SPE 77942

Smart Well Solutions for Thin Oil Rims: Inflow Switching and the Smart Stinger
Completion
* **
J.D. Jansen, SPE, Shell International E&P and Delft University of Technology, A.M. Wagenvoort , SPE, V.S. Droppert ,
SPE, Delft University of Technology, R. Daling, SPE, and C.A. Glandt, SPE, Shell International E&P.
* Now with Shell International E&P, ** Now with Wintershall

Copyright 2002, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the SPE Asia Pacific Oil and Gas Conference and
problems, in particular water or gas coning leading to early
Exhibition held in Melbourne, Australia, 8–10 October 2002. water or gas breakthrough.
This paper was selected for presentation by an SPE Program Committee following review of Horizontal wells are an attractive solution to reduce the
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
potential for coning because they require lower drawdown
correction by the author(s). The material, as presented, does not necessarily reflect any than vertical wells for the same production rates. However, a
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of possible problem of horizontal wells is the pressure drop over
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
the well bore caused by friction forces between the fluid and
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 the well bore (Fig. 1). As a result the drawdown at the heel of
words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O. the well becomes higher than the drawdown at the toe, which
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435. increases the tendency for water and gas coning at the heel and
thus partially cancels the beneficial effect of the horizontal
Abstract well. The reduced draw down near the toe of the well also
The production of oil from horizontal wells in thin rims lowers the effectiveness of increasing the well length.
sandwiched between gas and water is notorious for coning Pressure drop over horizontal oil wells has been modeled
problems. There is a strong tendency for early gas or water to various degrees of sophistication1-3. These studies show
breakthrough at the heel, especially if the pressure drop over that, roughly speaking, well bore pressure drop over a
the length of the well is in the same range as the drawdown. horizontal well becomes a problem when it is in the same
We present two conceptual solutions to counteract the order of magnitude as the drawdown at the heel. The ratio
negative effect of well bore pressure drop through the between pressure drop and drawdown increases for reducing
application of downhole measurement and control. One well diameter, increasing well length, and, importantly for
solution concerns inflow switching in a segmented well bore many oil rims, increasing reservoir permeability and reducing
and allows for coning control after breakthrough has occurred. oil viscosity.
The other solution aims at preventing breakthrough as long as
possible. This is achieved by flattening the drawdown profile Stinger completions
along the well through controlling inflow at one or more
Passive stinger completion
points in an extended stinger. The feasibility of the solutions
One of the solutions to the non-uniform drawdown
was demonstrated through numerical simulations over a range
problem is the use of an ‘extended stinger’ to shift the tubing
of reservoir and well bore parameters. Implementation would
inflow point from the heel of the well to somewhere near the
require further development of downhole water and gas
middle4,5. This effectively replaces the horizontal well by two
detection capabilities.
shorter ones (Fig. 2). However, such a ‘passive’ stinger has a
Introduction number of practical disadvantages:
Thin oil rims are relatively thin oil columns, in the order of a 1. Its dimensions are based on a fixed inflow profile along
few to tens of meters thick, sandwiched between water and gas the well. However, the inflow profile may change over
layers. They often occur in reservoirs with lightly compacted the life of the well due to reservoir pressure transients
sands having high porosities and high permeabilities of up to and due to the breakthrough of gas or water.
several Darcy, and they commonly contain light oils. These 2. It requires that the inflow profile along the well bore is
properties culminate in favorable reservoir flow conditions, known at the design stage. This is usually quite
reflected in a low drawdown required for production. unrealistic because of unpredictable reservoir
However, these properties can also cause production heterogeneities, in particular near-well bore permeability
fluctuations.
2 J.D. JANSEN, A.M. WAGENVOORT, V.S. DROPPERT, R. DALING, AND C.A. GLANDT SPE 77942

