ADVANCED WELL TEST

ANALYSIS
-------
Ch. 12 – Horizontal Wells
Djebbar TIAB, Ph.D.
Professor Emeritus, Petroleum Engineering
University of Oklahoma
[email protected] --- [email protected]
3709 Windover Drive
Norman, Oklahoma, 73072, USA

Copyright © 2016, Djebbar TIAB. All rights reserved.

No part of this manual maybe reproduced, stored in a retrieval

system, or transmitted in any or by any means, electronic,
mechanical, photocopying, recording or otherwise, without the
prior written permission of the author.
 3

CONTENT

1. VIDEO
2. Flow Regimes
3. Pressure Drawdown and Buildup Testing in HW –
Conventional Techniques
4. Modern Techniques: TCM, Regression, Deconvolution
5. Advanced Technique: Analytical/TDS Technique
6. Finite-Conductivity Fracture
7. Elliptical Flow Equations for HW
8. Determining Average Reservoir Pressure in HW
9. Practical Problems with HW Testing
10. Field Cases

1 – INTRODUCTION
 TDS Technique

VIDEO:
“Overview of HW
Technology”
 7

Horizontal wells are nothing new.

Engineers in the old Soviet Union drilled a number of them

in the 1950s, although they ultimately abandoned the
practice as uneconomical.

It wasn't until the early 1980s however, that two

petroleum companies (Elf and AGIP) established
horizontal wells as viable substitutes for vertical wells in
certain types of reservoirs.

To date, the industry has drilled several thousand

horizontal wells worldwide, in a large array of reservoirs.

Horizontal sections of 2000 to 8000 feet have become

common in the U.S.

8

The primary objective of horizontal drilling is to
drastically increase production.

It is reported that a horizontal well drilled across an oil

and gas formation in the Middle East increased
production by as much as 20 times (?!) more than that of
its vertical counterpart.

Horizontal drilling is any wellbore that exceeds 80

degrees, and it can even include more than a 90-degree
angle (drilling upward).

Ɵ<80o
 9
ADVANTAGES OF HORIZONTAL WELLS
(over Vertical wells)

The advantages of horizontal wells include:

1. Reduced water and gas coning because of reduced

drawdown in the reservoir for a given production rate,
thereby reducing the remedial work required in the
future.
2. Increased production rate because of the greater wellbore
length exposed to the pay zone.
3. Reduced pressure drop around the wellbore.
4. Lower fluid velocities around the wellbore.
5. A general reduction in sand production from a
combination of Items 3 and 4.
6. Larger and more efficient drainage pattern leading to
increased overall reserves recovery.

10

EXTENDED REACH WELLS

WORLDWIDE

QATAR 40,320 ft=12293 m (record!)

USA 40,000 ft=12195 m
RUSSIA 38,322 ft
CHINA 29,527 ft
ENGLAND 34,967 ft
CANADA 18,156 ft

1 m = 3.28084 ft
 11

Developments have been fast-moving in some aspects of
horizontal wells, and very gradual in others.

Spectacular growth has occurred in the areas of drilling,

directional control, reservoir engineering and formation
evaluation:

(A) Drilling and wellbore steering capabilities have matured

such that drillers can maintain very small thicknesses (less
than 20 ft) for ever-increasing lengths.

1. Using technologies such as bottom driven bits,

drillers are able to execute a sharp turn and drill
horizontally along any thin pay zone.

2. Thanks to real-time formation evaluation technology,

it's possible to steer within a desired layer even when
it is dipping.

12
(B) Reservoir engineering applications have likewise matured.
We have established, for instance, that horizontal wells are
generally attractive alternatives to vertical wells in
reservoirs having good vertical permeability (e.g., naturally
fractured carbonates).

1. For relatively small reservoir thicknesses, 20 ft or

less, vertical permeability is not as important.

2. We are also gaining a better understanding of

formation evaluation criteria.

3. We have identified a number of variables as crucial

for horizontal wells, including: vertical permeability,
horizontal permeability, stress anisotropy and, in
long horizontal wells, the presence of lateral
heterogeneities and discontinuities.
 13

4. To determine these variables, the industry has

developed new well testing and logging methods.

5. It has also made increasing use of borehole and 3D

seismic images and, very importantly, data integration
techniques.

6. Horizontal wells have themselves proven invaluable

formation evaluation tools. Many HW are drilled primarily
to characterize the reservoir.

7. They can provide data on lateral heterogeneities,

the presence of faults and the position and extent of fluid
contacts, shale lenses and so forth, which would be
difficult or impossible to obtain from vertical wells.

14
(C) Well completion and stimulation practices, by contrast,
have evolved less dramatically.

1. For example, it has become clear only in

relatively recent years that some stimulated
horizontal wells, or those completed without
regard to stimulation needs, may underperform
vertical wells in the same reservoir.

2. During the early stages of the 1980s' horizontal

drilling boom, several failures were attributed to
poor reservoir selection.

3. Today, while this issue has been largely resolved,

we can blame many lackluster horizontal wells
on inadequate completion or stimulation
practices, and failure to drill a pilot hole prior to
drilling a HW.
 15
SLANT DRILLING

Slant wells are drilled at an angle from
perpendicular (commonly 30° to 45° or higher
up to 75o).

This approach minimizes surface

environmental disturbance. For example, oil
reserves under a lake can be tapped by a slant
hole drilled from on shore.

More commonly in this area; four, six, even

eight slant wells are drilled from one "pad"
(i.e. well lease site).

This allows the oil reserves under a large land

area to be tapped by only one well site.

Thus, production of valuable oil reserves is

effectively harmonized with conserving the
environment.

16
DIRECTIONAL DRILLING

Drilling has advanced from slant and horizontal drilling to

drilling that can change direction and depth several times in
one well bore.
 17
In subsequent sections and chapters:

1. We review the geologic aspects of HW and briefly

discuss types of HW and HW completions (Video)

2. Evaluate pressure tests in pilot holes

3. Apply conventional and modern WTA techniques

to Slant and HW

4. Extensive application of various productivity

equations in oil and gas reservoirs

5. Evaluate matrix stimulation and hydraulic

fracturing of Horizontal Wells, and

6. Assess HW performance in naturally fractured

reservoirs.

18

2 – Horizontal
Well Applications
 19

OPTIMIZING PRODUCTION
SLANTED WELL OR HORIZONTAL WELL?

Reservoirs can be classified by drainage volume.

For each reservoir more than one well type may be

effective:
 vertical,
 hydraulically fractured vertical,
 slanted,
 horizontal,
 hydraulically fractured horizontal, and
 multiple or stacked laterals.

Depending on permeability (especially kV) and reservoir

characteristics, slanted and horizontal reentry drilling are
two methods for improving production and recovery.

General Guidelines:
SLANTED WELL OR HORIZONTAL WELL?
In channel sandstone reservoirs comprising several
reservoirs a vertical well may only find one or two target,
while a HW or Slanted well could reach several targets
(Fig.: Mukhaizna Field in Oman)
 25
Theoretically, all reservoirs can benefit from horizontal
wells, but often these reservoirs may be just as effectively
produced using conventional vertical wells.

For example, consider a vertical wellbore:

1. in a high permeability homogeneous oil reservoir.

2. no natural barriers traverse through the reservoir, and
3. the reservoir can be easily hydraulically fractured to
overcome the pressure drop near the wellbore caused by
drilling damage and gas turbulence.

The ultimate percentage of reserves recovered (by this

vertical well) will be high and the flow rate will deliver the
reserves in an acceptable time frame.

In this example, a HW will not significantly increase return

on investment, or improve on ultimate recovery from the
reservoir.

26
However, if the same reservoir is internally subdivided by shale
layers, as in a channel point bar system (Fig. 1), a horizontal well
will significantly increase the ultimate recovery of oil from the
reservoir.

To fully utilize the potential of a horizontal well, the

geologist must have a clear understanding of both vertical
and lateral reservoir characteristics.

The drive mechanism and fluid characteristics in the

reservoir system are equally important.
 27

Once a candidate has been selected, the geologist must
work closely with the drilling and completion engineer to
match hole azimuth (angle) to reservoir character and
geometry.

Although the driller has good control of the well path,

sudden changes of the well trajectory are difficult to achieve.

Effectively, the hole azimuth

must be chosen prior to kicking kick
the well off from vertical (Fig.).

If a vertical pilot hole is

drilled, data from an oriented
core and electric log run can
often be used to adjust the
azimuth (angle) of the final
hole.

Azimuth Angle (θ)

The azimuth angle is the angle of the horizontal
wellbore (or slanted well) with respect to the xy plane
as shown below.

An azimuth angle of θ = 0° positions the horizontal

wellbore in a plane parallel to the y-axis

28
 29

As the well is being drilled, an accurate geological model

of the reservoir and information regarding the lithology
that could be encountered are critical to operational
efficiency and a successful well.

Sample descriptions and mud gas logs are valuable

tools.

Using these tools, the geologist can provide a detailed

picture of the reservoir (i.e. a geologic model) as the well
is drilled laterally through the formation.

This picture represents valuable input when making

decisions regarding:
- completion techniques,
- future development drilling, and
- enhanced recovery methods in the field.

30

Horizontal drilling can be applied in any phase of

reservoir recovery: primary, secondary, and enhanced.

Horizontal wells can be considered as an alternative to

infill drilling (of vertical wells) and fracturing, with the
objective in all cases being to increase the economic
recovery of oil and gas.

Generally, stimulation and infill drilling is driven by the

need to overcome production problems.

These problems may include:

- low reservoir permeability,
- high formation fluid viscosity,
- high pressure drawdown in the near wellbore region, or
- heterogeneous reservoir characteristics resulting in low
recovery efficiency.
 31

3 – Horizontal
Wells With Zonal
Isolations

32

Candidates for zonal isolations?

 Where is water coming from?

Fault or
Fracture

33

34
ZONAL ISOLATION
Zonal isolation is as essential for horizontal wells
as it is for vertical wells.

Selective zonal control enables increased ultimate

oil recovery through effective management of:

1. water injection
2. gas and water breakthrough
3. individual zone productivity
4. Asphaltene deposition

34
 35
Information gained from evaluating
zonal contributions can be used
for:
1. Identifying economic
production intervals
2. Sizing stimulation treatment
3. Identifying prospective
injection intervals
4. Identifying zones requiring
isolation during permanent
completion
5. Verification of cement bond
integrity
6. Mitigating lost circulation Lost circulation in NFR

36
 37
NATURALLY FRACTURED CARBONATE RESERVOIRS

Horizontal wells greatly increase the probability of encountering a

fracture system and significantly reduce the dry hole risk.

However, extensive fracture systems which extend into underlying

water zones may be detrimental to production and should be avoided.

MAJOR CHALLENGE: ZONAL ISOLATIONS?

38

NF CARBONATE RESERVOIR MODEL?

Cement plugs?

Perforated intervals Zonal isolations

z
y

h
x
Lz LP
Lw
 39

Both Halliburton and Schlumberger offer “Smart-Well”

completion systems which consists of some combination of:

1. zonal isolation devices,

2. interval control devices,
3. downhole control systems,
4. permanent monitoring systems,
5. surface control and monitoring systems,
6. distributed temperature sensing systems,
7. data acquisition and management software.

39

40

To maintain effective zonal isolations for the well life, the

isolators should be designed for all conditions of harsh
environments.

However, many zonal isolators may not meet the necessary

requirements especially when several harsh conditions are
found in a single section of the formation that may lead to
zonal isolation failure.

The zonal isolators are challenged by harsh environments such

as high temperature, high pressure, and corrosive fluids as
well as stresses caused by production operation and/or
unstable formations.

Evaluating the performance of the isolators can be considered

as an important application for the well test analysis.

In addition, well test analysis can be used to detect the

locations of unsuccessful zonal isolations.
40
 41

4 – Vertical Well
Versus Horizontal
Well

42

The mechanical process of drilling highly accurate

horizontal wells is now a common occurrence throughout
the world.

While there will continue to be many innovations in the

realm of drilling technology, it is apparent that our
practical understanding of horizontal wells in petroleum
disciplines such as completion, production forecasting,
and well test interpretation is still slightly lagging
behind.

The scope of this chapter will be to discuss the current

theoretical understanding associated with pressure
transient analysis in horizontal wells and to assess the
practical application of these analytic solutions.
 43

The material presented here, especially with

respect to conventional techniques using the
pressure function, is based on a review of the
available literature.

Modern techniques that are based on the

pressure derivative function, and more
specifically the TDS (Tiab Direct Synthesis)
technique, will be emphasized.

There have been many groups of authors that

have published articles in the area of pressure
transient analysis for horizontal wells.

44

When one compares the mathematical models

introduced by these authors, it is apparent that
the analytic techniques necessary for pressure
transient analysis are still in the development
stage.

The majority of onshore horizontal wells are

being drilled in unconventional and complex
reservoirs and there is a growing awareness of
the necessity to complete these wells effectively.

One of the primary goals of pressure transient

testing is to provide a method of differentiating
between completion success and in situ
reservoir quality.
 45

The following question is often asked after a

horizontal well has been drilled successfully, but
completed for a production rate below what was
anticipated: Is the reservoir poorer than
anticipated or has the wellbore been damaged?

The value of properly assessing the condition of the

wellbore in relation to the reservoir through simple
pressure tests can be very significant.

The cost of stimulating a horizontal well is often

economically prohibitive and in most cases should
be considered well before the well is drilled, if
possible.

46
VERTICAL Vs. HORIZONTAL WELL
ANALYSIS

In this section, we will review methods to analyze
pressure-transient tests in horizontal wells and illustrate
application of the results of transient test analysis in
estimating productivity.

Productivity estimates in horizontal wells are subject to

more uncertainty than comparable estimates in vertical
wells, as we shall see in the next chapter.

Further, it is much more difficult to interpret well-test

data from a horizontal well than from a vertical well.

The major problem with a horizontal well is that the flow

geometry is 3D. We no longer have the radial symmetry
that is usually present in a vertical well.
 47

Several flow regimes can potentially occur in a horizontal

well test, and we need to consider these in analyzing test
data from horizontal wells.

Wellbore storage effects can be much more significant in

a horizontal wells than in vertical wells, and horizontal
wells will commonly exhibit partial penetration and end
effects that complicate interpretation.

In vertical wells, we are accustomed to dealing with

variables such as average permeability, net vertical
thickness, and skin.

In horizontal wells, we need more detail. Not only is

vertical thickness important, but the horizontal dimensions
of the reservoir, relative to the horizontal wellbore, need to
be known.

48

STEPS IN EVALUATING HORIZONTAL

WELL-TEST DATA

To evaluate pressure-transient data from a horizontal
well, we expect to follow three basic steps.

First, we must identify the specific flow regimes in the

test data.

Second, we need to be able to apply the proper analytical

and graphical procedures to the data.

Finally, we need to be able to evaluate the uniqueness

and sensitivity of the results to properties we have derived
from our analysis or simply assumed.
 49

IDENTIFY FLOW REGIMES

Typically, when engineers evaluate completion data
from a vertical wellbore, the techniques they use involve
a single flow regime, such as infinite-acting radial flow.

However, a pressure-transient test in a horizontal well

can involve as many as five major and distinct regimes
that need to be identified.

These regimes may or may not occur in a given test and

may or may not be obscured by well bore storage effects,
end effects, or transition effects between flow regimes.

50

APPLY THE PROPER PROCEDURES

Each flow regime can be modeled by an equation that
can be used to estimate important reservoir properties.

In some cases, only groups of analytical parameters

can be determined directly from equations.

Therefore, it is imperative that we apply the proper

analytical and graphical procedures to the data.

In many cases, when solving for specific parameters,

the application of these analytical expressions may
involve a complex iterative procedure.
 51

EVALUATE UNIQUENESS AND

SENSITIVITY

Experience indicates that results of horizontal well-test
analysis are seldom unique, so it is important that we
evaluate the uniqueness and sensitivity of the results to
assumed properties.

Simulation of the test using properties that have been

determined from the test can confirm that at least the
analysis is consistent with the test data.

A simulator can also determine whether other sets of

formation properties will also lead to a fit of the data

52
MATHEMATICAL MODELS

Goode and Thambynayagam developed generalized

mathematical response functions at the horizontal wellbore
for conditions of both pressure drawdown and pressure
buildup.

These general equations, assumed an effective pressure

point along the horizontal wellbore.

Later work by Kuchuk et al. was based on pressure

averaging.

Reservoir models used by Ozkan et al. are presented in

Appendix A at the end of this chapter.

This chapter uses the model and nomenclature used by

Babu and Odeh.
 53

5 – HW FLOW REGIMES

54
RESERVOIR MODELS

Babu-Odeh and Ozkan et al are two of several models used to

analyze pressure transient tests.

Typically, regardless of the model, the horizontal well is

assumed to be analogous to a vertical well with a fracture
producing from either the right or the left side.

FIG. 1.1 FIG. 1.2

 55
Several authors have developed FIG. 1.3
direct solutions for the HW. L

ye

Some models assume the HW is z y

zw
h
ye /2

parallel to the Y-axis where

X

x e /2
xe

kx>ky;
others assume the well is FIG. 1.4
parallel to the X-axis, where
ky>kx.

KAPPA Software:
HW is Parallel to the x-axis (same
as in this course). Therefore it is
assumed that the minimum
permeability direction in the
horizontal plane is parallel to the
x-axis, i.e. ky>kx)

56

The well length, formation thickness, and

vertical/horizontal permeability ratio are the most
important parameters affecting horizontal-well pressure-
transient responses.

These parameters govern the existence and duration of

three possible flow regimes during the infinite-acting
period:
 early-time (initial) radial flow (ERF),
 intermediate-time linear flow (ILF), and
 late-time radial (pseudoradial) flow (LRF).

When the effects of all the lateral boundaries are felt, a

pseudosteady-flow (PSF) regime develops.
 57

Permeability Anisotropy

kv=kz

kx
Horizontal ky / kx indicates
ky Plane magnitude of
anisotropy in
horizontal plane
kH = kx ky = effective horizontal perm

58

The usual assumptions apply in all productivity

equations, i.e., we have:

• a single phase flow of a slightly compressible fluid,

• a homogeneous reservoir with uniform thickness,
and
• gravity and capillary effects are negligible with
impermeable upper and lower boundaries for the
reservoir.
• Porosity and absolute permeability are
independent of position and pressure.
• We assume uniform flux along the wellbore.
• Initial condition is that throughout the reservoir
pressure is uniform.
 59
HORIZONTAL WELL CONDUCTIVITY

The conventional models of horizontal wells assume either

uniform flux or infinite conductivity.

The uniform-flux assumption is mainly a mathematical

convenience, with the only direct application being in the
case of high skin damage along the horizontal well (skin
damage causes more-uniform flux distribution along the
length of the horizontal well).

In practice, however, horizontal-well behavior is more

closely approximated by either infinite- or finite-conductivity
behavior.

For infinite- and finite-conductivity horizontal wells, flux

distribution along the well is determined by the interaction
between the wellbore and reservoir flow dynamics.

60
The figure shows schematics of the
flux distributions for uniform-flux,
infinite-conductivity, and finite-
conductivity conditions.

The infinite-conductivity condition

is characterized by a U-shaped flux
distribution that is symmetrical with
respect to the midpoint of the
horizontal well.

When the conductivity decreases and becomes finite, the

flux distribution becomes more skewed toward the heel
end of the horizontal well.
 61

In general, the infinite-conductivity-well assumption can
be justified based on the premise that the pressure drop
inside the well bore is small compared with the pressure
drop in the reservoir (Ozkan et al. 1995, 1999).

Because horizontal wells may be long and flow rates may

be high, frictional and acceleration effects may cause
considerable pressure drop and reduce the effective
conductivity of the well bore.

The effect of finite wellbore conductivity complicates the

analysis of pressure-transient responses of horizontal
wells.

Conventional analysis techniques, however, assume that

the effect of well bore hydraulics is negligible (i.e., that
the well is an infinite-conductivity conduit).

