SPE 80926
Effect of Non-Uniform Skin on Finite Conductivity Horizontal Well
Belhouchet Toufik, SONATRACH Djebbar Tiab and Sarfraz Jokhio University of Oklahoma
Copyright 2003, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the SPE Production and Operations Symposium
held in Oklahoma City, Oklahoma, U.S.A., 22–25 March 2003.
This paper was selected for presentation by an SPE Program Committee following review of
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
correction by the author(s). The material, as presented, does not necessarily reflect any
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
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
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 where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836 U.S.A., fax 01-972-952-9435.
Abstract
This paper investigates the performance of horizontal
wells under the combined influence of wellbore friction,
wellbore storage, and formation damage. A new model is
proposed wihich couples wellbore, reservoir, and non-uniform
skin, and takes into account the change in flow regime from
laminar to turbulent along the wellbore.
In the conventional perspectives, wellbore damage
may be viewed as an additional pressure drop, is correct only
for infinite-conductivity wellbore with uniform damage. It is
shown in this paper that the wellbore damage may not be
viewed merely as an additive pressure drop due to skin in the
presence of friction inside the long horizontal section of the
well. Wellbore hydraulics change the flux distribution (inflow
profile) along the wellbore and, thus, result in additional
pressure drop in the reservoir and across the skin zone.
Horizontal wells, for a number of reasons, are unlikely to have
smooth distributions of damage along their trajectories. Thus,
segmented well testing would be ideal tool to estimate the
skin, Si, of each segment especially in the zone where more
severe damage is expected, because it delivers a detailed
picture of the skin distribution along an extended horizontal
well.
Finite conductivity solution approaches to infinite
conductivity solution at high conductivity values and high
wellbore radius. The appearance of bi-linear, linear, and radial
flow regimes depends upon the length of the well; for long
wells, linear and bi-linear flow regimes are observed, and they
are masked in the case of short wells. The perforation number
effect appears only in the damaged case, when the open-hole
scenario deviates from the two other cases. Another important
factor, that affects the behavior of the pressure, is the
dimensions of altered permeability region; the higher the
radius of filtrate penetration, the higher the volume of affected
region and therefore, the higher the loss in pressure. The effect
of k/ks is more severe in small wells. Another correlation
which allows determine kx directly from the intersection point
with the half slope is presented using a new method of
interference test interpretation of different segments with each
other in the same well. Isolated segment testing would lead to
the estimate the local skin factor, a possible indication of
uneven damage distribution and a necessary variable for the
optimization of matrix stimulation. Thus designing of a
remedial treatment, both the type and the location of
permeability damage must be considered.
Several examples are included to illustrate the use of
the model developed.
Introduction
During the last three decades significant advances in
drilling technology have made it possible to drill horizontally.
Although slow, this technology is being applied worldwide
because of its affordable cost. Horizontal well due to its large
flow area, may be several times more productive than vertical
well. However, intuition tells that this large contact of the
horizontal well with the formation will result more formation
damage than the vertical well will. The residing mud filtrate
and also the mud particulates can cause damage to the zone by
reducing the permeability around the wellbore. This in turn
results in higher pressure losses in the vicinity of the wellbore
and thus reduces the productivity of the well. Any damage
around the wellbore is, therefore, never desired. Since
repairing the permeability damage is generally difficult and
expensive, all studies emphasize the importance of preventing
it. It is usually agreed that the uniform skin region is an
idealized case particularly for long horizontal wells. Since the
duration of the contact of the drilling fluid with the formation
decreases from the heel of the well to the toe, at best a conical
skin region with the apex at the toe should be expected.
Various physical forces such as grinding, crushing, bit
vibration, and fluid as well as particulate invasion are at work
at the well bore during the drilling of a long horizontal well
which cuase the changes in the radius and permeability of the
skin zone along the well length. This may even lead to
situations where some portions of the well are closed to inflow
because of highly damaged areas. The same effect would, of
course, result if the well were completed selectively.This
paper adresses issues concerning well pressure and flux
distribution under the combined influences of mechanical
damage and wellbore hydraulics.
In the first part of this study, we show that the pressure and
the inflow profile along the well length are affected by
2 SPE 80926

wellbore hydraulics and non-uniform damage; this, in turn, An equivalent radius with Vcone = Vcylinder thus can be defined
affects the well deliverability. The second part deals with the as:
estimation of the ratio (k/ks) along the well length using a new  1  2aSH, minaSV, min + aSH, maxaSV, min + 
0.5

correlation, and finally, a method allowing to evaluate the rS,equiv =    (9)

permeability kx in the case when the late radial is not  6  aSV, maxaSH, min + 2aSH, maxaSV, max 
observed. Particular attention is paid to the propper Because permeability anisotropy2 is crucial in flow regime
formulation of the wellbore model and the two correlations calculations for horizontal wells, this equation is modified as:
presented here.  1  a SH
2 
0 .5
 , max + 
Skin Distribution rS/ , equiv = (10)
 3I ani a 2 
Let’s assume a horizontal well with the following   SH , max a SH , min + a SH , min 
hypothesis: (1) the reservoir is horizontal, homogeneous, The skin effect caused by a zone of altered permeability
isotropic and has constant thickness h, constant porosity φ, (2) around a horizontal well is expressed as:
the well of length L and radius rw is assumed to run along the   A 2D , max + A D , max a D , min 
 k  1 4 (11)
X-axis at an elevation Zw with respect to the bottom of the /
S equiv =  − 1  ln   
I + a2 
formation, (3) the skin region of radius, rS, that may be a
k
 S   ani + 1 3
 D , min  

