You are on page 1of 15

Inflow Performance Relationship for

Perforated Horizontal Wells


Turhan Yildiz, SPE, Colorado School of Mines

Summary
This paper presents an investigation of perforated horizontal-well
performance in areally anisotropic reservoirs. Theoretical investigation is based on a 3D analytical IPR model. The analytical IPR
model considers an arbitrary distribution of perforations along the
completed segments. Changes in flow rate, pseudosteady-state
productivity, and cumulative production can be computed using
the solution.
The analytical IPR model was compared with the models available in the literature and verified. It was then used to investigate
the effects of well and reservoir parameters on the inflow performance of perforated horizontal wells.
Introduction
Horizontal wells may be perforated in selected intervals for several
reasons. The most common reasons for selective completion are
reducing the cost, delaying premature water/gas breakthrough, preventing wellbore collapse in unstable formations, and producing
multiple zones with large productivity contrast effectively.
Open-completed horizontal wells with negligible wellbore
pressure loss display a u-shaped influx profile; fluid velocities at
the heel and toe end of the well are higher than those at the
midsection of the well. Several simulation studies have shown that
water/gas prematurely breaks through at the heel end of the well
and causes inefficient sweep.
Uniform influx along the horizontal wellbore is desirable to
delay premature water/gas breakthrough and improve the sweep
efficiency. Water/gas breakthrough could be delayed by restricting
the flow and communication between reservoir and wellbore at the
intervals where local fluid velocities are higher.
Selective perforating with blank sections provides flexibility
for future intervention and workover options and for shutting off
the sections subject to excessive water/gas intrusion.
On the other hand, partial completion and enforcing uniform
inflow along the wellbore by variable shot density reduce the well
productivity. Therefore, a complete engineering analysis is required to weigh the gains from improved sweep provided by uniform inflow against the loss in well productivity.
The orientation of perforations is also a concern in optimizing
well productivity. Perforations aligned with minimum stress direction produce more sand. To reduce the risk of sand production,
it may be better to orient the perforations vertically. Additionally,
subsurface rocks exhibit horizontal permeabilities that are
higher than vertical permeabilities. Therefore, perforation tunnels
perpendicular to higher permeability would possess better
flow efficiency.
On the other hand, debris resulting from perforation process has
to be surged out of the tunnels to improve the productivity of the
perforated completions. It is more difficult to clean the perforations on the low side of the horizontal wells. Solid debris in the
low-side perforation tunnels may not be removed under the typical
underbalance pressures applied.
Vertically oriented perforation tunnels at the top side of the
horizontal wellbore are preferred for better perforation stability

Copyright 2004 Society of Petroleum Engineers


This paper (SPE 88987) was revised for publication from paper SPE 67233, presented at
the 2001 SPE Production Operations Symposium, Oklahoma City, Oklahoma, 2427
March. Original manuscript received for review 21 June 2002. Revised manuscript received
10 May 2004. Manuscript peer approved 17 May 2004.

September 2004 SPE Journal

and cleanup efficiency. However, if the perforations are to be


packed, it is difficult to transport the gravel into vertically oriented
tunnels at the top side.
Background
Perforating has been one of the most common completion methods
for vertical wells requiring sand control and prevention of wellbore
collapse and water/gas intrusion. The performance of perforated
vertical wells has been investigated extensively. However, only a
few modeling studies have dealt with the productivity of perforated horizontal wells. Field applications of the perforated horizontal wells have preceded the modeling efforts. Some of the field
applications and research studies are summarized below to acknowledge the previous contribution.
Field Applications. Horizontal wells are mostly completed openhole with slotted liners or prepacked screens. However, in some
fields, horizontal wells may need to be cased and perforated. Selective perforating has been implemented in the horizontal wells
drilled in many fields such as Bongkot,1 Joanne,2 Andrew,3
Oseberg,4,5 Wytch Farm,6 Statfjord,7 Elk Hills,8 Beryl,9 Dan,10
Alba,11 Yowlumne,12 and Prudhoe Bay.13 King14 pointed out the
significance of gun clearance and positioning in perforating and
gravel packing the horizontal wells.
Previous Modeling Studies. The performance of perforated vertical wells has been the subject of many modeling studies.1519
Perforation shot density and length and penetration ratio are found
to be the dominant parameters controlling the flow efficiency of
vertical wells.
Inflow performance models for horizontal open holes are relatively simple and well known.2022 However, these models ignore
the near-wellbore completion effects and consider that the total
drilled horizontal length is open to flow. In many cases, horizontal
wells may be completed at selected intervals along the well axis.
The horizontal wells partially completed at multiple producing
segments are referred as selectively completed wells. Several modeling studies on the performance and pressure behavior of selectively-completed horizontal wells have appeared in the literature.2326 Although these models account for selective completion
effect, they all ignore the details of flow convergence because of
the existence of perforations and slots in the near-wellbore region.
A few studies have addressed how the perforations and slots
influence the flow into horizontal wells.2734 Landman and Goldthorpe27 and Marett and Landman28 presented steady-state models
for flow into perforations distributed around a horizontal wellbore.
They also proposed the use of variable perforation shot density to
obtain a uniform influx along the wellbore. Gonzalez-Guevara and
Camacho-Velazques29 extended the model of Refs. 27 and 28 to
account for multiphase flow in the wellbore and reservoir and
non-Darcy flow around the perforations.
Asheim and Oudeman30 presented a simple algorithm to predict the perforation shot density yielding uniform influx along a
horizontal well. They concluded that optimal perforation distribution is rate-dependent, and enforcing uniform flux along the wellbore reduces the well productivity/injectivity.
Thomas et al.31 incorporated the near-wellbore skin term and
non-Darcy flow coefficient into their reservoir simulator and the
existing semianalytical IPR models for horizontal wells. They proposed the use of perforation pseudoskin nomograph developed for
vertical wells by Locke.16 They defined a global mechanical skin
factor (s), composed of perforation pseudoskin (sp), formation
265