62
If α denotes the contribution of the well bore pressure drop
to the total pressure drop, then (Yildiz and Ozkan 1998)

(1.1)

(1.2)

(1.3)

(1.4)
 63
And f is the Fanning friction factor and is based on NRe

The friction factor f may be obtained from correlations or

charts.

Many correlations are available to calculate the friction

factor. One of the correlations is given by Jain Eq.:

(1.5)

64
EXAMPLE – Effect of
Wellbore Hydraulics.

For the data in Table,

determine whether the the
infinite-conductivity
horizontal-well assumption
is appropriate .

SOLUTION
We have from Eqs. 1.2 through
1.4, respectively,
 65
The friction factor is (Eq. 1.5):

h=20 ft
Lh=20000 ft
q=5000 stb/d

Then, the condition given by Eq. 1.1 indicates that

(alpha<1%; that is, the contribution of the wellbore
pressure drop to the total pressure drop will be less than 1
%.
Therefore, the infinite-conductivity assumption is
appropriate for this horizontal well.

66

BABU AND ODEH MODEL

Babu & Odeh obtained a rigorous solution to the diffusivity equation
for a well in a box-shaped reservoir, subject to certain limiting
assumptions.

The assumptions include the following:

1. Fluid flows to the well uniformly at all points along the wellbore
(uniform flux)

2. The sides of the drainage volume are aligned with the principal
permeability direction.

3. The wellbore is parallel to the

sides of the drainage area and is
oriented parallel to one direction
of principal permeability ky (i.e.
parallel to the Y-axis) and
perpendicular to the other two.
 67
4. The boundaries of the reservoir are all no-flow boundaries, and the
well reaches stabilized, pseudosteady-state flow.

5. The formation damage around the wellbore is uniform at all points

along the wellbore.

The solution is quite complex but is approximated accurately with

an equation written in the same form as the pseudosteady-state
flow equation for a vertical oil well producing a single-phase,
slightly incompressible liquid.

68
OTHER ASSUMPTIONS & NOMENCLATURE

Figure 1.5 shows a horizontal well with length, Lw, located

within a reservoir that is assumed to be a rectangular
parallelepiped, or a “box reservoir” drainage area.

For now, the axes of the coordinate system are assumed to

coincide with the direction of principal permeability

Later, modifications to be
used when this statement
is not true will be
presented.

Also, it is assumed that the

well produces over its
entire length, Lw.
FIG. 1.5
 69
MODEL USED IN THIS COURSE

In this chapter, the HW is assumed to be parallel to the

X-axis, similarly to Goode and Thambynayagam and
Ozkan et al. AND KAPPA SOFTWARE.

All of the above assumptions made are applicable.

ye

ye /2
z y h
zw
X

x e/2
xe

70
Flow Regimes

Up to five different flow regimes can occur during a flow

test:

1. Early radial, hemiradial

2. early linear, bilinear or elliptical
3. late radial
4. late linear
5. pseudosteady state

We calculate different formation properties from the data

in each of these different flow regimes.

We need to remember that any flow regime may be

absent from a plot of test data because of geometry,
wellbore storage, or other factors.
 71
Flow Regimes

1 2

3a and/or 3b

5 ye

ye /2
z y h
zw
X

x e/2
xe

72
EARLY RADIAL
FLOW REGIME

FIG. 1.6
Consider a well producing at constant rate.

The early-radial flow regime occurs before the area drained or the
pressure transient caused by this production encounters either of the
boundaries of the reservoir.

So a radial flow pattern penetrates out into the reservoir, as shown in

Fig. 1.6.

However, this flow pattern is likely to be elliptical, moving further into

the reservoir at a given time in the higher-permeability x-direction
than in the lower-permeability z-direction.

This phenomenon causes no significant complications in our analysis.

 73
HEMIRADIAL FLOW REGIME

When the wellbore is much

nearer one vertical boundary
than the other, we may have
another flow regime, called
hemiradial flow (Fig. 1.7). FIG. 1.7
Hemiradial flow can occur
immediately following the early-
radial flow regime if the well is
much nearer one of the vertical
boundaries than the other.

Hemiradial flow ends when the

pressure data feel effects of the
closest boundary (Fig. 1.8).
FIG. 1.8

74
EARLY LINEAR FLOW REGIME

FIG. 1.9

Eventually, the area affected by the production will include

the entire thickness of the reservoir.

Once the flow pattern has encountered both the top and the
bottom of the reservoir, fluid will begin to flow linearly into
the well, as illustrated in Fig. 1.9.

This flow regime will be observed during the pressure test

only if the horizontal well is much longer than the formation
thickness.
 75
ELLIPTICAL FLOW REGIME

FIG. 1.10

If the horizontal well length is not long compared to the formation

thickness, then this flow regime (early linear) will not develop.

Instead, a lengthy transition zone will develop prior to the next

identifiable flow period.

Actually during this transition period the flow regime is elliptical,

which can be identified on the derivative curve by a straight line of
slope 0.36.

76
LATE RADIAL FLOW REGIME

FIG. 1.11

After a sufficiently long time, the pressure front will become

approximately radial in the x-y plane and a late radial flow pattern
will develop (Figure 1.11).

This period will not exist for wells with a gas cap or aquifer or if
other external boundaries are felt first.
 77
LATE LINEAR FLOW REGIME

The late linear flow period corresponds to late time, where the
pressure transient has reached the nearest vertical boundaries of a
finite rectangular (at least 4:1 or channel) reservoir. This flow
regime is rarely if ever observed.

Pseudosteady state is more likely to be observed in a long test.

Naturally, this flow period will develop only for a reservoir of finite
width.

FIG. 1.12

78

2 – Conventional
Techniques
 79

6 – Conventional
Techniques

80
PRESSURE DRAWDOWN
It is recommended that we identify all available flow regimes on the
diagnostic plot (loglog plot of ∆P and t*∆P’ versus time) and then plot
the bottom hole flowing pressure during the appropriate time range
against time on semilog coordinates (for radial flow) or Cartesian plot
(for linear flow regime).

100
t*ΔP'

10
ΔP

late radial

late linear

early radial
early linear
1
0.001 0.010 0.100 1.000 10.000 100.000 1000.000
time, hr

FIG. 1.13
The radial flow regimes are identified by a horizontal line and the
linear flow regimes by a half-slope on the derivative in the diagnostic
plot of pressure-test data.
 81
Again, for clarification purposes, the model used here is as
shown in Figures 2.1, which illustrates the horizontal well
configuration associated with the analytic solutions and well
test interpretation that will be presented.

ye

ye /2
z y h
zw
X

x e/2
xe

FIG. 2.1

KAPPA
nomenclature

82

Permeability Anisotropy

kv=kz

kx
Horizontal L

Plane
ky ye

ye /2
z y h
zw
X

x e/2
xe

kH = kx ky = effective horizontal perm

 83
1 – EARLY RADIAL FLOW REGIME

The flow is radial around the well. This is equivalent to a fully

penetrating, vertical well in an infinite reservoir.

its duration is very short when in thin reservoirs or high vertical

permeability reservoirs.

For this time period, the pressure response can be approximated by:

qµB   k y k z t  
Pi − Pwf = 162.6 log − 3 . 227 + 0 . 866 s  (2.1)
L k y k z   φ µ ct rw 
2 m

 

kz ye

ye /2
z y h
ky kx
X

x e /2
zw

xe

84
A semilog plot of ∆P versus t should yield a straight line of slope:

162.6qBµ
mER =
L kzk y
(2.2)

kz ye

ye /2
z y h
ky kx
X

x e /2
zw

xe
 85
The equivalent permeability in a
vertical plane around the wellbore
can be calculated from the slope of
the semilog straight line:

162.6qBµ
kzk y = (2.3)
LmER

(1) - Notice if kz=kv is known from a pilot hole test, we can

calculate ky. if kh, which is equal to sqrt(kxky), is also known
from a pilot-hole test, we can solve for kx.

Thus kx, ky and kz will be known from just observing the

early radial. That is why it is very important to run a pilot hole
test.

Observing all the remaining flow regimes is not very common.

(2) - Notice also that, to make this calculation, we must know the
effective completed length of the well, L.

This is not necessarily the same as the perforated or the

completed length of the well; some sections of the well may not
produce at all.

In horizontal well test analysis the length of the wellbore which is

contributing to flow is called the effective wellbore length.

It is almost always smaller than the actual horizontal wellbore

length (L).

For example, a well may be drilled for 5000 ft horizontally, but

only 3000 ft may be contributing to flow, either because the rest
of the wellbore is damaged or it may even be outside the
reservoir.

In this case, effective wellbore length would be 3000 ft and

horizontal wellbore length would be 5000 ft
 87
Extrapolating the straight line to t = 1 hr and rearranging equation
2.1, the mechanical skin factor (due to the alteration of near-
wellbore permeability during drilling and/or well completion) is:

 ∆P  k ykz  
− log  + 3.227 
( t =1hr )
sm = 1.151
 mER  φ µ ct rw2   (2.4)
   

It is helpful to check the expected duration of the early-radial

flow regime after we have estimated the parameters necessary
to make these calculations.

Sm

Semilog plot of dP versus time

 89
2 – HEMIRADIAL FLOW REGIME

The hemiradial flow period will occur only when the well is
close to one of the vertical boundaries (either the upper or
the lower boundary). The HRF regime is given by:

325.2qBµ   k z k y  
Pi − Pwf = log t − 3.23 + 0.8683S m  (2.5)
L k z k y   φµct rw 
2

 

The slope is used to calculate the sqrt(kzky):

325.2qBµ (2.6)
mHR =
L kzk y
(2.7)

325.2qBµ  162.6qBµ 
kzk y = = 2 
mHR L  m HR L 

90
3 – EARLY LINEAR FLOW REGIME

This flow regime may develop after the

effects of the upper and lower boundaries
are felt at the wellbore.

If the horizontal well is much longer than

the formation thickness, a period of linear
flow will be observed.

The ELF regime is given by:

µ t 141.2qµB
Pi − Pwf =
8.128qB
+ (s z + s m )
Lh k y φ ct L ky kz (2.8)
FIG. 2.4
 91
Pressure data identified as being in this flow regime can be plotted
against the square root of time.

The early-linear flow regime in a drawdown test corresponds to the

half-slope line on the derivative curve.

The slope of the straight line on the linear plot (∆P Vs. t^0.5) is:

8.128qB µ
m EL = (2.9)
Lh k y φct FIG. 2.5

Solving for ky gives:

2
 8.128qB  µ
k y =   (2.10)
m
 EL Lh  φct

Knowing ky from the early linear flow and sqrt(kykz) from the early
radial, we can solve for kz.

92
Extrapolating the straight line to t=0 we can solve for SZ (by
substituting t=1 in Eq. 2.8):

0.058 k z  ∆PEL1 
S z + Sm =  
h φµ ct  mEL 
(2.11)

where Sz is the pseudoskin resulting from partial penetration in the

VERTICAL direction.

This skin can also be estimated from:

(2.12)

h  ky    πz  
sz = ln  + 0.25 ln  − ln sin w   − 1.838
 rw   kv    h 

zw = vertical location of well, ft FIG. 2.6

 93
SZ is also referred to as the
convergence skin, which is an
additional pressure drop that acts like
a skin effect caused by flow moving
throughout the entire formation until
it converges down to the small
wellbore in the middle of the
formation (Fig. 2.7). FIG. 2.7

4200

As shown in Fig. 2.8, it is possible to 4150

m L1 =
8.128 qB µ
Data
linear flow(final)
Lh k yφ c t
draw several straight lines on the plot of 4100

4050
linear flow(1)
linear flow(2)

Pwf versus square root of time, but only 4000 m2 =-35.62 psi/hrs1/2 ????

Pwf, psi
one of them is the correct. 3950

3900
m1=-32.49 psi/hrs1/2
3850
????
To select the correct straight line on the 3800

3750
Correct value
mf=-38.81 psi/hrs1/2

linear plot is one of the main 3700

0 2 4 6 8 10 12 14 16 18
uncertainties of the conventional Square Root of Time, hrs 1/2

method.
FIG. 2.8

94
4 – LATE RADIAL FLOW REGIME FIG. 1.11

After a sufficiently long time, the

pressure front will become radial in the
x-y plane and a third radial flow pattern
will develop (Fig. 1.11).

This period will not exist for wells with a gas cap or aquifer or if
other external boundaries are felt first.

The diagnostic plot helps us identify the late-pseudoradial flow

regime with the characteristic (late-time) horizontal derivative (Fig.
2.9).
 95
For data in the appropriate time range, we prepare a semilog plot of
pressure against time according to:

162.6qµB   k x t  
 − 2.023 + 141.2qµ B (s z + s m )
Pi − Pwf = log (2.13)
hz k y k x   φµ c t L 
2
  L k y kz

FIG. 2.10

Semilog plot of dP versus time

 97
From equation 2.13 the slope of the line at late radial flow regime is:
162.6qBµ
mLR = (2.13a)
h kxk y

The geometric average

permeability in the horizontal
plane can be estimated from the
slope:
162.6qBµ
kH = kxk y = (2.13b)
hmLR
Knowing ky from the early linear flow and sqrt(kykx) from the late
radial, we can solve for kx.
Extrapolating the straight line to t = 1 hr and rearranging Eq. 2.13
gives the mechanical skin factor Sm or the convergence skin Sz:

1.151L k z  ∆P(t =1hr )  kx  

sm =  − log  + 2.023 − s z
2  (2.13c)
h k x  m LR  φµ ct L  

98
KAPPA SOFTWARE

The figure shows the nomenclature used in the software.

The LRF is used to calculate the radial permeability (kr),

which is referred to here as kH:

kr = kH = k x k y

The software uses hw

instead of Lw for the
HW length.
 99
5 – LATE LINEAR FLOW REGIME
The pressure transient has now
reached the lateral extremities of the
reservoir, thus yielding the second (or late)
linear (and probably final flow period) on a
Cartesian plot of Pwf Vs. t0.5

µt 141.2qµ B
This flow period is governed by the P − P = 8.128qB + (s x + s z + s m )
k y φ ct
i wf
following Eq.: hx h L ky kz

hx = formation thickness in the x-boundary, ft. (2.14)

Sx = SP = Skin due to partial penetration in the X-direction (Sx=0, if
well penetrates entire formation thickness in x-direction)

ye

ye /2
z y h
zw
X

x e/2
xe

hx

Cartesian plot of dP versus sqrt(t)

 101
The slope of the straight line 8.128qB µ
on the Cartesian graph (∆P mLL = (2.15)
hx h k yφct
Vs. t^0.5) is:

From the slope we can

calculate ky or hx:

2
µ  8.128qBo 
ky =   (2.16)
φc t  h x hm LL 

8.128qB µ
hx = (2.17) FIG. 2.11
mLL h k yφct

From the straight line at time t = 0, (i.e. the

intercept ∆P(t=0)) we can calculate Sx: L

k z  ∆P( t =0) 
ye

0.058Lw
Sm + S x + S z =  
ye /2
z y h

φµ ct  mLL 
zw
X

hhx x e/2
xe

(2.18) hx

102
PRESSURE DRAWDOWN PLOTS (summary)
Flow regime Reservoir Plot Calculation
Parameter
kzk y 162.6qµ B
Pwf, or (Pi-Pwf) vs. log t kzk y =
m Lw
Early time radial
 Pi − Pwf (1hr ) 
Pwf, or (Pi-Pwf) vs. t  − 
sm  m1r 
s m = 1.151 
 kykz 
log  + 3.227
  φ µ c t rw2  
   
where Pwf is at 1 hr
Intermediate time φ ct k y Pwf, or (Pi-Pwf) vs. t φ ct k y 8.128qB
linear =
µ µ mLw hz
Late time radial kzk y Pwf, or (Pi-Pwf) vs. log t 162.6qµB
kzk y =
mLw
φ ct k y 8.128qB
=
µ mhx hz
Late time linear φ ct k y Pwf, or (Pi-Pwf) vs. t
(Pseudosteady or hx or;
state) µ 8.128qB µ
hx =
mhz φ ct k y
 103
PRESSURE BUILDUP
In the case of a pressure buildup test, similar equations were
developed for two scenarios.

The first, and probably most common, is where the reservoir is

effectively infinite acting, i.e. hx → ∞, during the course of the test.

The second is where the pressure transient has reached the

reservoir boundaries, and the late time pseudo steady state was
established before the well was shut in.

EARLY RADIAL FLOW REGIME

Case 1: Infinite reservoir

162.6qµ B   to + ∆t  
Pi − Pwf = log ∆t  + γ 1 (2.19)
k y k z Lw    

 
γ 1 = Lw k z log k x t  − 2.023 − log t − log k y k z  + 3.227 + 0.868s
2    φµ ct r 2w  (2.20)
 φµ ct Lw 
z
hz kx    

104
EARLY RADIAL FLOW REGIME
Case 2: finite width

162.6qµ B   to + ∆t  
Pi − Pwf = log ∆t  + γ 2 (2.21)
k y k z Lw    

0.05 Lw k z t  k ykz t 
γ2 = − log  + 3.227 + 0.868( s + s ) (2.22)
h
hz x φ µ c  φµ c 2 
t rw 
z x
t 

EARLY LINEAR FLOW REGIME

Case 1: Infinite width

8.128q B µ∆t
Pi − Pwf = +γ3 (2.23)
h z Lw k y φ ct

162.6qµ B   k x t  
γ3 = log 
2 
− 2 . 023 (2.24)
k y k x h z   φµ ct L w  
 105
EARLY LINEAR FLOW REGIME
Case 2: finite width

8.128q B µ  t  141.2qµ B
Pi − Pws =  ∆t − Lw  + sx (2.25)
h z Lw k y φ ct  hx  Lw k x k y

LATE RADIAL FLOW REGIME

Case 1: Infinite width
162.6qµ B   t o + ∆t  (2.26)
Pi − Pwf = log 
k y k x h z   ∆t 

LATE RADIAL FLOW REGIME

Case 2: finite width

162.6qµ B   t o + ∆t  
Pi − Pws = log ∆t  + γ 4  (2.27)
k y k x hz    

γ4 =
0.05 Lw kx t − log k x t  + 2.023 + 0.868s
 φµ c 2  (2.28)
φµ ct
x
hzhx  t Lw 

106
LATE LINEAR FLOW REGIME (pseudosteady state)
Only one case: finite width

Pi − Pwf =
8.128q B µ
( t − ∆t ) (2.29)
hz hx k y φ ct

---------------------------------------------------------------------------------
In general, the pressure transient plotted as the wellbore
pressure versus the logarithm of time will yield two straight
lines corresponding to the early time and late time radial flow
regimes.

The pressure transient plotted as the wellbore pressure versus

the square root of a time could also yield two straight lines,
corresponding to the early time and late time linear flow (or
pseudosteady state) regimes.

Odeh and Babu correctly stressed that not all four flow
regimes will automatically exist during every pressure test.
 107
PRESSURE BUILDUP PLOTS
Flow regime Reservoir Plot Calculation
Parameter
Early time radial kz k y  t + ∆t  162.6qµ B
∆P vs. log o  kz k y =
 ∆t  m Lw
Intermediate time µ ∆P vs. ∆t µ 8.128qB
linear =
φ ct k y φ ct k y mLwhz
Late time radial kz k y  t + ∆t  162.6qµB
∆P vs. log o  kz k y =
 ∆t  mhz
Generally, the initial part of this Horner plot will be a straight line. If the reservoir is of finite width, and the
pressure transient has reached the final linear flow period before shut-in, and when the inequality to+∆t becomes
invalid, this Horner relationship will no longer produce a straight line
Late time linear φ ct k y P vs. t − ∆t Extrapolate to Pi
(Only exists for a or
µ
finite width
∆P vs. t − ∆t φ ct k y 8.128qB
reservoir) hx or Pi =
µ mhx hz

108

EXERCISE 1
 109
EXAMPLE - Well Prf-3
A pressure drawdown in a horizontal well yielded the data shown
in Table 2.1. The following data are known:
q = 800 STB/D µ = 1.0 cp
B = 1.25 RB/STB rw = 0.25 ft
Ø = 0.2 ct = 15 x 10-6 psi-l
h = 150 ft Lw= 900 ft
hY=aH=5280 ft hX=bH=5280 ft
The well is approximately
in (finite) drainage volume.

