Notes Hydraulic Fracturing
Notes Hydraulic Fracturing
Notes Hydraulic Fracturing
Fundamental of Hydraulic Fracturing
by:
Amit Verma
B.Tech (Applied Petroleum Engineering)
MBA (Oil & Gas management)
M.Tech (Petroleum Engineering)
1
lOMoARcPSD|3236323
Hydraulic Fracturing
Objectives and Economic Considerations
Hydraulic fracturing (HF) is a type of well stimulation treatment designed to improve the
fluid flow path from the formation to the well, and hence enhance the productivity of a well.
It also can be used to restore the skin effects due to near wellbore damage. Figure 1 shows
the typical hydraulic fracture treatment facilities.
Figure 1: Typical the hydraulic fracturing treatment
HF is an expensive and complicated engineering activity, which essentially requires
understanding of the fundamental principles of:
reservoir geology and reservoir engineering,
fluid flow in porous media,
rock mechanics,
Rock fracture mechanics,
fluid rheology and fluid mechanics,
solids transport, and
issues related to gel chemistry.
Since its first use in 1946 as a means of improving production from marginal wells in Kansas,
hydraulic fracturing has become a dominant completion technique. By 1993, fracture
treatments were being carried out on 40% of new, completed oil wells and 70% of gas wells
in the USA. Of course, fracture treatments are not restricted to new wells; they are also
widely used to stimulate older wells and to re‐stimulate already treated wells. It becomes
one of key technologies for the production from very low permeable gas (tight and/or ultra
2
lOMoARcPSD|3236323
tight gas) reservoir. It also gains popularity in case of production from high permeable oil
reservoir, especially reservoir having a tendency of excessive sand production.
During drilling and completion of a hydrocarbon producing well, even if the best practices
are followed, it is inevitable that certain "damage" would be created around the wellbore as
shown in Fig. 2 (mechanical damage).
Figure 2: Skin effect interpreted as altered permeability zone
This damage takes the form of an additional, uninvited resistance to flow of hydrocarbon.
Since in most cases, the flow of fluids converges toward the wellbore (radial flow), this extra
resistance causes a very large loss of pressure that decreases the overall well productivity. In
a hydraulic fracturing treatment, a high‐viscosity fluid is injected into the well at treating
pressures that are higher than the so‐called formation breakdown pressure (practically
speaking, the minimum horizontal stress, in case of vertical well). These high pressures
typically result the propagation of a fracture (both side of the well), usually known as two‐
wing vertically oriented (longitudinal) fracture. Fluid injection continues for some time
beyond this initial propagation; and when the created fracture is wide enough to accept
fluids, solid particles (sand or some other type of proppant material having very high
permeability) are injected simultaneously with the carrying fluid. The former fluid is called
pad fluid which is used to initiate and propagate fracture. The proppant material gradually
fills up the fracture, and help keeping the fracture open by supporting the overburden
pressure when the pumps are stopped and the pad fluid bleed off from the fracture.
The propped‐open, vertically oriented fracture that results from a successful fracture
treatment might be several dozen or several hundred feet high and possibly several
thousand feet long. Although it will typically be only a fraction of an inch wide, it will
drastically change the streamline structure of flow pattern in the formation. Not only will it
bypass near‐wellbore damage, but it will also impart a bilinear rather than a radial flow
structure (as illustrated in Figure 3). The net effects of such changes reduce the frictional
resistance to flow such that it can enhance well productivity by as much as four to ten times
that of the initial stabilized rate.
3
lOMoARcPSD|3236323
Flow pattern in HF
Figure 3: Flow pattern and pressure distribution around the HF
The objective of any fracture stimulation treatment is to improve the well’s productivity (or
injectivity, in case of injection well) index. The production engineer can then use this
productivity index improvement to achieve certain economic goals such as increasing
production rate at a certain pressure drawdown, or decreasing pressure drawdown while
maintaining an economic production rate. The latter is the basis for such non‐traditional
applications of fracturing, as sand control and condensate dropout control, which are seeing
ever‐increasing use.
Strictly speaking, fracture stimulation affects only the rate at which hydrocarbons are
withdrawn from the reservoir at a certain pressure drawdown. It does not increase the total
amount of petroleum that can be produced from the reservoir, provided time and
economics are not relevant factors. But in the real world, time and economics are relevant.
Once economics enters the picture, it is readily apparent that a large number of currently
producing oil and gas wells (even entire fields) could not have been produced at all without
being fractured, because of their uneconomical rates of natural productivity. In this sense,
we can consider fracturing also as a means of increasing industrial reserves. It is estimated
that over 25% of the total hydrocarbon reserves in the United States would not have been
recovered without the advent of hydraulic fracturing. Similarly beneficial results from
fracturing treatments have been realized in many low permeability and/or already mature
oil and gas fields outside the United States.
For a long time, however, formations having an effective permeability of more than 1 md
were rarely selected as candidates for extensive fracture stimulation programs. The
situation has changed dramatically with the rise of high‐permeability fracturing (frac & pack,
tip screen‐out technique) Recently propped hydraulic fracturing has displaced other
stimulation and/or sand‐control methods at a tremendous pace in high‐permeability
reservoirs.
Productivity Index and Skin Factor
The Productivity Index of a specific well can be considered as the proportionality constant, J
in the deliverability equation between production rate and driving force (pressure
drawdown):
4
lOMoARcPSD|3236323
q Jp (1)
During its life span, a well is subject to several changes with respect to flow conditions. For
the production engineer, the most important flow regime is pseudo‐steady state in a
bounded circular reservoir, where the pressure drawdown is defined as the difference of
average reservoir pressure and well flowing pressure, i.e.
p p pwf (2)
and the Productivity Index is approximately given by
2kh 1
J (3)
B ln 0 . 472 r
e
s
rw
Where, k is the formation permeability, h is the formation thickness, B is the formation
volume factor, is the viscosity of the oil, re is the drainage radius and rw is the well radius
and S is the radial flow skin factor representing the effect of near‐wellbore damage.