function of X, is assumed to be concentric with the horizontal Where
well. Let kr denotes the effective permeability in the « Y.Z » a SH , max a SH , min
plane of the reservoir and kS the permeability of the skin zone. A D, max = A D, min = (12)
qh and qhS are the fluxes entering and leaving the skin zone. rw rw
The pressure drop between the skin boundary and the wellbore Concept of Infinite Conductivity Transient
at some point X is: Pressure Solution
141 .2µ r (1) The solution for infinite-conductivity model is achieved by
PS ( rS , x , t ) − PS ( rw , x , t ) = q hS ( x , t ) ln S
kS rw applying the Green’s function and the product solution method
Similarly, if the permeability of the skin zone were the same with source function, presented by Gringarten and Ramey8.
as that of the reservoir, then the pressure drop across the skin The pressure distribution is:
 
zone would be: tD 
kH
+ xD
kH
− xD 
π kH  kX kX   y 2D 
141.2µ r
P (rS , x, t ) − P(rw , x, t ) = q h ( x, t ) ln S (2) PD ( x D , y D , y wD , z D , z wD , t D ) =
4 kY ∫ erf
 2 τ
+ erf
2 τ 



 exp − 4τ 

kr rw 0
 
 
Then the additional pressure drop at the wellbore because of
 ∞  dτ
the skin region is given by:
141.2µ
1 + 2

∑ exp( − n 2 L 2D π 2 τ ) cos( n π z D ) cos( n π z wD ) 
 τ
(13)
∆PS (rw , x, t) = ∆PS (rS , x, t) + q hS (x, t)Sm (x, t) (3) n =1
kr The pressure derivative with respect to the logarithm of
Where dimensionless time is given as follows:
∆PS (r, x , t ) = P(r, x , t ) − PS (r, x , t ) , r = rw ..or..rS (4)  kH kH 
 + xD − xD 
And δPD π kH  k X k X   y2D 
 *
τPD/ (x D , yD , ywD, z D , z wD, t D ) = = erf + erf  * exp −
δ ln t D 4 k Y  
k q ( x , t )  rS 2 τ 2 τ   4τ 
Sm ( x , t ) =  r − h  ln (5)  
 
 k S q hS ( x , t )  rw
 ∞
 (14)
The skin factor given above may be expressed in terms of flux 1+ 2 ∑
n =1
exp( − n 2 L2D π 2τ )cos( n π z D )cos( n π z wD )  τ

distribution as:
kh
∆PS Horizontal Well Model
Pf (rw , x , t ) − PS (rw , x, t ) 141.2qBµ (6)
Shm = = In most models, the wellbore hydraulics of horizontal
Lk r  dP  q hSD
r  wells is simulated to that of horizontal pipe. This is because of
hk  dr  r = rw , x
the geometrical similarities between a horizontal well and a
Where horizontal pipe.
q hSD =
q h ( x, t )L
=
Lk r  dp 
r (7)
q 141.2qBµ  dr  r = rw , x
Friction Factor for a Pipe with Flow Influx or Out-
Development of an Analytical Expression for the flux
Skin in a Horizontal Well Ouyang et al23 developed wall-friction-factor correlation
During horizontal well drilling, the exposure of the well to for fluid flow through pipe with fluid influx and out-flux. The
drilling and completion fluids results not in a uniform extent different correlations are based on that of solid pipe friction
of the invaded zone but rather in a truncated elliptical cone factor and influx or out-flux Reynolds number. For production
with the larger base near the vertical section of the well as wells, the well friction factor is given by the following
shown in Fig. F-1 and F-3. The volume of this cone with an correlations:
elliptical base is:
Lπ  2a SH , min a SV , min + a SH , max a SV , min +  (8)
Laminar flow f m =
64
N Re
1 + 0.04304 N 0Re,
.6142
w (15)( )
V =
cone  
6 a SV , max a SH , min + 2a SH , max a SV , max 
SPE 80926 3

(
Turbulent flow f m = f m 0 1 − 0.0153N 0Re,
.3987
w ) (16) The dimensionless horizontal well conductivity is defined as11:
1 π d 4w
where N Re, w = N Re, j CH D = (22)
nd p 4 16k H Ze L
Combining these equations, the dimensionless frictional
1 pressure drop can be written as:
and f m0 = 2
(17)
 21.25   j 
2
1.17 − 2Log R p + 0.9  π

 N Re, j  PDfj ( t m ) = N Re f j  q (t ) 
16n CH D  k = i
/ D
 ∑ (23)

Finite Conductivity Model n/ is the number of segments.
The semi-analytical model described in this section is
based on the assumptions made for infinite conductivity Reservoir Model
horizontal well model as well as the assumptions made for The infinite conductivity solution for horizontal well
finite conductivity model, these assumptions are: pressure behavior in dimensionless variables is given by the
• The reservoir is horizontal, homogeneous or anisotropic, following equation9:
and has constant horizontal kH, and vertical kV, tD
π kH  xD + j / n xD −(j −1) / n/ 
/
xD +(j −1) / n/ xD − j / n/
permeabilities ‘kH = kV’.
• It has constant thickness h, porosity φ, and effective
PD(xD, yD, zD, tD) =
4 kY 
0
∫
erf
2τ
−erf
2τ
−erf
2τ
−erf
2τ
*


formation compressibility Cf, and infinite lateral  ∞  n  y 2D  dτ
extension.