Interpret this drawdown test using the conventional technique.

kz
L

ky
hy y=aH
e

ye /2
z y h
X
zw
kx
x e/2
xe
hX=bH

110
Table 2.1a – Pressure Drawdown Data
t Pwf t Pwf t Pwf t Pwf
0.000 3000 0.119 2978.18 1.707 2971.336 29.136 2959.721
0.001 2996.745 0.134 2977.859 1.913 2971.031 32.633 2959.104
0.002 2994.507 0.151 2977.543 2.143 2970.719 36.550 2958.48
0.003 2992.58 0.170 2977.23 2.401 2970.397 40.936 2957.848
0.004 2990.776 0.192 2976.921 2.690 2970.063 45.850 2957.209
0.006 2989.243 0.216 2976.616 3.014 2969.715 51.353 2956.562
0.007 2987.956 0.242 2976.312 3.376 2969.353 57.516 2955.906
0.009 2986.87 0.272 2976.011 3.782 2968.976 64.419 2955.242
0.011 2985.939 0.306 2975.712 4.237 2968.582 72.150 2954.573
0.013 2985.206 0.343 2975.415 4.746 2968.172 80.808 2953.901
0.016 2984.492 0.386 2975.12 5.317 2967.744 90.506 2953.237
0.019 2983.856 0.433 2974.826 5.956 2967.3 101.368 2952.572
0.022 2983.284 0.485 2974.533 6.671 2966.84 113.533 2951.902
0.025 2982.761 0.545 2974.242 7.473 2966.365 127.158 2951.228
0.029 2982.278 0.611 2973.951 8.370 2965.875 142.418 2950.546
0.033 2981.848 0.685 2973.662 9.376 2965.37 159.509 2949.853
0.038 2981.418 0.768 2973.374 10.502 2964.852 178.651 2949.147
0.044 2981.009 0.861 2973.085 11.763 2964.323 200.090 2948.423
0.050 2980.617 0.965 2972.797 13.175 2963.782 224.101 2947.676
0.057 2980.239 1.082 2972.509 14.757 2963.231 250.994 2946.9
0.064 2979.878 1.213 2972.22 16.529 2962.668 281.114 2946.088
0.073 2979.523 1.359 2971.929 18.514 2962.096 314.849 2945.231
0.083 2979.176 1.523 2971.635 20.736 2961.515 352.632 2944.318
0.094 2978.836 23.225 2960.925 394.949 2943.340
0.106 2978.506 26.013 2960.327 437.596 2942.385
 RADIAL FLOW REGIMES
Semilog plot of dP versus time

LINEAR FLOW REGIMES

Cartesian plot of dP versus sqrt(t)
 113
Table 2.1b – Pressure Derivative data
t ∆P t*∆P' t ∆P t*∆P'
0.119 21.82 2.672
0.001 3.255 2.605 0.134 22.141 2.665
0.002 5.493 3.894 0.151 22.457 2.654
0.003 7.42 4.745 0.170 22.77 2.632
0.004 9.224 4.982 0.192 23.079 2.623
0.006 10.757 4.924 0.216 23.384 2.601
0.007 12.044 4.7 0.242 23.688 2.583
0.009 13.13 4.402 0.272 23.989 2.578
0.0110 14.061 4.502 0.306 24.288 2.572
0.0130 14.794 3.965 0.343 24.585 2.575
0.0160 15.508 3.582 0.386 24.88 2.564
0.019 16.144 3.495 0.433 25.174 2.525
0.022 16.716 3.311 0.485 25.467 2.534
0.025 17.239 3.164 0.545 25.758 2.536
0.029 17.722 3.045 0.611 26.049 2.544
0.033 18.152 3.085 0.685 26.338 2.540
0.038 18.582 3.009 0.768 26.626 2.538
0.044 18.991 2.943 0.861 26.915 2.538
0.050 19.383 2.882 0.965 27.203 2.512
0.057 19.761 2.831 1.082 27.491 2.525
0.064 20.122 2.833 1.213 27.78 2.537
0.073 20.477 2.798 1.359 28.071 2.558
0.083 20.824 2.765 1.523 28.365 2.579
0.094 21.164 2.734
0.106 21.494 2.702

114
Table 2.1b – Pressure Derivative data (continued)
t ∆P t*∆P' t ∆P t*∆P'
1.707 28.664 2.578 29.136 40.279 5.338
1.913 29.136 2.712 32.633 40.896 5.438
2.143 32.633 2.801 36.550 41.52 5.539
2.401 36.550 2.899 40.936 42.152 5.628
2.690 40.936 2.967 45.850 42.791 5.708
3.014 45.850 3.112 51.353 43.438 5.767
3.376 51.353 3.259 57.516 44.094 5.78
3.782 57.516 3.404 64.419 44.758 5.876
4.237 64.419 3.547 72.150 45.427 5.949
4.746 72.150 3.7 80.808 46.099 5.903
5.317 80.808 3.821 90.506 46.763 5.879
5.956 90.506 3.981 101.368 47.428 5.905
6.671 101.368 4.131 113.533 48.098 5.92
7.473 113.533 4.27 127.158 48.772 5.953
8.370 127.158 4.398 142.418 49.454 6.065
9.376 142.418 4.471 159.509 50.147 6.186
10.502 159.509 4.609 178.651 50.853 6.312
11.763 178.651 4.731 200.090 51.577 6.49
13.175 200.090 4.837 224.101 52.324 6.699
14.757 224.101 4.931 250.994 53.1 7.007
16.529 250.994 4.966 281.114 53.912 7.34
18.514 281.114 5.071 314.849 54.769 7.785
20.736 314.849 5.164 352.632 55.682 8.318
23.225 352.632 5.27 394.949 56.66 8.955
26.013 394.949 5.351 437.596 57.615
 115
SOLUTION
The diagnostic plot confirms the presence of four flow regimes in
Well Prf-3 : ER, EL, LR and LL.
100

t*ΔP'

10
ΔP

late radial

late linear

early radial
early linear
1
0.001 0.010 0.100 1.000 10.000 100.000 1000.000
time, hr

FIG. 2.12 No smoothing, using dP/d(lnt)=t*dP’)

116

Derivative values in table 5.1b are simulated (no noise in WBS

hump)
 117
A plot of pressure change Δp vs. the logarithm of time (Fig. 2.13 )
confirms two radial flow regimes.

Early radial - A straight line fits the data from 0.1 to 1 hour; the slope
of the line, mER is 6.0 psi/cycle, and Δplhr= 27.35 psi

Late radial - A straight line fits the data from 72 to 142 hours; the
slope of the line, mLR is 13 psi/cycle, and Δplhr= 18.8 psi

FIG. 2.13

118
A plot of pressure change Δp vs. the square root of time (Fig.
2.14 ) confirms two linear flow regimes.

Early linear - A straight line fits (approximately) the data from 2

to 4 hr^0.5; the slope of the line, mEL is 3.3 psi/t0.5 and ∆P(t=0)=
26.5

Late linear - A straight line fits the data starting 15 hr^0.5; the
slope of the line, mLL is 0.89 psi/t0.5 and ∆P(t=0)= 40

FIG. 2.14
 119
Calculations

1 – Early Radial

Average permeability in
the vertical plane:
162.6qBµ
kzk y =
LmER
162.6 × 800 ×1.25 × 1.0
= = 30.11 md
900 × 6

Mechanical skin:

 ∆P  k y kz  
sm = 1.151 wf (t =1hr ) − log  + 3.227 
 mER  φµ ct rw2  
   
 27.3  30.11  
= 1.151 − log −6 2 
+ 3.227  = −0.636
 6  0.2 × 1.0 ×15 ×10 × 0.25  

120
2 – Early Linear
Permeability in the y-direction (parallel to the HW):
2
 8.128qB  µ  8.128 × 800 ×1.25 
2
1
k y =   =  = 111 md
 mEL Lh  φct  3.3 × 900 ×150  0.2 ×15 ×10
−6

Calculate Sz (pseudoskin resulting from partial penetration in the

vertical direction).

0.058 k z  ∆P(t =0)  0.058 14.47  26.5 

SZ + Sm =   = −6   = 6.82
h φµ ct  m EL  150 0.2 × 1.0 × 15 × 10  3.3 

S Z + S m = 6.82 ⇒ S z = 6.82 − S m = 6.82 + 0.636 = 7.45

ye

y e /2
z y h
zw
X

x e /2
xe
 121
3 – Late Radial

162 . 6 qB µ
k xk y =
hm LR
162 . 6 × 800 × 1 . 25 × 1 . 0
= = 83 . 4
150 × 13

kx =
( k xk y ) 2

=
83 . 4 2
= 62 . 6 md
ky 111

kz =
( k zk y )
2

=
30 . 11 2
= 8 md
ky 111

122
1.151L k z  ∆P(1hr )  kx  
sm + sz =  − log   + 2.023
2 
h k x  m LR  φµ c t L  
1.151 × 900 8 18.8  62.6  
=  − log  −6 2 
+ 2.023 = 6.83
150 62.6  13  0.2 × 1.0 × 15 × 10 × 900  

S Z + S m = 6.83
S z = 6.83 − S m = 6.83 + 0.636 = 7.46

ye

y e /2
z y h
zw
X

x e /2
xe
4 – Late Linear
8.128qB µ
hx =
mLL h k yφct
8.128 × 800 ×1.25 1
=
0.89 ×150 111× 0.2 ×15 ×10 −6
= 2670 ft

0.058L w kz  ∆P( t =0) 

Sm + S x + Sz =  
hh x φµ c t  m LL 
0.058 × 900 8  40 
= −6   = 16
150 × 2670 0.2 × 1.0 × 15 × 10  0.89 
S Z + S m + S x = 16 ⇒ S x = 16 − ( S m + S z ) = 16 − 6.8 = 9.2

124

Note: L

In the book (p.233) they obtained ye

ye /2
(simulated, HW parallel to Y-axis): z y
zw
h
X

x e/2
xe
From EL: kx = 100 md
From ERF: kz = 10 md
ky = 50 md

The results calculated here from

Semilog analysis
(HW parallel to X-axis)

From EL: kx = 62.5 md

From ERF: kz = 14.5 md
ky = 111 md
 125

7 – STARTING AND
ENDING TIMES OF
FLOW REGIMES
(useful for design purposes)

126
STARTING AND ENDING TIMES OF FLOW REGIMES

It is important to estimate the times relating to each of the flow

regimes because each flow regime has a unique mathematical
solution relating to it, and certain reservoir parameters can only be
approximated during particular flow regimes.

Method 1 – Goode and Thambynayagam’s Equations

1. Early-Time Radial Flow

The end of the early-radial flow regime may occur either when the
transient reaches a vertical boundary or when flow comes from
beyond the end of the wellbore.

The early-time flow ends:

190d Z2.095rw−0.095φµo ct
teER = (3.1)
kv
dz = the shortest distance between the well and the z-boundary, ft
 127
2. Early-Time Linear Flow
The early-time linear flow ends at:

20.8φµo ct L2
teEL = (3.2)
kv

The intermediate-time linear flow may not develop if the time

estimated from Eq. 3.1 is less than the time calculated for the early
time radial flow to end.

3. Late-Time Radial Flow or Pseudo-Radial Flow

If late-time radial develops, it will begin at approximately:

1230.0L2φµo ct
teLR = (3.3)
kv

128
Method 2 – Odeh and Babu's Equations

1. Early-Time Radial Flow

The duration of this period may be approximated by the
lowest of the following two terms:

1800d z2φµo ct 125L2φµo ct

teER = (3.4) teER = (3.5)
kv kx

Hemiradial Flow
In the absence of wellbore storage, the start of hemiradial flow is
approximately given by:

1800d z2φµo ct
t sHR = (3.6)
kv

the end of hemiradial flow occurs at:

1800Dz2φµo ct FIG. 1.14

teHR = (3.7)
kv
 129

2. Intermediate-Time or Early Linear Flow

Time durations for the start and end of this linear flow can be
found by:

1800Dz2φµo ct
t sEL = (3.8)
kv

OR

160L2φµo ct
teEL = (3.9)
kx

DZ = h-dZ = the longest distance between the well and the z-

boundary, ft

130
3. Late-time radial flow
This flow period starts at

1480L2φµo ct
t sLR = (3.10)
kx

This flow period ends at

1650φµo ct d x2
teLR = (3.11)
kx

2000φµo ct (d x − 0.25L )
2

OR teLR = (3.12)
kx

kx = permeability in x-direction, mD
dx = the shortest distance between the well and the x-boundary, ft
Dx = the longest distance between the well and the x-boundary, ft.
 131
4. Late Linear flow
The start of this time period is the maximum of two equations. The
first depends on the time at which the transient front reach the
boundary, Dx, beyond the end of the horizontal well.

It also depends on the permeability, 4800φµo ct (Dx − 0.25L )

2

kx, in the direction parallel to the t sLL =

kx
wellbore.
(3.13)
Another requirement for the start
of the late-linear flow regime is the 1800Dz2φµo ct
t sEL = (3.14)
time at which we reach the kv
maximum vertical distance, Dz:

Usually, the start of the late-linear flow regime is dictated by the time
to reach the boundaries in the y-direction.

1650φµo ct d y2
The end of this period is given by: teLL = (3.15)
ky

132

EXERCISE 2
 133

Calculating the Time Required to End Early-Time Radial Flow

A horizontal oil well, which is 1000-ft long, is drilled in a

reservoir with the following characteristics:

h = 120 ft rw = 0.354 ft
Ø = 15.0% Bo = 1.235 rb/stb
µo= 0.35 cP ct= 10.0x 10-6psi-1
k = 8 mD (horizontal kH, from pilot-hole test)
kv = 0.2 mD (from core data)

The Well is in the central elevation of the reservoir, and the

distance from the upper reservoir boundary to center of
horizontal well is 20 ft.

Estimate the time required to end initial radial flow.

134
Solution

Method 1 (Goode)- using Eq. 3.1

dz = (120/2) - 20 = 40 ft

190d Z2.095rw−0.095φµo ct
teER =
kv
190 × 402.095 × 0.354−0.095 × 0.15 × 0.35 ×10.0 ×10−6
= = 0.29 hr
0.2
Method 2 (Babu)- using Eqs. 3.4 and 3.5 (assuming kx=kh)
1800d z2φµo ct 125L2φµo ct
teER = teER =
kv kx
1800 × 402 × 0.15 × 0.35 ×10.0 ×10−6 125 ×10002 × 0.15 × 0.35 ×10.0 ×10−6
= = 7.56 hr = = 8.2 hr
0.2 8

The minimum of these two values is 7.56 hr.

Thus (according to Babu’s method), the initial radial flow period will
end in 7.56 hr.
 135

8 – ADVANCED
TECHNIQUE
-
TDS* TECHNIQUE

*Tiab Direct Synthesis

136
TDS TECHNIQUE IN ANISOTROPIC MEDIA

CASE 1 – ALL FLOW REGIMES ARE OBSERVED

A log-log plot of dimensionless pressure and pressure derivative
versus time is shown in Figure 5.3.

Identified on this figure are the characteristic points and lines

available for evaluating the well test.

100

late
linear
t*ΔP'

10 late
t*ΔPLR radial
ΔP

early
t*ΔPER radial early
linear

t*ΔPEL1hr t*ΔPLL1hr
1
0.001 0.010 0.100 1.000 10.000 100.000 1000.000
time, hr FIG. 5.3
 139
1 – EARLY RADIAL FLOW REGIME y =∞ Z

The first flow regime represents X

Z=
hz

the infinite-acting, early time radial y=0

Lw

flow regime in the yz plane Lza

Lzb
rw 0
Z=
z=zw=hs
X=Lxd X=Lx1

The dimensionless flow equation X=0 X=hx

which describes this flow regime is:

1 k y   k z  
PD = ln t D + 0.80907 + 2 S m 
2 k z   k y  

(5.1)
Where: 100

k y Lw ∆P 0.0002637 k y t
PD = (5.2) tD =
φµ ct rw 2
late

141.2q µ B
linear

t*ΔP'
10 late
t*ΔPLR radial

ΔP
(5.2a) early
The derivative of Eq. 5.1 is: t*ΔPER radial early
linear

t*ΔPEL1hr t*ΔPLL1hr
1 ky
t D × P' D =
1

(5.3) 0.001 0.010 0.100 1.000 10.000 100.000 1000.000

2 kz time, hr

140
100
Substitution of the dimensionless
variables in Eq. 5.3 and
rearranging the results we arrive late
linear
t*ΔP'

at the expression for the kykz 10 late

t*ΔPLR radial
permeability product:
ΔP

early
t*ΔPER radial early
linear

70.6qµB (5.4)
t*ΔPEL1hr t*ΔPLL1hr

ky kz = 1

L (t × ∆P ' ) ER 0.001 0.010 0.100 1.000

time, hr
10.000 100.000 1000.000

where (t*ΔP’)ER is the derivative value corresponding to the early

time radial flow period on the graph.

If kz is known from a partial penetration pilot-hole test or core

data, then we can solve for ky:

2
1  70.6qµB 
ky =   (5.4a)
k z  L (t × ∆P' ) ER 
 141
100
The early radial flow period is only
influenced by the mechanical skin
attributed to drilling and completion late
linear

t*ΔP'
operations. 10 late
t*ΔPLR radial

ΔP
early
The ratio of the dimensionless pressure t*ΔPER radial early
linear
and pressure derivative expressions (Eq. t*ΔPEL1hr t*ΔPLL1hr

5.1 and 5.3) provides a method of 1

0.001 0.010 0.100 1.000 10.000 100.000 1000.000
calculating this skin factor. time, hr

In field units, the mechanical skin is determined by:

1  ∆ p ER  k y k z t ER  
Sm=  - ln   + 7.43
(5.5)
2  (t × ∆P ' ) ER  φµ ct r 2w  
  

where ∆PER, (t*∆

∆P’)ER, and tER are read from the graph at a
convenient point during the early radial flow period.

The WBS coefficient C is calculated as discussed  qB  t

earlier, i.e. by taking any point on the early-time C = 
unit slope line and using the following Equation:  24  ∆P

142
2 – EARLY LINEAR FLOW REGIME

As discussed in the previous section, this flow regime occurs in

response to the pressure transient reaching the upper and lower
boundaries simultaneously.

This flow period occurs when the horizontal well length is sufficiently
long in comparison to the thickness and/or the vertical permeability is
greater than the horizontal permeability.

The dimensionless flow equation

which describes this flow regime is:

2rw ky 100

PD = π t D + ( s z + sm ) (5.6)
hz kz late
linear
t*ΔP'

10 late
radial

The derivative of Eq. 5.6 is:

t*ΔPLR
ΔP

early
t*ΔPER radial early
linear

rw t*ΔPEL1hr t*ΔPLL1hr

t D × P'D = π tD (5.7)
1
0.001 0.010 0.100 1.000
time, hr
10.000 100.000 1000.000

hz
 143
Substituting for dimensionless terms and taking the logarithm of
both sides of Eq. 5.7 gives:

log(t × ∆P' ) = 0.5 log t + log(0.5mEL ) (5.8)

2(4.064) qB µ
m EL = (5.9)
Lh z k y φ ct
100

At time t = 1 hr, Eq, 5.8 becomes:

late

(t × ∆P' ) EL1 = 0.5mEL (5.10)

linear

t*ΔP'
10 late
t*ΔPLR radial

ΔP
early
Substituting for mEL and solving for t*ΔPER radial early
linear

ky or for the effective well length L t*ΔPEL1hr t*ΔPLL1hr

(if ky is known from the early radial) 1

0.001 0.010 0.100 1.000 10.000 100.000 1000.000

gives: time, hr

µ
2
 4.064qB  µ 4.064qB
k y =  
 (5.11) L= (5.11a)
 h z L (t × ∆P' ) EL1  φct h z (t × ∆P' ) EL1 k y φ ct

144
The early linear flow period is
influenced by a combination of
mechanical skin and partial 100

penetration in the z-direction

pseudoskin.
late
linear
t*ΔP'

Dividing the dimensionless pressure 10

t*ΔPLR
late
radial
ΔP

equation (Eq. 5.6) by the early

radial
dimensionless pressure derivative t*ΔPER early
linear

equation (Eq. 5.7) and expressing in t*ΔPEL1hr t*ΔPLL1hr

1
field units, results in an expression 0.001 0.010 0.100 1.000 10.000 100.000 1000.000
time, hr
for Sz:

0.029 k z t EL  ∆PEL 
Sz=  − 2 − S m (5.12)
hz φµ ct  (t × ∆P' ) EL 

Where:
∆PEL, (t*∆
∆P’)EL, and tEL are read from the graph at a convenient point
during the early linear flow period, and

Sm is obtained from the early radial flow.