For gas wells, we can define a similar relation in terms of the squared pressures and gas
properties as:
2
p pwf
2
p , (4)
psc
and
2khTsc 1
J (5)
ZT ln 0.472re s
rw
Where, is average viscosity; Z is average compressibility factor; T is temperature and the
index sc refers to standard conditions.
Even using optimum drilling and completion practices, some kind of near‐well damage is
present. In Equations 3 and 5, the skin‐factor is dimensionless and can be considered as the
proportionality constant between the excess pressure drawdown (due to damage) and
production rate, provided both are cast into dimensionless form. For an undamaged well,
the skin factor is zero; for a damaged well, it has a positive value. One way to visualize the
skin (Figure 2) is to consider a damaged cylindrical region around the well, with outer radius,
rs, and impaired permeability, ks. Then, according to Hawkins (1956):
5
lOMoARcPSD|3236323
k r
s 1 ln s (6)
ks rw
The Hawkins formula shows that if the permeability impairment is several‐fold, a relatively
small damage radius is enough to cause significant skin effect. In other words, radial flow is
very sensitive to near‐well damage.
Well Stimulation to Improve Productivity Index
Well stimulation was introduced into the petroleum industry to eliminate the effect of near‐
well damage or, in terms of the skin factor, to "restore" zero skin. A typical well stimulation
technique of this type is sandstone acidizing, where the solid particles causing the near‐well
damage are dissolved by the acid and the original permeability is restored in the "skin
region". Obviously, such a treatment does not change the structure of the streamlines for
the flowing hydrocarbon.
In carbonate reservoirs, matrix acidizing may not only restore the hypothetical original
state, but it may even establish larger‐than‐original permeability in a finite region near the
well. The streamline structure remains intact following the stimulation treatment, but the
skin factor may attain a value of less than zero. Negative skin indicates that the treatment
has created a flow situation that is better than that of the hypothetical undamaged
formation. In fact, it is not even necessary to eliminate all the damage. It is enough that the
treatment creates enough flow capacity near the well to bypass the damaged zone.
In a hydraulic fracturing treatment, we create a conductive fracture by driving a "fluid
wedge" through the rock. We then place a solid propping agent in the created void to
prevent the fracture from healing and to provide the desired conductivity. Alternatively, we
sometimes achieve the post‐treatment fracture conductivity by using a low‐pH fluid to
dissolve a portion of the rock on the fracture face. Then the two etched surfaces are unable
to close and seal properly, and therefore a high‐conductivity conduit remains in the
formation. This technique is termed acid fracturing. Because of the relative significance of
propped fractures, much of our remaining discussion will concentrate on them.
With propped fracturing, we not only bypass the damaged zone, but by superposing a highly
conductive planar conduit on the formation we change the geometric structure of the
streamlines. The Productivity Index thus increases significantly. One way to numerically
characterize the effect of a propped fracture is to introduce the pseudo‐skin factor, sf
(Cinco‐Ley and Samaniego, 1979). We can then express the post‐treatment Productivity
Index as:
2kh 1
J (7)
B ln 0.472re s
f
rw
6
lOMoARcPSD|3236323
Having placed a propped fracture that really bypasses damage, the pseudo‐skin factor will
be negative. The question is, of course, how to predict sf, knowing the relevant formation
and fracture properties. Once we can answer this question, we can optimize the treatment.
Dimensionless Fracture Conductivity
Following Cinco‐Ley and Samaniego (1979) we assume a fully penetrating rectangular
fracture, in which the fracture height and formation thickness are equal (Figure 3). The
fracture half‐length xf is the length of one wing of a fracture.
Figure 3: Fully penetrating rectangular fracture
Most authors assume that two symmetrical fracture wings are created simultaneously
during a fracture operation, with the total overall length equal to twice the half‐length.
To obtain better insight, we rewrite Equation 7 into an expanded form:
2kh 1 2kh 1
Jf (8)
B 0.472re x B ln 0.472 r
f
s f ln f
e
ln
xf
xf rw
where the fracture half‐length, xf, is introduced directly into the Productivity Index. Instead
of the pseudo‐skin factor, we use the dimensionless factor f, which does not contain
reference to the wellbore radius. While the two forms of the Productivity Index are
algebraically equivalent, the expanded form is physically more meaningful. The expanded
form of the denominator involves three terms. The first term, ln(0.472) (= ‐0.75), is present
because the pressure drawdown is defined in terms of average pressure in semi‐steady
state. The second term xf, represents the effect of the fracture half‐length. The third term,
f, represents the effect of a combination of fracture variables called dimensionless fracture
conductivity, CfD, which is defined as:
7
lOMoARcPSD|3236323
kf w
C fD (9)
kx f
where k is the reservoir permeability, xf is the half‐length of the propped fracture, kf is the
permeability of the proppant pack and w is the average fracture width. The dimensionless
fracture conductivity, CfD should not be confused with the dimensioned fracture
conductivity, kf w. The dimensionless fracture conductivity expresses how the two functions
of the fracture are related. The two functions are to:
conduct the hydrocarbon inside the fracture to the well,
collect the hydrocarbon from the surrounding matrix rock
If CfD << 1, flow within the fracture is restricted, but flow into the fracture from the
surrounding matrix is unrestricted.
If CfD >> 1, flow within the fracture is unrestricted, but flow from the surrounding matrix into
the fracture is restricted.
The expanded form of the post‐treatment Productivity Index (Equation. 8) indicates that
post‐treatment performance is not related to the wellbore radius and the original skin
factor. That is how it should be, because the radial streamline structure has been changed
and the fracture bypasses the pre‐treatment damage.
The key to understanding hydraulic fracturing is that the dimensionless factor, f, depends on
CfD only. The most well‐known graphical representation of the function f (CfD) was given by
Cinco‐Ley and Samaniego, (1981) as shown in Figure 4.
Figure 4: Relationship of factor f and CfD (after Cinco‐Ley and Samaniego, 1981)
Following correlation can be used as an alternative:
8
lOMoARcPSD|3236323
When the dimensionless fracture conductivity is high (e.g., CfD > 100, which is possible for
low permeability formations having undergone a massive hydraulic fracturing treatment)
the behavior is similar to that of an infinite conductivity fracture (Gringarten and Ramey,
1974). For an infinite conductivity fracture, the dimensionless factor f is equal to ln(2),
indicating that the fractured well produces similarly to a hypothetical well of enlarged radius
equal to xf/2. Such an infinite conductivity behavior is, however, impossible to achieve in
most formations except for those with very low permeability. In medium and high
permeability formations, the propped fracture is always of finite conductivity.