damage skin (sd), and skin caused by rock crushing around the
perforations (sdp). They treat the global mechanical skin factor as
the sum of three individual skins listed. However, it can be shown
that three skin terms (sp, sd, and sd) are not additive.
Ozkan et al.32 presented a 3D transient-flow model for perforated horizontal and inclined wells and compared the pressure
transient responses of openhole, partially open (selectively completed), and perforated horizontal wells. The simulated results
have shown that the transient component of convergent flow
around perforations subsides very fast. Beyond early-time flow
regime, the convergent flow near perforations results in a stabilized additional pressure drop around the wellbore. The model
presented in Ref. 32 may require extensive CPU time for long
horizontal wells.
Recently, Goktas and Ertekin33 developed a numerical simulator for perforated horizontal wells. Their model relies on flexible
and locally refined grids to capture the details of the convergent
flow around the perforation accurately. They stated that their numerical model is limited to low shot densities, otherwise demanding extensive computation time.
Most recently, Tang et al.34 extended the perforation pseudoskin model proposed by Ozkan et al.32 and investigated the
impact of perforation parameters on horizontal well performance.
They observed that the perforation densities higher than 0.5 spf
yield a marginal increase in well productivity.
In a previous paper,35 it has been shown that, in fully perforated
vertical wells, an ideal combination of perforation length, shot
density, and phasing angle may result in a negative perforation
pseudoskin. Under steady-state flow conditions, this negative
pseudoskin caused by perforating may yield a 10 to 15% productivity improvement over openhole completed vertical wells. However, in most cases, productivity improvement diminishes with
time if the reservoir is sealed at the outer lateral boundaries and the
well produces at constant wellbore pressure.
The objectives of the present study are: (1) to build a model for
selectively perforated horizontal wells in bounded reservoirs with
no-flow external boundaries, (2) to examine the flow rate and
cumulative production responses of perforated horizontal wells
under constant wellbore pressure, and (3) to investigate the impact
of different perforating schemes on the long-term performance of
horizontal wells.
Flow Model for Perforated Horizontal Well
In a previous study,32 it was shown that the transients caused by
the convergence of streamlines around perforations dies out very
fast. Hence, the energy loss caused by convergent flow into perforation tunnels can be represented as an additional stabilized pressure drop beyond very early times.
To construct a simple and computationally fast model, the 3D
flow into a perforated horizontal well is decomposed into two
smaller subproblems: a transient 3D model for flow into selectively completed horizontal well, and a perforation totalpseudoskin model for accounting the flow convergence around the
tunnels in near-wellbore region. This model will be referred to as
the decoupled model. The modeling concept is illustrated in
Fig. 1. The decoupled model concept has previously also been
applied to selectively perforated vertical wells.19,35
Selectively Completed Horizontal Well (SCHW) Model. A multisegment horizontal well in a rectangular parallelepiped reservoir
with impermeable external boundaries is considered. Multisegmentation allows us to account for the local changes around the
wellbore. The SCHW assumes that the completed intervals are
fully open to flow all around the perimeter of the segment. A
variable local skin around each segment is also incorporated into
the SCHW model. The additional pressure change caused by perforations is superimposed on SCHW in terms of local skin. A
schematic of reservoir model for multisegment horizontal well is
given in Fig. 2. The mathematical treatment of the SCHW model
is presented in the Appendix.
266