 145
3 – LATE RADIAL FLOW REGIME

This flow regime may not be present if the

horizontal length is long compared to the
reservoir thickness or if the reservoir width is
small relative to the horizontal length.

The dimensionless flow equation which describes this flow regime is:

ky   16 k x r 2w   (5.13)
PD =
L
ln  + 0.80907 + k y ( S z + S m)
2 tD
2 hz kx   k y L   kx kz

100

And the derivative is:

late
linear

t*ΔP'
10 late
t*ΔPLR radial

ΔP
1L ky
t D × P' D =
early

(5.14) t*ΔPER radial early

2 hz kx t*ΔPEL1hr t*ΔPLL1hr
linear

1
0.001 0.010 0.100 1.000 10.000 100.000 1000.000
time, hr

146
Substituting for the dimensionless terms in Eq. 5.14, we can
solve for kxky product:

70.6qµB
100

kx ky = (5.15)
h z (t × ∆P ' ) LR late
linear
t*ΔP'

10 late
radial
Knowing ky from the previous flow t*ΔPLR
ΔP

early
regimes, we can calculate kx: t*ΔPER radial early
linear

t*ΔPEL1hr t*ΔPLL1hr
2
1  70.6qµB  1

kx=   (5.15a) 0.001 0.010 0.100 1.000 10.000 100.000 1000.000

k y  h z (t × ∆P ' ) LR
time, hr

Dividing PD (Eq. 5.12) with tD*PD’ (Eq. 5.13) and substituting for
the dimensionless terms, and then solving for Sz yields:

L kz  ∆ p LR  k t  
Sz + Sm =  - ln  x LR 2  + 4.659 (5.16)
2 hz kx  (t × ∆P ' ) LR  φµ ct L  

Where: ∆PLR, (t*∆

∆P’)LR, and tLR are read from the graph at a
convenient point during the late radial flow period
 147

The system or total skin (St) is also called apparent skin (Sa)
and is defined as the difference between PwD of HW and VW
corresponding to the LRF (Ozkan et al.):

k HW
S a = PwDHW − PwDVW (5.16a)
kz

k = k HW = 3 k x k y k z For HW (5.16b)

kh = k x k y For VW (5.16c)

The pseudoskin Sp is related to the apparent skin and

mechanical skin (van Everdingen formation damage skin
factor) as follows:

k HW
S p = Sa − Sm (5.16d)
kz

148

The Besson correlation (one of many) is used to estimate

pseudoskin factor:

 
  h k /k 2
 1
 w
4 r h k / k 2 h k / k
 −  z 
S p = ln +
z
ln  z z
 + 2( w − 0.5) 2 
 L  L   zw    L  6
 h 
 2πrw (1 + k / k z ) cos π ( h − 0.5)  
  

L kz / k (5.16e)
for ≥ 0.4 (5.16f)
h
Note: The following equation may be used to estimate
potential mechanical skin of HW if mechanical skin of a VW
is know:

 k h
S mHW = S mVW  
h
(5.16g)
L
 kykz 
 149
4 – LATE LINEAR FLOW REGIME 100

The late-linear flow regime (which is late

linear
not very likely to be observed)

t*ΔP'
10 late

normally starts after the pressure t*ΔPLR radial

ΔP
early

transient has reached the boundaries t*ΔPER radial early

linear

in the z- and x-directions and the flow 1

t*ΔPEL1hr t*ΔPLL1hr

behavior with regard to these 0.001 0.010 0.100 1.000

time, hr
10.000 100.000 1000.000

directions has become pseudosteady-

state (finite system).

The dimensionless pressure function

for this flow regime is:

 L rw  ky
P D = 2  π t D + ( S x + S z + S m ) y=∞ Y

 hx hz  kz X

hz
Z=
Z
(5.17) y=0

And its derivative is: Lw

Lza z=hs
Lzb
rw 0
Z=
 L rw  X=Lxd X=Lx1

t D × P ' D =   π tD (5.18)
 X=0 X=hx
 hx h z 

150
Substituting for dimensionless terms in Eq. 5.18, and taking the
logarithm of both sides of the resulting equation gives:

log(t × ∆P' ) = 0.5 log t + log(0.5mLL ) (5.19)

Where:
100

2(4.064)qB µ
m LL = (5.20)
k y φ ct h x
late
2 linear
hz
t*ΔP'

10 late
t*ΔPLR radial
ΔP

early
radial
At time t = 1 hr. t*ΔPER early
linear

t*ΔPEL1hr t*ΔPLL1hr

(t × ∆P')LL1 = 4.064qB µ
1
0.001 0.010 0.100 1.000 10.000 100.000 1000.000
time, hr

hx h z k y φ ct (5.21)
 151
hx = formation thickness in the x-boundary, ft.

Sx = SP = Skin due to partial penetration in the

X-direction (Sx=0, if well penetrates entire
formation thickness in x-direction)

hx ye

ye /2
z y h
zw
X

x e/2
xe

152
Eq. 5.21 can be solved for hx (where 100

ky is known from the EL flow regime,

or for ky (if hx is known and the EL late
linear
flow is not well defined):
t*ΔP'

10 late
t*ΔPLR radial
ΔP

2
 4.064qB   µ  early

k y =     radial

2
(5.22) t*ΔPER early

 h z (t × ∆P' ) LL1   φ ct h x 
linear

t*ΔPEL1hr t*ΔPLL1hr
1
0.001 0.010 0.100 1.000 10.000 100.000 1000.000

µ
time, hr

4.064qB
hx = (5.22a)
h z (t × ∆P' ) LL1 k y φ ct
Substituting for dimensionless terms in
Eq. 5.17 and 5.18, and taking the ratio of
the resulting equation yields an equation
for Sx:
0.029 L k z t LL  ∆ p LL  (5.23)
Sx + Sm+ Sz =  − 2
hx hz φµ ct  (t × ∆P' ) LL 
Where:
∆PLL, (t*∆
∆P’)LL, and tLL are read from the graph at a convenient
point during the late linear flow period.
 153
VERIFICATION
The ratio of the early radial and late radial pressure derivative lines results in
the following expression:

70.6qµB 70.6 qµB (t × ∆P ' ) ER hz kx

(t × ∆P ' ) ER = (t × ∆P ' ) LR = =
Lw k y k z hz k x k y (t × ∆P ' ) LR Lw kz

(5.24a) (5.24b) (5.24c)

The ratio of the early linear and late linear pressure derivative lines at time t
= 1 hr results in the following expression:

µ µ (t × ∆P' ) EL1 hx
(t × ∆P ' ) EL1 =
4.064qB (t × ∆P')LL1 = 4.064qB =
(t × ∆P ' ) LL1 Lw
h z Lw k y φ ct hx h z k y φ ct
(5.25b) (5.25c)
(5.25a)

(t × ∆P ' ) EL1 1 kz
Combining Eq. 5.24a and 25a: = (5.25d)
(t × ∆P ' ) ER 17.37 h µφct

(t × ∆P ' ) LL1 1 kx
Combining Eq. 5.24b and 25b: = (5.25e)
(t × ∆P ' ) LR 17.37 hx µφct

154
The time of intersection of the  h2 
early radial and early linear flow t ERiEL = 301.77φµ ct  z  (5.26)
periods is given by:  kz 

The time of intersection of the  L2 

late radial and early linear flow t LRiEL = 301.77φµ ct  w  (5.27)
periods is given by:  kx 

The time of intersection of the  h2 

late radial and late linear flow t LRiLL = 301.77φµ ct  x  (5.28)
periods is given by:  kx 

The time of intersection of the  h 2  h 

2

t ERiLL = 301.77φµ ct  z  x  (5.29)

early radial and early linear
flow periods is given by:  k z  Lw 

Equations 5.24 - 5.29 can be used to verify the calculated values of

kx, ky, kz, and hx.
 155
2
Combining these times t ERiEL t L 
= LRiEL =  w  (5.30)
of intersection yields: t ERiLL t LRiLL  hx 

Eq. 5.30 can be used for verification or well test design purposes.

For example, given horizontal well length and reservoir width, the
equation provides a quantitative measure of the time span to
capture the anticipated flow regimes.

If the left-hand-side of Eq. 5.30 is small (~1), then the late radial
flow period will not be observed and the start time of the late
linear flow period will be dependent only on the magnitude of the
early linear flow period.

If the ratio is large (~100) then the late radial period is observed
but the test design will have to extend a sufficient period of time
to capture the late linear flow period.

156
 1004.064qB  µ
k y = 
2

 kz =
(
SOLUTION
k y kz )2
4.064qB µ

L (t × ∆P' )EL1  φct hx =
The hz diagnostic plot confirms kthe
y h z (t × ∆P' ) LL1 k y φ ct
presence of all
k z tfour flow
∆PELregimes
0.029 EL   0.029 L  ∆ p LL 
+ SWell
S z in m = Prf-3.  − 2 S x + Sm+ S z =
k z t LL
 − 2
hz φµ ct  (t × ∆P' ) EL  hx hz φµ ct  (t × ∆P' ) LL 
FIG. 3.3
late
2 4 linear
t*ΔP'

10 late
t*ΔPLR radial
ΔP

early
t*ΔPER radial early
linear

t*ΔPEL1hr
3
t*ΔPLL1hr
70.6qµB 70.6 qµB
ky kz =
1
1 kx ky =
(t × ∆P ' ) ER0.010
L0.001 0.100 1.000 10.000 100.000 (t × ∆P ' ) LR
h z 1000.000

Sm =
1  ∆ p ER
 - ln
 k y k z t ER 
  + 7.43


time, hr
kx =
( kx k y )2

2  (t × ∆P ' ) ER  φµ ct r 2w   ky
  
 157

EXERCISE 3

158

EXAMPLE - Well Prf-3

A pressure drawdown in a horizontal well yielded the data shown

in Table 5.1 (same as Table 2.1). The following data are known:

q = 800 STB/D µ = 1.0 cp

B = 1.25 RB/STB rw = 0.25 ft
Ø = 0.2 ct = 15 x 10-6 psi-l
h = 150 ft Lw= 900 ft
hY=aH=5280 ft hX=bH=5280 ft

The horizontal well is approximately centered in (finite) drainage

volume.

Interpret this test using the TDS Technique.

 159
Table 5.1a – Pressure Drawdown Data
t Pwf t Pwf t Pwf t Pwf
0.000 3000 0.119 2978.18 1.707 2971.336 29.136 2959.721
0.001 2996.745 0.134 2977.859 1.913 2971.031 32.633 2959.104
0.002 2994.507 0.151 2977.543 2.143 2970.719 36.550 2958.48
0.003 2992.58 0.170 2977.23 2.401 2970.397 40.936 2957.848
0.004 2990.776 0.192 2976.921 2.690 2970.063 45.850 2957.209
0.006 2989.243 0.216 2976.616 3.014 2969.715 51.353 2956.562
0.007 2987.956 0.242 2976.312 3.376 2969.353 57.516 2955.906
0.009 2986.87 0.272 2976.011 3.782 2968.976 64.419 2955.242
0.011 2985.939 0.306 2975.712 4.237 2968.582 72.150 2954.573
0.013 2985.206 0.343 2975.415 4.746 2968.172 80.808 2953.901
0.016 2984.492 0.386 2975.12 5.317 2967.744 90.506 2953.237
0.019 2983.856 0.433 2974.826 5.956 2967.3 101.368 2952.572
0.022 2983.284 0.485 2974.533 6.671 2966.84 113.533 2951.902
0.025 2982.761 0.545 2974.242 7.473 2966.365 127.158 2951.228
0.029 2982.278 0.611 2973.951 8.370 2965.875 142.418 2950.546
0.033 2981.848 0.685 2973.662 9.376 2965.37 159.509 2949.853
0.038 2981.418 0.768 2973.374 10.502 2964.852 178.651 2949.147
0.044 2981.009 0.861 2973.085 11.763 2964.323 200.090 2948.423
0.050 2980.617 0.965 2972.797 13.175 2963.782 224.101 2947.676
0.057 2980.239 1.082 2972.509 14.757 2963.231 250.994 2946.9
0.064 2979.878 1.213 2972.22 16.529 2962.668 281.114 2946.088
0.073 2979.523 1.359 2971.929 18.514 2962.096 314.849 2945.231
0.083 2979.176 1.523 2971.635 20.736 2961.515 352.632 2944.318
0.094 2978.836 23.225 2960.925 394.949 2943.340
0.106 2978.506 26.013 2960.327 437.596 2942.385

t dP t*dP'
0 0.1190 21.82 2.762
0.0010 3.255 0.1340 22.141 2.675
0.1510 22.457 2.643
0.0020 5.493 4.190
0.1700 22.77 2.591
0.0030 7.42 5.641
0.1920 23.079 2.565
0.0040 9.224 5.237
0.2160 23.384 2.633
0.0060 10.757 7.091 0.2420 23.688 2.626
0.0070 12.044 6.818 0.2720 23.989 2.557
0.0090 13.13 4.498 0.3060 24.288 2.571
0.0110 14.061 4.502 0.3430 24.585 2.551
0.0130 14.794 3.965 0.3860 24.88 2.529
0.0160 15.508 3.582 0.4330 25.174 2.571
0.0190 16.144 3.809 0.4850 25.467 2.540
0.0220 16.716 4.003 0.5450 25.758 2.521
0.0250 17.239 3.704 0.6110 26.049 2.537
0.0290 17.722 3.294 0.6850 26.338 2.523
0.0330 18.152 3.194 0.7680 26.626 2.523
0.0380 18.582 2.921 0.8610 26.915 2.527
0.0440 18.991 2.938 0.9650 27.203 2.521
0.0500 19.383 2.977 1.0820 27.491 2.523
0.0570 19.761 3.008 1.2130 27.78 2.545
0.0640 20.122 2.921 1.3590 28.071 2.570
0.0730 20.477 2.701 1.5230 28.365 2.601
0.0830 20.824 2.718 1.7070 28.664 2.649
0.0940 21.164 2.739 1.9130 28.969 2.713
0.1060 21.494 2.783 2.1430 29.281 2.790
 2.4010 29.603 2.886 45.8500 5.672
42.791
2.6900 29.937 2.999
51.3530 43.438 5.748
3.0140 30.285 3.126
3.3760 30.647 3.256 57.5160 44.094 5.823
3.7820 31.024 3.394 64.4190 44.758 5.880
4.2370 31.418 3.541 72.1500 45.427 5.916
4.7460 31.828 3.691 80.8080 5.894
46.099
5.3170 32.256 3.840
90.5060 46.763 5.863
5.9560 32.7 3.985
6.6710 33.16 4.121 101.368 47.428 5.889
7.4730 33.635 4.253 113.533 48.098 5.929
8.3700 34.125 4.386 127.158 48.772 5.982
9.3760 34.63 4.508 142.418 6.066
49.454
10.5020 35.148 4.616
11.7630 35.677 4.719 159.509 50.147 6.172
13.1750 36.218 4.816 178.651 50.853 6.309
14.7570 36.769 4.912 200.090 51.577 6.490
16.5290 37.332 5.004 224.101 52.324 6.719
18.5140 37.904 5.085 250.994 53.1 7.006
20.7360 38.485 5.165
23.2250 39.075 5.240 281.114 53.912 7.363
26.0130 39.673 5.310 314.849 54.769 7.809
29.1360 40.279 5.394 352.632 55.682 8.343
32.6330 40.896 5.474 394.949 56.66 8.989
36.5500 41.52 5.541
437.596 57.615
40.9360 42.152 5.607

162
SOLUTION
Interpretation of plot (no noise in WBS hump), derivative values
in table 5.1b are smoothed.
 163

The diagnostic plot confirms the presence of all four flow regimes in
Well Prf-3.

100

late
linear
t*ΔP'

10 late
t*ΔPLR radial
ΔP

early
t*ΔPER radial early
linear

t*ΔPEL1hr t*ΔPLL1hr
1
0.001 0.010 0.100 1.000 10.000 100.000 1000.000
time, hr

164
Values needed for evaluation using TDS Technique
 165
EARLY RADIAL
Step 1 – From the early radial:
(t*dP’)ER = 2.45
The geometric average permeability
in the vertical yz (or ER) plane is:

70.6qµB
ky kz =
L (t × ∆P ' ) ER Z
70.6 × 800 × 1.0 × 1.25
= = 32 md
900 × 2.45
Y
Step 2 – From the early radial:
at tER = 0.768, dPER = 26.63 X

1  ∆ p ER  k y k z t ER 
 
 y=∞ Y

Sm =  - ln + 7.43 X

2  (t × ∆P' ) ER  φµ ct r 2w   z

   y=0
Z
Z=h

Lw

 26.63  32 × 0.768   Lza

rw z=hs

= 0. 5  − ln + 7.43 ≈ 0.45

Lzb
Z=0

2 
X=Lxd X=Lx1
−6
 2.45  0.2 × 1.0 × 15 × 10 × 0.25   X=0 X=hx

y =∞
166
Y

EARLY LINEAR y=0

Z
X

Z=
hz

Lw
Lza z=hs
Lzb
rw 0
Z=

Step 3 – Extrapolating the ELF straight X=Lxd X=Lx1

X=0 X=hx

line to t=1 hr gives (t*dP’)EL1 = 1.73

2
 4.064qB  µ
k y =  

 h z L (t × ∆P ' ) EL1  φc t

 4.064 × 800 × 1.25 

2
1
=  −6
= 100 md
 150 × 900 × 1.73  0.2 × 15 × 10

Step 4 – Vertical permeability (k

in the z-direction): kz

kV = k z =
( ky kz ) 2

=
32 2
= 10.2 md ky
ky 100
kx
 167
Step 5 – From early linear at any tEL=4.237, dPEL=31.42,
(t*dP’)EL=3.547 (from Table)

0.029 k z t EL  ∆PEL 
S z + Sm =  − 2
hz φµ ct  (t × ∆P' ) EL 
0.029 10 × 4.237  31.42 
= −6 
− 2  = 4.5
150 0.2 × 1.0 × 15 × 10  3.547 

S Z + S m = 4.5 ⇒ S z = 4.6 − S m = 4.5 + 0.48 = 5

where Sz is the pseudoskin

resulting from partial
penetration in the VERTICAL
direction.

168
LATE RADIAL
Step 6 – From the late radial:
(t*dP’)LR = 6.0

The geometric average permeability in

the horizontal xy (or LR) plane is:

70.6qµB
kH = k x k y =
h z (t × ∆P ' ) LR
70.6 × 800 × 1.0 × 1.25
= = 78 md
150 × 6

Step 7 – Permeability in the x- 10

direction (parallel to HW):

kx =
( kx k y )
2

=
782
= 60 md
100

ky 100
60
 169
Step 8 – Calculate Sm+Sz

From late radial: at tLR = 101.4, dPLR =

47.43, (t*dP’)LR = 5.9 (from Table)

L kz  ∆p  k t  
Sz + Sm =  LRtlr
- ln  x LR 2  + 4.659 
2 hz kx  (t × ∆P ' ) LRtlr  φµ ct L  
 
900 10  47.43  60 × 101.4  
=  − ln −6 2 
+ 4.659 = 5.6
2 × 150 60  5.9  0.2 × 1.0 × 15 × 10 × 900  
S Z + S m = 5.6 ⇒ S z = 5.6 − S m = 5.6 + 0.48 = 6

The early linear and late radial do not yield (approximately) the
same value of Sz. The difference is very likely due to the fact that
early linear is not well defined.