In a finite conductivity fracture, we have a case of two players (fracture length and fracture
width) competing for the same resource: an incremental amount of propping agent. In other
words, we can use the propping agent to increase fracture length or width.
Before the advent of tip screenout (TSO) techniques, fracture extent and width were
difficult to influence separately. The TSO technique has brought a significant change to this
design philosophy. Now, fracture width can be increased without increasing the fracture
extent. In this context, we can formulate a strictly technical optimization problem:
How should we select the optimum fracture length and width when the proppant volume
(constraint) is given?
The solution to this problem is of primary importance in understanding hydraulic fracturing.
Surprisingly, it was already found as early as 1961 by Prats, but, unfortunately, has since
been somewhat forgotten. Prats assumed that the volume of one wing, Vf, the fracture
height, hf, and the two permeabilities, k, and kf, are given and wanted to find the optimum
width and half‐length.
We can use the same propped volume to create a narrow, elongated fracture or a wide,
short fracture as shown in Figure 5.
Figure 5: Two ways to place the same amount of proppant into the formation
9
lOMoARcPSD|3236323
It is convenient to select CfD as the decision variable, and then to express the fracture half‐
length using the propped volume of one wing, Vf as:
V k
xf f f (11)
C hk
fD
Substituting Equation 11 into Equation 8, we obtain
2kh 1 2kh 1
J f (12)
B ln 0.472re
f B 1 ( 0.472 r ) 2
hk 1
ln e
ln C fD f
Vf k f 2 V k 2
f f
C hk
fD
in which the only unknown is CfD . Since the drainage radius, formation thickness, two
permeabilities and the propped volume are fixed, the maximum productivity index occurs
when the quantity y reaches its minimum, where
1
y ln C fD f (13)
2
The quantity y is also shown in Figure 4. Since it depends only on CfD, the optimum CfD,opt =
1.6 is a given constant for any fixed amount of proppant.
The optimum dimensionless fracture conductivity corresponds to the best compromise
between the fracture’s capacity to conduct hydrocarbons and the reservoir’s capacity to
deliver hydrocarbons, for fixed fracture volume.
Once the volume of proppant that can be placed into one wing of the fracture, Vf, is known,
the optimum fracture dimensions can be calculated as:
V k 1.6V f k
x f f f , and w
1.6hk hk
f
Moreover, since:
1
yopt ln(1.6) f (1.6) 1.619 (15)
2
10
lOMoARcPSD|3236323
and yopt +ln(0.472) = 0.869, we obtain from Eq. 12, the optimum productivity index:
2kh 1
J f , opt (16)
B 1 r 2 hk
0.869 ln e
2 V f k f
There are several implications of the above results.
There is no theoretical difference between low and high permeability fracturing.
In both cases, technically there exists an optimally fracture conductivity, and in
both cases it should have a dimensionless fracture conductivity of 1.6. In a low
permeability formation, this requirement results in a long and narrow fracture. In
high permeability formations, a short and wide fracture will provide the same
dimensionless conductivity.
Increasing the volume of proppant or the permeability of the proppant pack by a
given factor (for example, 2) has exactly the same effect on the productivity if
otherwise the proppant is placed optimally.
The skin improvement depends on the amount of proppant (or on the proppant
pack permeability) according to a log‐square‐root relation.
To achieve the same post‐treatment skin factor in a low and a high permeability
formation, the volume of proppant should be increased by the same factor as the
ratio of the formation permeabilities, provided all the other formation and
proppant parameters remain same.
The above relations shed light on the role of the individual variables, and provide for the
theoretical optimum placement of a given amount of proppant. In practice, however, there
may be several factors forcing us to depart from this theoretical optimum.
Since not all proppant will be placed into the permeable layer, the optimum length
and width should be calculated with the effective volume, subtracting the proppant
placed in the non‐productive layers.
In low permeability formations, the indicated fracture width might be too small
(when the permeability of the proppant pack cannot be considered constant).
Therefore, a minimum width limit should be applied.
In high permeability formations, the indicated fracture length might not be enough
to bypass the damaged zone, therefore a minimum length should be applied.
Considerable fracture width can be lost because of proppant embedment into soft
formations.
For gas wells, non‐Darcy effects may create a dependence of the apparent
permeability of the proppant pack on the production rate itself.
Transient flow regime, high penetration ratio of the fracture with respect to the drainage
area, and other phenomena may also modify the optimum compromise between width and
length, but these issues are of secondary importance.
11
lOMoARcPSD|3236323
Having settled the optimization of fracture length vs. width for a fixed proppant volume, the
remaining task is to optimize treatment size, which is best characterized by proppant
volume placed into the formation. The currently preferred method of minimizing net
present value by varying propped length but using an arbitrary CfD of 10 or 30 is
theoretically wrong. This has to be kept in mind when studying older publications, along
with the fact that several other definitions of the dimensionless fracture conductivity that
are in use contain an additional factor of or /2.
The pseudo‐skin factor is not the only possible indicator for visualizing the effect of a
fracture treatment. We could also derive all of the above results using the concept of
equivalent wellbore radius. We have to be very cautious not to use both the pseudo‐skin
factor and the equivalent wellbore radius at the same time, however, because that might
lead to accounting for the same effect twice. Therefore, in this treatise we use only the
pseudo skin concept.
Variable proppant size:
In above calculation, it is assumed that a fixed volume of proppant will be placed into the
pay zone to calculate corresponding optimum fracture sizes (length and width)
Romero and Valko (SPE73758, 2002) developed a physical optimization technique to
determined the performance of a fractured well by proppant number (.i.e, the volume of
proppant placed into the pay and the permeability contrast). They defined the
dimensionless proppant number as:
4k f x f w 4k f x f whp 2k f V p
N prop I x2C fD 2
2
kx f kx h
e p kVr
Where Ix is the penetration ratio, CfD is the dimensionless conductivity, Vr is the reservoir
drainage volume, Vp is volume of proppant in the pay (i.e. total volume of proppant times
the ratio of net height to the fracture height), kf is the proppant pack permeability, k is the
reservoir permeability, xe is the drainage length of fracture, hf is the height of fracture, and h
is the reservoir thickness.