Model for Perforation Total Pseudoskin (PTP). Here, we make


use of the PTP model developed for perforated vertical wells. The
justification for the use of vertical well perforation pseudoskin is
that perforation length is very small compared to horizontal well
length and the distance to the external reservoir boundaries. Therefore, 3D convergent flow around the perforations is not influenced
by far field-flow conditions. However, PTP calculated for a
vertical well has to be properly scaled for use in horizontal
well models.
PTP model takes care of the flow into perforation tunnels, flow
in the damaged zone around the wellbore, and flow across the
crushed/compacted zone around the tunnels. A schematic of the
PTP model is displayed in Fig. 3. The mathematical model for PTP
has been described in two previous publications.19,35 The reader
is referred to Refs. 19 and 35 for a complete description of the
PTP model.
The perforation pseudoskin is calculated for only a unit formation thickness (for example, 1 ft of thickness). Therefore, the PTP
model is computationally very efficient.
Verification of the Decoupled Model. First, we measured the
SCHW model up to the single-segment models proposed by Babu
and Odeh20 and Goode and Kuchuk21 under pseudosteady-state
flow conditions and observed excellent agreement. Then, the
SCHW model was compared against the partially open horizontal
well model presented by Goode and Wilkinson.23 The comparison
is made in terms of productivity ratio for an 800-ft partially open
horizontal well producing at constant flow rate under pseudosteady-state flow conditions. Productivity ratio (PR) is defined
as the ratio of the productivity index of a partially open well to the
productivity index of a fully open horizontal well. Three different
completion scenarios are considered. The comparison is displayed
in Fig. 4. As can be seen, an excellent agreement between the
models has been established. Also, the proposed model was compared against the model presented in Ref. 25. The current model is
for SCHW in bounded reservoirs. The well-testing model described in Ref. 25 is for SCHW in infinite reservoirs. The comparison, depicted in Fig. 5, is for a 900-ft-long horizontal well with
five 100-ft-long open segments producing at a constant flow rate.
The models agree very well until no-flow reservoir boundaries are
felt at the wellbore. The proposed model was also compared with
the model of Retnanto et al.24 However, the results from our model
and the model of Ref. 24 significantly deviated.
The PTP model has already been previously compared against
the experimental data and the Karakas and Tariq18 model, showing
good agreement. The details on the verification of PTP model can
be obtained in Ref. 19.

Fig. 1Decoupled model for perforated horizontal wells.


September 2004 SPE Journal

Fig. 2Multisegment selectively-completed horizontal well model.

After verifying the individual components of it, the decoupled


model itself was tested against the results presented in Refs. 32 and
34. Unlike the decoupled model, the pressure transient analysis
model of Ozkan et al.32 considers all the individual perforations
around the wellbore. The comparison with the Ozkan et al.32
model is shown in Fig. 6. Except for the early-time flow period at
which only the formation around individual perforations contributes to flow, very good agreement is observed between the models.
Finally, the proposed model was compared with the results given
by Tang et al.34 We calculated the transient productivity index of
the well described in the data set given in Table 2 of Ref. 34. The
PI values predicted from the decoupled model are identical to the
values given in Table 3A of Ref. 34.
The decoupled model is computationally efficient. It takes less
than a minute (typically 15 to 20 seconds) on a high-end PC to
simulate the 10-year rate and cumulative production responses of
the cases displayed in Figs. 7 through 20.

Limitations of the Model. There are several other factors that


may influence the performance of a perforated well but are not
considered in the model. The pressure losses inside the perforations may be important in high-rate wells. However, this factor has
not yet been incorporated into the model.
In several numerical studies on perforated vertical wells, it has
been shown that non-Darcy flow around the perforations could
substantially reduce the flow efficiency. The impact of non-Darcy
flow is not considered in this study.
Near-wellbore heterogeneity and accumulation of fines around
perforations may also contribute to perforation efficiency. Heterogeneity and fines movement are not accounted for in the model.
Discussion
In this section, using the decoupled model, we simulate different
perforation schemes and discuss the impact of perforating param-

Fig. 3Perforation total pseudoskin model.19


September 2004 SPE Journal

267

Fig. 4Comparison of SCHW model with Goode and Wilkinson model.23

Fig. 5Comparison of SCHW models in bounded and infinite reservoirs.

Fig. 6Comparison of the decoupled model to the exact perforated horizontal well model of Ozkan et al.32

Fig. 7Effect of shot density on flow-rate decline.


268

September 2004 SPE Journal

Fig. 8Effect of shot density on cumulative production.

eters on rate decline, cumulative production, and the productivity


during boundary dominated flow period. The data set given in
Table 1 is used in the simulations.
Effect of Shot Density. To examine the impact of perforation shot
density on well performance, we considered shot densities of 0.25,
0.5, 1, 2, 4, 8, 12, and 16 spf. The other perforation parameters
used in the simulations are given in Figs. 7 and 8. First, we calculated perforation total skin for a fully perforated vertical well
(sptV). Then sptV was scaled for horizontal well by multiplying it
with h/Lh. The scaled perforation total pseudoskin for horizontal
wells will be referred to as sptH; sptV and sptH for each perforation
scheme are listed in Table 2. Shot densities higher than 2 spf
create a negative pseudoskin. After predicting sptH, we introduced
it into our SCHW model. We simulated how the flow rate and
cumulative production change as a function of time when the well
produces at a constant wellbore pressure drop of 500 psi. The
results are shown in Figs. 7 and 8. Initially, there is a considerable

difference in flow rates. However, initial high flow rates, because


of negative perforation pseudoskin resulting from high shot density, diminish with time. For each perforation scheme, the cumulative productions at the end of 1 and 5 years are given in
Table 3. The cumulative productions from a horizontal well perforated with 2 spf are almost identical to those from openhole.
Higher shot densities improve the cumulative production slightly.
For each scenario, we also determined productivity ratio. The
productivity ratio is defined as the ratio of the productivity index
of a perforated well to the productivity index of an open hole under
pseudosteady-state flow conditions when the well produces at a
constant flow rate. The productivity ratios for different cases are
listed in Table 3. The productivity ratio for a perforated well with
2 spf is 0.98.
The results in Table 3 and Figs. 7 and 8 show that a perforated
well with a shot density of 2 spf performs as well as an open hole.
Shot densities larger than 2 spf marginally improve the long-term
well performance.