170
LATE LINEAR
Step 9 – Extrapolating the late linear straight line to t=1 hr gives
(t*dP’)LL1 = 0.28 (from graph). Since ky is known from the EL flow
regime, the LL flow will be used here to calculate hx

4.064qB µ 4.064 × 800 × 1.25 1

hx = = = 5280 ft
hz ( t × ∆P' ) LL1 k y φ ct 150 × 0.28 100 × 0.2 × 15 × 10−6

y=∞ Y

hz
Z=
Z
y=0

Lw
Lza z=hs
Lzb
rw 0
Z=
X=Lxd X=Lx1

X=0 X=hx

hx
 171
Step 10 – Calculate total skin: Sm+Sz+Sx

From the late linear: at tLL=395, dPLL=56.66, t*dP’LL=8.955 (from

Table)

0.029 L k z t LL  ∆ p LL 
Sx + Sm + Sz =  − 2
hx hz φµ ct  (t × ∆P' ) LL 
0.029 × 900 10 × 395  56.66 
=  8.955 − 2 = 5.2
150 × 5214 0.2 × 1.0 × 15 × 10 −6

Step 10 – Calculate Spp=Sx (partial

penetration skin in the x-direction)
a) From late radial:

S Z + S m + S x = 5.2 ⇒ S x = 5.2 − ( S m + S z ) = 5.2 − 5.6 ≈ 0

b) From early linear:

S Z + S m + S x = 9.2 ⇒ S x = 5.2 − ( S m + S z ) = 5.2 − 4.52 = 0.68 ≈ 0

172
Table 5.3
Table 5.3 compares the results
obtained by the conventional semilog SLA TDS Simulated
analysis (SLA), TDS and simulatedf kx 62.6 60 50
values. ky 111 100 100
kz 14.5 10 10
There is a reasonable agreement hx 2670 5280 5200
between TDS and Simulation in the Sm -0.63 0 0
values of kx, ky, hx, Sm and Sz.
Sz 7.4 5.6 5
Note that these values are obtained
Spp 9.2 4.7
from the radial flow regimes, which
appear to be well defined.

The main difference is in kx (which is

obtained from the early linear).

The differences are very likely due to

the fact that either the early linear or
late linear, or both, are not well
defined or do not exist.
 173
VERIFICATION
kx 60
The calculated values of kx and kz yield: = = 2.45
kz 10

From the ratio of the early radial and late radial pressure derivative lines we
obtain (Eq.5.24c):

k x Lw (t × ∆P ' ) ER 900 × 2.45

= = = 2.45
k z hz (t × ∆P ' ) LR 150 × 6

Thus, the kx and kz values are correct.

From the ratio of the early linear and late

linear pressure derivative lines at time t =
1 hr we can calculate hx (Eq.5.25c):

(t × ∆P' ) EL1 1.73

h x = Lw = 900 = 5560 ft
(t × ∆P' ) LL1 0.28
This confirms the test was correctly interpreted.

174

ANALYZING THE
SEMILOG PLOT OF
dP AND t*dP’
 176
EARLY RADIAL
The skin factor can also be estimated
from the intersection point of the semilog
straight line of slope mER and (t*dP’)RER

At the intersection point, ∆PERi = (t*∆P’)ERi

70.6qµB
( t × ∆P' )ERi = (1)
L k y kz

qµB   k y k z 
 − 3.227 + 0.866s 

∆PERi = 162.6 log t (2)
L k y k z   φ µ c t rw 
2 ERi m
  

Setting these two equations equal and solving explicitly for mechanical
(Sm) gives:

 kykz 
S m = 4.2277 − 1.1515 log t  (3)
 φµc t rw 2 ERi

 
 VERIFICATION 1
Calculate Sm from semilog plot of dP and derivative. The
following equation is only applicable to the ER (tERi =
intersection of straight lines corresponding to ER on the
semilog and derivative curves)

 kykz 
S m = 4.2166 − 1.15 log t 
 φµc t rw 2 ERi 
 

178
LATE RADIAL
The skin factor (Sz+Sm) can also be
estimated from the intersection point of
the semilog straight line of slope mLR and
(t*dP’)LR

At the intersection point, ∆PLRi = (t*∆P’)LRi

70.6qµB 70.6qµB
( t × ∆P' ) LR = (4) ( t × ∆P' )ER = (5)
h kx ky L k y kz

162.6qµB   k x   141.2qµ B
∆PLRi = log t  − 2.023 + (s z + s m ) (6)
h k x k y   φµ c t L
2 LRi 
 
 L k y k z

Substituting Eqs. 4 and 5 into Eq. 6 gives:

  kx  
∆PLRi = 2.303( t × ∆P' ) LR log  t LRi  − 2.023 + 2( t × ∆P' ) ER (S z + S m ) (7)
 
  φµ c t L
2
 
At the intersection point, i.e. setting Eq.4=Eq.7, i.e. ∆PLRi = (t*∆P’)LRi
yields:

  kx  
( t × ∆P' ) LR = 2.303( t × ∆P' ) LR log

t 

− 2 . 023 + 2( t × ∆P' ) ER (S z + S m )
  φµ c t L
2 LRi
 

  kx   ( t × ∆P' ) ER
1 = 2.303log
 φµ c L2
t 

− 2 . 023+2 (Sz + Sm )
( t × ∆P' ) LR
LRi
  t  

  
1 − 2.303log  k x t  − 2.023  = 2( t × ∆P' ) ER (S + S )
  2 LRi 

  φµ c t L   ( t × ∆P' ) LR
z m
 

Solving for skin:

  
Sz + Sm =
( t × ∆P' ) LR 1 − 2.303log  t LRi k x 
 − 2.023 
  
2( t × ∆P' ) ER   φµ c t L  
2
 

VERIFICATION 2
Calculate Sm+Sz from semilog plot of dP and derivative. The
following equation is only applicable to the LR (tLRi =
intersection of straight lines corresponding to LR on the
semilog and derivative curves)
  
Sz + Sm =
( t × ∆P' ) LR 1 − 2.303log  t LRi k x 
 − 2.023 
2( t × ∆P' ) ER    
  φµ c t L
2
   
 181

9 – Finite
Conductivity
Horizontal Well

182
FINITE CONDUCTIVITY
HORIZONTAL WELLS
Most well test analysis techniques presented in the literature are
based on the assumption that horizontal wells are best modeled as
having infinite conductivity.

However, it has been observed in actual field cases that, after the
early radial flow regime, the bilinear flow may also occur during the
transient period of a horizontal well pressure test in homogeneous
reservoirs.

Quarter slopes, which normally indicate finite conductivity behavior,

have been observed for low conductivity and high rate horizontal
wells, especially for low dimensionless horizontal well length (h/L).

In the following section a description of the bilinear flow that occurs

in horizontal well transient behavior under the finite conductivity
assumption is presented.
 183

HORIZONTAL WELL

h
L/2
Zw
Z

Y X

FIG. 6.1
Fig.3.2. A Horizontal Well Model.

184

Type curves

bilinear

FIG. 6.2
 185
Figure 6.3 shows a comparison
between two different type curves, one
for infinite conductivity horizontal well in
dashed lines, the other is for finite
conductivity horizontal well in unbroken
lines.

We observe that the effect of the

effective well length on finite
conductivity is pronounced at early times
where radial flow is dominant for long
horizontal wells. FIG. 6.3

The two solutions tend to overlap as transient proceeds.

We can conclude that, at early times of the transient period, the longer
the horizontal well is, the more sensitive the pressure solution is to finite
conductivity.

For example, finite conductivity dimensionless pressure solution, for

LD=100, is equal to the infinite conductivity dimensionless pressure solution,
for LD=50.

186
Interpretation methods
The sequence of flow regimes is typically as shown in Figure.
Notice the bilinear flow (finite conductivity HW) has replaced linear
flow (infinite conductivity HW).

Only the bilinear flow regime will be analyzed in this section.

The equations and analysis of the other flow regimes that may be
observed are exactly as discussed previously.

1000
ΔP t*ΔP'

100 bilinear
slope=0.25 late
radial
early
radial

constant
pressure
10
0.001 0.01 0.1 1 10 100
time, hr
FIG. 6.4
 CONVENTIONAL
TECHNIQUE

188
CONVENTIONAL TECHNIQUE

Based on the analogy of a finite horizontal well model to a finite

conductivity hydraulically fractured well model, the equation governing the
bilinear flow regime is as follows:

31.128qµB 141.2qµB
∆Pw = t
1/ 4
+ (S m + S z ) (6.1)
k y hLw (φµc t )
3/ 4 1/ 2 1/ 4
k y k z Lw

OR

∆Pw = mBLt 1/ 4 + I BLS (6.2)

31.128qµB
mBL = (6.3)
k y3 / 4 hL1w/ 2 (φµct )1/ 4

141.2qµB (6.4)
I BLS = (S m + S z )
k y k z Lw
 189
Thus a Cartesian plot of ∆P vs. t1/4 will have straight line of slope mBL which
can be used to calculate ky.
4/3
 31.128qµB 
k y =  
1/ 4  (6.5)
m hL
 BL w
1/ 2
(φµct ) 

The intercept at t = 0 hr can be used to calculate Sm+Sz, assuming kz is

known from a pilot-hole test or core analysis, and Lw=Leff.

I BLS k y k z Lw 400

Sm + S z = (6.6) 350 FIG. 6.5

141.2qµB 300
250

ΔP
200
bilinear
150 slope=mbl
If Sm is obtained from the
100
early radial, then we can
50 IblS
calculate Sz.
0
0 0.5 1 1.5 2
t^0.25

MODERN
TECHNIQUE
 191

According to the TDS method, the second flow regime to appear

(after the early radial) is the intermediate time bilinear flow.

The dimensionless equation that describes this flow regime is:

ky
PD = 1.73t 1D/ 4 + (S m + S z ) (6.7)
kz

kyh
PD = ∆P (6.8)
141.2qµB

0.0002637 k y t
tD = (6.9)
φµct L2w

The corresponding pressure derivative is given by:

t D × PD' = 0.4325t 1D/ 4 (6.10)

192
Substituting for the dimensionless parameters and solving for the
derivative of well pressure, we obtain:

t × ∆P ' = 0.25mBL t 0.25 (6.11)

31.128qµB
1000

mBL = 3 / 4 1/ 2 (6.12)
k y hLw (φµct )1/ 4
ΔP t*ΔP'

100 bilinear
slope=0.25 late
radial

Taking the logarithm of both sides of early

radial

the equation gives: (t*ΔP')1hr

constant
pressure

( )
10

Log t × ∆P ' = 0.25 log(t ) + log(0.25mBL )

0.001 0.01 0.1 1 10 100
time, hr

(6.13)

Thus a loglog plot of t*dP’ vs. time should yield a straight line of
slope 0.25 during the bilinear flow regime.
 193
1000
At time t = 1hr Eq. 6.13 becomes:

 31.128qµB 
(t × ∆P )
'
= 0.25mBL = 0.25 3 / 4 1/ 2 

t*ΔP'
BL1hr  k hL (φµc )1/ 4  100 bilinear

 y  slope=0.25

ΔP
w t late
radial
early
radial

(6.14) (t*ΔP')1hr
constant
pressure
10
0.001 0.01 0.1 1 10 100
time, hr

This equation can be used to determine ky from the straight line

(extrapolated if necessary) of slope=0.25 at t=1hr:

4/3
 0.25  31.128qµB 
ky =   1/ 2 
1/ 4 
(6.15)
 (t × ∆P')BL1hr hL
 w (φµct ) 

194
Combining Equations 6.2 and 6.11 Yields: 1000

0.0551 k z Lw  ∆PBL1hr 
Sm + S z =  − 4 
 (t × ∆P')BL1hr
ΔP t*ΔP'

h(φµct k y )1/ 4  (6.16) 100 bilinear

slope=0.25 late
radial
early
radial

(t*ΔP')1hr
constant

At time t=1hr, Eq. 6.2 gives: 10

pressure

0.001 0.01 0.1 1 10 100

∆Pw = mBL t 1/ 4 + I BLS time, hr

(∆Pw )BL1hr = mBL + 141.2qµB ( S m + S z ) (6.17)

k y k z Lw
Thus
31.128qµB
Lw k y k z
[(∆P ) mBL =
Sm + S z = − mBL ] (6.18) k hL1w/ 2 (φµct )1/ 4
3/ 4
141.2qµB
w BL1hr
y

Where ∆PBL1hr and (t*∆P’)BL1hr are read from the straight line of
slope = 0.25 (extrapolated if necessary). If the derivative is too
noisy use the ∆P curve.
 195

1000
ΔP t*ΔP'

100 bilinear
slope=0.25 late
radial
early
radial

(t*ΔP')1hr
constant
pressure
10
0.001 0.01 0.1 1 10 100
time, hr

196
The distance to the LINEAR constant-pressure boundary
maybe estimated from a point on the negative unit-slope
line:

d = 1.932 × 10 −3
k h
(t cp ) × (t × ∆P')cp
µ qBφc t
Or, at tcp=1hr

d = 1.932 × 10 −3
k h
(t × ∆P')CP1
µ qBφc t
1000
ΔP t*ΔP'

100 bilinear
slope=0.25 late
radial
early
radial

(t*ΔP')1hr
constant
pressure
10
0.001 0.01 0.1 1 10 100
time, hr
 197
 100
k y = 
4.064qB  µ
2

 kz =
(
k y kSOLUTION
z )
2

hx =
4.064qB µ

(t × ∆P' )EL1  φplot
 hz Ldiagnostic
The ct k y the
confirms h z (t × ∆P' ) LL1 k y φ ct
presence
0.029of all four flow regimes
k z t EL  ∆PEL  0.029 L k z t LL  ∆ p LL 
+
in
Sz SmWell
= Prf-3.  − 2 Sx + Sm+ Sz =  − 2
hz φµ ct  (t × ∆P' ) EL  hx hz φµ ct  (t × ∆P' ) LL 
FIG. 3.3
late
4 linear
t*ΔP'

4/3
 100.25  31.128qµB  late
ky =   1/ 2 
1/ 4 
 (t × ∆P ')BL1hr
radial
φµ t*ΔPLR
ΔP

 hLw ( ct ) 
0.0551 k z Lw  ∆PBL1hr  early 2
Sm + S z =  − 4 
h(φµct k t*ΔP
y)
1/ 4
ER
(t × ∆P')BL1hr  radial early
linear

t*ΔPEL1hr
3
t*ΔPLL1hr
1 70.6qµB
1 70.6 qµB
kx ky =
ky kz =
(t × ∆P ' ) ER0.010
L0.001 0.100 1.000 10.000 × ∆P ' ) LR
100.000 h z ( t1000.000

1  ∆ p ER

 k y k z t ER 
 


time, hr
kx =
( kx ky )
2

Sm = - ln + 7.43
2  (t × ∆P ' ) ER  φµ ct r 2w   ky
  

198

EXERCISE 4
 199

EXAMPLE

Table 6.1 presents the buildup data

measured in a horizontal well (Hassi
Messaoud oilfield, Algeria).

Other data:
Lw=1585 ft hz=121.92 ft
Ø=0.09 Bo=1.76 RB/STB
ct=3.27*10-6 qo=1192.52 bbl/day
rw=0.25 ft tp=10 hours
µ = 0.25 cp

Interpret this test.

200

Table 6.1 – Hassi Messaoud (Algeria)

pressure test in HW

t ∆P tx∆P' t ∆P tx∆P'
0.0042 20 20.1 0.4 198.4 50.4
0.0083 40 38.5 0.5 210.2 52.9
0.0125 58 44.4 0.6042 220.1 53
0.0167 69 38.4 0.7083 228.8 53.5
0.0208 78 36.8 0.8042 235.6 55.6
0.025 84 34.7 0.9125 242.7 57
0.0292 89 36 1.0042 248.1 56.1
0.0333 94 37.1 1.45 269.1 58.1
0.0375 98 35.5 2.0208 288.3 57.5
0.0417 102 36.0 3.0417 311.8 57.5
0.05 109 36.3 4.0458 327.6 53.7
0.0708 122 38.8 5.025 339 51.5
0.0792 126 39.3 6.0625 348.5 50.9
0.0917 132 40.8 6.9875 355.4 45.5
0.1 136 38.9 7.875 361 46.2
0.1542 154 43.3 8.9417 366.6 42.9
0.1958 164 43.3 10.004 371.5 39.7
0.3042 185 48.4 11.067 375.7 31.4
12.125 379.3 38.5
 201
DIAGNOSTIC PLOT

This plot reveals 5 distinct 1000

flow periods:

1. Wellbore storage
2. Early radial

ΔP t*ΔP'
3. Bilinear 100 bilinear
slope=0.25 late
4. Late radial radial

5. Boundary effect
early
3 4
radial
5
(constant pressure) 1 2 (t*ΔP')1hr
constant
pressure
10
0.001 0.01 0.1 1 10 100
time, hr

202
TDS TECHNIQUE

t ∆P tx∆P' t ∆P tx∆P'
0.0042 20 20.1 0.4 198.4 50.4
0.0083 40 38.5 0.5 210.2 52.9
0.0125 58 44.4 0.6042 220.1 53
0.0167 69 38.4 0.7083 228.8 53.5
0.0208 78 36.8 0.8042 235.6 55.6
0.025 84 34.7 0.9125 242.7 57
0.0292 89 36 1.0042 248.1 56.1
0.0333 94 37.1 1.45 269.1 58.1
0.0375 98 35.5 2.0208 288.3 57.5
0.0417 102 36.0 3.0417 311.8 57.5
0.05 109 36.3 4.0458 327.6 53.7
0.0708 122 38.8 5.025 339 51.5
0.0792 126 39.3 6.0625 348.5 50.9
0.0917 132 40.8 6.9875 355.4 45.5
0.1 136 38.9 7.875 361 46.2
0.1542 154 43.3 8.9417 366.6 42.9
0.1958 164 43.3 10.004 371.5 39.7
0.3042 185 48.4 11.067 375.7 31.4
12.125 379.3 38.5
 -1 ?

204
EARLY RADIAL
Step 1 – From the early radial:
(t*dP’)ER = 36

70.6qµB
k y kz =
L (t × ∆P' )ER
70.6 × 1192.5 × 0.25 × 1.76
= = 0.65
1585 × 36

Step 2 – From the early radial:

at tER = 0.03, dPER = 90

1  ∆ p ER  k y k z t ER  
Sm=  - ln   + 7.43
2  (t × ∆P' ) ER  φµ ct r 2w  
  
 90  0.65 × 0.03  
= 0.5 − ln −6 2 
+ 7.43 = −2.66
 36  0.09 × 0.25 × 3.27 × 10 × 0.25  
 205

BILINEAR
Step 3 – From bilinear straight line at t=1 hr, (t*dP’)BL1 = 63

4/3
 0.25  31.128qµB 
ky =   1/ 2 
1/ 4  
 (t × ∆P')BL1  hLw (φµct ) 
4/3
 0.25  31.128 ×1192.52 × 0.25 ×1.76 
ky =   
− 6 0.25  
= 0.756
 63  121.92 ×1585 × (0.09 × 0.25 × 3.27 ×10 ) 
0 .5

Step 4 – Permeability in the z-

direction:

kz =
( ky kz )2

=
0.652
= 0.558 md
ky 0.756

206
Step 5 – From bilinear: at tBL=1, dPBL1=250, (t*dP’)BL1=63

0.0551 k z Lw  ∆PBL1hr 
Sm + S z =  − 4 
h(φµct k y )1/ 4  (t × ∆P')BL1hr 
0.0551× 0.558 ×1585  250 
=  − 4 = 0
121.92 × (0.09 ×1.76 × 3.27 ×10 −6 ) 0.25  63 

S Z + S m = 0 ⇒ S z = 0 − S m = 0 + 2.665 = 2.665
 207
LATE RADIAL
Step 6 – From the late radial:
(t*dP’)LR = 57

70.6qµB
kH = k x k y =
h z (t × ∆P ' ) LR
70.6 ×1192.52 × 0.25 × 1.76
= = 5.33 md
121.92 × 57

Step 7 – Permeability in the x-

direction (parallel to HW):

kx =
( kx ky)=
2
5.332
= 37.6 md
ky 0.756

208

Step 8 – Calculate Sm+Sz From late radial:

at tLR = 2, dPLR = 288, (t*dP’)LR = 57

L kz  ∆ p LR  k t  
S z + Sm =  - ln  x LR 2  + 4.659
2 hz kx  (t × ∆P ' ) LR  φµ ct L  
1585 0.558  290  37.6 × 2  
=  − ln −6 2 
+ 4.659 = 2.96
2 ×121.92 37.6  57  0.09 × 0.25 × 3.27 × 10 × 1585  

S Z + S m = 2.96 ⇒ S z = 2.96 − S m = 2.96 + 2.66 = 6

The bilinear and late radial do not

yield the same value of Sz.