∑
n =1
2 2 2
1 + 2 exp(−n LD π τ) cos(n πz D ) cos(n πz wD )

∑Q 1

DJ (t D ) exp −


 4τ  τ
• The production occurs through a well of length Lw, and a
radius rw located at a distance Zw from the lower (24)
boundary. In the above equation, the unknowns, qDj are the local
• A single-phase fluid, of small and constant dimensionless production rates at the dimensionless
compressibility c, constant viscosity µ, and FVF “B”, coordinates (xD, yD, zD, tD). The solution that is considered in
flows from the reservoir to the well at a constant rate q. this study for the infinite-conductivity idealization is obtained
from the uniform flux solution by dividing the well length into
• The reservoir is bounded by upper and lower
small segments each with uniform flux per unit area. The
impermeable formations.
procedure used to determine the flow rate distribution at each
• The effect wellbore storage is neglected, and uniform
time-step level is as follows:
initial pressure is assumed in the reservoir.
• The transient period is divided into, m, different intervals
• The relative pipe roughness is assumed constant along the
and dimensionless length is divided into, n, segments
wellbore.
containing the different trances of the skin as presented in the
• Flow regime can change from laminar to turbulent wellbore-reservoir system (Fig. F-3).
depending on the flow rate inside the wellbore.
• Each of these segments produces a dimensionless flow
• The pressure gradient at any cross sectional area along the rate equal to qDj per unit length.
wellbore is negligible.
• Assuming that the flux entering each segment is uniform
and has stepwise distribution in time and space; i.e., qDj, is
Wellbore finite-conductivity model constant for segment, j, and at time step level, m, and
As shown in Fig. F-2, the wellbore with half-length, Lf, is independent of production history.
divided into, n, equal segments. Fluid flow through the
• Assuming also that the skin at each segment has a conical
segment, j, creates a pressure drop that can be evaluated as
shape then the resulting system of n equations can be solved.
follows:
2
Hence, the dimensionless pressure for each segment can be
8Lρf j  j  expressed as:
∆Pfj ( t m ) = ∑
Nπ 2d 5  k = i
q k ( t ) 

(18)
π
tD
 1 + ( x D − x DU ) 1 − ( x D − x DU ) 
 PDj ( x Dj , y D , z D , t D ) =
4 ∫
q Dj erf
0
2 τ
+ erf
2 τ
*

 ∞  dτ
exp − D * 1 + 2∑ exp(−n 2 L2D π 2 τ) cos(n πz D ) cos(n πz wD )
Using the following dimensionless variables, we obtain the y2
dimensionless expression of pressure loss in each segment at  4 τ   n =1  τ
time-step level, m. (25)
2πk H Ze Model Coupling
PDf ( t m ) = ∆Pf (19)
q oµ The effect of a skin around the well should be incorporated
j into the analytical solution as follows4
N Re
N Re, j =
qo ∑q
k =1
k (t m ) (20) ∆Pwith skin = ∆Pwith no skin + ∆Pdue to skin
Introducing PD concept:
 j
2
 4k Z L  q S( x ) 
PDWS = PDNS + D (26)
 ∑
PDfj ( t m ) = N Re f j  q k ( t )  H e
 n / d4
(21)  x D LD 

 k =i 
4 SPE 80926
q hS 2x
Where q D = , xD = , (27)
q L
LD =
L
(Isotropic medium) and LD = L kV (Anisotropic
2h 2h kH
medium).
And
( )
S ( x ) = S equiv = k −1 ln  1
kS
4 (A 2
 I ani +1 3
D , max + AD , max a D , min + a D , min )
Where
a SH, max a SH , min
A D, max = , A D, min =
rw rw
Calculation Procedure
In order to calculate the changing production rate along the
wellbore during the transient behavior of a horizontal well,
one haa to solve simultaneously the pressure drop equations.
In other words, one has to equate pressure drop in each
segment with the reservoir drawdown, assuming the continuity
of pressure between the wellbore system and the reservoir
system. The combination of these two systems results in a set
of systems of, n, non-linear equations.
Model Procedure Solution
To solve this system of non-linear equations, we adopted
the Newton-Raphson method. We assume that the flow regime
within the horizontal wellbore changes from laminar to
turbulent at the value of 2300, in order to avoid the
discontinuity in the friction factor at the boundary between the
laminar and the turbulent flow regimes. This system requires
an intensive mathematical computation, so a computer
program was written to evaluate the solution numerically.
Methodology of the Study: This study performs
calculation of pressure loss and flux distribution. Single-phase
oil flowing within both the reservoir and the horizontal
wellbore is assumed. This simplicity enables us to use an
equation of fluid flow in a conduit coupled with an equation of
reservoir inflow and including the skin effect. To simplify the
problem here, a cone-like skin distribution (Fig. F-3) is
assumed. The skin radius is highest at the heel and shortest or
negligible at the tip end. The changes in permeability within
the damage zone along the horizontal wellbore are also
assumed the same or same k/ks, within the entire damage zone
for a given case. k here represents the original permeability,
and ks is the reduced permeability within the damage zone.
Data Used
A homogeneous, isotropic reservoir system is employed,
different dimensionless wellbore lengths are used “LD=0.5, 1,
5, 10, and 50”. The borehole is located in the middle of the
drainage volume, different dimensionless horizontal well
conductivity values are used CHD=0.01, 0.5, 5, 50. For the
given reservoir system, wellbore size and damage severity are
both varying. Conventional horizontal wellbore sizes, i.e. 95/8,
7, 5 in. diameter, and a horizontal slim hole of 3.5-in diameter
are used. A well roughness of 0.3 is used for different
scenarios of open hole and liner perforated with different
perforation densities.
Effect of CHD
From all plots indicating the behavior of the pressure
for different LD values, in the case when the skin is not
involved, the pressure and the pressure derivative are fairly
2 