Fig. 9The influence of perforation length on transient-rate response.


September 2004 SPE Journal

269

Fig. 10Effect of perforation length on cumulative production.

Influence of Perforation Length. To investigate the impact of


perforation length on well performance, we considered perforation
lengths of 3, 6, 12, 18, and 24 in. and shot density of 4 spf. sptV and
sptH for each perforation scheme are listed in Table 4. 3- and 6-in.
perforations cause positive perforation pseudoskin of 0.2114 and
0.0677, respectively. Longer perforations yield small negative
pseudoskin values. Flow rate and cumulative production responses
are displayed in Figs. 9 and 10. Longer perforations initially produce at somewhat higher flow rates. However, except for the
completion with 3-in. perforations, all the completion schemes
result in approximately the same production rate after one month.
The cumulative production at the end of 1 and 5 years and productivity ratios are tabulated in Table 4. The completion with
24-in. perforations recovers 10% more fluid than that with 3-in.
perforations at the end of 5 years. The cumulative production and
productivity for the horizontal well shot with 6-in. perforations are
slightly lower than those for openhole. Marginally higher recov-

eries and productivity ratios are obtained if the well is shot with
longer perforations.
Impact of Phasing Angle. To examine the effect of phasing angle,
we considered 0, 180, and 90 phasing. The other perforation
parameters were fixed at spf4, Lp12 in., and dp0.2 in. Figs.
11 and 12 display the insensitivity of transient flow rate and cumulative production to phasing angle. The cumulative productions
at the end of 1 and 5 years and productivity ratio under pseudosteady-state flow conditions are listed in Table 5. All the results
indicate that phasing angle has negligible impact on the performance of perforated horizontal wells in isotropic formations.
Although not shown here, we also investigated the effect of
perforation diameter and observed that the productivity of perforated horizontal wells is insensitive to perforation tunnel diameter.
It should be stated that this conclusion applies to the case for which
the pressure losses inside the perforation are negligible.

Fig. 11Impact of phasing angle on transient rate response.


270

September 2004 SPE Journal

Fig. 12Effect of phasing angle on cumulative production.

Combinations of Perforation Parameters. In the discussion


above, we only scrutinized the effect of one parameter by keeping
the rest of the perforation parameters constant. Here, we change all
the perforation parameters and consider pessimistic, reasonable,
and optimistic combinations of them. The perforation parameters
of each case are given in Table 6. The simulated rates and cumulative productions for each case are displayed in Figs. 13 and 14,
respectively. Table 7 lists perforation pseudoskin, productivity
ratio, and cumulative productions for each case. The optimistic
case yields a small negative pseudoskin, while the pessimistic case
generates a positive pseudoskin of 1.37. There is a significant
difference in rate and cumulative production responses of pessimistic and optimistic combinations of perforation parameter. The
optimistic perforation scheme recovers 44% more oil than the
pessimistic case at the end of the 5-year period. The rate, cumulative production, and productivity ratio responses to reasonable
combination of perforation parameters are almost identical to those
for openhole. This exercise shows that although well productivity

may be weakly sensitive or insensitive to a given perforation parameter, if all the perforation parameters act in the same direction,
then well performance and cumulative production may be strongly
influenced by perforation design.
The Impact of Formation Damage. Many studies have concentrated on how formation damage and rock compaction impact the
flow efficiency in perforated vertical wells. Here, we investigate
the influence of permeability reduction caused by formation damage around the horizontal wellbore. We considered that the permeability in the damaged zone was 25% of the original. Additionally, spf2, Lp12 in., dp0.2 in., and p180. The radius of
damaged zone around the wellbore was varied from 6 to 15 in. The
results are shown in Figs. 15 and 16 and Table 8. It is observed
that as long as the perforation penetrates beyond damaged zone
(rwd6, 9, and 12 in.), the loss in productivity is below 10%.
Otherwise, if the perforation tunnel does not extend beyond the
damaged zone, the productivity impairment is more severe. When

Fig. 13Transient-rate response to pessimistic, reasonable, and optimistic combinations of perforation parameters.
September 2004 SPE Journal

271

Fig. 14Cumulative productions for pessimistic, reasonable, and optimistic combinations of perforation parameters.

the whole tunnel is inside the damaged zone (rwd15 in.), then the
loss in productivity is 20%. Therefore, the perforations should
be designed to penetrate beyond the damage zone around
the wellbore.
The Impact of Crushing Around Perforation Tunnels. Many
experimental studies on the API RP 43 setup have concluded that
there exists a crushed zone around the perforation tunnel.36 The
permeability of the crushed zone is about 10% of the original
formation permeability. Although the extent of the crushed zone
varies along the perforation tunnel, an average and constant
crushed-zone thickness of 0.5 in. is accepted. The results showing
the combined impact of rock crushing around the tunnels and
formation damage around the wellbore are summarized in Figs. 17
and 18 and Table 9. We considered kcp/k0.1 and kd/k0.25 in
the simulations. It can be observed that the rock crushing reduces

the well productivity much more severely. When the perforation


has a crushed zone and does not extend beyond damaged zone
around the wellbore, the total loss in well productivity is 32%. The
rock crushing alone results in a 25% productivity impairment.
Thus, it can be stated that as long as there exists a crushed zone
around the perforation tunnel with a permeability lower than damaged-zone permeability, then the productivity impairment caused
by the crushed zone is more severe and formation damage has a
lesser influence. The productivity loss caused by formation damage is marginal, especially if the perforation tunnel extends beyond
the damaged zone.
Effect of Phasing in Anisotropic Formations. We investigated
the impact of perforation phasing angle in an anisotropic formation
with kz/kx0.1 and kykx. Additionally, we considered spf2,
Lp1 ft, and dp0.2 in. The results are given in Figs. 19 and 20