Sizing of Fracturing Treatments
Optimizing the treatment size (i.e., finding the optimum proppant volume) is an economic
rather than a technical optimization issue. The more proppant that is placed in the
formation (otherwise optimally), the better the performance of the well will be. At this
point, economic considerations must take over. The additional revenue from a larger
propped volume at some point becomes marginal compared to the more‐than‐linearly
increasing costs. This situation is properly treated by net present value (NPV) analysis.
The NPV is the difference between the present value of all receipts and costs (both current
and future) resulting from the stimulation treatment. Future receipts and costs are
converted into present value using a discount rate and taking into account the year in which
12
lOMoARcPSD|3236323
they will appear. The NPV (as other equivalent indicators are also) is sensitive to the
selected discount rate and to the predicted future hydrocarbon prices.
When using NPV or any related economic indicator, we understand that for any given
treatment size, we find the technically optimal way to place the proppant. If we neglect that
issue, the result of the NPV optimization will be wrong. We notice that there is no reason to
fill technically oriented publications with details of the NPV technique, because that subject
is better covered in relevant economics and accounting textbooks.
Although hydraulic fracturing was originally developed to improve oil well productivity, it
has since been found to have significant application to gas wells. The magnitude of the
hydraulic fracturing operations required in tight gas reservoirs has led to the development
of a special stimulation service termed massive hydraulic fracturing. Such a treatment
typically involves pumping very large volumes of frac fluid and proppant in a single
treatment to create an exceptionally deep‐penetrating propped fracture. Following this type
of treatment, we may produce the reservoir at much higher rates from a limited number of
wells, thus avoiding the expense of extensive infill development drilling. In view of the
above derivations, however, it should be obvious that the additional performance
improvement is physically limited. In fact the incremental benefit from an additional
incremental amount of proppant is even less if the fracture already penetrates a significant
portion of the drainage area. Therefore, oversized treatments are likely to be attractive only
in periods of high gas prices.
Formation and Fracture Properties Affecting the Performance
At this point it is useful to make a list of properties of the formation and fracture that
directly affect well performance (Table 1).
Formation Fracture
Pay thickness Extent
Permeability Proppant Permeability
Fluid viscosity Effective Propped Volume
Drainage radius
Table 1: Primary formation and fracture variables affecting performance
Once the optimum dimensionless fracture conductivity is understood, fracture length and
width are no longer freely selectable design parameters. From the production engineer’s
point of view, the amount of proppant (propped volume) should be the primary variable
characterizing the treatment size. Fracturing engineers, however, traditionally prefer to
think about fracture half‐length as the main variable.
13
lOMoARcPSD|3236323
Fracture Initiation Orientation and Growth
Fracture Initiation
In most cases, a fracture may be initiated by applying hydraulic pressure to an exposed
formation. Prior to fracture initiation, a positive differential pressure will cause the fluid to
enter the formation in a radial flow pattern, with the rate of fluid flow through the rock
limited to a rate that is in compliance with Darcy's law.
Maintaining the injection rate of a fluid above the maximum matrix flow capacity of the
exposed formation area will continually increase the flowing pressure at the wellbore.
Finally, the pore pressure will be increased to the point at which the rock ruptures in tension
in a direction perpendicular to the least principal stress present in the formation.
Fracture Orientation
Fracture orientation is directly related to the formation’s far field stress state. The fracture
will be oriented perpendicular to the direction of the least principal stress (Figure 1).
Least Principal Stress Least Principal Stress
Horizontal fracture Vertical fracture
Figure 1: Fracture orientation
This is most easily understood by realizing that for a fracture to form, a portion of the
reservoir must undergo physical deformation. The direction in which it is easiest to push or
deform the rock is the one exerting the least resistance (least stress). Thus, the fracture will
be oriented at a 90‐degree angle to this stress.
The vertical stress stems from the weight of the overburden, and it is partially translated to
horizontal stresses. At sufficient depths (usually below 1000 m or 3000 ft) the minimum
principal stress is horizontal; therefore, the fracture faces will be vertical. For shallow
14
lOMoARcPSD|3236323
formations, where the minimum principal stress is vertical, horizontal (pancake) fractures
will be created.
The above picture of fracture orientation is somewhat simplistic. Perforations and pre‐
existing flaws and microfractures may "guide" fracture orientation, at least at the initiation
stage. Moreover, in the near‐wellbore region the original stress state is disturbed and the
minimum principal stress direction might be different from the far field direction. It is
important to remember that such disturbances are localized in the vicinity of the wellbore
(say within a distance of two‐to‐three times the wellbore diameter). Once the fracture
extent is large enough, the far field stress state dominates its orientation.
Fracture Growth
After breakdown, the fluid entering the fracture partly leaks off through the exposed faces
of the fracture. The other part of the fluid continues to enlarge the fracture as long as
sufficient hydraulic pressure is maintained and the injection rate is kept above the rate at
which the injected fluid continues to leak off into the formation. Growth is generally
confined to a single plane (perpendicular to the least principal stress), and continues equally
in all directions of the fracture plane until it encounters some barrier limiting the growth
rate in that direction (Figure 2).
Figure 2: Fracture initiation and growth
Simultaneously with fracture propagation, the fracture’s average width is also expanding. In
fact, lateral propagation, height growth and width inflation are competitive processes. The
created fracture geometry depends on how these processes share the fluid volume left in
the fracture after fluid leakoff.