3. It requires that the well bore pressure drop can be few fixed positions, or even an on-off ICV that could be
accurately predicted. However, effects of well bore frequently switched to produce, on average, about the same
roughness, multi-phase flow, radial inflow and well bore result as the variable ICV. Driving this idea further we
undulations can only be predicted to a limited extent. developed the concept of the Inflow Switching Process (ISP).
4. It somewhat increases the pressure drop over the Fig. 6 depicts a well completed to accommodate the ISP. It
completion, either through reducing the perforation contains an extended stinger with multiple packers, on-off
density near the heel, or through addition of a stinger ICVs, downhole pressure gauges and measurement equipment
with a smaller diameter than the casing. This deteriorates to detect gas or water breakthrough. The basic idea of the ISP
the vertical lift performance of the well. is to control coning through varying the on-off setting of the
Stinger completions could in theory be applied to cemented valves alternatingly. Closure of the valves is triggered by gas
casing or to uncemented pipe. e.g. slotted screens or or water breakthrough, but after a certain period, allowing the
perforated pipe. However, in an uncemented completion the gas or water cones to recede, the valves are re-opened until the
additional annular flow path between casing and reservoir next breakthrough occurs.
drastically increases the uncertainty, because it is usually
Numerical simulation
unknown to what extent the annulus has collapsed. In the
remainder of this paper we will tacitly assume that the Reservoir model
completion has been cemented in place. To assess the feasibility of the smart solutions we numerically
We performed several field trials with extended stinger simulated their performance in draining a thin oil rim, with the
completions in our operations. They revealed that in particular aid of a proprietary reservoir simulator. Fig. 7 depicts the test
the changing inflow profile over the life of the well and the configuration: a simple box model with no-flow boundaries
unpredictable reservoir heterogeneity hampered the and an analytical aquifer model. The horizontal well is
effectiveness of a fixed-geometry stinger completion, and this positioned mid in between the gas-oil contact (GOC) and the
stimulated the development of an extended stinger with an oil-water contact (OWC). The aquifer strength and response
intermediate inflow point (Fig. 3). Not only does this result in time were chosen such as to obtain simultaneous water and
an even further flattened drawdown profile, it also offers an gas breakthrough when producing from a frictionless (infinite
opportunity to introduce flexibility: by equipping the conductivity) well. Making use of symmetry, we actually only
intermediate inflow point with a variable-position sliding simulated half of the model, which contained 28290 grid
sleeve, the degree of inflow through the end of the stinger or blocks. We used a logarithmic grid refinement in the
through the intermediate inflow point can be influenced. The directions perpendicular to the well. Oil and gas fluid
flexibility, however, is rather limited because it requires properties were modeled using a Standing black oil model.
well intervention. Table 1 gives the key dimensions and reservoir and fluid
properties. We used a fully implicit solution scheme with
Smart Stinger Completion
automatically varying time steps.
A logical extension of the passive extended stinger with
adjustable intermediate inflow is a stinger where the sliding Completion model
sleeve can be actively controlled and acts as an Inflow Control Pressure drop over the horizontal well bore and the casing-
Valve (ICV) (Fig. 4). We named this conceptual solution the stinger annulus was taken into account with the aid of pre-
Smart Stinger Completion (SSC). Two similar concepts were defined ‘pressure drop tables’ that specify the pressure drop
described in Ref. 6, however with the difference that in those per unit length as a function of oil, gas and water flow rates. A
cases an additional packer was incorporated in the completion fully implicit simultaneous solution of well bore and reservoir
(Fig. 5). In combination with the ability to measure downhole flow was performed at each time step. Input for the tables was
pressures, the ICV in the SSC allows for a careful tuning of generated with the aid of a proprietary multi-phase well bore
the amount of intermediate inflow such that the drawdown flow simulator. Vertical lift performance was not taken into
profile along the well becomes as flat as possible. It also account, and the well was operated on a mixed bottom hole
allows for adjustments during the life of the well, in particular pressure and surface rate constraint as discussed below in
after gas or water breakthrough. In addition, it allows for more detail. We used a Peaceman well inflow model in the
switching from breakthrough prevention to coning control: y-z plane.
once breakthrough has occurred we can manipulate the ICV to Modeling of the SSC requires the computation of well bore
maintain ‘critical’ flow such that the cone does not grow flow in a network configuration: fluid flowing from the
further. This of course requires means to detect gas or water reservoir into the casing-stinger annulus may subsequently
breakthrough downhole, which are currently only emerging flow to either the stinger end or the ICV at the intermediate
technologies. However, for the purpose of this study, we inflow point. The annulus was modeled as a ‘mini-sidetrack’
assumed that some form of downhole phase detection is from the end of the stinger back towards the ICV. To model
already available. the ICV we initially used a workaround: Two reservoir grid
blocks, isolated from the true reservoir model, were used to
Inflow Switching Process
create a second connection between the annulus and the
The SSC as described above has a continuously variable
stinger (Fig. 8). The transmissibility between the two blocks
ICV. Alternatively we could use a cheaper ICV with only a
SPE 77942 SMART WELL SOLUTIONS FOR THIN OIL RIMS: INFLOW SWITCHING AND THE SMART STINGER COMPLETION 3