Fig. 15Effect of formation damage on transient-rate response of perforated horizontal wells.


272

September 2004 SPE Journal

Fig. 16Impact of formation damage on long-term cumulative productions from perforated wells.

and also in Table 10. First, formation anisotropy significantly


impacts the performance of open completed horizontal well. While
the horizontal open hole in isotropic formation produces 310,000
STB oil in 12 months, the same well in anisotropic formation
recovers 226,200 STB oil in 13 months. In general, when perforations are oriented vertically, well performance is enhanced because the perforations are normal- to high-permeability directions.
When all the perforations are vertical and phasing angle is 180,
the well performs the best. In this case, the productivity of the
perforated well is approximately 15% better than that of the openhole. 90 phasing, which orients half of the perforations vertically
and half of them horizontally, yields the second-best performance.
Vertical perforations with 0o phasing result in the third-best recovery. However, these results are only valid for spf2. We also
simulated the effect of phasing in anisotropic formations with less
dense perforations. For spf0.125, vertical perforations with 0

phasing produce more fluid than 90 phasing. At low shot densities, perforations do not interfere with each other. In such a case,
negative impact of 0 phasing diminishes.
The well shows the worst performance when all the perforations are horizontally oriented with a phasing angle of 0.
Conclusions
Based on the discussion presented above, the following conclusions are reached. It should be emphasized that the conclusions
listed pertain to the range of the data used in the simulations.
1. An analytical IPR model for perforated horizontal wells has
been developed. The model is compared against the several
models in the literature and verified.
2. The perforation pseudoskin model for vertical wells can be used
to predict the performance of perforated horizontal wells. How-

Fig. 17Influence of crushed zone and formation damage on transient-rate decline in perforated wells.
September 2004 SPE Journal

273

Fig. 18Cumulative production affected by crushed zone and formation damage.

ever, the perforation pseudoskin for vertical wells has to be properly rescaled for use in horizontal-well performance calculations.
3. In undamaged formations, a horizontal well perforated with a
shot density of 2 spf and 1-ft-long tunnels performs as effectively as the well-completed open hole. Denser perforations
yield slightly better cumulative recovery and productivity.
4. As long as it is not too short (longer than 3 in.), perforation
length has marginal impact on the long-term cumulative recovery and well productivity in undamaged formations.
5. In isotropic formations, perforation phase angle has an insignificant effect on transient rate, cumulative production, and productivity of perforated horizontal wells.
6. Even when well performance may be weakly sensitive to a
given single perforation parameter, if all the perforation param-

eters act in the same direction, then well performance and cumulative production may be strongly influenced by perforation design.
7. If the perforation tunnel terminates inside the damaged zone
around the wellbore, up to 20% loss in the productivity may be
encountered. If the tunnel created by perforating penetrates beyond the damaged zone, then the detrimental effect of formation
damage is considerably less.
8. The productivity of perforated horizontal well is significantly
reduced because of permeability impairment in the crushed zone
around the perforation tunnels. As long as there exists a crushed
zone around the perforation tunnel with a permeability lower
than damaged-zone permeability, then formation damage has a
lesser influence.
9. Formation anisotropy has a significant impact on well produc-

Fig. 19Effect of permeability anisotropy on transient-rate decline in perforated wells.


274

September 2004 SPE Journal

Fig. 20Effect of permeability anisotropy on cumulative production.

tivity. The well has the best performance when all the perforations are vertically oriented with a phasing angle of 180. The
well productivity is the lowest when all the perforations are
horizontally oriented, with a phasing angle of 0.

September 2004 SPE Journal

275

Nomenclature
Bo formation volume factor, RB/STB
ct total compressibility, psi1
h height, ft
k permeability, md
kcp permeability of crushed zone around perforation
tunnel, md
kd permeability of damaged zone, md
K0 zero order modified Bessel function of the first kind
L reference length
Lh horizontal well length
Ls segment length
nseg number of selectively completed segments
p pressure, psi
pi initial reservoir pressure, psi
PR productivity ratio
q flow rate, STB/D
qj flow rate at the jth segment, STB/D
qw total well flow rate, STB/D
r radius, ft
rw wellbore radius, ft
rwe equivalent wellbore radius, ft
s Laplace space variable
sptH perforation total pseudoskin rescaled for horizontal
wells
sptV perforation total pseudoskin for vertical wells
t time, hours
xe length of the reservoir in x-direction
xs location of the segment center in x-direction
ye width of the reservoir in y-direction
yw location of the well in y-direction
z vertical direction, ft
zw location of the well in vertical plane, ft
viscosity, cp
porosity
Subscripts
d wellbore damage
D dimensionless variable
i initial