We may define a fracture growth barrier as anything that limits the extension of a fracture
in any direction. Barriers may be overlying or underlying zones having significantly different
properties of elasticity than the zone being fractured (Young's modulus of elasticity and
Poisson's ratio). They may be rocks having a higher tensile stress, high‐stress loadings, or
stress loadings in which the least principal stress is in a different direction than at the
15
lOMoARcPSD|3236323
wellbore. They may be rocks having higher frac gradients, or zones having lower pore
pressures. They may be slippage planes—unique bedding planes having no vertical bonding,
in which the adjacent surfaces act almost as if they are lubricated, and which dissipate the
dynamic growth energy of a fracture. Barriers may also be physically intruded solids, such as
propping agents. In short, fracture barriers may be combinations of any or all of these
factors. Many of these factors are difficult to measure or even estimate. The variation of
minimum principal stress from layer to layer is, however, more accessible and is considered
to be the main factor controlling height containment and growth.
Role of Formation Properties in Fracturing
The formation properties that are known to influence a fracture’s growth pattern, including
its height, are:
Young's modulus
Poisson's ratio
Tensile strength
Fracture toughness
Permeability
Porosity
Poroelasticity constant
Along with these material properties, the actual state of the formation also affects the
evolution of the fracture. A detailed discussion regarding the effect of each of these
parameters is given below.
Rock Properties
A rock’s elastic properties are most often described by two different terms: Young’s
modulus and Poisson’s ratio.
Young's Modulus: Young's modulus (E) is essentially an index of the rock’s resistance to
external force. It is defined as the ratio of the applied stress to the resulting strain:
E (1)
In other words, it is a coefficient of proportionality between uniaxial stress and strain. It has
the same dimensions as pressure, and is typically measured in units of Pa or psi. Its value
can be determined from a uniaxial stress test as shown in Figure 1.
16
lOMoARcPSD|3236323
Figure 1 (Uniaxial stress test for determining Young’s Modulus and Poisson’s ratio).
Higher values of Young's modulus indicate greater stiffness. Therefore, a given amount of
fracturing fluid will create a relatively long, narrow fracture in a rock having a high Young's
modulus value. The same amount of fluid will create a shorter but wider fracture in a rock
with a low Young's modulus value, provided all other properties are the same. This is one of
the main differences in fracturing hard versus soft formations. If the Young's modulus varies
from layer to layer, it might cause a complex width profile, with reduced widths in the layers
of higher modulus values. Soft formations are characterized by E values as low as 105 psi
(Diatomite), while hard formations can have E values as large as107 psi (hard Limestone).
Poisson's Ratio: Poisson's ratio, , is defined as the ratio of the lateral strain demonstrated
by a rock when subjected to a longitudinal load, divided by the amount of longitudinal strain
caused by the same loading. It is a dimensionless quantity, usually ranging from 0.15 to 0.35.
It can also be measured in the laboratory as shown in Figure 1
From a hydraulic fracturing standpoint, Poisson's ratio is primarily responsible for translating
vertical stress into horizontal stresses. It also has some (limited) influence on fracture width.
Plane Strain Modulus: Most of the equations used in fracturing contain only a certain
combination of Young's modulus and Poisson's ratio, denoted by E':
E
E (2)
1 2
The plane strain modulus is numerically very near to the Young's modulus, because the
square of the Poisson's ratio can usually be neglected with respect to one.
Shear modulus: Some authors prefer to use the shear modulus, G, which can be easily
calculated from the Young's modulus and the Poisson's ratio according to
17
lOMoARcPSD|3236323
E
G (4)
2(1 )
It is important to understand that the above properties are related to the elastic behavior of
the rock. They can be measured on a core sample using static or dynamic measurement
methods, or in‐situ, using dynamic (mostly sonic) methods. The "static" and "dynamic"
properties may be somewhat different.
Tensile Strength: The maximum stress that a material can tolerate without rupture in a uni‐
axial tensile experiment is the tensile stress. Though the effect of the tensile strength is
minimal during fracture extension, it affects the fracture initiation (formation breakdown)
pressure.
Fracture Toughness: The critical value of the stress intensity factor, or fracture toughness,
characterizes a rock’s resistance to the propagation of an existing fracture. It is measured in
Pam0.5, psift0.5, or psi inch0.5, because the stress intensity factor at the tip of a fracture is the
product of the pressure loading and the square root of a characteristic length (such as
fracture half‐length). Fracture propagation occurs when the stress intensity factor reaches
its critical value. It is easier to propagate a larger fracture than a smaller one, provided the
pressure loading on the faces is the same. When the stress intensity factor at the tip equals
the fracture toughness, a special equilibrium state is reached. Vertical fracture extension
(i.e., height growth) is often considered as a process passing through such equilibrium
states. Sometimes the lateral extension of the fracture is also considered as a sequence of
such equilibrium states. The concept itself is important, but the actual value has limited
effect on our calculations. Laboratory measurements indicate fracture toughness values
ranging from 500 to 2000 MPa‐m0.5 (almost the same limits are obtained if expressing the
values in psi‐inch0.5).
Permeability: The larger the fluid leakoff, the less driving force is available for fracture
growth. Formation permeability is one of the main factors controlling fluid leakoff, and
hence, indirectly affects fracture propagation. Porosity and total compressibility have a
limited influence on fluid leakoff as well.
Poroelastic Constant The Poroelastic constant is defined by the relation:
K
1 (4)
Ks
where K is the bulk modulus (ratio of hydrostatic pressure to volumetric strain) of the dry
rock material and Ks is the same measured in a saturated sample. The saturated sample is
more resistive to compression because the fluid carries part of the load; therefore the
poroelasticity constant is less than one. Most fracturing calculations assume a value
between 0.7 and 1.
All the above properties are related to the material behavior of the rock matrix.
18
lOMoARcPSD|3236323
State Variables
The following properties are related to the actual state of the rock matrix, as determined by
the in‐situ conditions.
Bedding planes, layer structure: Existing large discontinuities in the rock are usually two‐
dimensional structures, hence the name plane. Slippage that occurs along bedding planes
tends to dissipate the energy required for fracture propagation, and thus reduce fracture
growth in that direction. The presence of intersecting fractures or planes of weakness
hinders further fracture growth in that direction, even if growth is not stopped completely.
The least principal stress in the reservoir (also referred to as closure stress or closure
pressure) equals the fluid pressure required to hold open an induced fracture. It is often
convenient to report this quantity as a fracture gradient, defined as the least principal stress
divided by the depth. The fracture gradient is normally calculated from pressure
measurements taken during injection tests. If it is considered a constant for all wells in the
same reservoir, the stress state around one well can be extrapolated to another well. As the
reservoir pressure varies, the least principal stress and the fracture gradient vary
simultaneously.
