life a reservoir. Thus, it is necessary
to study the robustness of SICD configuration with different flow rates.
The pressure drop in a SICD has a q2 dependency on the flow rate (Eq. 5), and because the fanning factor is almost constant
for high flow rates (Eq. 4b), the frictional pressure drop will also depend on the flow rate as approximately q2 (Eq. 2). We
would therefore, expect that the optimal ICD settings will be robust to rate changes. This has been checked by simulations;
after running the algorithm for a specific well and reservoir conditions, three different flow rates have been tested in the same
reservoir model with the same SICD setting to evaluate the oil inflow profile.
Fig. 14, Fig. 15 and Fig. 16 shows the oil inflow rate at different segments numbering from heel to toe of the well versus
simulation time with different well head oil flow rates of: 90 Sm3/day (Fig.14), 140 Sm3/day (Fig. 15) and 250 Sm3/day
(Fig. 16).

Fig. 14 Oil inflow profile in homogeneous

generic reservoir model including SICD
3
setting with 90 Sm /day total oil flow rate

Fig. 15 Oil inflow profile in homogeneous

generic reservoir model including SICD
3
setting with 140 Sm /day oil flow rate

Fig. 16Oil inflow profile in homogeneous

generic reservoir model including SICD
3
setting with 250 Sm /day oil flow rate

Testing the SICDs configuration in three different oil flow rates shows the SICD setting could equalize the oil inflow in a
wide range of oil production for a well.
Simulation results for applying the settings in a heterogeneous reservoir
We do not want the selection of ICD settings to be dependent upon detailed knowledge of the permeability distribution along
the wellbore. Thus we need to check the performance of ICD settings derived from our semi analytical method in a reservoir
with more realistic heterogeneities. We do not expect the method to work for reservoirs with arbitrary large heterogeneity,
but based on conceptual understanding of the geology of some currently producing sandstone oil rim reservoirs we have
constructed a shoe box reservoir model with sinusoidal layering of 2000 mD and 200 mD permeability zones. The rest of the

OTC 22963

rock and fluid properties are the same as the homogeneous reservoir model mentioned in the Table 2. The permeability map
of the reservoir model with relative location of the horizontal well with Gas-Oil Contact (GOC) and Oil-Water Contact
(OWC) is shown in the Fig. 17.

Fig. 17Permeability map of a reservoir model with sinusoidal layering of high and low permeability zones

The ICD configuration derived from the semi-analytical algorithm is tested in this reservoir model. The simulation was run
using the same total gas rate constraints as the homogeneous model and the results are shown in Fig. 18 and Fig. 19.

Fig. 18 Oil inflow profile in sinusoidal generic

reservoir model including SICD setting
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time(days)

Fig. 19 Gas inflow profile in sinusoidal generic

reservoir model including SICD setting
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time(days)

Fig. 18 and 19 show non-uniform oil and gas inflow in this case. The amount of inflow is correlated with average absolute
permeability close to the well. In addition, early gas breakthrough in the toe section of the well because of high average
permeability in the toe is obvious. Heterogeneous permeability in this reservoir model gives higher average drawdown in the
sand phase than what was pre-assumed in the algorithm for derivation of SICD setting in a horizontal well. In other words,
the frictional pressure drop in the well is not the dominating pressure drop in this reservoir model. In order to fulfill the
assumption for using the algorithm, an extra pressure drop calculated by Eq. 1 using lower case permeability in the model
should be added to whole frictional pressure drop profile. And then, the best match of SICD configuration can be found for
the given new pressure profile in the well. With the new setting of SICD configuration the pressure drop in the well is
dominated and fulfills the algorithm assumption.

10

OTC 22963

Fig. 20Oil inflow profile in sinusoidal generic

reservoir model including ICD setting with additional
pressure drop
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time(days)

Fig. 21 Gas inflow profile in sinusoidal generic

reservoir model including ICD setting with additional
pressure drop
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time(days)

Fig. 20 and Fig. 21 show the results of simulation for the model with sinusoidal permeability layering including ICD setting
with additional pressure drop. As can be seen from the figures, having modified SICD configuration creates more uniform oil
inflow and later gas breakthrough in the reservoir model.
The SICD configuration modification will affect the total oil production as well. Total oil production in three years period
was compared in three different cases: open hole horizontal well (no ICD setting), a horizontal well including ICD
completion with no extra pressure drop in the well, and finally a horizontal well including ICD setting with extra pressure
drop. Fig. 22 shows total oil production results for first 1000 days of the simulation in all cases.

Fig. 22Total oil production for three different cases: open hole horizontal well (red line), horizontal well including SICD
completion with no extra pressure drop in the well (green line), horizontal well including ICD setting with extra pressure drop in the
toe (blue line)

The total oil production is increased by 10% and the gas breakthrough was also delayed by 36 days for the case of horizontal
well including ICD setting with extra pressure drop in the toe compare with open hole horizontal well. While, the case of
long horizontal well including ICD setting with no extra pressure drop shows only 4% increase in total oil production
compare with the open hole well conditions.
Common practice of full coverage 3.2 SICD throughout the long horizontal well is tested in the same reservoir model with
the same reservoir conditions. Oil and gas inflow profile are shown in the Fig. 23 and Fig. 24 respectively. Early gas
breakthrough in the heel section of the well where the well is penetrating through high permeable zone is obvious.