The difference is probably due to the

fact that the bilinear is not well
defined because of the constant
pressure line (WOC line).
 209

Step 9 – The total skin (Sm+Sz+Sx) cannot be calculated because

the late linear is not observed due to the constant pressure
boundary.

The HW was drilled too close to the WOC line.

The risk of water coning is high.

Step 10 – Distance to constant pressure boundary

t cp = 10 (t × ∆P')cp = 39.7

d = 1.932 × 10 −3
kH
µ
h
qBφc t
(t cp )× (t × ∆P')cp = 364.6 ft

-1 ?

t cp = 10 (t × ∆P' )cp = 39.7

 211

SIMILAR BEHAVIOR?

Algeria

Nigeria
10

½ slope
Bilinear
0 .1
1 E- 4 1E- 3 0.0 1 0 .1 1 10

L o g -L o g p lo t: p -p @ d t= 0 a nd d e riv a tiv e [p s i] v s d t [hr]

Pseudo-Steady state
Late-time unit-slope behavior caused by pseudosteady state occurs only
during drawdown.

Steady state
Steady state implies that pressure in the well drainage volume does not
vary in time at any point and that the pressure gradient between any two
points in the reservoir is constant.
This condition may occur for wells in an injection-production scheme.
In buildup and falloff tests, a steeply falling derivative may represent
either pseudosteady or steady state.
 213

10 – ELLIPTICAL
FLOW REGIME

214

Elliptical Flow Regime

For a HW with infinite-conductivity, the elliptical flow regime may
occur in lieu or immediately after the early linear flow.

Figures 7.1 and 7.2 present a schematic description of this flow

regime

F ig u r e a . E a r ly tim e r a d ia l F ig u r e b . E a r ly tim e l in e a r
flo w in th e y z p la n e . f lo w i n th e y z p la n e .

r1 r1
r2

F ig u r e c . E llip tic a l (b ir a d ia l)
flo w in th e x y p la n e .

F i g u r e d . L a te tim e p s e u d o r a d ia l F ig u r e e . L a te t im e l in e a r fl o w
flo w in th e x y p la n e . in th e x y p la n e .

Figure 7.1 Figure 7.2

 215
Its main characteristic on a log-log plot of the pressure derivative
versus time is a straight line of slope = 0.36.

1000

The elliptical
flow occurs when 100
the horizontal well

∆P, (t *∆P'), psi

penetrates
between 10 and
70% of the 10 Slope = 0.36
formation width:
0.1<L/hx<0.7
(t*∆P') Ell1hr = 3.5 psi

1
0.001 0.01 0.1 1 10 100 1000 10000 100000

Time, hr

Figure 7.3

216
The equation of the straight lines corresponding to the elliptical
flow regime is:
0.72
h h  tD
0.36
t D * PD ' =  x z 
 L rw  108519 (7.1)
 

Substituting for the dimensional terms and rearranging Eq. 7.1

becomes

0.64
 µ  0.72 0.36
 qB   hx hz   t  (7.2)
t * ∆P = 
'
    
 φc r 2 
 14930.4hz  k x k y 
  Lrw   t w 

Taking the logarithm of both sides of Eq. 7.2 gives:

log(t * ∆P ' ) = 0.36 log t + log(mELL ) (7.3)

Where mELL is:

 217
0.64
 µ  0.72
 qB   hx hz 
mELL =  
2 0.36 
  (7.4)
 14930.4hz (φct rw )  k x k y 
  Lrw 

At time t – 1hr, Eq. 7.2 becomes:

0.64
 µ  0.72
 qB   hx hz 
(t * ∆P ) ELL1hr = mELL
'
=  
2 0.36 
  (7.4)
 14930.4hz (φct rw )  k x k y 
  Lrw 

Where (t*ΔP)ELL1hr is the value of

pressure derivative read at t =1hr 1000

(extrapolated if necessary as shown ∆Ppss=295 psi

in Figure) 100
(t*∆P')pss=103 psi
∆Pr=181 psi

(t*∆P')r=29 psi

∆ P, (t *∆P'), psi
Eq. 7.4 can be solved for the horizontal
permeability kh=sqrt(kxky) or hx. 10
Slope = 0.36
tpss=10259 hr

(t*∆P') er = 4.8 psi

Let: (t*∆P')i=29 psi
Elliptical flow
tRPi=2770 hr
ti = 0.0033 hr

tr=740 hr

 qB( µ 0.64 ) 
(t*∆P') Ell1hr = 3.5 psi
1

E =  
0.001 0.01 0.1 1 10 100 1000 10000 100000

2 0.36  (7.5) Time, hr

 14930. 4 hz (φc r
t w ) 

218
If hx is known and the radial flow lines are 1000

not well defined, the elliptical flow line can ∆Ppss=295 psi

also be used to calculate average horizontal ∆Pr=181 psi

(t*∆P')pss=103 psi

permeability kh= sqrt(kxky) from: 100

(t*∆P')r=29 psi
∆ P, (t *∆P'), psi

1
 E  h x hz 
0.72
 0.64 10
Slope = 0.36
tpss=10259 hr

kh = kx k y = 
(t*∆P') er = 4.8 psi
   (7.6) (t*∆P')i=29 psi
Elliptical flow
tRPi=2770 hr

 (t × ∆P' ) ELL1  Lrw  

ti = 0.0033 hr

tr=740 hr

(t*∆P') Ell1hr = 3.5 psi

1
0.001 0.01 0.1 1 10 100 1000 10000 100000
0.72 3.125
1   h x hz   Time, hr
E
ky =    
k x  (t × ∆P' ) ELL1  Lrw  


If kh is known (from late radial), the y=∞ Y

formation lateral thickness in the x- X

direction is: Z=
hz
Z
y=0

Lw
1
 (t × ∆P' ) ELL1
( ) 
Lza z=hs
Lr 0.64 0.72 Lzb
rw
Z=
0

hx = w  kxk y 
X=Lxd X=Lx1

hz  E  X=0 X=hx

hx
(7.7)
 100
2
µ  4.064qB 
ky =  
φc t  hx hz ( t × ∆P' ) LL1 
late linear
slope = 0.5

t*ΔP'
2 10 late

1  70.6qµB  t*ΔPLR radial

ΔP
kz =  
× ∆
early
ky  L ( t P ' ) ER  t*ΔPER radial

t*ΔPELL1hr elliptical flow

slope = 0.36
2
1  70.6qµB
1
 0.001 0.010 0.100 1.000 10.000 100.000 1000.000

k x =   time, hr

k y  h z (t × ∆P ' ) LR 
OR from the ELL flow regime:

 qB( µ 0.64 ) 
3.125
1   hxhz  
0.72

kx = 
E
   
E = 
2 0.36 
k y  ( t × ∆P' ) ELL1  Lrw    14930.4 h (φc r ) 
  z t w

220

Simulated case Exercise 3

1000 100

100
late linear
slope = 0.5
t*ΔP'
∆P, (t *∆P'), psi

10 late
t*ΔPLR radial
ΔP

10 early
Slope = 0.36
t*ΔPER radial

t*ΔPELL1hr elliptical flow

(t*∆P') Ell1hr = 3.5 psi slope = 0.36
1
1
0.001 0.010 0.100 1.000 10.000 100.000 1000.000
0.001 0.01 0.1 1 10 100 1000 10000 100000
time, hr
Time, hr

Figure 7.4 - Example of how to read (t*ΔP’)ELL1hr

on the pressure derivative in oilfield units.
 221

SUMMARY

CASE 1 (Ideal) - OBSERVED FLOW REGIMES: ER, EL, LR & LL

1 - FROM EL & LL
2
 4.064qB  µ
k y =  

 z
h L (t × ∆P ' ) EL1  φct

2
OR µ  4.064qB 
ky =  
φc t  hx hz ( t × ∆P' ) LL1 
2 - FROM ER & LR
2
1  70.6qµB 
kz =    qB( µ 0.64 ) 
ky L ( t × ∆ P ' ) E =  
 ER 
 14930.4 h z (φc r
t w
2 0.36 
) 
2 3.125
1  70.6qµB    hxhz  
0.72
1 E
k x =   kx =    
 ( t × ∆P' ) ELL1  Lrw  
k y  h z (t × ∆P ' ) LR 
OR ky
 CASE 2 - OBSERVED FLOW REGIMES: ER, BL, LR & LL
1 - FROM BL & LL
4/3
 0.25  31.128qµB 
ky =   1 / 2 
1 / 4 
 (t × ∆P ' ) h L
BL1  z w (φµc t ) 
2
OR µ  4.064qB 
ky =  
φc t  hx hz ( t × ∆P' ) LL1 
2 - FROM ER & LR
2
1  70.6qµB 
kz =  
ky  L ( t × ∆ P ' ) ER 

2
1  70.6qµB 
k x =  
k y  h z (t × ∆P ' ) LR 

CASE 3 - OBSERVED FLOW REGIMES: ER, ELL, LR & LL

1 - FROM LL

2
µ  4.064qB 
ky =  
φc t  hx hz ( t × ∆P ' ) LL1 

2 - FROM ER & LR
2
1  70.6qµB 
kz =  
ky  L ( t × ∆ P ' ) ER 

3.125
2   hxhz  
0.72
1  70.6qµB  OR kx =
1

E
  
kx=   ky  ( t × ∆P' ) ELL1  Lrw  
k y  h z (t × ∆P ' ) LR 
 qB( µ 0.64 ) 
E =  
2 0.36 
 14930.4 h z (φc r
t w ) 
CASE 4 - OBSERVED FLOW REGIMES: ER, (EL, BL or ELL) & LR (No LL)

1 - FROM EL, BL or ELL

2
 4.064qB  µ
k y =  

 hz L (t × ∆P ' ) EL1  φct

4/3
 0.25  31.128qµB 
OR k y =   1 / 2 
1 / 4 
 (t × ∆P ' )BL1  hz Lw (φµct ) 
3.125
  hxhz    qB( µ 0.64 ) 
0.72
1 E  
OR kx =     E = 2 0.36 
ky  ( t × ∆P' ) ELL1  Lrw    14930.4hz (φct rw ) 

2 - FROM ER & LR
2 2
1  70.6qµB  1  70.6qµB 
kz =   and k x =  
ky  L ( t × ∆P' )ER  k y  h z (t × ∆P ' ) LR 

CASE 5 - OBSERVED FLOW REGIMES: (EL, BL or ELL) & LR (No LL)

WBS mask ER
1 - FROM EL, BL or ELL
2
 4.064qB  µ
k y =  

 h z L (t × ∆P ' ) EL1  φct

4/3
 0.25  31.128qµB 
OR k y =   1 / 2 
1 / 4 
 (t × ∆P ' )BL1  hz Lw (φµct ) 
 qB( µ 0.64 ) 
E =  
2 0.36 
2 - FROM LR  14930. 4 h z (φc r
t w ) 
2
1  70.6qµB  3.125
  hxhz  
0.72

k x =   OR kx =
1

E
  
k y  h z (t × ∆P ' ) LR  ky  ( t × ∆P' ) ELL1  Lrw  

3 – kV=kz only from Pilot – hole Test (or core) k sp3

kV =
k H2
 227
CASE 6 - OBSERVED FLOW REGIMES: ER and (EL or BL)
(No LR, ELL & LL)
1 - FROM EL or BL
2
 4.064qB  µ
k y =  

 h z L (t × ∆P ' ) EL1  φct

4/3
 0.25  31.128qµB 
OR k y =   1 / 2 
1 / 4 
 (t × ∆P ' )BL1  hz Lw (φµct ) 
2 - FROM ER
2
1  70.6qµB 
kz =  
ky  L ( t × ∆ P ' ) ER 

k 2H
3 – kH FROM PILOT-HOLE kH = k x ky ⇒ kx =
ky

CASE 7 - OBSERVED FLOW REGIMES: ER, LR and ELL (No EL, BL & LL)

1 - FROM PILOT HOLE: kH and kV=kz k sp3

kV =
2 - FROM ER k H2
2
1  70.6qµB 
ky =  
kz  L ( t × ∆ P ' ) ER 

3 - FROM LR or ELL
2
1  70.6qµB 
k x =  
k y  h z (t × ∆P ' ) LR 
3.125
1   hxhz    qB( µ 0.64 ) 
0.72
E
OR kx =     E =  
2 0.36 
k y  ( t × ∆P' ) ELL1  Lrw    14930.4 h (φc r ) 
  z t w

k 2H
OR kH = k x ky ⇒ kx =
ky
CASE 8 - OBSERVED FLOW REGIMES: ER and ELL (No LR, EL, BL & LL)

1 - FROM PILOT HOLE: kH and kV=kz

k sp3
kV =
k H2
2 - FROM ER

2
1  70.6qµB 
ky =  
kz  L ( t × ∆ P ' ) ER 

3 - FROM ELL

qB( µ 0.64 )
3.125
1   hxhz  
0.72
 
kx = 
E
   
E = 
2 0.36 
k y  ( t × ∆P' ) ELL1  Lrw    14930. 4 h (φc r ) 
  z t w

k 2H
OR
kH = k x ky ⇒ kx =
ky

230

EXERCISE 5
 231

EXAMPLE

A pressure drawdown test was run in horizontal well

located in the middle of the (vertical) formation thickness.

Other known reservoir and well data are:

q = 1200 BPD Ø = 0.17 µ = 0.77 cp
ct = 14x10 psi
-6 -1 B = 1.2 rb/STB hz = 92 ft
rw = 0.35 ft L = 2100 ft
hx = 9800 ft (from geologic map)
KZ = 20 md (from Pilot hole test)

Interpret this test.

232
 233
The diagnostic plot confirms the presence of at least three flow
regimes: ERF, Elliptical Flow and Late LRF. The LLF is not well
defined.

234
TDS Technique:

Step 1 – From the early radial:

(t*dP’)ER = 4.9.
The average permeability in the
vertical plane is:

70.6qµB
kV = k y k z =
L (t × ∆P' ) ER
70.6 × 1200 × 0.77 × 1.2
= = 7. 6
2100 × 4.9

Step 2 – From the early radial: at tER = 1.0, dPER = 90

1  ∆ p ER  k y k z t ER 
 

Sm =  - ln + 7.43
2  (t × ∆P ' ) ER  φµ ct r 2w  
  
 90  7.6 × 1.0  
= 0. 5  − ln −6 2 
+ 7.43 = 4.2
 4.9  0.17 × 0.77 × 14 × 10 × 0.35  
 235
Step 3 – From the late radial:
(t*dP’)LR = 24.
The average permeability in the
horizontal plane is:

70.6qµB
kH = k x k y =
h z (t × ∆P ' ) LR
70.6 × 1200 × 0.77 × 1.2
= = 35.45 md
92 × 24

Step 4 – Permeability in the y and x-directions

ky =
( kz ky
=
)
2
7.62
= 2.9 md
kz 20

kx =
( kx k y
=
)
2
35.452
= 433.4 md
ky 2.9

236
Step 5 – Elliptical Flow
The formation lateral thickness is obtained from the elliptical flow
line and Eq. 2.5.6, where (t*dP’)ELL1 = 3.7:

qBµ 0.64
E=
14930.4hz (φct rw ) 0.36
2

(2100)(1.2)(0.77) 0.64
= 0.21645
[
(14930.4)(92) (0.17)(14 ×10 −6 )(0.352 ) ]
0.36

hx = formation thickness in the x-

direction, ft.

1
Lr
hx = w
 (t × ∆P' ) ELL1 
 (
 kxk y )
0.64  0.72

hz  E  
1
2100 × 0.35  3.7  0.64 
0.72
hx = 
 0.21645 35.45  = 9822 ft
92   
This test confirms the formation thickness in the x-direction obtained
from geologic map.
 237

Step 6 – Calculate Sm+Sz

From the graph: tLR=800, dPLR=190 and t*dP’LR=24

L kz  ∆ p LR  k t  
S z + Sm =  - ln  x LR 2  + 4.659 
2 hz kx  (t × ∆P ' ) LR  φµ ct L  
2100 20 190  448.8 × 800  
=  − ln −6 2 
+ 4.659 = 4.5
2 × 92 448.8  24  0.17 × 0.77 × 14 × 10 × 2100  
S Z + S m = 0.19 ⇒ S z = 4.5 − S m = 4.5 − 4.19 = 0.31

The test was stopped right after

the late radial.

The partial-penetration skin Sp (or

Sx) cannot be calculated since the
late linear is not well defined (not
enough points to draw with
confidence the half-slope line).

238

11 – HW IN FINITE
SYSTEMS & RESERVES
 239
Finite System:
FAULT BLOCK

Rectangular Triangular
Finite System Finite System

240
Infinite-Conductivity HW inside a Closed System
 241
DRAINAGE AREA
The average pressure for a reservoir without water influx is
defined as the pressure the reservoir would reach if all wells were
shut-in for an infinite time.
In a strong water drive reservoir Pavg ≈ Pi.

The average reservoir pressure is used:

1) To characterize a reservoir:

a) If the pressure drop (Pavg-Pwf) is small per unit of

production, a clue is given to the unexpected existence of a
water drive or to drainage from a large reservoir volume.

b) If the pressure drop is large for a given amount of

production, this may indicate drainage from a small lens or
fault block.

2) To compute oil in place.

3) To predict future reservoir behavior.

242
The pseudo-steady state solution in well test analysis assumes

1. closed reservoir boundaries; and

2. the material balance, based on the cumulative production,

provides an initial approach to predict well performance under
depletion conditions at late time of the transient test.

MATERIAL BALANCE FORMULATION

For slightly compressible fluids and a depletion-type oil reservoir (i.e.

oil production by fluid expansion), the MBE is (A is in ft2):

V p ct ( Pi − P ) Ahφct ( Pi − P )
N pB = = (8.1)
5.615 5.615

If the well is producing at a constant flow rate q, STB/D, for a given

time period, t days, then the cumulative oil produced, Np is:

N p B = qt (8.2)
 243
Combining Eq. 8.1 and 8.2 and solving for the average reservoir
pressure yields:

 5.615qB 
P = Pi −  t (8.3)
 Ah φct 

Converting time to hours and differentiating Eq. 8.3 with respect to

time t, assuming constant flow rate, gives:

∂∆P  5.615qB  qB
∆P' = =   = = m* (8.4)
∂t  24 Ahφct  4.27 Ahφct
Notice, the right side of equation 8.4 implies that the variation of
pressure with respect to time is constant, which corresponds to the
definition of pseudosteady state.

Multiplying both sides of equation 6.4 by t yields:

 qB 
t × ∆P' =  t (8.5)
 4.27 Ahφct 

244
Eq. 8.5 can be solved for A (ft2) :

 qB  t PSS
A =  
 4.27φct h  (t × ∆P')PSS
(8.6)

Where tpss and (t*ΔP’)pss are obtained at any on the pseudosteady

state straight (of slope = 1) on the loglog plot of the pressure
derivative t*ΔP’ versus test time, as shown in Figure 6.1.