Pore pressure: Pore pressure influences the effective stresses in a formation via the
poroelastic constant. The change of pore pressure in the zone of interest (due to depletion)
brings about a change of the same direction (but lesser magnitude) in the least principal
stress.
The presence of higher pore pressure in an adjacent formation increases the tensile forces
present in that zone, thereby requiring a lower internal hydraulic pressure to initiate failure
caused by rupturing, which can actually cause a fracture to grow into the adjacent
formation. Conversely, an adjacent low‐pressure zone (or an area of lower pressure within a
reservoir, such as that surrounding an old producing well) will put that formation in
compression and cause it to serve as a fracture barrier and stop continued growth, or
possibly divert fracture growth in another direction. Accurate analyses of the pore pressures
in the zone of interest and the surrounding formations are one way to predict height
containment of vertical fractures.
The stress‐state of the formation is of primary importance to the fracturing engineer. It is
not necessary, however, to know every little detail. The most important issue is that of the
least principal stress and its variation with depth, often referred to as vertical profile of
the minimum stress.
Vertical Stress
At any point in the formation, the total vertical stress due to overburden, v, is simply the
weight of the material above that point:
D
v g dz (5)
0
19
lOMoARcPSD|3236323
where g is the acceleration due to gravity, is the density of the rock (possibly varying with
depth z) and D is the true vertical depth.
The total stress is carried by both the "grains" and the pore fluid in a porous medium (Figure
2)
Figure 2: Poroelasticity
The effective stress, ’, is the absolute stress minus the pore pressure (p) weighted by the
poroelastic constant ( ):
v v p (6)
We can estimate the minimum effective horizontal stress due to the overburden by
"translating" the effective vertical stress, i.e., multiplying it by /(1‐):
h v p (7)
1
Where, is the Poisson ratio. Finally, we calculate the total horizontal stress by adding back
the poroelastic term:
h v p p v 1 2 p (8)
1 1 1
In addition to the overburden, tectonic forces created by geological events have also
induced stresses in the formation, as evidenced by the many structural features present in
the formation (e.g., faults, folds, natural fractures and inhomogeneity.) Since it is not
possible to fully describe the origins and causes of these geologic events, it is also not
possible to accurately predict the magnitudes or directions of the resulting stresses.
Nevertheless, we can use Equation 6 to quantitatively explain several important
phenomena:
20
lOMoARcPSD|3236323
It is often observed that the least principal stress decreases with depletion.
Indeed, Equation. 8 shows that a unit decrease of pore pressure causes
(1 2 ) /(1 ) decrease in the fracturing pressure. On the contrary, injection of
fluid and temporary increase of pore pressure may increase least principal stress.
Equation 8 explains the existence of large stress contrasts between adjacent
layers. While the overburden is almost the same, the difference between the
Poisson's ratios can cause higher stresses in the layer characterized by larger
Poisson's ratio.
The fact that the least principal stress is vertical at shallow depths can be
explained if we accept that horizontal stresses are "frozen traces" of some prior
geological state, but erosion of the surface have decreased the overburden.
Consequently, for shallow formations, the vertical stress can be less than the
"frozen" horizontal stress. For deeper formations, however, the ratio of
horizontal stress to vertical stress approaches, and hence the former is less than
the latter.
The above theory is more qualitative than quantitative. Tectonic stresses due to
geologic movements can cause a large scatter around the theoretically calculated
stresses. It is generally accepted that a fracture gradient of less than 0.7 psi/ft in
a tectonically relaxed reservoir indicates that a vertically oriented fracture will be
created, because it is easier to part the earth than to lift it. A fracture gradient of
greater than 1.1 psi/ft (which is in agreement with the generally accepted value
for the normal overburden gradient) indicates a serious anomaly specific to the
formation.
Knowledge of the vertical profile of the minimum horizontal stress is essential for two
reasons. First, the value at the center of the perforations will be the base to which net
pressure is added to obtain the fracture propagation pressure. Second, a positive stress
contrast in the neighboring layers is believed to be the most important controlling factor for
height containment (Figure 3). A detailed depth‐to‐depth determination of the minimum
stress might be expensive and time‐consuming; therefore, several methods have been
suggested for using available well logs to predict vertical stress variations.
21
lOMoARcPSD|3236323
Figure 3: Vertical profile of minimum horizontal stress (Economides, et 1994)
Most of these methods are reliable only if used after calibration. The calibration process
involves correlating the log with known values of the minimum stress for at least one well in
the formation. The known values of the minimum stress are obtained from calibration
fracturing treatments, several variations of which are called micro fracturing, step‐rate test,
pump‐in flow‐back test, and minifrac closure pressure determination.
In addition to the magnitude of the least horizontal stress, its orientation is also of interest,
because it governs the orientation of the induced fracture. Several methods are available for
estimating the principal stress orientation. One group of measurements uses oriented core
samples, such as an elastic strain relaxation. Another group includes tilt meter or similar
measurements in open hole. Acoustic measurements on the oriented core sample and
acoustic logs of the open hole can also provide information on the orientation of the least
principal stress.
Fracturing Pressure
Concepts such as fracturing pressure are not well defined, because they are not just
properties of the formation, but are also a function of how they are determined. Except
when the injection rate is very low (at least on the order of several gallons per minute) we
have to distinguish fracture initiation pressure, fracture propagation pressure and fracture
closure pressure. These pressures, which are expressed as bottomhole values referenced to
the center of the perforations, may or may not be equal.
Fracture Initiation Pressure or breakdown pressure is the peak value of the pressure
appearing when the formation breaks down and a fracture starts to evolve. Usually it is
approximated by
22
lOMoARcPSD|3236323
Where min is the minimum horizontal stress, max is the maximum horizontal stress, T is the
tensile stress of the rock material, is the poroelasticity constant and po is the pore
pressure. In the above equation, only the tensile stress is a material constant that can be
measured in a laboratory. The principal stresses and the pore pressure are state properties.