affected by CHD (Fig. 1), but once, the skin is introduced, we
observe a big difference in pressure plots for each CHD value.
(28) Fig. 2 shows that the skin has a considerable effect going from
low to high values of CHD, in the pressure plots. While for the
pressure derivative plots (Fig. 3), CHD presents a fair effect
(29) comparing with pressure response. As a result, we can
conclude that the CHD presents a distinguishable effect
especially in the pressure plots, when this effect is combined
with the skin. The scenario of skin is presented for kD=k/ks=8
and AD=AShmax/rw=10.
Effect of rw
First, we discuss the first scenario. We choose here
the results of the case 3.5-in. wellbore diameter as a reference.
As shown in Fig. 5, systems with larger wellbore diameters
and no damage approach characteristics of an infinite
conductivity well, while the 3.5.in. wellbore exhibits a finite
conductivity well. For these particular cases, the pressure loss
in a 3.5 in-hole is higher than that in the other well sizes, and
the pressure at the tip of 3.5 in hole is higher but much lower
at the heel end than that in the other well sizes. Fig. 4 shows
that the specific influx increases toward the heel in the case of
no skin. The presence of damage around and along the
wellbore can change in particular the distribution of the
specific influx, for our model of non-uniform skin distribution
(Fig. 1). One can expect that the specific productivity index
increases toward the toe because the damage zone reduces
toward the tip end. As compared to the undamaged wells, the
pressure distribution is similar (Fig. 6), but the presence of the
damage reduces the pressure at any point as much as 1.2. Also
we see that the influx is reduced in the vicinity of the heel
because of the large value of the skin at this section (Fig. 7).
Furthermore, following interesting results are observed:
• For larger diameters; i.e. 5, 7 and 95/8-in used here,
the specific influx at the heel end decreases while it
increases toward the toe.
• For smaller diameters, the influxes at the heel and the
toe are higher than that at other portions of the well,
as demonstrated in Fig. 7.
This reveals that the presence of non-uniformly damage
with k/ks=8 reduces influx at the heel end and this situation is
balanced by an increase in influx at the toe to achieve the
given dimensionless rate ‘ΣQd=1’.
Another scenario treated in the case where all the wellbore
systems have a fixed skin radius (Ash, max = constant) at the
heel end. This means that a smaller wellbore has a thicker
damage zone. This scenario can represent conditions caused
by lower penetration rate resulting in a longer exposure time
to the drilling mud when drilling small wellbore diameters.
Fig. 8 demonstrates the distribution of flowing pressure for
four different wellbore sizes. Since the damage zone is thicker
for the smaller wellbore diameter, then the flowing pressure is
also lower along the entire length of a smaller wellbore.
SPE 80926
Effect of ks
In order to understand the effect of this non-uniform
shape of skin on the pressure distribution, we have
investigated the effect of damage severity on pressure and
pressure derivative response, and inflow behavior. An increase
in k/ks results in higher PD “lower flowing pressure” at any
point along the wellbore. For LD = 0.5, the increase in PD is 95
as a result of skin damage from zero “k/ks = 1” to k/ks = 10,
and 335 when k/ks = 50 (Fig. 9). For longer wells LD = 50,
increase in PD “reduction in flowing pressure” due to presence
of damage is slightly smaller as shown in Fig. 10. Concerning
the pressure derivative P’D, the response on the derivative plot
appears clearly in shorter wells (LD=0.5) and take a slight
variation for long wells (LD=5), Fig. 11 and 12 show this
effect. Fig. 13 demonstrates the redistribution of the influxes
as the damage becomes severe. The resistance to flow around
the downstream portion of the well increases, resulting in
influx reduction.
Comparison between Solution in Case when Skin
Damage is considered and Neglected
To show clearly the effect of non-uniform skin on finite
conductivity solution, we compare the pressure drop solutions
considering the skin (conical shape) to that where we neglect
this effect (S=0). The three scenarios were considerd::
1. Np=574, ∆p=0.2813 2. Np=497, ∆p=0.3559 and 3 open
hole completion.
For short wells (LD=1) we observe that there is a big
difference between the finite conductivity solutions in the case
when considering the effect of skin and when neglecting it
(Fig. 14), however, for long wells of (LD=5) this difference
tend to be minimal and neglected, otherwise, there is a big
difference comparing the first two solutions to that of open
hole solution in all plots (Fig. 15).
As presented in plots of QD (Figs. 14 and 15), the
open hole solution gives a high quantity of the fluid flowing
from the heel, but it decreases sharply towards the toe
especially for small wells. Concerning the pressure plots, the
openhole response presents a high PD (low flowing pressure)
which enables the well to receive large quantities of fluid (Fig.
16), the pressure derivative presents a distinguishable effect
(Fig. 17).
Effect of LD
The investigation of LD effect leads us to have a more
understanding about the development of the different flow
regimes in a horizontal well. Plots giving pressure and
pressure derivative responses for short wells (LD=1) indicates
the development of the early and the late radial flow regime,
but the linear flow regime is masked, which is due to the
length of the well. While in the presence of the skin (non-
uniform), a bi-linear flow regime is developed for LD=1 (Fig.
18). For other well lengths, we observe the development of the
bi-radial flow for LD=10 with the presence of early and late
radial (Figs. 19 & 20). For LD=50 the linear flow is developed
in addition to the early and the late radial flow regime
(Fig. 21).
We can also conclude that the effect of the same
value of skin for different LD becomes negligible in long
5
wells, and the distinction becomes difficult in pressure
derivative plots.
Importance of Near-Wellbore Permeability
The effect of permeability in the skin zone has an
extremely disproportionate effect on well productivity thus
any damage around the wellbore is, therefore, never desired
because its two fold financial implication: (1) revenue loss due
to lost production and (2) remedial cost.
Skin Correlation
Damage mechanism identification requires a
systematic approach to research, planning, and evaluation of
all available information. Because, the evaluation of reservoir
formation sensitivity due to changing conditions during
petroleum reservoir exploitation requires multi-disciplinary
team effort and the integration of various instrumental and
analytical approaches (Kersey, 1998). Some methods such as
well test interpretation may be used to infer for limited
information on a few critical parameters of reservoir
formation. However, direct measurements of core properties at
reservoir conditions are preferred, because they provide the
most realistic information about the petroleum-bearing
formations. The intent of this paper is also to get familiarize
with a new correlation leading us to use the well test data and
interpretation to calculate the ratio (k/ks) for any well at any
location without going to laboratory.
A.The Non-uniform Case
The equation that correlates the different parameters is:
S = 0.6489A D + 0.1696 (30)
Once AD is determined, we can use the following equation to
determine the ratio k/kS
 k   1 4 2 
/
Sequiv =  − 1 ln (AD, max + AD, maxa D, min + a 2D, min )
 kS   Iani + 1 3 
(31)
B.Uniform Case
The equation that correlates the different parameters is
S = AD (32)
A graphical form for the determination of all parameters from
the same plot is shown in the Figs. 22 & 23.
Horizontal Well testing of Isolated Segments
Horizontal wells, for a number of reasons, are unlikely to
have smooth distributions of damage along their trajectories.
Thus, segmented well testing would be an ideal too to estimate
the individual skin, Si, especially in the zones where more
severe damage is expected. Several stimulation techniques
(e.g. employing coiled tubing) could use this information to
optimize the distribution of stimulation fluids.
Testing of Isolated Segments
The main criterion for selection of the length of the
tested interval is the presence of flow regimes of interest. It is,
for example useful to see both linear and late radial flow in a
well test. This would require a certain minimum well length to
develop linear flow as well as a certain maximum segment
6 SPE 80926