If sqrt(kxky) is known, A (ft2) can

also be calculated from the
following equation, where tLRPi is
the time of intersection of the PSS
line and the late-radial line:

 kxk y 
A= t (8.7)
 301.77φµct  LRPi
 
 245
PORE VOLUME & OIL-IN-PLACE
The pore volume VP (ft3) is calculated from the following
equation:
(8.8)

 qB  t pss Where (tx∆ ∆Pw’)pss is

VP = Ahφ =  
 (t ×∆P ' ) pss obtained from the PSS
 4.27 ct  straight line at any time
tpss.
The oil in place N (STB)
contained in this drainage area
or finite reservoir is then
estimated from:

(1 − S wi )V p
N= (8.9)
5.615Boi

246

AVERAGE RESERVOIR PRESSURE

(a) The average reservoir pressure can be calculated from
the following equation:

 (tx∆Pw ' ) pss   A 

P = Pi − (tx∆Pw ' ) LR    ln 2 − 2.645 + 2S  (8.10)
 (∆Pw ) pss − (tx∆Pw ' ) pss 
  rw 

Where:

(∆Pw)pss , (tx∆
∆Pw’)pss at tpss,
∆Pw’)R are read from
and (tx∆
the graph.
 247

ESTIMATING RESERVES
Quantitatively, Reserves are estimated from:

Re serves = EUR − N P (8.11)

Np = the cumulative oil produced , STB

EUR = Estimated Ultimate Recovery:

EUR = N × ER

N = Initial Oil in Place, STB

ER = Recovery efficiency (strongly affected by
reservoir geology and drive mechanism)

248
If no oil is produced, i.e. Np = 0, Eq. 8.11
becomes:

Re serves = EUR = N × ER (8.12)

Gas Reserves are also estimated from:

Re serves = G × ER − GP (8.13)

G = Initial gas in place (IGIP), SCF

If no gas is produced, i.e. Gp = 0, Eq. 8.13

becomes:

Re serves = EUR = G × ER (8.14)

 249
φ
RECOVERY EFFICIENCY OF VOLUMETRIC
UNDERSTATURATED OIL RESERVOIR
The recovery efficiency (by expansion Fault blocks
drive, Pi>Pb) from a volumetric black
oil reservoir (undersaturated Reservoir)
from Pi to Pb is calculated from:

ER = ce ( Pi − Pb ) Boi / Bob (8.15)

ce = ( S o co + S wcw + c f ) / S o (8.16)

undersaturated
So = oil saturation
Sw = water saturation
Co = oil compressibility, 1/psi
Cw = water compressibility, 1/psi
Cf = formation compressibility, 1/psi
saturated
Bob = oil FVF at Pb
Boi = Oil FVF at Pi 249

250
SHAPE FACTOR
Figure 8.2 presents the influence of the size of the drainage area on
behavior of the pressure derivative curve.

As expected the smaller the drainage area the earlier the late-time PSS
straight line of slope = 1 occurs, and the less likely the late radial will
be observed.

The equation of this PSS straight 10000

line is: Drainage area

(acres)
a: 83
b: 212
1000 c: 1188
d: 2899

t DA × PwD = 2π t DA
e: 5743

/ (8.17)
f: 8523
g: 12437
∆ P, (t *∆ P'), psi

e
100
f
a b c d g

10

The dimensionless pressure

during pseudo-steady state flow
1
is a linear function of 0.001 0.01 0.1 1 10 100 1000 10000 100000

Time, hr
dimensionless time.
Figure 8.2
 251
The equation corresponding to this regime in a hydraulically fractured
vertical well is:

1  xe  2.2458 
2

PwD = 2π t DA + ln    (8.18)
2  x f 
  C 
 A

Modeling a HW as a hydraulic fracture, Eq. 8.18 becomes:

1  hS   2.2458 
2

PwD = 2π t DA + ln     (8.19)
2  L   C A 

The ratio of equations 8.19 and 8.17 gives:

PwD 1  hx  2  2.2458 
= 1 + ln     (8.20)
'
(t DA * PwD ) 4πt DA  L   C A 

252
Substituting for the corresponding dimensionless terms
and solving for the shape factor yields:

h 
2   0.0033137 k x k y t pss  (∆Pw ) pss  
C A = 2.2458  x  exp  1-   (8.21)
φµ ct A  
 (t * ∆Pw ) pss
/
L    

Where: tpss is any convenient time during the pseudo-steady state

line portion of the curves; (∆Pw)pss and (t*∆∆P/w)pss are values of
(∆Pw) and (t*∆ ∆P w) corresponding to tpss, respectively.; the drainage
/

area A is in ft2):
 253
AVERAGE RESERVOIR
Using a similar approach, the average reservoir pressure in the
horizontal well system is:

qµB  0.116867 k x k y t pss  (∆Pw ) pss   hx  2  2.2458  

P = Pi −    − 35.3 ln      (8.22)
φµct A  (t * ∆P / ) 
hz k x k y   w pss   L   C A  

Eq. 8.22 can also be related to the late radial:

(t × ∆P' ) LR  0.116867 k x k y t pss  (∆Pw ) pss   2

  
P = Pi −    − 35.3 ln  hx   2.2458   (8.23)
φµct A  (t * ∆P ) 
/  
70.6   w pss   L   C A  

Or to the early radial by substituting

the following equation:

 L 
(t × ∆P' )LR =   (t × ∆P ' ) ER (8.24)
h k /k 
 z x z 

254

EXERCISE 6
 255

EXAMPLE

A pressure buildup test was run in horizontal well located

in the middle of the formation (vertical) thickness. The well
produced for 4 years.

Other known reservoir and well data are:

q = 5000 BPD Ø = 0.24 µ = 1.2 cp
ct = 5x10 psi
-5 -1 B = 1.2 rb/STB hz = 84 ft
rw = 0.35 ft L = 2626 ft
Pi = 7000 psia Pwf(Δt=0) = 6459.58

Using TDS technique, calculate the directional

permeability values, wellbore storage, skin factors, drainage
area and average reservoir pressure.

256
Fig. 8.6 - pressure buildup test
 257
SOLUTION
The diagnostic plot confirms the presence of four flow regimes.

As shown in the plot of the

pressure and pressure
derivative, the curves exhibit
five well defined flow periods:
1. Wellbore storage flow
(slope=1)
2. early radial flow
3. Early linear (slope=0.5)
4. late radial flow and
5. pseudosteady state flow
(slope=1)

The late linear flow period is

not available.
FIG. 8.7

258
TDS Technique:

Step 1 - Wellbore Storage

From the log-log plot of ΔP vs. t,

presented in Figure 6.6, one of the
points on the early unit slope line is
identified at the coordinates:

ti= 0.0033 hr
ΔPi = 29 psi

Substituting into the WBS equation

gives:

 qB  t  (5000)(1.12)  0.0033
C =  =  = 0.026 STB / psi
 24  ∆P  24  29
 259
Step 2 – Early Radial

The early-radial flow (horizontal)

line yields:

(Δt*ΔP’)ER = 2.95 psia

Therefore, the geometric average

(horizontal) permeability is:

70.6qµB (70.6)(5000)(1.2)(1.12)
kz ky = = = 61.5 md
L(t × ∆P' ) ER (2626)(2.95)

260
Step 3 – Late Radial

The late-radial flow (horizontal

line) yields:

(Δt*ΔP’)LR = 28.7 psia

Therefore, the geometric average

(horizontal) permeability is:

70.6qµB
kx ky =
h z (t × ∆P' ) LR
(70.6)(5000)(1.2)(1.12)
= = 196 md
(84)(28.7)
 261
Step 4 - Elliptical

at t = 1 hr, (t*dP’)EL1=1.84

2
 4.064qB  µ
k y =  

 h z L (t × ∆P' ) EL1  φct
 4.064 × 5000 ×1.12 
2
1.2
=  −5
 84 × 2626 ×1.84  0.24 × 5 × 10
= 314 md

Since: → kx =
( kx ky ) 2

=
196
= 122 md
k x k y = 196 md
ky 314

And since: → kz =
( kz ky ) 2

=
61.5
= 12 md
k z k y = 61.5 md
ky 314

This confirms the value of kz obtained during the pilot-hole test.

262
Step 5 - VERIFICATION

The ratio of the early radial and late

radial pressure derivative lines results
in the following expression:

(t × ∆P ' ) ER h z kx
=
(t × ∆P ' ) LR L kz 12

From the graph:

(t*dP’)LR = 26.7
(t*dP’)ER = 2.95
314
k z h z (t × ∆P' )LR 84 28.7
= = = 0.312
k x L (t × ∆P' )ER 2626 2.95 122

Since: k = 12 md → k = kz 12
= = 122 md
z x
( kz / kx )2
0.312 k x k y = 196 md = k H
This confirms the value of kx obtained from the late radial.
 263
Step 6 – Pseudosteady State

Using the time of intersection, the

drainage area is (Eq. 8.6a):

 kxk y 
A= t
 301.77φµct  LRPi
 
(196)(2770)
=
(301.77)(0.24)(1.2)(5 x10−5 )
= 1.249 x108 ft 2 = 2868 acres
  t pss
A =  
qB
 4.27φ ct h  (t × ∆P ' ) pss
In the absence of the late radial, the 
drainage area can also be calculated from
Eq. 8.6, where tpss=5084.9 hr and the 5000×1.12×5084.9
corresponding derivative value (t*dP’)pss =
= 52.43 psia (this point is selected from
4.27×0.24×5×10 −5 ×84×52.43
the graph but the values are actually read 8 2
from the table): = 1.28 × 10 ft = 2896 acres

264
Substituting for the corresponding values and solving for the
shape factor yields (Eq. 8.21):

h 
2   0.0033137 k x k y t pss  (∆Pw ) pss  
C A = 2.2458  x  exp  1-  
φµ ct A  
 (t * ∆Pw ) pss
/
L    

 12538    (0.0033137)(196 )(10259)  295  

2

C A = 2.2458   exp  −5  1-   = 0.052

 2626    (0.24)(1.2)(5 x10 )(124173861)  103  

Where tpss, dPpss and

t*dP’pss are obtained
from the graph or table.
 265
Substituting for the corresponding values and solving for
the average pressure yields (Eq. 8.22):

qµB  0.116867 k x k y t pss  (∆Pw ) pss   2

  
P = Pi −    − 35.3 ln  hx   2.2458  
φµct A  (t * ∆P / )   
hz k x k y   w pss   L   C A  

 (0.116867)(196)(10259)(295) 
 (0.24)(1.2)(5 x10 −5 )(124173861)(103) 
(5000)(1.2)(1.12)  
P = 7000 −   = 6958 psia
(84)(194.8)   12538  2  2.2458 
− 35.3 ln     
  2626   0 .052   
   

Where tpss, dPpss and

t*dP’pss are obtained
from the graph or table.

266

EXERCISE 7*

*Let students solve the problem.

 267
EXAMPLE - Well LLF-4

A pressure drawdown in a horizontal well yielded the data shown in

Table 5.2 (simulated). The following data are known:

q = 800 STB/D µ = 1.0 cp

B = 1.25 RB/STB rw = 0.25 ft
Ø = 0.2 ct = 15 x 10-6 psi-l L

h = 150 ft Lw= 2500 ft

hY=6968 ft hX=2000 ft ye

Dz = 85 ft dz = 65 ft z y h
ye /2 hy
Kx=50 md, ky=100 md, kz=10 md. X
zw

x e/2
xe
hx
The horizontal well is centered
in (finite) drainage volume.
Interpret this test using the TDS
Technique (as in Ex. 5 with HW is // dz
to X-axis).
Dz

268
Table 5.2a t (hrs) P, psi dP, psi t*dP'
0 3000
0.001 2994.053 5.947 3.063
0.002 2991.735 8.265 3.691
0.003 2990.026 9.974 3.837
0.004 2988.755 11.245 3.382
0.006 2987.818 12.182 2.994
0.007 2987.091 12.909 2.762
0.009 2986.487 13.513 2.64
0.011 2985.958 14.042 2.567
0.013 2985.487 14.513 2.588
0.016 2985.046 14.954 2.549
0.019 2984.635 15.365 2.516
0.022 2984.249 15.751 2.481
0.025 2983.881 16.119 2.447
0.029 2983.53 16.47 2.416
0.033 2983.192 16.808 2.401
0.038 2982.865 17.135 2.389
0.044 2982.548 17.452 2.375
0.05 2982.239 17.761 2.364
0.057 2981.937 18.063 2.345
0.064 2981.64 18.36 2.314
0.073 2981.348 18.652 2.313
0.083 2981.062 18.938 2.314
0.094 2980.779 19.221 2.315
0.106 2980.5 19.5 2.307
0.119 2980.223 19.777 2.297
0.134 2979.949 20.051 2.27
Different from Ex. 1 0.151 2979.677 20.323 2.278
 t (hrs) P, psi dP, psi t*dP'
0.17 2979.407 20.593 2.28
Table 5.2b 0.192 2979.14 20.86 2.282
0.216 2978.874 21.126 2.278
0.242 2978.609 21.391 2.244
0.272 2978.345 21.655 2.251
0.306 2978.082 21.918 2.265
0.343 2977.821 22.179 2.27
0.386 2977.561 22.439 2.268
0.433 2977.301 22.699 2.223
0.485 2977.042 22.958 2.244
0.545 2976.783 23.217 2.251
0.611 2976.525 23.475 2.257
0.685 2976.268 23.732 2.266
0.768 2976.011 23.989 2.264
0.861 2975.752 24.248 2.24
0.965 2975.494 24.506 2.269
1.082 2975.235 24.765 2.283
1.213 2974.974 25.026 2.314
1.359 2974.71 25.29 2.341
1.523 2974.441 25.559 2.352
1.707 2974.167 25.833 2.432
1.913 2973.887 26.113 2.512
2.143 2973.597 26.403 2.599
2.401 2973.298 26.702 2.697
2.69 2972.986 27.014 2.776
3.014 2972.661 27.339 2.914
3.376 2972.322 27.678 3.06
3.782 2971.967 28.033 3.203

270
t (hrs) P, psi dP, psi t*dP'

Table 5.2c 4.237

4.746
2971.596
2971.208
28.404
28.792
3.359
3.507
5.317 2970.803 29.197 3.636
5.956 2970.381 29.619 3.794
6.671 2969.942 30.058 3.949
7.473 2969.486 30.514 4.109
8.37 2969.014 30.986 4.251
9.376 2968.525 31.475 4.362
10.502 2968.02 31.98 4.517
11.763 2967.501 32.499 4.619
13.175 2966.972 33.028 4.712
14.757 2966.433 33.567 4.809
16.529 2965.882 34.118 4.906
18.514 2965.32 34.68 5.027
20.736 2964.744 35.256 5.133
23.225 2964.158 35.842 5.232
26.013 2963.559 36.441 5.337
29.135 2962.947 37.053 5.454
32.632 2962.323 37.677 5.57
36.549 2961.685 38.315 5.698
40.936 2961.032 38.968 5.827
45.849 2960.362 39.638 5.987
51.351 2959.674 40.326 6.169
57.514 2958.963 41.037 6.379
64.417 2958.228 41.772 6.607
72.148 2957.464 42.536 6.878
80.807 2956.667 43.333 7.189
90.504 2955.833 44.167 7.53
101.366 2954.957 45.043 7.922
113.53 2954.036 45.964 8.331
127.155 2953.064 46.936 8.807
142.414 2952.037 47.963 9.311
159.505 2950.951 49.049 9.86
178.647 2949.802 50.198 10.439
200.085 2948.583 51.417 11.074
224.096 2947.289 52.711 11.772
240.481 2946.442 53.558 12.225
 271
SOLUTION
Notes:
1. TXTBK uses Babu-Odeh model which has HW parallel to y-
axis. In our model (used in previous exercises) HW is parallel
the x-axis.

2. TXTBK treated the transition period Between ERF and LRF

as ELF. Actually the slope is NOT 0.5 but 0.36, therefore the
transition period corresponds to elliptical flow

272
Workshop/homework
 273

12 – MULTIPHASE
FLOW IN
Horizontal Wells

27
ANISOTROPIC RESERVOIR 4

ALL FLOW REGIMES ARE OBSERVED

A log-log plot of dimensionless pressure and pressure derivative
versus time is shown in Figure 3.3.

Identified on this figure are the characteristic points and lines

available for evaluating the well test.

100

late
linear
t*ΔP'

10 late
t*ΔPLR radial
ΔP

early
t*ΔPER radial early
linear

t*ΔPEL1hr t*ΔPLL1hr
1
0.001 0.010 0.100 1.000 10.000 100.000 1000.000
time, hr FIG. 3.3
 275
Using the same procedure discussed previously, we can calculate
Oil, Gas and Water mobilities in the x, y and z directions, and skin
values.

1 – OIL MOBILITY IN THE X, Y AND Z DIRECTIONS

2
   4.064qo Bo  1
λoy =  k o  = 
100

 µo  x  h z L (t × ∆P ' )EL1  φct
late
linear

t*ΔP'
2
  1  70.6 qo Bo  10 late

λox =  k o  =   t*ΔPLR radial

ΔP
 µo  y λoy  hz ( t × ∆ P ' ) LR 
early
t*ΔPER radial early
linear

t*ΔPEL1hr t*ΔPLL1hr
1
2
  1  70.6 qo Bo 
0.001 0.010 0.100 1.000 10.000 100.000 1000.000

λoz =  k o  =  
time, hr

 µo  z λoy  L (t × ∆P ' )ER 

276
2 – WATER MOBILITY IN THE X, Y AND Z DIRECTIONS

2 100

   4.064qw Bw  1
λwy =  k w  =  
µ
 w  y  hz L ( t × ∆P ' ) EL1  φct
late
linear
t*ΔP'

10 late
t*ΔPLR radial
ΔP

2
k  1  70.6 qw Bw  early

λwx =  w  =   t*ΔPER radial early

 µw  y λwy
linear

 hz ( t × ∆ P ' ) LR  t*ΔPEL1hr t*ΔPLL1hr

1
0.001 0.010 0.100 1.000 10.000 100.000 1000.000
time, hr

2
  1  70.6 qw Bw 
λwz =  k w  =  
 µw  z λwy  L ( t × ∆ P ' ) ER 
 277
3 – GAS MOBILITY IN THE X, Y AND Z DIRECTIONS

Qg = qg − 0.001qo Rs
100

2
 k   4.064Qg Bg  1
λgy =  g  =   late
µ
 y  h L ( t × ∆P ' ) EL1  φct
linear

t*ΔP'
g z
10 late
t*ΔPLR radial

ΔP
early
2 radial
k 
early
 70.6Qg Bg 
t*ΔPER
1 linear

λgx =  g  =   t*ΔPEL1hr

 µ g  x λgy
t*ΔPLL1hr

 z
h ( t × ∆ P ' ) LR  1
0.001 0.010 0.100 1.000 10.000 100.000 1000.000
time, hr

2
k  1  70.6Qg Bg 
λgz =  g  =  
µ
 g z λ gy  L ( t × ∆ P ' ) ER 

278
1  ∆ p ER  λ ty λ tz t ER 
 

4 – SKIN Sm =  - ln + 7.43
2  (t × ∆P' ) ER  φ ct r 2w  
  

0.029 λ tz t EL  ∆PEL 
S z + Sm =  (t × ∆P' ) − 2 
hz φ ct  EL 
0.029 L λ tz t LL  ∆ p LL 
S x + Sm + S z =  (t × ∆P' ) − 2
hx hz φ ct  LL 
100

late
linear
t*ΔP'

10 late
t*ΔPLR radial
ΔP

early
t*ΔPER radial early
linear

t*ΔPEL1hr t*ΔPLL1hr
1
0.001 0.010 0.100 1.000 10.000 100.000 1000.000
time, hr
 279

13 – Practical Problems
With Horizontal Well
Testing

280
The measurements in horizontal wells are usually made
above the well bore with the pressure gauge still in the
vertical section, as shown in figure.

Often, the test string may be too rigid to pass through

the well-bore.

However, in most cases,

conventional hardware
can be used for
horizontal well tests.
 281
With longer horizontal well bores, wellbore storage is an
inherent problem for testing, even for buildup tests with
downhole shut-in.

It is not recommended to conduct buildup tests with short-

duration production periods because superposition is
inappropriate; therefore, Horner plots which depend on
superposition being applicable, is often inappropriate.

Another problem in conducting buildup tests is that significant

pressure gradients along the length of the well bore may
cause cross-flow within the wellbore during shut-in, so fluid
may flow from one region to another in the well bore.