OTC 22963

11

Fig. 23 Oil inflow profile in sinusoidal generic

reservoir model including full coverage of 3.2 bar SICD
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time(days)

Fig. 24 Gas inflow profile in sinusoidal generic

reservoir model including full coverage of 3.2 bar SICD
x-axis: Segment numbers, numbering from heel to toe
y-axis: Time(days)

Comparing the cumulative oil production in this case with the equal inflow SICD setting and openhole conditions in first
1000 days of simulation time laps shows the total oil production will increase by 10% and 15% in the case of equal inflow
SICD setting from full coverage of 2.3 bar SICDs throughout the well and openhole conditions respectively (Fig. 25).

Fig. 25Total oil production for three different cases: open hole horizontal well (red line), horizontal well including full coverage of
3.2 bar SICD throughout the well (green line), horizontal well including even inflow SICD setting (blue line).

Comparison between AICDs and SICDs

In addition to SICDs, AICDs may optionally be used in completion design to control water and gas production when the
chance of early breakthrough is high. Both AICDs and SICDs create extra pressure drop for unwanted fluid and partially
choke back water and gas in the well bore. However, the way of making the pressure drop is different between the valves. In
special type of AICD Eq. 9 is used to calculate pressure drop through the AICDs (Vidar Mathiesen, 2011).
y

PAICD

2
q
mix . cal .a AICD
, (9)
cal mix
qcal .n

where mix ,

cal , cal , mix , q and n have the same definition as Eq. 5. a AICD is AICD strength which is specifically

defined for each type of AICD. x and y are volume flow rate exponent and viscosity function exponent respectively. They
are non-zero values which can define how a specific AICD is sensitive to viscosity and flow rate. a AICD , x and y are predefined AICD characteristics for each type of AICDs. Practically, in applying the AICDs, only one type of AICD with
different distribution of the valves is used.

12

OTC 22963

Autonomous valves have been tested for a generic reservoir model that potentially has the high chance of water breakthrough
in early time steps. In this case shown in the Fig. 26, 30 degree dipping layers of high and low permeable zones was made.
Oil and water relative permeabilities are considered linear with respect to water saturation to get the chance of higher fluid
movement in high permeable zones. The rest of rock and fluid properties are mentioned in Table 2.

Fig. 26Generic reservoir model with 30 dipping of 2000 mD and 200 mD permeable layers. OGC and OWC are oil-gas contact and
oil-water contact respectively.

The described algorithm was used to find the best ICD setting with the purpose of equalizing the inflow in the horizontal well
without detailed knowledge of the permeability distributions. Fig. 27, Fig. 28 and Fig. 29 show the simulation results for oil,
water and gas inflow profile through SICDs during the simulation time. As can be seen from the figures, oil, water and gas
inflow are correlated with the average permeability of the region close to the well in each section. In additions, SICDs are not
able to sufficiently choke back water and gas. However, the water and gas inflow become more stable in the later time steps.
The same study has been done for the case of having AICDs instead of SICDs to be able to compare the results for the two
cases in the generic reservoir model. Fig. 30, Fig. 31 and Fig. 32 show the simulation results for the same reservoir model
with AICDs in place. The results for the case of autonomous valves show the water and gas inflow can be limited to a higher
degree compared with the same conditions with Spiral valves. Also, later gas breakthrough is obvious in comparison between
Fig. 29 and Fig. 32.

Fig. 27 Oil inflow profile in dipping

layers reservoir model including
SICD setting
x-axis: Segment numbers, numbering
from heel to toe
y-axis: Time(days)

Fig. 28 Water inflow profile in dipping

layers reservoir model including
SICD setting
x-axis: Segment numbers, numbering
from heel to toe
y-axis: Time(days)

Fig. 29 Gas inflow profile in dipping

layers reservoir model including
SICD setting
x-axis: Segment numbers, numbering
from heel to toe
y-axis: Time(days)

OTC 22963

Fig. 30 Oil inflow profile in dipping

layers reservoir model including
AICD setting
x-axis: Segment numbers, numbering
from heel to toe
y-axis: Time(days)

13

Fig. 31 Water inflow profile in dipping

layers reservoir model including
AICD setting
x-axis: Segment numbers, numbering
from heel to toe
y-axis: Time(days)

Fig. 32 Gas inflow profile in dipping

layers reservoir model including
AICD setting
x-axis: Segment numbers, numbering
from heel to toe
y-axis: Time(days)