The above theoretical breakdown pressure might be masked by other factors such as the
microstructure and material behavior of the borehole wall, the geometry of the perforations
and the properties of the fluid. The most serious obstacle in applying Equation 9 is that we
rarely know the maximum horizontal stress. In fact, Equation 9 is more often used in a
reverse manner to approximate the maximum stress from known minimum stress and
observed breakdown pressure.
Fracture Propagation Pressure is the stabilized value of the injection pressure for a longer
period of time during which the fracture is evolving. Obviously, it is not a material or even
state property of the formation itself, because the fluid type, the injection rate, and most
important, the leakoff process may dramatically affect its value. To understand pressure
behavior, one has to be familiar with at least the simplest mathematical models of fracture
propagation.
In a narrower sense, fracture propagation pressure is associated with the so‐called step‐rate
test as shown in Figure 4
Figure 4 : Detection of formation breakdown from a step‐rate test.
During a step‐rate test, the fluid is injected into the formation while increasing the injection
rate in discrete steps. At each step, the stabilized pressure is recorded. The plot of stabilized
injection pressures versus injection rates typically shows a break point. At low injection
rates, the subsequent steps increase the pressure according to Darcy's law. After a critical
pressure is reached, a fracture is created (and propagated), and the subsequent change in
pressure with each rate step is usually much less than in the Darcy region. The break point is
usually determined by fitting two straight lines: one through the low, and another through
the high injection rate points. The procedure calls upon the engineer’s subjective judgment,
and therefore requires a careful examination of all circumstances. Since the fracture
propagation pressure is a state property, its value might change during the life of a well,
mostly because of pore pressure depletion.
23
lOMoARcPSD|3236323
Fracture Closure Pressure. After a fracture calibration treatment, which is carried out
without injecting proppant material, the fracture volume gradually decreases because of
leakoff (and also because of possible back flow, if the injected fluid is flowed back through
the well). At the same time, the pressure decreases. Eventually the fracture will close (that
is, the fracture faces will contact). The bottomhole value of the pressure at the moment of
closure is the fracture closure pressure. Usually it is determined by careful examination and
processing of a pressure fall‐off curve. The basis for selecting the closure point is that before
closure, the leakoff process and its combination with the elastic behavior of the formation
dominate the pressure falloff, but after closure the pressure is governed by the general laws
of fluid flow in porous media. Even if the exact quantitative description of these processes is
not readily available, a marked change in the character of the pressure fall‐off may be
interpreted as the closure point. This is the reason why several transformation plots are in
use.
Often, a simple pressure versus time plot or pressure versus square root of time plot is
suitable to pick the closure pressure. Unfortunately, the closure pressure may be masked by
non‐ideal (stochastic) events during the closure process and especially by the particular way
in which the fracture faces approach each other. The terms fracture closure pressure and
minimum principal stress are used interchangeably in the technical literature. Figure 5
shows some of the "strategic locations" on the pressure response curves of typical fracture
calibration tests
Figure 5 : Typical Fracture Calibration test showing fracture‐related pressure points:
(1)breakdown pressure; (2) fracture propagation pressure; (3) instantaneous shut‐in
pressure; (4) closure pressure; (5) fracture reopening pressure; (6) closure pressure from
flow‐back; (7) asymptotic reservoir pressure; (8) rebound pressure
The pressure in a propagating fracture is higher than the closure pressure. The difference
between the actual and closure pressures is called the net pressure. In everyday usage, the
net pressure is meant at the wellbore, but in fracture propagation models the net pressure
varies along the length.
24
lOMoARcPSD|3236323
Quantitative Description of Fracture Growth
The increase in fracture volume over the course of a treatment is determined by the
amount of injected fluid that does not leak off.
Leakoff
The key to the material balance is fluid leakoff. Fluid leakoff is controlled by a continuous
build‐up of a thin layer, or filter cake, which manifests an ever‐increasing resistance to flow
through the fracture face. In reality, the actual leakoff is determined by a coupled system, of
which the filter cake is only one element. A fruitful approximation dating back to Carter
(Appendix to Howard and Fast, 1957), is to consider the combined effect of the different
phenomena as a material property. According to this concept, the leakoff velocity, vL, is
given by the Carter equation:
CL
VL (1)
t
Where CL is the leakoff coefficient (length/time0.5) and t is the time elapsed since the start of
the leakoff process. The ideas behind Carter's leakoff coefficient are that
if a filter‐cake wall is building up, it will allow less fluid to pass through a unit area in
unit time; and,
the reservoir itself can take less and less fluid if it has been exposed to inflow.
Both of these phenomena can be roughly approximated as "square‐root time behavior". The
integrated form of the Carter equation is:
VLost
2CL t S p (2)
AL
where VLost is the fluid volume that passes through the surface AL during the time period
from time zero to time t. The first term, 2C L t can be considered as width of the fluid
passing through the surface during the main part of the leakoff process. (The factor 2
appears because the integral of 1 / t is 2 t ).
The integration constant, Sp, is called the spurt loss coefficient. It can be considered as the
width of the fluid body passing through the surface instantaneously at the very beginning of
the leakoff process. The two coefficients, CL and Sp, can be determined from laboratory
tests.
A vertical hydraulic fracture has two wings. For modeling purposes, we usually assume that
these are identical, thus making it possible to model just one wing. Suppose qi is the
injection rate, Vi the injected volume, V the volume of fluid contained and A is the surface
25
lOMoARcPSD|3236323
area of one face, all corresponding to one wing of the fracture and a given time t during the
treatment as shown in Figure 1.
Figure 1, Definition of injection rate, fracture area and permeable height.
We will use the subscript e to denote the end of pumping in order to distinguish the
quantities corresponding to the time instant te. If the injection rate, qi is constant, Vi = qit .
The fluid efficiency defines the fraction of the fluid remaining in the fracture: = V/Vi. The
fracture surface, A, is the area of one face of one wing and the average width, w , is defined
by the relation: V Vi Aw .
It is often assumed that the created fracture remains in a well‐defined lithological layer
(mostly the producing formation), and the fracture is therefore characterized by a constant
height, hf.