276

r
s
w
x
y
z

radial direction
segment
wellbore
x direction
y direction
vertical direction

References
1. Horn, M.J., Plathey, D.P., and Ibrahim, O.: Multilateral Horizontal
Well Increases Liquids Recovery in the Gulf of Thailand, SPEDC
(June 1998) 78.
2. Thomson, D.W. and Nazroo, M.F.: Design and Installation of a CostEffective Completion System for Horizontal Chalk Wells Where Multiple Zones Require Acid Stimulation, SPEDC (September 1998) 151.
3. Kusaka, K. et al.: Underbalance Perforation in Long Horizontal Wells
in the Andrew Field, SPEDC (June 1998) 73.
4. Sognesand, S., Skotner, P., and Hauge, J.: Use of Partial Perforations
in Oseberg Horizontal Wells, paper SPE 28569 presented at the 1994
SPE European Petroleum Conference, London, 2527 October.
5. Sognesand, S.: Evaluation of Oseberg Horizontal Wells After 4 Years
Production, paper SPE 36864 presented at the 1996 SPE European
Petroleum Conference, Milan, 2224 October.
6. Blosser, W.R. et al.: Unique ESP Completion and Perforation Method
Maximizes Production in World Record Step-Out Well, SPEDC
(March 2000) 7.
7. Kostl, P. and stvang, K.: Completion and Workover of Horizontal
and Extended-Reach Wells in the Statfjord Field, SPEDC (December
1995) 211; Trans., AIME, 299.
8. Gangle, F.J. et al.: Improved Oil Recovery Using Horizontal Wells at
Elk Hills, California, SPEDC (March 1995) 27; Trans., AIME, 299.
9. Sadek, H., Williams, C.R., and Smith, M.: Production Strategy Yields
Unique Perforating, Cost-Efficient Multilateral Drilling Solution in the
Beryl Field, paper SPE 50123 presented at the 1998 SPE Asia Pacific
Oil and Gas Conference, Perth, Australia, 1214 October.
10. Damgaard, A.P. et al.: A Unique Method for Perforating, Fracturing,
and Completing Horizontal Wells, SPEPE (February 1992) 61.
11. Alexander, K., Winton, S., and Price-Smith, C.: Alba Field CasedHole Horizontal Gravel Pack: A Team Approach to Design, SPEDC
(March 1996) 31.

September 2004 SPE Journal

12. Marino, A.W. and Shultz, S.M.: Case Study of Stevens Sand Horizontal Well, paper SPE 24910 presented at the 1992 SPE Annual
Technical Conference and Exhibition, Washington, DC, 47 October.
13. Pucknell, J.K. and Broman, W.H.: An Evaluation of Prudhoe Bay
Horizontal and High-Angle Wells After 5 Years of Production, JPT
(February 1994) 150.
14. King, G.E.: Perforating the Horizontal Well, JPT (July 1989) 671.
15. Bell, W.T., Sukup, R.A., and Tariq, S.M.: Perforating, SPE Monograph Vol. 16, Richardson, Texas (1995).
16. Locke, S.: An Advanced Method for Predicting the Productivity Ratio
of a Perforated Well, JPT (December 1981) 2481.
17. Tariq, S.M.: Evaluation of Flow Characteristics of Perforations Including Nonlinear Effects With the Finite-Element Method, SPEPE
(May 1987) 104; Trans., AIME, 283.
18. Karakas, M. and Tariq, S.M.: Semianalytical Productivity Models for
Perforated Completions, SPEPE (February 1991) 73; Trans., AIME,
291.
19. Yildiz, T.: Productivity of Selectively Perforated Vertical Wells,
SPEJ (June 2002) 158.
20. Babu, D.K. and Odeh, A.S.: Productivity of a Horizontal Well,
SPERE (November 1989) 417.
21. Goode, P.A. and Kuchuk, F.J.: Inflow Performance of Horizontal
Wells, SPERE (August 1991) 319.
22. Renard, G. and Dupuy, J.M.: Formation Damage Effects on Horizontal-Well Flow Efficiency, JPT (July 1991) 786.
23. Goode, P.A. and Wilkinson, D.J.: Inflow Performance of Partially
Open Horizontal Wells, JPT (August 1991) 983.
24. Retnanto, A. et al.: Optimization of the Performance of Partially
Completed Horizontal Wells, paper SPE 37492 presented at the 1997
SPE Production Operations Symposium, Oklahoma City, Oklahoma,
911 March.
25. Yildiz, T. and Ozkan, E.: Transient Pressure Behavior of Selectively
Completed Horizontal Wells, paper SPE 28388 presented at the 1994
SPE Annual Technical Conference and Exhibition, New Orleans, 25
28 September.
26. Kamal, M.M. et al.: Pressure-Transient Analysis for a Well With
Multiple Horizontal Sections, paper SPE 26444 presented at the 1993
SPE Annual Technical Conference and Exhibition, Houston, 36 October.
27. Landman, M.J. and Goldthorpe, W.H.: Optimization of Perforation
Distribution for Horizontal Wells, paper SPE 23005 presented at the
1991 SPE Asia Pacific Conference, Perth, Australia, 47 November.
28. Marett, B.P. and Landman, M.J.: Optimal Perforation Design for
Horizontal Wells in Reservoirs With Boundaries, paper SPE 25366
presented at the 1993 SPE Asia Pacific Oil and Gas Conference, Singapore, 810 February.
29. Gonzalez-Guevara, J.A. and Camacho-Velazquez, R.: A Horizontal
Well Model Considering Multiphase Flow and the Presence of Perforations, paper SPE 36073 presented at the 1996 Latin American and
Caribbean Petroleum Engineering Conference, Port of Spain, Trinidad
and Tobago, 2326 April.
30. Asheim, H. and Oudeman, P.: Determination of Perforation Schemes
To Control Production and Injection Profiles Along Horizontal Wells,
SPEDC (March 1997) 13.
31. Thomas, L.K. et al.: Horizontal Well IPR Calculations, SPEREE
(October 1998) 392.
September 2004 SPE Journal