was used directly to control the pressure drop over the ICV. and optimized drawdown profiles for a conventional stinger
Later we eliminated the need for the workaround by adapting completion (as in Fig. 2) and a stinger with intermediate
the simulator solver to enable modeling of ‘networked’ well inflow or an SSC (as in Fig. 3 or Fig. 4). The semi-analytical
bore configurations directly. method is based on a steady-state, single-phase flow model,
The completion for the ISP was represented as a number of and is therefore only a strong approximation to the transient,
‘mini-sidetracks’, each of which contained a second branch, to multi-phase flow behavior that governs our numerical
model the casing-stinger annuli (Fig. 9). Annular friction was simulation. However the method proved very useful in quickly
modeled with the aid of pressure drop tables. determining the optimal stinger configuration. It is noted that
we may include the stinger diameter as a third optimization
Base case
parameter, in addition to the stinger length, and the ICV flow
As a base case, to which we could compare the smart
area. This allows us to not only minimize the difference in
solutions, we simulated the behavior of a conventional
drawdown along the well bore, but also the pressure drop over
horizontal well as depicted in Fig. 1. In addition we simulated
the ICV (which, in the optimized case, equals the pressure
a fictitious frictionless well, representing the theoretically
drop over the stinger). The ICV pressure drop does not
maximum achievable pressure drop reduction with any (smart)
directly influence the well inflow performance along the well,
solution. Tables 2 and 3 give the well dimensions and the
but it does affect the well’s outflow performance, i.e. its
production constraints. Fig. 10 depicts the production profiles
lift capacity.
for the base case and the frictionless case. It can be seen that
the effect of friction is to rapidly cause a drop in the oil ICV operating strategy
production from an initial plateau rate of 20,000 bpd. Gas As discussed above, the optimum steady-state drawdown
breaks through after 36 days and in order to keep the gas-oil- profile is obtained when the pressure drop over the stinger
ratio (GOR) below its constraint of 1500 scf/stb (i.e. about 1.2 equals the pressure drop over the ICV. We used this criterion
times the solution GOR) production is beaned back. Water to control the ICV setting until the first signs of gas or water
breaks through after 218 days and causes a further drop in breakthrough. In practice this would require three pressure
production, until the simulation is terminated because the gauges: two at the ICV (one in the annulus and one in the
minimum oil rate of 400 bpd is reached. Fig. 11 depicts the stinger), and one at the end of the stinger. In theory it should
typical friction-induced saturation profile with severe gas and be possible to control the ICV based on only the difference
water coning at the heel. In the frictionless case, gas and water between the annulus pressures at the stinger end and the ICV,
break through simultaneously after 399 days. The cumulative which would require only two pressure gauges, but this
production from both cases is identical, although the resulted in a less stable ICV behavior in our simulations.
frictionless case reaches it final production almost a year After the occurrence of gas or water breakthrough we
earlier. Clearly the key potential benefit of friction reduction switched from breakthrough prevention to coning control. We
in this case is production acceleration through maintaining used an algorithm to maintain a ‘critical’ flow rate through the
plateau production. As illustrated in Table 4 an additional ICV to keep the gas and water cone development within
potential benefit is a drastic reduction in total volumes bounds. At predefined time intervals the phase rates in the
produced gas (19 % reduction) and water (42 % reduction). flow through the ICV were checked. When more than 25 %
water cut or more than a 5% increase in GOR at downhole
Smart Stinger Completion
conditions were detected the ICV pressure drop set point was
Optimal drawdown profile multiplied with a factor greater than one. When the water rate
The objective of the SSC is to create a drawdown profile dropped below 25% or when the downhole GOR dropped
along the well that is as flat as possible. For a given stinger more than 5% again, the set point was multiplied with a factor
diameter we can vary the stinger length and the ICV flow area between zero and one. By choosing a more aggressive
to influence the position of the maxima and minima in the shrinkage factor than growth factor, we obtained cycles of
drawdown profile. Fig. 3 illustrates that there are two local rapid closure at breakthrough and more gradual re-opening
minima and two local maxima. The flattest profile is achieved thereafter, which proved to be an effective strategy for coning
by minimizing the difference between the global maximum control. Note that because our configuration only includes one
and the global minimum. This is equivalent to ensuring that ICV, at the beginning of the stinger, coning control at the
the two local minima are equal and the two local maxima are stinger end is not possible. This would require a second ICV,
equal. Note that this also implies that in the optimum situation which would obviously increase the costs of the completion.
the pressure drop over the ICV equals the pressure drop over
Numerical simulation results
the stinger. The optimization is a trial and error process that
Numerical simulation showed that the SSC performed very
can be quite time-consuming if performed with a numerical
effectively for the reservoir and well parameters in our base
reservoir simulator. Therefore we developed a semi-analytical
case example. Water and gas broke through simultaneously at
formulation based on the classic well bore pressure drop
the points of maximum drawdown after about 310 days. This
model of Dikken1. Details of the semi-analytical method are
breakthrough time is not much shorter than the theoretical
presented in Ref. 7. Fig. 12 gives an example of the drawdown
maximum of 399 days for a frictionless completion, and much
profile of a conventional horizontal completion (as in Fig. 1),
longer than the base case breakthrough times of 36 days (gas)
4 J.D. JANSEN, A.M. WAGENVOORT, V.S. DROPPERT, R. DALING, AND C.A. GLANDT SPE 77942