These gradients can exist in new wells but are even more
probable in developed fields.

Heterogeneities cause uneven pressure gradients and uneven

fluxes (flow rate per unit well length) along the wellbore.

282
A number of factors may affect the transient response of a
horizontal well test:
1. Horizontal permeability (normal and parallel to well
trajectory).
2. Vertical permeability.
3. Drilling damage.
4. Completion damage.
5. Producing interval that effectively may be much less
than drilled length.
6. Variations in standoff along the length of the well.

In summary, many factors may affect interpretation for

horizontal wells in homogeneous reservoirs before the effects
from boundaries.

The problem is complex, so test results are frequently

inconclusive.
 283
Furthermore, wellbore storage inhibits determination of
properties associated with early-time transient behavior such
as vertical permeability and damage from drilling and
completion;

ERF regime and especially LRF regime may require several

hours, days, or months to appear in transient data.

We can take some practical steps to help ensure that we will

obtain interpretable test data.

First, it is helpful to run tests in the pilot hole before kicking

off to drill the horizontal borehole section.

From a test in the vertical section, it is possible to get usable

estimates of horizontal and vertical permeabilities with
modern wireline-test tools.

284

Furthermore, wellbore storage inhibits determination of

properties associated with early-time transient behavior
such as vertical permeability and damage from drilling and
completion.

Second, a good directional drilling survey can frequently

provide us with an adequate estimate of standoff.

A production-log flow survey conducted with coiled tubing

can determine which part of the wellbore is actually
producing and, therefore, help give us an estimate of
effective productive length.

A flow survey is generally essential for valid test

interpretation.
 285
For high-rate wells, continuous borehole pressure and flow-
rate measurements acquired during production can be used to
interpret the pressure-drawdown transient response.

If the downhole rates are not measured, the buildup test

should be conducted with downhole shut-in to minimize
wellbore-storage distortion of test data.

Estimating the productivity of a horizontal well accurately is

even more difficult than estimating the productivity of a
vertical well, mostly because of two fundamental problems:

1. the theoretical models that we have available have a

number of simplifying assumptions, and
2. the data required for even these simplified models are not
likely to be available.

286

Still, we must make estimates and decisions

based on those estimates.

The productivity models of HW depend on the

accurate interpretation of pressure-transient
test analysis in horizontal wells.

These models illustrate the minimum data

requirements for productivity estimates.

Actual productivity may differ significantly

from the predictions of these simple problems.
 287

Conventional models usually assume that the horizontal

well is parallel to one of the principal permeability
directions (preferably the minimum permeability
direction in the horizontal plane).

In many cases, this is not true; in fact, in many cases we

do not even know what the principal permeability
directions are.

When the principal permeability directions are known,

corrections to length are possible; if they are not known,
there is no way to correct the analysis.

Contemporary models show that the error in

permeability estimates approaches 50% when the
deviation angle exceeds 50°.

288

Unfortunately, the models also indicate that there is

nothing in a well's response that provides any indication
that the assumption that the well is parallel to a
principal permeability direction is incorrect.

The damaged region around a horizontal well probably

is nonuniform with distance (perhaps with the greatest
damage near the heel of the well and the least near the
toe because filtrate invasion is of much longer duration
near the heel).

If there is variable permeability along the path of the

well, the situation is even more complicated.
 289

It is a common practice to complete horizontal wells

selectively.

Also, in other cases, some segment of the well may not

be open to the flow of reservoir fluids because of
relatively low permeabilities or relatively large local skin
effects.

The absolute amount of the well that is open to flow and

the location of the open intervals affect the pressure
response in the well.

Ultimately, numerical models are usually required for

definitive test analysis in horizontal wells.

290

Analytical models, such as the simpler ones described in

this chapter and the more comprehensive ones
discussed by Ozkan, include many simplifying
assumptions.

Realistic assessments of the effects of damage and

permeability that vary along the wellbore usually make
numerical models necessary.

Beliveau presented a helpful discussion that points out

the danger of relying on simplified models to predict the
productivity of horizontal wells.

Rarely does actual productivity equal predicted

productivity.
 291

11 – FIELD CASES

292

FIELD CASE 1

Offshore, Brazil: Deep-Water HW

 293
A pressure buildup test in a horizontal well (deep-water
offshore Brazil) yielded the data shown in the following
tables.

The following well, fluid and rock characteristics are

known:
q= 1800 STB/D
µ= 28 cp
B = 1.14 RB/STB
rw = 0.354 ft
Ø = 0.3
ct = 3 x 10-5 psi-l
h= 98.425 ft
Lw= 2050.525 ft
Hx = 9842.52 ft
Pwf(Δt=0) = 4175 psia

The Horizontal well is approximately centered in the

drainage volume.
Interpret this test using the TDS Technique.

294
BUILDUP TEST

t Pws ∆P t*∆P' 0.078 4229.103 54.103 3.119

0 4175 0.081 4229.199 54.199 2.869
0.017 4225.618 50.618 0.189 0.083 4229.323 54.323 3.181
0.019 4225.671 50.671 0.949 0.086 4229.469 54.469 3.542
0.022 4225.868 50.868 1.376 0.089 4229.565 54.565 3.535
0.025 4226.020 51.020 1.559 0.092 4229.600 54.600 3.526
0.028 4226.209 51.209 2.133 0.094 4229.772 54.772 3.431
0.031 4226.405 51.405 2.539 0.097 4229.889 54.889 3.318
0.033 4226.704 51.704 2.502 0.100 4229.988 54.988 3.209
0.036 4226.915 51.915 2.727 0.103 4230.045 55.045 3.674
0.039 4227.062 52.062 2.460 0.106 4230.137 55.137 3.402
0.042 4227.265 52.265 2.549 0.108 4230.233 55.233 3.512
0.044 4227.406 52.406 2.913 0.111 4230.289 55.289 3.336
0.047 4227.595 52.595 3.199 0.114 4230.401 55.401 3.645
0.050 4227.784 52.784 3.009 0.117 4230.491 55.491 3.836
0.053 4228.006 53.006 3.045 0.119 4230.573 55.573 3.944
0.056 4228.082 53.082 2.868 0.122 4230.701 55.701 3.910
0.058 4228.252 53.252 2.691 0.128 4230.842 55.842 4.090
0.061 4228.340 53.340 3.028 0.133 4231.046 56.046 4.403
0.064 4228.489 53.489 3.184 0.139 4231.203 56.203 4.535
0.067 4228.652 53.652 3.088 0.144 4231.376 56.376 4.772
0.069 4228.810 53.810 3.098 0.150 4231.575 56.575 5.030
0.072 4228.908 53.908 3.021 0.156 4231.776 56.776 5.050
0.075 4229.024 54.024 2.991 0.161 4231.935 56.935 4.935
0.167 4232.154 57.154 5.276
 295
0.172 4232.288 57.288 4.957 0.344 4235.762 60.762 5.848
0.178 4232.433 57.433 4.966 0.353 4235.870 60.870 5.717
0.183 4232.541 57.541 5.046 0.361 4235.991 60.991 5.537
0.189 4232.779 57.779 4.454 0.369 4236.148 61.148 5.563
0.194 4232.876 57.876 4.970 0.381 4236.278 61.278 5.519
0.200 4233.020 58.020 4.569 0.392 4236.448 61.448 5.318
0.206 4233.152 58.152 5.262 0.403 4236.607 61.607 5.446
0.211 4233.219 58.219 4.473 0.414 4236.751 61.751 5.095
0.217 4233.423 58.423 4.919 0.425 4236.897 61.897 5.328
0.222 4233.460 58.460 4.671 0.436 4236.990 61.990 5.334
0.228 4233.682 58.682 4.777 0.447 4237.151 62.151 5.172
0.233 4233.774 58.774 4.389 0.458 4237.246 62.246 5.099
0.239 4233.839 58.839 5.100 0.469 4237.413 62.413 5.117
0.244 4233.994 58.994 4.561 0.481 4237.506 62.506 5.039
0.253 4234.136 59.136 4.477 0.492 4237.661 62.661 5.292
0.261 4234.213 59.213 4.221 0.506 4237.774 62.774 5.620
0.269 4234.459 59.459 5.018 0.519 4237.914 62.914 4.961
0.278 4234.550 59.550 4.777 0.533 4238.032 63.032 5.504
0.286 4234.704 59.704 5.207 0.547 4238.192 63.192 5.348
0.294 4234.758 59.758 5.609 0.561 4238.400 63.400 6.135
0.303 4235.025 60.025 5.280 0.575 4238.412 63.412 6.287
0.311 4235.131 60.131 5.718 0.589 4238.696 63.696 5.837
0.319 4235.334 60.334 5.570 0.603 4238.761 63.761 5.893
0.328 4235.458 60.458 6.198 0.619 4238.900 63.900 5.754
0.336 4235.612 60.612 5.527 0.636 4239.176 64.176 6.264

296

0.653 4239.255 64.255 5.359 1.200 4242.615 67.615 6.005

0.669 4239.364 64.364 5.761 1.228 4242.787 67.787 5.801
0.686 4239.482 64.482 5.467 1.258 4242.908 67.908 5.936
0.703 4239.638 64.638 5.489 1.289 4243.081 68.081 6.099
0.719 4239.738 64.738 4.802 1.319 4243.201 68.201 6.156
0.736 4239.877 64.877 5.208 1.350 4243.357 68.357 6.247
0.756 4240.049 65.049 5.617 1.383 4243.444 68.444 6.257
0.775 4240.101 65.101 5.557 1.417 4243.629 68.629 6.398
0.794 4240.256 65.256 5.688 1.450 4243.748 68.748 6.220
0.814 4240.348 65.348 5.750 1.486 4243.931 68.931 6.224
0.833 4240.514 65.514 5.517 1.522 4244.072 69.072 6.276
0.853 4240.710 65.710 5.655 1.558 4244.250 69.250 6.751
0.875 4240.819 65.819 5.987 1.594 4244.393 69.393 6.676
0.897 4241.011 66.011 5.631 1.633 4244.525 69.525 6.849
0.919 4241.130 66.130 6.100 1.672 4244.646 69.646 6.504
0.942 4241.219 66.219 5.970 1.711 4244.812 69.812 6.925
0.964 4241.411 66.411 5.676 1.753 4245.008 70.008 6.539
0.989 4241.543 66.543 5.911 1.794 4245.172 70.172 6.765
1.014 4241.612 66.612 5.500 1.836 4245.327 70.327 6.591
1.039 4241.818 66.818 5.708 1.881 4245.426 70.426 7.408
1.064 4241.953 66.953 5.841 1.925 4245.660 70.660 7.708
1.089 4242.079 67.079 5.780 1.969 4245.744 70.744 7.108
1.117 4242.241 67.241 5.753 2.017 4245.943 70.943 7.255
1.144 4242.329 67.329 6.224 2.064 4246.028 71.028 7.366
1.172 4242.497 67.497 5.886 2.114 4246.336 71.336 7.578
 297
2.164 4246.572 71.572 7.629 3.869 4251.141 76.141 7.926
2.214 4246.623 71.623 7.776 3.961 4251.304 76.304 8.043
2.267 4246.819 71.819 7.634 4.053 4251.455 76.455 8.429
2.319 4246.998 71.998 8.144 4.147 4251.606 76.606 8.393
2.375 4247.143 72.143 7.631 4.244 4251.706 76.706 8.677
2.431 4247.386 72.386 7.395 4.344 4251.849 76.849 8.208
2.489 4247.508 72.508 8.070 4.447 4252.121 77.121 8.285
2.547 4247.669 72.669 7.967 4.550 4252.319 77.319 8.581
2.608 4247.873 72.873 8.106 4.656 4252.556 77.556 9.119
2.669 4248.057 73.057 8.385 4.764 4252.719 77.719 9.330
2.733 4248.241 73.241 8.160 4.875 4252.917 77.917 9.817
2.797 4248.444 73.444 8.037 4.989 4253.093 78.093 9.653
2.864 4248.616 73.616 8.074 5.106 4253.309 78.309 10.132
2.931 4248.823 73.823 8.460 5.225 4253.574 78.574 9.865
3.000 4249.029 74.029 8.241 5.347 4253.717 78.717 9.957
3.069 4249.215 74.215 8.431 5.472 4253.960 78.960 10.029
3.142 4249.307 74.307 8.648 5.600 4254.190 79.190 9.807
3.217 4249.478 74.478 8.434 5.731 4254.487 79.487 10.084
3.292 4249.761 74.761 8.629 5.864 4254.665 79.665 10.025
3.369 4249.897 74.897 8.202 6.000 4254.843 79.843 10.476
3.447 4250.114 75.114 8.744 6.139 4255.051 80.051 10.742
3.528 4250.361 75.361 9.059 6.281 4255.176 80.176 10.888
3.611 4250.481 75.481 9.013 6.428 4255.448 80.448 10.337
3.694 4250.729 75.729 8.422 6.578 4255.694 80.694 10.703
3.781 4250.834 75.834 8.276 6.731 4255.925 80.925 11.100

298
6.886 4256.216 81.216 11.310 13.092 4262.914 87.914 11.470
7.047 4256.471 81.471 12.134 13.394 4263.163 88.163 10.916
7.211 4256.648 81.648 11.793 13.706 4263.385 88.385 9.735
7.378 4256.896 81.896 11.969 14.022 4263.622 88.622 9.393
7.550 4257.153 82.153 12.126 14.347 4263.726 88.726 8.637
7.725 4257.401 82.401 11.782 14.681 4263.947 88.947 8.560
7.906 4257.695 82.695 11.998
15.019 4264.120 89.120 8.417
8.089 4257.887 82.887 12.401
15.367 4264.170 89.170 8.168
8.278 4258.163 83.163 12.423
15.722 4264.250 89.250 8.086
8.469 4258.420 83.420 12.038
8.667 4258.636 83.636 12.024 16.086 4264.324 89.324 7.918
8.867 4258.926 83.926 11.234 16.458 4264.406 89.406 9.040
9.072 4259.177 84.177 11.716 16.839 4264.627 89.627 8.728
9.283 4259.425 84.425 11.674 17.228 4264.794 89.794 9.305
9.500 4259.598 84.598 11.815
9.719 4259.835 84.835 11.569
9.944 4259.963 84.963 11.613
10.175 4260.155 85.155 11.653
10.411 4260.434 85.434 11.686
10.653 4260.699 85.699 11.780
10.900 4261.017 86.017 12.194
11.153 4261.150 86.150 12.297
11.411 4261.506 86.506 12.651
11.675 4261.685 86.685 12.952
11.944 4261.934 86.934 12.756
 299
SOLUTION
The diagnostic plot confirms the presence of at least two flow
regimes: EARLY RADIAL and ELLIPTICAL.

300
The ER (Layer1) and LATE RADIAL could not be confirmed based on
available seismic and geologic information.
Thus, the interpretation of this test will only concentrate on the
analysis of the system-ER (all layers) and Elliptical flow regimes.
 301
TDS Technique:

Step 1 – System Early Radial

The early-radial flow (horizontal)

line yields:

(Δt*ΔP’)ER = 5.4 psia

Therefore, the geometric average

permeability in the (vertical) yz
plane is:

70.6qµB
kV = k z k y =
L(t × ∆P' ) ER
(70.6)(1800)(20)(1.14)
= = 366 md
(2050.525)(5.4)

302
Step 2 – Elliptical Flow
(slope = 0.36)

The geometric average

permeability in the
horizontal plane is
obtained from the elliptical
flow regime, where at
t=1hr, (t*dP’)ELL1 = 5.8:

qBµ 0.64 (1800)(1.14)(28) 0.64

E= = = 1.63
[
14930.4hz (φct rw ) 0.36 (14930.4)(98.425) (0.3)(3 ×10 −5 )(0.354 2 )
2 0.36
]
1 1
 E  hx hz  
0.72 0.64
1.63  9842.52 × 98.425  0.72  0.64
kH = kxk y =     =    = 451 md
 (t × ∆P' ) ELL1  Lrw    5.8  2050.525 × 0.354  
 303
Step 3 – Mechanical Skin
Sm

From the early radial: at

tER = 0.5, dPER = 62.77

1  ∆ p ER  k y k z t ER  
Sm=  - ln   + 7.43
2  (t × ∆P' ) ER  φµ ct r 2w  
  
 62.77  366 × 0.5  
= 0. 5  − ln −5 2 
+ 7.43 = 3
 5.4  0.3 × 28 × 3 × 10 × 0.345  

304
IS ALTERNATE INTERPRETATION POSSIBLE?
(i.e. IS LATE RADIAL PRESENT?)
 305
Late Radial

The late-radial flow (horizontal

line) yields:

(t*ΔP’)LR = 12.44 psia

Therefore, the geometric average

(horizontal) permeability is:

70.6qµB (70.6)(1800)(28)(1.14)
kH = k x k y = = = 3313 md
h z (t × ∆P' ) LR (98.425)(12.44)

If the LR flow regime does indeed exist, then the transition

period between ER and LR does not correspond to elliptical
flow (since from this flow regime kH = 451 md!).

Does the transition zone then correspond to the early linear

flow regime?

306
Early Linear: at t = 1 hr, (t*dP’)EL1=5.2
 307
Permeability in the x, y and z-directions
From EL at t=1hr:

2
 4.064qB  µ  4.064 × 1800 ×1.14 
2
28
k y =  
 =  = 196 md
 h z L (t × ∆P' ) EL1  φct  98.425 × 2050.525 × 5.2  0.3 × 3 ×10
−5

kx =
( )
kx ky
2

=
33132
= 55873md
y=∞ Y

ky 196 y=0
Z
z
Z=h

( )
Lw
Lza z=hs
2 Lzb
rw
Z=0
2 X=Lx1
kz ky 388 X=Lxd

kz = = = 683 md X=0 X=hx

ky 196

The permeability in the x-direction cannot be 55873 md, under no

circumstance (HW is drilled perpendicular to ky).
Also the model assumes ky>kx always (i.e. HW drilled perpendicular
to ky!)

Therefore the original interpretation based on the elliptical

flow regime is correct.

308

APPENDIX A
-
BASIC MODEL
 309

MODELS USED IN HORIZONTAL-WELL-TEST

ANALYSIS

Many solutions have been presented in the literature to

model the pressure-transient behavior of horizontal wells
(Ozkan et al. 1989; Goode and Thambynayagam 1987; Odeh
and Babu 1990; Kuchuk et al. 1990, 1991).

These solutions differ in their definitions and in the solution

techniques used.
In general, the differences are minor, and all solutions yield
the same fundamental characteristics.

310

DEFINITIONS
Wells that are parallel to the bedding plane of the
producing formation are called horizontal wells (although
the bedding plane and, therefore, the well may not be truly
horizontal).

Fig. shows the geometry of a horizontal-well/reservoir

system (the dip of the formation is assumed to be zero).
 311
For the discussion of the general characteristics of
horizontal-well responses, the following definitions of
dimensionless variables are used in this chapter:

A.1

A.2

A.3

k = (k x k y k z )
1/ 3
A.4

312
In Eq. A.4 kx, ky and kz are the principal permeabilities
(assumed to be in the directions of the coordinate axes,
with the x-axis in the direction of the well).

The other dimensionless variables are defined by

A.5

A.6

The equivalent radius of the horizontal well in an

anisotropic reservoir as given by Peaceman (1983):

 k 
0.25
k 
0.25

rweq = 0.5rw    + Z  
y

k k   A.7
 Z   y  
 313
The pressure-transient solution obtained by using the
source-function approach is given by (Ozkan and Raghavan
1990):

A.8

For applications such as variable-rate problems and naturally

fractured and layered reservoirs, it is useful to express the
solution in the Laplace-transform domain as follows (Ozkan
and Raghavan 1991):

314

A.9
 315

Most of the analytical models used in horizontal-

well-test analysis assume uniform flux distribution
along the well length.

As mentioned before, however, horizontal-well

behavior may be more closely approximated by
either infinite- or finite-conductivity behavior.

If the influence of wellbore hydraulics is not

expected to play a role in the performance of the
well, then the infinite-conductivity assumption is
more appropriate.

316

End of Chapter 12