Conclusion
For long horizontal wells in reservoirs with high average permeability, wellbore frictional pressure drop is the
dominating pressure drop. This will result in early gas and water breakthrough in the heel and very little inflow in the toe
section of the well. Using ICDs can help to avoid the early breakthrough, and equalize oil flow after break through. In
this study, a semi-analytical approach was proposed to choose the best ICD setting for compensating the frictional
pressure drop regardless of permeability distributions in the reservoir. The following conclusion can be drawn based on
simulation study in this work:
An algorithm for selecting well completions with ICDs such that the oil inflow is equalized along a horizontal well
has been developed and tested in generic homogeneous and heterogeneous cases. The algorithm input is independent
of detailed information about the rock property variations along the wellbore in the reservoir.
Compared with full coverage constant strength ICD completions, equalizing the oil inflow will enhance the sweep
efficiency and total oil production in homogenous reservoirs, and simulations on cases with layered sinusoidal
heterogeneity also shows improved well performance and reservoir drainage.
In a heterogeneous reservoir having ICDs in the toe enhance production compare with open hole toe.
Selecting optimal ICD settings for cases with significant permeability trends along the wellbore remains an open
question.
The proposed algorithm can be used to compare different types of passive inflow control devices such as
comparison between SICDs and AICDs.
Results for the comparison between SICDs and AICDs in a dipping sequence of low and high permeable layers
show that AICDs can choke back water much better than SICDs. However, the total oil production for both cases is
the same.
Nomenclature
aAICD = spiral inflow control devices strength, bars/kg/m3
aSICD = autonomous control devices strength, bars/(m3/day)2
D = well bore diameter, m
e
= absolute roughness of tubing, m
f
= fanning friction factor, dimensionless
k
= absolute permeability, md
n
= number of inflow control devices per section of the well, dimensionless
Pf
= frictional pressure, bar
Pr
= reservoir pressure, bar
qcal = calibration flow rate, m3/day
Q = volumetric flow rate, m3/day
re
= oil drainage radius, m
= wellbore radius, m
rw
Re
= Reynolds number, dimensionless
x
= volume flow rate exponent for AICDs, dimensionless
y
= viscosity function exponent for AICDs, dimensionless

14

OTC 22963

Greek symbols
PAICD = pressure drop caused by autonomous inflow control device, bars
PSICD = pressure drop caused by spiral inflow control device, bars
PDD

= draw down pressure drop, bars

P Friction = frictional pressure loss, bars

PICD

cal
mix

cal
mix

= pressure drop caused by inflow control devices, bars

= viscosity, cp
= calibration viscosity, cp
= mixture viscosity, cp
= fluid density, kg/m3
= calibration fluid density, kg/m3
= mixture fluid density, kg/m3

Subscripts
AICD = autonomous inflow control device
SICD = spiral inflow control device
f
= friction factor
e
= external radius
w
= well bore
cal = calibration
mix = mixture
ICD = inflow control device
DD = draw down
OWC = oil-water contact
GOC = gas-oil contact
References
Are Mjaavatten, Robert Aasheim, Steinar Saelid, Oddvar Gronning, Prediktor a.s. 2006. A Model for Gas Coning and Rate-Dependent
Gas/Oil Ratio in an Oil-Rim Reservoir. Paper SPE 102390-RU presented at SPE Russian Oil and Gas Technical Conference and
Exhibition, Moscow, Russia, 3-6 October.
Birchenko, V.M. 2010. Analytical Modeling of Wells with Inflow Control Devices, submitted for Phd degree, Institute of Petroleum
Engineering, Heriot-Watt University.
Bourgoyne, T. Adam. 1986. Applied Drilling Engineering, Society of Petroleum Engineers, Vol. 2.
Dake, L.P. 1978. Fundamentals of Reservoir Engineering, Amsterdam: Elsevier
Daugherty, Robert L., Franzini, Joseph B., Finnemore, E. John, 1985, Fluid Mechanics With Engineering Applications, Eighth Edition,
Mcgraw-Hill College
Journal, Haaland, S.E. 1983. Journal of Fluid Engineering, Fulid mechanic, volume 105
K.H. Henriksen, E.I. Gule, J. Augustine. 2006 .Case Study: The Application of Inflow Control Devices in the Troll Field, Paper SPE
100308-MS presented at SPE/EAGE Annual Conference and Exhibition, Vienna, Austria, 12-15 June.
Manning, Francis, S., Thompson, Richard E. 1991. Oilfield Processing of Petroleum, Volume 1: Natural Gas, page 293, PennWell Books.
Preston Fernandes, Zhuoyi Li, and D. Zhu. 2009 . Understanding the Roles of Inflow-Control Devices in Optimizing Horizontal-Well
Performance. Paper SPE 124677-MS presented at Annual Technical Conference and Exhibition, New Orland, Louisiana 4-7 October.
Streeter, Victor L. and Wylie, E.B. 1975. Fluid Mechanics, Eighth Edition, Mcgraw-Hill College
Vidar Mathiesen, Haavard Aakre, Bjrnar Werswick, Geir Elseth. 2011. Autonomous Valve, A Game Changer Of Inflow Control In
Horizontal Wells. Paper SPE145737-MS presented at SPE offshore Europe Oil and Gas Conerence held in UK. 6-8 September 2011.