and 218 days (water). Fig. 13 depicts the production profile, This reflects the effect of reservoir pressure transient behavior.
and it can be seen that the SSC results almost correspond to The major benefit of an SSC over a passive extended stinger
the, ideal, frictionless case. Figure 14 depicts the completion with intermediate inflow is therefore the flexibility
corresponding saturation profile at breakthrough, and Table 5 to maintain the optimum drawdown profile during the entire
gives the production statistics. The cumulative amounts of oil period prior to breakthrough. An added benefit is that after
gas and water are about the same as for the base case and the breakthrough some degree of coning control can be exercised
frictionless case, but the oil is produced much earlier, leading at the inflow point. Optimal design of the SSC, however,
to a significant acceleration benefit. To quantify this benefit requires a reasonable knowledge of the permeability
we performed a differential Net Present Value (NPV) analysis, distribution along the well.
compared to the base case, using the economic parameters
Inflow Switching Process
given in Table 6. The differential NPV of the SSC amounted
to about 5.3 million $. Obviously this figure strongly depends Switching algorithm
on the various assumptions made and is therefore of limited The basic idea behind the ISP is to switch off well
value on its own. However, it served as a useful reference segments whenever gas or water breaks through, and to switch
figure to perform a sensitivity analysis of various them back on again after a certain period during which the
model parameters. water or gas cone has had a chance to retreat from the well
bore. Various sequences of opening and closure were tested,
Sensitivity analysis
which resulted in varying degrees of coning control. A
We repeated the SSC simulation with different values for
particularly successful algorithm was as followings:
the reservoir and fluid parameters, in particular permeability,
vertical/horizontal permeability ratio and oil gravity. In all 0. Initialize the current liquid rate, GOR and
these cases it appeared that the scope for benefit of the SSC is watercut constraints to the default values (see
larger when the original problem, i.e. uneven drawdown Table 3.) Set the current gas rate constraint
distribution, is worse. For example an increased permeability equal to the GOR times the liquid rate
leads to a relatively larger difference in drawdown between constraint.
heel and toe, and therefore to an increased scope for 1. Produce the well with all segments open, at the
breakthrough time reduction with the SSC and an associated liquid rate constraint. Reduce the rate if
increased NPV. Conversely, low permeabilities lead to low or necessary to maintain, gas, GOR or watercut
even negative NPVs; in our example down to –2 million $ for constraints.
a reduction in permeability from 2000 to 200 md. We also 2. Close the first segment as soon as water or gas
investigated effects of reservoir tilt and of various reservoir breaks through. Increase the drawdown to
one cycle

heterogeneities. In all these case the SSC performed maintain maximum production honoring the
reasonably well to very good (in terms of NPV), depending on constraints.
matters like position of the stinger-end with respect to the 3. Repeat step 2) for the other segments until all
aquifer and the degree of heterogeneity. segments are closed.
The NPV analysis also allowed us to assess the sensitivity 4. Reset the current gas rate constraint to the last
of the completion design and the cost effectiveness of a oil production rate at breakthrough in the last
variable position ICV versus a fixed position on-off one. The section times the maximum allowable GOR.
performance of the SSC appeared to be rather insensitive to 5. Repeat steps 1-5 until the minimum oil rate is
stinger length, at least for homogeneous reservoirs. reached.
Furthermore, the production performance of the SSC with Two key aspects of the algorithm are 1) the increase in
variable ICV was clearly superior to SSC with on-off ICV in drawdown whenever a segment closes to maintain the
all cases. However, due to the large price difference between production rate within a cycle, and 2) the decrease in gas rate
the two types in our economic assumptions, the on-off constraint at the end of each cycle to ensure a gradually
configurations sometimes resulted in higher NPVs. We also decreasing drawdown over the life of the reservoir. Because
looked into the benefit of an ICV to counteract the effects of the gas rate constraint is determined based on the last observed
unexpected reservoir heterogeneities compared to a passive oil rate before breakthrough in a cycle, we are also gradually
stinger completion with a fixed-size intermediate inflow point. constraining the oil production. Note that the base case
The ICV proved not always to be the better choice. If, for example reflects a gas-dominated situation: gas breaks through
example, the stinger end is located in a zone with an before water. Therefore the closure of a segment in steps 2)
unexpected high permeability, there is little that can be done to and 3) of the algorithm will always be triggered by gas
improve the drawdown profile anyway, once the completion is breakthrough. For a water-dominated case the algorithm
installed, and in that case the high costs of the ICV resulted in probably needs to be revised.
a low NPV.
In all of the SSC simulations, the ICV pressure drop set- Numerical simulation results
point slowly changed in order to maintain the flattest possible Fig. 15 depicts saturation plots of the various stages within
a single cycle. Frame a) shows the first breakthrough at the
drawdown profile until breakthrough eventually occurred.
SPE 77942 SMART WELL SOLUTIONS FOR THIN OIL RIMS: INFLOW SWITCHING AND THE SMART STINGER COMPLETION 5