length in order to minimize the time after which late radial 947φµc t L2g
develops (Figs. 24 & 25). For long horizontal wells, this kx = (40)
consideration favors segmented testing. Segmented testing can t hump C 2Lw
Lg
be optimized17 by placing a pressure gauge in an isolated non-
flowing segments, this test configuration would result in an Where C Lw is a shape factor defined as follows:
Lg
interference test of the well with itself. It would be, however, a 2 3 4
prerequisite to keep the well segment between the tested Lw L     
C Lw = 1 − 0.4273 + 0.1642 w  − 0.03069 L w  + 0.002229 L w 
section and the pressure gauge imperforated. Lg Lg  Lg   Lg   Lg 
     
(41)
Development of the Pressure Response
Case Study
The dimensionless expression for the pressure derivative
The case study shown here is divided into two main parts.
measured at a gauge located at a distance from the tested
The first case, a field case, which allows us to check the
segment with open length Lw in a thin reservoir of thickness
reliability of the correlation used to determine the ratio k/kS,
h, and lateral boundaries are assumed to be far away is
three samples are presented. The second case is a comparative
developed as follows:
example between the correlation developed in this study and
The instantaneous point source solution is:
the one presented by Frick17.
hD  x 2D 
PD = exp −  with h D = h and x D = x
4(πt D ) 1/ 2   L L First Case
 4t D 
The pressure response can be described, after integrating along The data presented below are measured in the laboratory
the open segment by: in order to determine the damage degree k/kS. This is a
recapitulative table for different well results:
h  L gD L gD + L wD  ∆Pi = ∆Pf =30 atm
PD = D erfc − erfc  (33)
4π  2 tD 2 t D 
Well L 1/Qi 1/Qf Ki Kf Ki/kf Cor.1 Cor.2
For the continuous solution: (cm) s/CC s/CC mD mD
tD
hD  L gD L gD + L wD  MD401 4.8 25 55 0.494 0.224 2.2 2.16 2.08
PD = ∫
4π 
0
erfc
2 tD
− erfc
2 t D 
dt D (34) GB3 3.67 12 78 0.787 0.412 1.91 2.01 1.94
MD486 2.8 34 67 0.721 0.366 1.97 1.93 1.86
Concerning the pressure derivative, we obtain the expression: MD46 4.9 29 52 0.435 0.242 1.79 2.17 2.09
∂P h  L gD L gD + L wD  OM522 3.2 20 44 0.411 0.187 2.2 1.96 1.89
P' = D = D erfc − erfc  (35)
∂t D 4π  2 tD 2 t D 
Second Case-2
This derivative exhibits a characteristic hump, the time when The second example, which is analyzed to investigate
this hump occurs can be related directly to the permeability in the application of the isolated segment testing, uses the
the direction of the well, kx, by: following data:
947φµc t L2w φ = 0.1, µ = 2 cP, Ct = 10-4 Psi-1, Lw = 1000 ft, Kx = 100 mD
kx = (36)
t hump TLg −1 Reservoir width = 200 feet. In this case two gauges are placed
Lw at 50 and 300 feet apart. The intersection points with the linear
The function T is defined as follow: flow (half slope) are as follow (Fig. 29): 400 and 305 hours
2 For 50 feet perforated
 Lg  L 
TLg = 1.89  + 0.60747 g  + 0.244 (37) Lg =1050 , Lw = 50, t i.linear = 400 hr
  Lw  • Using our correlation
Lw  Lw   
The maximum (Fig. 27) occurs at: 947φµc t L2w
kx = −1
t hump, D = 0.4726(L gD / L wD ) 2 + 0.5517(L gD / L wD ) + 0.061 t hump TLg
Lw
(38) 2
With a correlation coefficient equals to 1.  Lg  L 
TLg = 1.89  + 0.60747 g  + 0.244 ⇔
On a log-log plot of the derivative versus time, a negative half   Lw 
Lw  Lw   
slope for t>thump can be observed. This is shown in Fig. 27.
2
Fig. 28 is the graphical presentation of the function T. A  1050   1050 
dimensionless pressure derivative presented by Frick17 has the TLg = 1.89  + 0.60747  + 0.244 =881.0593
Lw  50   50 
following form:
 Then
∂P 1 L gD L gD + L wD 
P' = D = erfc − erfc  (39) 947 * 0.1 * 2 * 0.0001 50 2
∂t D 4h D L D πt D  2 tD 2 t D  kx = =104 mD
400 881.059−1
This derivative exhibits also a characteristic hump, with a unit
slope. The permeability can be related to the hump’s time by
the relation:
SPE 80926 7