32. Ozkan, E., Yildiz, T., and Raghavan, R.: Pressure-Transient Analysis
of Perforated Slant and Horizontal Wells, paper SPE 56421 presented
at the 1999 SPE Annual Technical Conference and Exhibition, Houston, 36 October.
33. Go ktas, B. and Ertekin, T.: Performances of Openhole Completed and
Cased Horizontal/Undulating Wells in Thin-Bedded, Tight Sand Gas
Reservoirs, paper SPE 65619 presented at the 2000 SPE Eastern Regional Meeting, Morgantown, West Virginia, 1719 October.
34. Tang, Y. et al.: Performance of Horizontal Wells Completed With
Slotted Liners and Perforations, paper SPE/CIM 65516 presented at
the 2001 SPE/CIM International Conference on Horizontal Well Technology, Calgary, Alberta, 68 November.
35. Yildiz, T.: Impact of Perforating on Well Performance and Cumulative Production, Journal of Energy Resources Technology (September
2002) 163.
36. Karacan, C.O. and Halleck, P.M.: Mapping of Permeability Damage
Around Perforation Tunnels, paper ETCE 2000-10036 presented at
the 2000 Engineering Technology Conference on Energy, New Orleans, 1417 February.

AppendixTransient Model for Multisegment


Horizontal Well
The 3D diffusivity equation governs the flow into a multisegment
horizontal well. The governing differential equation is as follows:
2pD
xD2

2pD
yD2

SD = GD

2pD

zD2

nseg

qiD

i=1

siD

+ SD =

pD
. . . . . . . . . . . . . . . . . . . . . . . . (A-1)
tD

HxD i HxD i+ . . . . . . . . . . . . . (A-2)

GD = 2hD kD zD zwDyD ywD. . . . . . . . . . . . . . . . . . . (A-3)


One initial and six boundary conditions are needed to solve the
differential equation given above. The initial and boundary conditions are given in the following:
pD xD ,yD ,zD,0 = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-4)
pD
0,yD ,zD ,tD = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-5)
xD
pD
x ,y ,z ,t = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-6)
xD eD D D D
pD
x ,0,z ,t = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-7)
yD D D D
pD
x ,y ,z ,t = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-8)
yD D eD D D
pD
x ,y ,0,t = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-9)
zD D D D
pD
x ,y ,h ,t = 0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-10)
zD D D D D
277

Additionally, for constant flow rate at the wellbore,


nseq

= 1, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-11)

iDtD

i=1

and for the production under constant wellbore pressure condition,


nseg

qwD tD =

iD tD

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-12)

i=1

pwD tD = 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-13)
The dimensionless variables are defined in the following:
For constant flow rate,
pD =

kh
p px,y,z,t . . . . . . . . . . . . . . . . . . . . . . (A-14)
141.2qBo i

qD = qsx,y,z,tLh qw. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-15)


For constant wellbore pressure case,
pD =

pi px,y,z,t
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-16)
pi pw

qwDtD = 141.2qwtBo kh pi pw . . . . . . . . . . . . . . . . . (A-17)


NpDtD = Np Bo 1.119 ct hL2 pi pw . . . . . . . . . . . . . . . (A-18)
tD = 2.63679 104kt ct L2 . . . . . . . . . . . . . . . . . . . . . . . (A-19)
thD = 2.63679 104kt ct Lh 22 . . . . . . . . . . . . . . . . . . (A-20)

mnkl = Ko

yD = yk ky L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-23)
zD = zk kz L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-24)
hD = hk kz L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-25)
kD = k2 kx ky . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-26)
LhD = Lhk kx L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-27)

i = xsiD LsiD 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-28)


i+

= xsiD + LsiD 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-29)


nseg

LtD =

iD

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-30)

rw
rweD = kz ky0.25 + ky kz0.25. . . . . . . . . . . . . . . . . . . . . (A-31)
2L
In the definitions listed above, k and Lthe characteristic permeability and characteristic length, respectively. In this study, we
choose k(kxkykz)1/3 and LLh/2.
We have solved Eqs. A-1 through A-13 using the Laplace and
Finite Fourier Cosine transformations. First, we assume uniform
but unknown and different flux along each completed segment;
then, we use the pressure-averaging technique. The final Laplace
space solution for segment j is as follows:
nseg