first interval after 60 days, yielding a profile as would be that gas penetrates the oil zone in the reservoir, it is trapped in
obtained using a conventional completion. Frame b) shows gas the pores until the critical gas saturation has been reached after
breakthrough in the second interval after 74 days. It can also which it becomes mobile and flows further. This results in a
be seen that the gas cone in the first interval has almost fully delay in the progress of the gas oil contact towards the well.
receded. Frame c) shows the breakthrough in the third interval However, when the GOC retreats, the residual gas saturation is
50 days later with the cone in the second interval almost above the critical gas saturation. As a result the gas is almost
receded. Frame d) shows the fourth interval being closed due perpetually mobile after it has retreated for the first time. This
to gas breakthrough 38 days later with the cone in the former results in a far quicker gas breakthrough during subsequent
interval almost receded. Only 12 days after this, the fifth cycles.
interval is shut in, as shown in frame e), and the cone in the To get a first indication of the effect of hysteresis on the
fourth interval is halfway in its recession. Frame f) shows the ISP we analyzed two cases with imbibition curves with a 20%
situation 34 days after the closure of the fifth interval, with gas and 30% residual gas saturation, and compared them to the
breakthrough at the first interval, completing the cycle. It can base case where the primary drainage and imbibition curves
also be seen the GOC and to a lesser extent the OWC have are the same with a 5% residual gas saturation. When the gas
fully recovered to their new levels. saturation of a particular grid block increased, the primary
Fig. 13 depicts the oil production profile. As was the case drainage curve was used, whereas for decreasing gas
for the SSC, the ISP results in a profile that comes close to saturation the imbibition curves were used. The simulations
that of an ideal frictionless well. However, unlike the SSC, the showed that the increased gas mobility leads to a more rapid
cumulative production of gas and water is much lower bean back of the liquid production in order to stay within the
compared to a frictionless well: 28 % less gas and 45 % less GOR constraint. This leads to an increase in total production
water, that is 44 % less gas and 68% less water compared to time and therefore to a reduction in NPV. In our example the
the base case of a horizontal well with friction; see Tables 4 NPV dropped to a negative value of about –1.3 million $ for
and 5. This illustrates that the ISP goes beyond breakthrough both the 20% and the 30% residual gas saturation case.
prevention and is also capable of keeping water and gas influx Cumulative oil and gas production remained almost equal
under control after breakthrough has occurred. Using the compared to the case without hysteresis; cumulative water
economic parameters of Table 6, the benefits of the ISP production went somewhat down.
amount to a positive differential NPV of 6.9 million $. In conclusion, it appears that gas saturation hysteresis has
a significant negative effect on the performance of the ISP as
Sensitivity analysis
the total production time increases sharply when hysteresis is
We repeated the ISP simulation for varying numbers of
taken into account. The effect of the degree of hysteresis, the
segments. Table 7 depicts the results in terms of operating
residual gas saturation, appears to be less of an influence.
parameters and NPV. The economic optimum in our case is
However, our hysteresis model was a simple one and further
three segments. However, this number may be different for
research in this area is required to reach more firm
other reservoir parameters, production constraints or economic
conclusions. It is noted that in the study reported in Ref. 6 no
conditions. Table 7 also shows that a larger number of
negative effects of hysteresis were experienced, although
segments results in a more quiet switching pattern as
details of the modeling were not presented.
expressed in the total number of valve operations, i.e. the total
number of cycles times the number of intervals. Furthermore, Discussion
a larger number of segments will allow more flexibility to This paper is based on work performed in 1999 and 2000.
react to unforeseen reservoir heterogeneities. A similar study was performed in parallel, but independently,
We also repeated the ISP simulation with different values and has been reported in Ref. 6. Not surprisingly many of the
for the reservoir and fluid parameters, which lead to the same conclusions of the studies are the same.
conclusion as for the SSC: the larger the effect of well bore Details of the completions in the studies are slightly
pressure drop, the larger the scope for improvement. However, different as illustrated in Figs. 4 to 6. This involves in
unlike the SSC, the ISP still leads to somewhat lower gas and particular the additional packer in the stinger completions
water production in cases of relatively low frictional pressure depicted in Fig. 5. Although the packer makes it easier to
drop. We furthermore investigated the performance of the ISP numerically simulate the well performance in a reservoir
in various slanted and heterogeneous reservoir configurations. simulator, it appears to offer no additional functionality, at
The results were situation-specific and cannot easily be least not in a reasonably homogeneous reservoir. The
generalized. In most cases the ISP maintained its capability to additional ICV depicted in Fig. 5 (bottom) offers increased
strongly reduce gas and water production, but sometimes at a flexibility for coning control after breakthrough has occurred,
price of strongly reduced oil rates compared to the but obviously at an expense.
homogeneous case, leading to a lower NPV. Both studies treated breakthrough prevention, through
flattening of the drawdown profile or ‘flow equilibration’ as it
Hysteresis
was named in Ref. 6, and coning control after breakthrough
An important aspect of the ISP is the repetitive approach
has occurred. The ISP is slightly different in this respect as it
and retreat of cones near the well bore. This may involve
focuses completely on coning control, and does not rely on
hysteresis effects in the saturations at pore level. The first time
6 J.D. JANSEN, A.M. WAGENVOORT, V.S. DROPPERT, R. DALING, AND C.A. GLANDT SPE 77942