• Using the correlation presented by Frick17 • Another correlation allowing us to determine kx directly
Thump is difficult to locate, let’s take 270 C2 = 0.96 and from the intersection point with the half slope is presented
9470.1 * 2 * 0.0001 (1050 )
L2g
2 using a new method of interference test interpretation of a
947φµc t
kx = = =80. well segment with another one.
t hump C 2Lw 270 0.74 • Permeability damage can occur anywhere along the flow
Lg
path, from formation to perforation and into the wellbore.
When we take t hump = 240hr kx = 90.6 mD, and for t hump =210 Isolated segment testing would lead to the calculation of
hr, then kx = 104mD the local skin factor, a possible indication of uneven
For 300 feet perforated damage and a necessary variable for the optimization of
Lg =800 , Lw = 300, t i.linear = 305 hr matrix stimulation, thus when designing a remedial
• Using our Correlation treatment, both the type and location of permeability
 Lg 
2
 Lg  damage must be considered.
TLg = 1.89   + 0. 60747   + 0.244 ⇔
  Lw 
Lw  Lw    Nomenclature
2
 800   800 
TLg = 1.89  + 0.60747  + 0.244 =19.589 Iani Permeabilty anisotropy
Lw  300   300  h Thickness
Then φ Porosity
2 Sm (x) Mechanical skin factor
947 * 0.1 * 2 * 0.0001 300
kx = =109 mD asv Half-axes of the elliptical base
305 19.589 −1 AD,max a SH, max
A D, max = dimensionless half-axes of the
• Using the correlation presented by Frick17 rw
Thump is 184 C2=0.74 and, elliptical base
9470.1 * 2 * 0.0001 (1050)
2 2
947φµc t L g LD Dimensionless well length
kx = 2
= =88.8 ρ Fluid density
t hump C Lw 184 0.74
Lg q Flow rate
µ Fluid viscosity
When we take t hump = 170 kx = 96.1 mD.
D Internal pipe diameter
ε Absolute pipe roughness
Remarks
n Number of perforation
As presented in the examples, the correlations developed
x Distance measured along the well length
in this work allow us to estimate k / kS and kx with high qw Well rate at any point along the well length
precision in comparison with classical methods. J Jacobian matrix
Re Reynolds number
Conclusions CHD Horizontal well dimensionless conductivity
• The developed model couples wellbore, reservoir, and Bo Formation volume factor
non-uniform skin, and takes into account the change in Ct Total compressibility
flow regime from laminar to turbulent along the wellbore. QjD Dimensionless flow rate at the segment j
• Finite conductivity solution approaches to that of infinite tD Dimensionless time
conductivity solution at high conductivity values and high rw Wellbore radius
wellbore radius. kD Dimensionless permeability
• The appearance of bi-linear, linear and radial flow
regimes depends upon the length of the well; for long Vertical section
wells, linear and bi-linear are observed, and they are aSV,max dv/dx
masked in the case of short wells. aSV,min
• The perforation number effect appears only in damaged
case, when the openhole scenario deviates from the two aSH,max aSH,min
other cases.
• Another important factor, that affects the behavior of the Fig. F-1 Damage shape around vertical and horizontal wells.
pressure, is the dimensions of altered permeability region;
the higher the radius of filtrate penetration, the higher the
volume of affected region and therefore the higher the total flow
q1
loss in pressure. q +q
1 2
Σ qk ∑qk ∑q q
k o
• The permeability ratio (k/kS) is directly proportional to the x=0
skin, and its effect is more severe in small wells.
• The correlation from which we calculate the ratio (k/kS) q1 q2 qj qn-1 qn
presents a new reliable way to use in order to minimize
time, effort, and costs involved in laboratory work.
Fig.F-2 Wellbore finite conductivity presentation.
8 SPE 80926