A s q
ji

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-32)

iDs

i=1

Ajis = F1jis + F2jis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-33)


4

kD hD
F1jis =
xeD n=1
F2jis =

2kD hD xeD

0nkl

. . . . . . . . . . . . . . . . . . . . . . . (A-34)

k=0 l=
4

LhiD LhjD n=1


2

278

. . . . . . . . . . . . . . . (A-37)

2
22 = 2ywD 2l yeD2 + zkn
. . . . . . . . . . . . . . . . . . . . . . . . . . . (A-39)

zk1 = 2hD k + rweD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-40)


zk2 = 2hD k + 1 rweD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-41)
zk3 = 2hD k + 2zwD + rweD . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-42)
zk4 = 2hD k + 1 2zwD + rweD . . . . . . . . . . . . . . . . . . . . . . (A-43)

m = s + m xeD2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-44)
Rxmi = sin mi+ xeD sin mi xeD. . . . . . . . . . . . . . . (A-45)
The additional pressure change caused by perforating, formation
damage, and rock crushing around each segment is incorporated
into the solution by assuming that the flow around each segment is
normal to segment axis. This yields
qjBo

psj = 141.2

kz ky Lsj

sptVj, . . . . . . . . . . . . . . . . . . . . . . . . (A-46)

where sptVj=the perforation total pseudoskin across the jth segment. The dimensionless pressure change caused by local skin
effect is
k

h
sptVj = qjD sptHj . . . . . . . . . . . . . . . . . (A-47)
L
kz ky sj

psjD = qjD
sptHj =

h
sptVj. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-48)
L
kz ky sj
k

If the permeability anisotropy is accounted for in the calculation of


perforation pseudoskin, then, in Eq. A-48, k/(kzky)1/21 should be
set.
Adding the additional pressure-change expression into the solution in Eq. A-32, we obtain
nseg

pjDs =

A s q
ji

iDs

+ qjDssptHj. . . . . . . . . . . . . . . . . . . (A-49)

i=1

The solution up to this point is general and does not include any
assumption related to inner-wellbore flow condition. For infiniteconductivity and constant flow rate at the wellbore,
pjDtD = pwDtD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-50)

RxmiRxmj
. . . . . . . . . (A-35)
m2

iDtD

k=0 l= m=1

= 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-51)

i=1

Rewriting Eq. A-49 for each completed segment, we obtain


aji = Aji. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-52)
ajj = Ajj sptHj . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-53)

0 1

1 a11 a12 a1ns


1 a21 a22 a2ns

1 ans1 ans2 ansns


pwD

1s

q1D

q2D

qnsD

. . (A-54)

The solution of the matrix in Eq. A-54 gives the wellbore


pressure and fractional-flow rates across each segment.
For constant pressure in an infinite-conductivity wellbore, the
solution becomes
pwDtD = 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-55)

mnkl

s + K s . . . . . . . . . . . . . . . . . . (A-36)

0nkl = Ko

nseg

i=1

pjDs =

1 + Ko

2
21 = 2l yeD2 + zkn
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-38)

txeD = 2.63679 104kt ct xe2 . . . . . . . . . . . . . . . . . . . . . (A-21)


xD = xk kx L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-22)

nseg

qwDtD =

iDtD .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-56)

i=1

September 2004 SPE Journal

Combining Eqs. A-49, A-55, and A-56, the solution for constantwellbore pressure drop is found:

N ptDs = q wDs s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-63)

aji = Aji . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-57)


ajj = Ajj + sptHj . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-58)

1 1 1 1
0 a11 a12 a1ns
0 a21 a22 a2ns

0 ans1 ans2 ansns


q wD

q 1D

1s

q 2D

q nsD

1s

. (A-59)

1s

1s

The solution of the matrix described in Eq. A-59 results in


wellbore flow rate and rate distribution along the wellbore.
Once the wellbore and segment flow rates are known, the cumulative production from the wellbore and each segment could be
predicted as follows:
NpiDtD =

tD

qiDd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-60)

N piDs = q iDs s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-61)


NptDtD =

tD

qwDd . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-62)

September 2004 SPE Journal

SI Metric Conversion Factors


bbl 1.589 873
E01 m3
cp 1.0*
E03 Pas
ft 3.048*
E01 m
E02 m3
ft3 2.831 685
in. 2.54*
E+00 cm
lbf 4.448 222
E+00 N
lbm 4.535 924
E01 kg
md 9.869 233
E04 mm2
*Conversion factor is exact.

Turhan Yildiz is an associate professor in the Petroleum Engineering Dept. at the Colorado School of Mines. e-mail:
tyildiz@mines.edu. Previously, he worked for the U. of Tulsa,
Simulation Sciences, Istanbul Technical U., and Louisiana State
U. Currently he is involved in modeling of complex reservoir
flow problems and in intelligent/multilateral well design. Yildiz
holds a BS degree from Istanbul Technical U. and MS and PhD
degrees from Louisiana State U., all in petroleum engineering.
He serves on the SPE Editorial Committee.

279

You might also like