delayed breakthrough. In addition to eliminating the effect of SSC

well bore pressure drop, the ISP can further reduce gas and • The primary function of the SSC is breakthrough
water production to levels below those obtained with an prevention through active control of the drawdown
‘ideal’ frictionless well. profile along the well.
Other new aspects in our study are the detailed • The secondary function of the SSC is partial coning
optimization procedure for the SSC, based on a semi- control at the intermediate inflow point after
analytical approach, the practical optimization criterion for the breakthrough has occurred.
SSC based on pressure drop over the stinger and the ICV, and • The major benefit of the SSC compared to a passive
the ISP algorithm. Furthermore we performed various extended stinger completion is its flexibility to
sensitivity studies, using differential NPV as a yardstick, to maintain the optimum drawdown profile during the
assess the robustness of the completion concepts under entire period prior to breakthrough.
different reservoir, fluid and completion parameters. • In our example, the SSC gave an oil production
Both studies assumed the availability of downhole phase profile closely resembling that of an ideal frictionless
measurement devices. In reality these are still very much well.
under development. Potential phase detection techniques • Gas and water production were higher than for a
include distributed temperature sensing (DTS), differential friction less well, but still well below those of the
pressure and temperature measurements over restrictions, base case of a conventional horizontal completion
acoustic techniques and behind-casing permanent resistivity with friction.
measurements. Sometimes information from surface-based • Maintaining the pressure drop over the ICV equal to
multiphase flow meters would be sufficient, for example to the pressure drop over the stinger is an effective
operate the ISP in a homogeneous reservoir where the closure control criterion to keep the drawdown profile as flat
sequence of the valves in a cycle is predictable. An interesting as possible.
option would be to base coning control on data obtained from
• The SSC offers little additional flexibility to
permanent behind-casing resistivity measurements, which
counteract the effects of unexpected reservoir
would allow for ‘pro-active’ water coning control, i.e. we
heterogeneities compared to a passive extended
could then increase the ICV pressure drop set point even
stinger completion with a fixed-size intermediate
before the water reached the well bore. inflow point.
Conclusions • Optimal design of the SSC, requires knowledge of
the permeability distribution along the well.
General
• If frictional pressure drop causes early water or gas ISP
breakthrough in a horizontal well, both the SSC and • The primary function of the ISP is to keep water and
the ISP can provide improvements in production gas influx under close control after breakthrough has
behavior through breakthrough prevention and/or occurred.
coning control. The larger the effect of the pressure • In our example, the ISP gave an oil production
drop, the larger the scope for improvement. profile closely resembling that of an ideal frictionless
• The economic feasibility of the SSC and the ISP well.
strongly depends on reservoir parameters, production • Furthermore, total gas and water production figures
constraints and economic boundary conditions. With were much lower that those for the friction less well
the parameters used in our study we achieved (44 % less gas, 68 % less water).
differential NPVs in a range from –2 to +8 million $ • The higher the number of segments, the lower is the
compared to a conventional horizontal completion. total number of valve switching operations. The
• Both completion types result in an increased pressure optimum number of segments for our example
drop over the completion for a given oil production was three.
rate, which reduces the lift potential of the well. In a • Gas saturation hysteresis appears to have a significant
gas-dominated production scenario, as considered in negative effect on the performance of the ISP
our paper, this is usually not a problem; in a water- because it increases the gas mobility. However, our
dominated case the effect will be more pronounced. hysteresis model was a simple one and further
• The future technical and economic feasibility of research in this area is required to reach more
breakthrough prevention and in particular coning firm conclusions.
control with techniques such as the SSC and the ISP
will strongly depend on development of cost- Acknowledgments
The work described in this paper was inspired by contacts
effective and reliable downhole phase detection
technology. with Duncan Green Armitage who pioneered the concept of
the extended stinger completion with intermediate inflow in
one of our operating units, and Rob Kleibergen, who brought
the concept to our attention.
SPE 77942 SMART WELL SOLUTIONS FOR THIN OIL RIMS: INFLOW SWITCHING AND THE SMART STINGER COMPLETION 7