9. Adalberto J.R. and Carvalho, S.: “A mathematical

model for pressure evaluation in an infinite-
ASH,max conductivity horizontal well”, Paper SPE 15967
presented at the SPE Formation Evaluation,
aSH,min
Dec.1989.
X=L 10. Beggs, H.D.: Production optimization using nodal
analysis, OGCI Publications., Tulsa,1991, 57-67.
11. Ozkan, E, Sarica, C. and Haciislaoglu, M.: “Effect
Zw of conductivity on horizontal well pressure
Y behavior”, Paper SPE 24683 presented at the 67th
Annual Technical Conference and Exhibition of
X
SPE, Washington, DC, October 4-7, 1992.
Z 12. Clonts and Ramey, H.J.: “Pressure transient
analysis for wells with horizontal drainholes”,
Fig. F-3 Reservoir model. Paper SPE 15116 presented at the 56th California
Wall friction shear Regional Meeting of SPE, Okland, April 2-4, 1986.
13. Permadi, P. and Wibowo, W.: “Effect of non-
uniform skin distribution on horizontal well inflow
performance”, Paper SPE 68952 presented at the
SPE European Formation Damage Conference,
r
Inlet x g Outlet Netherlands, May 21-22, 2001.
Gravity Flodirection 14. Ozkan, E.: “Analysis of horizontal-well responses:
Contemporary vs.Conventional”, Paper SPE 72494
presented at the 1999 SPE Mid-Continent
Fig. F-4 Horizontal well model configuration Operations Symposium, Oklahoma, March 28,
2001.
15. Renard, G. and Dupuy, J.M.: “Formation damage
References effects on horizontal–well flow efficiency”, JPT
1. Engler, T., Osisanya, S. and Tiab, D.: “Measuring jul.1991.
skin while drilling”, Paper SPE 29526 presented at 16. Baba, A.: Finite Conductivity Horizontal Well
the Production Operation Symposium, Oklahoma, Transient Behavior, MS thesis University of
2-4 Apr 1995. Oklahoma, Norman, Oklahoma, 2001.
2. Frick, T.P. and Economides, M.J.: “Horizontal well 17. Boussis, H.: An Integrated Finite Conductivity
damage characterization and removal”, Paper SPE Transient Behavior Model for Horizontal Wells,
21795 (Feb.1993). MS thesis, University of Oklahoma, Norman,
3. Byrom, T.G.: “Some mechanical aspects of Oklahoma, 2001.
formation damage and removel in horizontal 18. Frick, T.p., Brand, C.W., Shlager, B. and
wells”, Paper SPE 31145 presented at the SPE Economides, M.J.: “Horizontal well testing of
Formation Damage Control Symposium, Louisiana, isolated segments”, Paper SPE 29959 presented at
14-15 Feb. 1996. the International Meeting on Petroleum
4. Ozkan, E.: “Estimation of formation damage in engineering, China, November 14-17, 1995.
horizontal wells”, Paper SPE 37511 presented at 19. Civan, F.: Reservoir Formation Damage
the 1997 SPE Production Operations Symposium, Evaluation, Gulf publishing Company, 2000.
Oklahoma, 9-11 March. 20. Tiab, D. and Donaldson. E.C.: Petrophysics:
5. Dikken, B J.: “Pressure drop in horizontal wells and Theory and practice of measuring reservoir rock
its effect on production performance”, JPT and fluid transport properties, Gulf publishing
Nov.1990. Company, 1996, 495-499.
6. Goode and Thambinayagam: “Pressure drawdawn 21. Raghavan A, R.: Well Test Analysis, PTR Prentice-
and buildup analysis of horizontal wells in Hall, Inc. New Jersey pp 209-211 (1993).
anisotropic media”, Paper SPE 14250 Presented at 22. Joshi, S.D.: Horizontal Well Technology, Penwell
the 60th Annual technical Conference and Publishing Company, Oklahoma, 1991.
Exhibition of SPE, Las Vegas, 22-25. Sep.1985 23. Tiab, D.: Well Test Analysis, Lecture Note,
7. Daviau, F., Mouronval, G., Bourdarot, P.C.: Volume I & II, Norman, Oklahoma, 2001.
“Pressure analysis for horizontal wells”, Paper SPE 24. Ouyang, L.B, Arabi, S and Aziz, K.: “General
14251 presented at the SPE Formation Evaluation, Wellbore Flow Model, Vertical, and Slanted Well
Dec.1988. Completions”, paper SPE presented at the 1996
8. Gringarten, A.C. and Ramey, H.J.: “The use of SPE Annual Technical Conference and Exhibition,
source and Green’s functions in solving unsteady Colorado, October 6-9, 1996.
flow problems in reservoirs”, SPEJ(Oct.1973) 285-
96;
SPE 80926 9

Appendix: Figures

1.E+01 5.E-01
3.5 k1
5.E-01
5 k1
PD & tD*P'D

4.E-01
7 k1
4.E-01 9 k1
3.E-01

QD
1.E+00 3.E-01
Chd=0.01
Chd=0.01 2.E-01
Chd=0.5
Chd=0.5
2.E-01
Chd=5 1.E-01
Chd=5
5.E-02
Chd=50
1.E-01 Chd=50 0.E+00
1.E-02 1.E-01 tD 1.E+00 1.E+01 Inf.Con 1.E+02 0.E+00 1.E-01 2.E-01 3.E-01 4.E-01 5.E-01 6.E-01 7.E-01 8.E-01 9.E-01 1.E+00
Inf.Con
Fig.1--PD & tD*P'D plot for Ld=5-no skin XD

Fig.4--Qd plot for each section Ld=10--no skin--for td=1E2

1.E+02
Chd=0.01
Chd=0.5
2.E+00
Chd=5
2.E+00 3.5 k1
PD

Chd=50
2.E+00 5 k1
7 k1
1.E+00
9 k1
1.E+01 1.E+00
PD 1.E+00
8.E-01
6.E-01
4.E-01
2.E-01
1.E+00
0.E+00
1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E-01 2.E-01 3.E-01 4.E-01 5.E-01 6.E-01 7.E-01 8.E-01 9.E-01 1.E+00
tD XD
Fig.2--PD plot for Ld=5--kd=8--different Chd Fig.5--PD for each section with S=0 at td=1E2

1.E+01
1.E+01

1.E+00
3.5 k1
tD*P'D

Chd=0.01
3.5 k8
Chd=0.5 1.E+00
5 k1
PD

Chd=5 5 k8
1.E-01
Chd=50 7 k1
7 k8
95/8 k1
95/8 k8
1.E-02 1.E-01
1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02
tD tD

Fig.3--tD*P'D plot for Ld=5--kd=8--different Chd Fig.6--PD plot for Ld=10--with & without skin for different rw
10 SPE 80926

3.E-01 1.E+01
3.5 k8
5 k8
3.E-01
7 k8
9k8
2.E-01
QD

2.E-01 1.E+00
kd=1
kd=2

t D*P' D
kd=4
1.E-01 kd=6
kd=8
kd=10
5.E-02 kd=20
kd=30
kd=50
0.E+00 1.E-01
0.E+00 1.E-01 2.E-01 3.E-01 4.E-01 5.E-01 6.E-01 7.E-01 8.E-01 9.E-01 1.E+00 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02
tD
XD
Fig.11--t D*P' D plot for Ld=0.5--Chd=1
Fig.7--Qd plot for each section Ld=10--with skin--for td=1E3