References Vertical permeability 200 md

1. Dikken, B.J.: “Pressure Drop in Horizontal Wells and Its Effect Porosity 0.30 -
on Production Performance”, JPT (Nov. 1990) 1426-1433. Matrix compressibility 4.00*10-6 1/psi
2. Penmatcha, V.R., Arbabi, S. and Aziz, K.: “Effects of Pressure Initial pressure at GOC 3430 psi
Drop in Horizontal Wells and Optimum Well Length”, SPEJ Temperature 160 °F
(Sept. 1999) 215-223. Capillary pressure 0 psi
3. Ozkan, E., Sarica, C. and Haci, M.: “Influence of Pressure Drop Pressure regime: hydrostatic
Along the Well bore on Horizontal-Well Productivity”, SPEJ Relative permeabilities
(Sept. 1999) 288-301. Water-oil, residual oil saturations 0.20 -
4. Brekke, Kristian and Lien, S.C.: “New and Simple Completion Water-oil, relative oil permeability 1.00 -
Methods for Horizontal Wells Improve the Production Water-oil, Corey exponent oil 2.00 -
Performance in High-Permeability, Thin Oil Zones”, paper SPE Water-oil, residual water saturation 0.20 -
24762 presented at the 1992 SPE Annual Technical Conference Water oil, relative water permeab. 0.40 -
and Exhibition, Washington DC. Water-oil, Corey exponent water 2.00 -
5. Permadi, P., et al.: “Horizontal Well Completion With Stinger Gas-oil, residual gas saturation 0.05 -
for Reducing Water Coning Problems”, paper SPE 37464 Gas-oil, relative gas permeability 1.00 -
presented at the 1997 SPE Production Operations Symposium, Gas-oil, Corey exponent gas 2.00 -
Oklahoma City. Gas-oil, residual oil saturation 0.30 -
6. Sinha, S., Kumar, R., Vega, L. and Yalali, Y.: “Flow Gas oil, relative oil permeability 0.40 -
Equilibration Towards Horizontal Wells Using Downhole Gas-oil, Corey exponent oil 2.00 -
Valves”, paper 68635 presented at the SPE Asian Pacific Oil Linear iso-perms
and Gas Conference and Exhibition, Jakarta, Indonesia, April
Fluid properties
17-19, 2001.
7. Jansen, J.D.: “A semi-analytical model for calculating pressure
Oil gravity 35 °API
drop along horizontal wells with extended stinger Oil density at reservoir conditions 38.886 lbm/ft3
completions”, submitted for publication. Also available as Oil dynamic viscosity at res. cond. 0.0422 cp
“Numerical modeling of the flow in extended stinger Oil compressibility at res cond. 2.04*10-5 1/psi
completions”, report CTG-PW-00-003, Centre for Technical Bubble point pressure 3430 psi
Geoscience, Dept. of Applied Earth Sciences, Delft University Solution gas-oil ratio 1255 scf/stb
of Technology, Delft, The Netherlands, Dec. 2000. Water density at reservoir cond. 62.428 lbm/ft3
Water dynamic visc. at res. cond. 0.4000 cp
Gas specific gravity 1 -
Metric Conversion Factors Gas density at reservoir conditions 20.126 lbm/ft3
°API 141.5/(131.5+°API) = g/cm3 Gas dynamic viscosity at res. cond. 0.0350 cp
bbl × 1.589 873 E–01 = m3 Aquifer
cp × 1.0
*
E–03 = Pa s Strength 8.93*103 bbl/psi
Characteristic time 232 days
°F (°F-31)/1.8 = °C
*
in. × 2.54 E+00 = cm
*
ft × 3.048 E–01 = m TABLE 2 – WELL DIMENSIONS
lbm × 4.535 924 E–01 = kg
Variable Value Units
md × 9.869 233 E–04 = µm2
Well length 6562 ft
psi × 6.894 757 E + 00 = kPa
Stinger length for SSC 2871 ft
*Conversion factor is exact Number of sections for ISP 5 -
Casing outside diameter 7.0 in
TABLE 1 – BOX MODEL PROPERTIES Casing inside diameter 6.4 in
Casing roughness 0.0231 in
Variable Value Units Stinger outside diameter 4.5 in
Geometry Stinger inside diameter 4.0 in
Length (x direction) 6562 ft Stinger roughness o.0021 in
Width (y direction, full model) 820 ft
Height (z direction) 344 ft
Number of grid blocks in x direction 23 - TABLE 3 – PRODUCTION CONSTRAINTS
Number of grid blocks in y direction 60 -
Number of grid blocks in z direction 41 - Variable Value Units
Accumulation Minimum bottomhole pressure 1500 psi
Depth of GOC 7867 ft Maximum GOR 1500 ft3/bbl
Oil column 115 ft Maximum watercut 0.9 bbl/bbl
Depth of OWC 7982 ft Maximum liquid rate 20,000 bpd
Reservoir Minimum oil rate 400 bpd
Horizontal permeability 2000 md
8 J.D. JANSEN, A.M. WAGENVOORT, V.S. DROPPERT, R. DALING, AND C.A. GLANDT SPE 77942

TABLE 4 – PRODUCTION STATISTICS BASE CASE

AND FRICTIONLESS CASE
Variable Base Frictionless Units
case case
Cumulative oil 1.2*107 1.2*107 bbl
Cumulative gas 1.8*1010 1.4*1010 ft3
Cumulative water 3.4*106 2.0*106 bbl
Gas breakthrough time 36 399 days
Water breakthr. time 218 399 days
Total production time 6.2 5.4 years

TABLE 5 – PRODUCTION STATISTICS SSC AND ISP

Variable SSC ISP Units
Cumulative oil 1.2*107 1.2*107 bbl
Cumulative gas 1.5*1010 1.0*1010 ft3
Cumulative water 2.3*106 1.1*106 bbl
Gas breakthrough time 310 60 days
Water breakthr. time 310 not recorded days
Total production time 6.1 6.6 years

TABLE 6 – ECONOMIC PARAMETERS FOR NPV

ANALYSIS
Variable Value Units
Discount rate 8 %
Oil price 14 $/bbl
Lifting costs 0.05 $/bbl
Oil processing costs 0.17 $/bbl
Water processing costs 0.025 $/bbl
Gas processing costs 0.001 $/bbl
Variable Opex (as % of Capex) 10 %
Capex variable ICV 1,000,000 $
Capex on/off ICV 150,000 $
Capex pressure gauges 200,000 $
Capex phase monitoring equipment 150,000 $

TABLE 7 – ISP PERFORMANCE AS FUNCTION OF

NUMBER OF WELL SEGMENTS
Variable 2 3 4 5
Number of cycles 61 35 25 20
Total number of valve operations 122 105 100 100
Differential NPV (million $) 7.0 7.9 7.8 6.9
SPE 77942 SMART WELL SOLUTIONS FOR THIN OIL RIMS: INFLOW SWITCHING AND THE SMART STINGER COMPLETION 9

drawdown
additional packer

distance along well

well bore flow

heel toe
Fig. 1 – Horizontal well with skewed drawdown profile, decreasing
from heel to toe because of well bore pressure drop. Fig. 5 - Stinger completions as described in Ref. 6.

drawdown

distance along well Fig. 6 – Completion for the Interval Switching Process.

gas
oil
well
Fig. 2 – Extended stinger completion with somewhat flatter water
drawdown profile compared to Fig. 1.

drawdown

z x
half -model y
full -model
distance along well
Fig. 7 – Box model with fully penetrating, centered horizontal well.

2 isolated gridblocks to model ICV

main bore section to model casing
Fig. 3 – Extended stinger completion with intermediate inflow,
showing a further flattened drawdown profile compared to Fig. 2. main bore section to model stinger

heel toe

ICV
side-track to model casing-stinger annulus

branch point
perforated well segment
unperforated well segment
Fig. 4 – Smart Stinger Completion with remotely controllable
Inflow Control Valve (ICV).
Fig. 8 – Completion model for the Smart Stinger Completion.
10 J.D. JANSEN, A.M. WAGENVOORT, V.S. DROPPERT, R. DALING, AND C.A. GLANDT SPE 77942

20000
main bore to model stinger frictionless
ISP
heel toe
15000

oil rate [sbbl/day]

double side-track to model casing-stinger annulus
10000

primary branch point (can be switched on and off) basecase

secondary branch point
5000
perforated well segment SSC
unperforated well segment
0
Fig. 9 – Completion model for the Inflow Switching Process. 0 1 2 3 4 5 6 7
time [years]

20000 Fig. 13 – Production profiles for the base case, frictionless case,
SSC (red) and ISP (blue).
frictionless

15000
oil rate [sbbl/day]

10000

basecase
5000

0
0 1 2 3 4 5 6 7
time [years]
Fig. 10 – Production profiles for base case and frictionless case.

Fig. 14 – Saturation profile for the SSC. Gas and water are about
to break through simultaneously at the points of maximum
drawdown after about 310 days.

a) d)

Fig. 11– Saturation profile for base case after gas and water
breakthrough at the heel. b) e)
25

20
conventional
drawdown [psi]

15 stinger
SSC
10
c) f)
5

0
0 1000 2000 3000 4000 5000 6000
distance along well [ft]

Fig. 12 – Optimized drawdown profiles of three different

completions with the same production rate obtained with the
semi-analytical method of Ref. 7.
Fig. 15 – Saturation profiles during a complete cycle of the ISP.