1.E+01
5.E+00
3.5 k8
4.E+00
5 k8
4.E+00 7 k8
9 k8 1.E+00
3.E+00 kd=1
3.5 k8

t D*P' D
kd=2
3.E+00 5 k8
kd=4
PD

2.E+00 7 k8
kd=6
9 k8
1.E-01 kd=8
2.E+00
kd=10
1.E+00 kd=20
5.E-01 kd=30
kd=50
0.E+00 1.E-02
1.E-01 2.E-01 3.E-01 4.E-01 5.E-01 6.E-01 7.E-01 8.E-01 9.E-01 1.E+00 1.E-05 1.E-04 1.E-03 1.E-02 t 1.E-01 1.E+00 1.E+01 1.E+02
D
XD
Fig.12--t D*P' D plot for Ld=5-Chd=1
Fig.8--Pd plot for comparaison Ad=constant & Ad varies for each rw

kd=1
kd=2 1.E+00
1.E+03 kd=1
kd=4
kd=6 kd=10
kd=8 kd=50
1.E+02 kd=10
kd=20
kd=30
PD

1.E-01
kd=50
QD

1.E+01

1.E+00
1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03
1.E-02
tD 1.E-01 1.E+00
XD
Fig.9--P D plot for Ld=0.5--Chd=1
Fig.13--Q D plot for Ld=0.5--td=1E2--different kd

kd=1
1.E+02 kd=2
1.E +00
kd=4 k=1 V1
k=8 V1
kd=6
k=1 op
1.E+01 kd=8 k=8 op
kd=10 k=1 V3
k=8 V3
kd=20
1.E+00 1.E -01
kd=30
QD

kd=50
PD

1.E-01

1.E -02
1.E-01 X D 1.E+00
1.E-02
Fig.14--Q plot for Ld=1--kd=1 & 8--different perforation number
1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 D

tD
Fig.10--P D plot for Ld=50-Chd=1
SPE 80926 11

1.E+00 1.E+01

slope 0.36

P D & t D*P' D
1.E-01

k=1 V1 1.E+00
QD

k=8 V1
kd=1
k=1 op
1.E-02 kd=1
k=8 op
kd=10
k=1 V3
kd=10
k=8 V3

1.E-01
1.E-03 1.E-05 1.E-04 1.E-03 1.E-02 t D 1.E-01 1.E+00 1.E+01 1.E+02
1.E-01 XD 1.E+00

Fig.15--Q D plot for Ld=5--kd=1 & 8--different perforation number td=1E1 Fig.19--P D & t D*P' D plot for Ld=5-with & without skin

1.E+02
kd=8 V1
kd=8 op
1.E+01
kd=8 V2
PD

1.E+01
1.E+00 slope 0.36

P D & t D*P' D
kd=1
kd=1
kd=10
1.E-01 kd=10

1.E+00
1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03
tD
1.E-02
Fig.16--P D plot for Ld=10--kd=8--Chd=0.5
1.E-05 1.E-04 1.E-03 1.E-02 t D 1.E-01 1.E+00 1.E+01 1.E+02
1.E+01
kd=8 V1 Fig.20--PD & tD*P'D plot for Ld=10-with & without skin
kd=8 op
kd=8 V2

1.E+00
1.E+01
t D*P' D

1.E-01
1.E+00
P D & t D*P'D

slope 0.5

1.E-02 kd=1
1.E-05 1.E-04 1.E-03 1.E-02 t D 1.E-01 1.E+00 1.E+01 1.E+02 1.E-01 kd=1
kd=10
Fig.17--t D*P' D plot for Ld=10--kd=8--Chd=0.5
kd=10

1.E+02 1.E-02
1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02
tD
Fig.21--PD & tD*P'D plot for Ld=50-with & without skin

1.E+01
P D & t D*P' D

1.E+02
slope 0.25
y=x
R2 = 1

1.E+00
kd=1
kD
kd=1
S & kD

kd=10 1.E+01
kd=10
1.E-01
1.E-05 1.E-04 1.E-03 1.E-02 t D 1.E-01 1.E+00 1.E+01 1.E+02

Fig.18--P D & t D*P' D plot for Ld=1-with & without skin

1.E+00
1.E+00 1.E+01 A 1.E+02
D

Fig.22--Skin correlation for uniform case

12 SPE 80926

1.E+02 2.5
y = 0.6489x + 0.1696
y =0.4726x2 +0.5517x +0.061
R2 = 1 2
2 R =1

1.5
S&kD

kD

t hump
1.E+01

0.5

1.E+00
1.E+00 1.E+01 1.E+02 0
AD
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Fig.23--Skin correlation--non-uniform case
Lg/Lw

Fig.27--Correlationt hump versus Lg/Lw

1.E+03
1.E+02
Xe = Ye
1.E+02 h = 0.02Xe
L = 0.8Xe
PD & tD*P'D

1.E+01 PSS flow

1.E+01

T Lg/Lw
1.E+00
Linear flow

1.E-01 1.E+00
Early

1.E-02
tD
1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02
Fig.24--Pd and tD*P'D response for a horizontal well with considerable lateral 1.E-01
penetration 1.E-01 1.E+00 1.E+01
Lg/Lw

Fig.28--TLg/Lw function versus Lg/Lw

1.E+03
Xe = Ye
h = 0.02Xe
1.E+02 L = 0.08Xe 1.E-01
PSS flow P' for Lp=50
PD & tD*P'D

1.E+01 P' for Lp=100

P' for Lp = 300
1.E+00 P' for Lp = 500
Pseudo-Radial flow
Linear flow

1.E-01 Early radial 1.E-02

P'

1.E-02
1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02
tD
t-inters 305 t-inters 400
Fig.25--Pd and tD*P'D response for a partially penetrating laterally horizontal well

1.E-03
1.E+00 Lg/Lw = 0.2 1.E+02 1.E+03 t 1.E+04
Lg/Lw = 0.4
0.2 to 1.6 Lg/Lw = 0.6 Fig.29--P' plot for for different perforation lengths
Lg/Lw = 0.8
Lg/Lw = 1
1.E-01 Lg/Lw =1.2
Lg/Lw =1.4
Lg/Lw =1.6
P'D

1.E-02

1.E-03
1.E-02 1.E-01 tD 1.E+00 1.E+01 1.E+02

Fig.26--P'D plot for isolated segments