A hydraulic fracturing operation may last from tens of minutes up to several hours. Points of
the fracture face nearest to the well are opened at the beginning of pumping, while the
points at the fracture tip are "younger". To apply the following equation:
we must track the opening‐time of the individual fracture face elements. If we divide the
injected volume by the surface area of one face of one wing, A, we obtain the so‐called
"would‐be" width. The would‐be width can be broken down into average fracture width,
leakoff width and spurt loss width:
Vi
w 2CL te S p (3)
A
where the factor 2 is introduced because the fluid leaks off through both faces of one wing.
The dimensionless factor, , is the opening‐time distribution factor. It reflects the effect of
26
lOMoARcPSD|3236323
the distribution of the opening‐time. If all the surface is opened at the beginning of the
injection, then reaches its absolute maximum, = 2.
To obtain an analytical solution for constant injection rate, Carter considered a hypothetical
case where the fracture width remains constant during the fracture propagation (the width
"jumps" to its final value in the first instant of pumping.) Nolte (1986) postulated a basically
similar, but mathematically simpler assumption. He assumed that the fracture surface
evolves according to a power law,
AD t D (4)
Where AD A / Ae , t D t / te and the exponent remains constant during the whole injection
period.
If we accept this assumption, we can easily obtain the opening time distribution factor from
the exponent. Selected values are given in Table 1.
Table 1: Opening time distribution factor
Once we know the opening time distribution factor, it is easy to make material balance
calculations.
Width Equations
Three dimensions compete for the fluid volume remaining in the fracture: lateral extent,
height and width.
Simple 2‐D design models assume either that the fracture height is a given value (e.g., PKN
and KGD models), or that the fracture is of penny shape (e.g., Radial model). Once the
problem is reduced to two dimensions, additional assumptions are applied in order to
obtain a relation between fracture extent and width. The final equations are obtained from
the conceptual model that the viscous resistance to flow gives rise to a net pressure that is
exerted on the fracture faces and keeps the fracture open. Therefore it is not surprising that
fracture width is related (among other things) to elastic modulus (E'), injection rate to one
wing (qi), viscosity of the fracturing fluid () and half‐length (xf).
Perkins‐Kern‐Nordgren (PKN) Width Equation
The visual representation underlying the PKN model is a two‐wing rectangular fracture of
constant height. The vertical cross section of the fracture is an ellipse as shown in Figure 2.
27
lOMoARcPSD|3236323
Figure 2 L PKN geometry (after Perkins and Kern, 1961. Economides, etl, 1994).
The maximum width of the ellipse at a certain distance from the well is related to the
height, plane strain modulus and net pressure at that location. Since the net pressure is also
related to injection rate and fluid viscosity, a width equation not containing the pressure can
be derived. Of particular interest is the maximum width of the ellipse located at the
wellbore, ww,0:
q x
1/ 4
ww,0 3.27 i f (5)
E
(The constant was originally 3.57 in the Perkins‐Kern form of the equation, but the above
form given by Nordgren has become more accepted.) The average width of the fracture is
related to the maximum width according to:
w 0.628ww,0 (6)
Kristianovich‐Zheltov‐Geertsma‐DeKlerk (KGD)
The visual representation behind the KGD model is also a two‐wing rectangular fracture of
constant height. The vertical cross section is considered, however, to be rectangular as
shown in Figure 3.
28
lOMoARcPSD|3236323
Figure 3: KGD geometry (After Geertsma and deKlerk, 1969, Economides, etl, 1994).
Physically, this means that the fracture faces slip freely at the upper and lower boundary of
the pay layer. The fracture width at the wellbore is given by
1/ 4
qi x 2f
ww 3.22 (7)
E h
f
(Notice that we use only one width index, because the width does not change vertically.)
The average fracture width is calculated from
w 0.785ww (8)
Radial (Penny‐shaped) Width Equation
By analogy, a radially expanding fracture as shown in Figure 4,
29
lOMoARcPSD|3236323
Figure 4: Radial Penny Shape Fracture Geometry has
Radial Penny Shape Fracture Geometry has a maximum width at the wellbore:
qi R f
1/ 4
ww,0 4.20 (9)
E
and the average fracture width is calculated from:
w 0.533ww,0 (10)
Figure 5 shows that given all the same parameters, the curves of width versus fracture
extent for the PKN and KGD models cross each other.
Figure 5 (Comparison of PKN and KGD width equations)
At smaller extent, the PKN width equation predicts larger width. The crossover occurs
approximately at the point at which a "square fracture" has been created, i.e., when 2xf is
approximately equal to hf.. While this fact has been used to argue for one or the other
equation, the truth is that the physical assumptions behind the KGD equation are more
30
lOMoARcPSD|3236323
realistic for the small fracture extent situation. For larger fracture extents, however, the PKN
width equation is physically more sound.
No‐leakoff Behavior of Width Equations
The so‐called width equations relate fracture width and extent. Thus, if we know the
fracture volume, we can use the width equation to obtain the fracture dimensions. In the
particular case of negligible leakoff, the fracture volume is simply equal to the injection rate
multiplied by the injection time. Using this fact, we can derive the time behavior of a
propagating fracture for the no‐leakoff case, as summarized below for the Perkins‐Kern‐
Nordgren, Geertsma and deKlerk, and Radial models.
1. Perkins‐Kern‐Nordgren model
Fracture Extent:
1/ 5
q E
x f 0.572 i 4 t 4 / 5
h
f
(the constant is 0.524 for the original PK equation)
Width:
1/ 5
q2 1/ 5
w 1.75 i t
E h
f
(the constant is 1.91 for the original PK equation)
Net Pressure:
1/ 5
E 4 q 2i 1 / 5
pn , w 1.39 t
h6
f
(the constant is 1.52 for the original PK equation)
2. Geertsma and deKlerk model
Fracture Extent:
1/ 6
q E
x f 0.539 i 3 t 2 / 3
h
f
Width:
1/ 6
q 3i 1/ 3
w 1.85 t
Eh3
f
31
lOMoARcPSD|3236323
Net Pressure:
pn , w 1.09 E 2 1 / 3 1 / 3
t
3. Radial model
Fracture Extent:
1/ 9
q 3i E 4 / 3
R f 0.572 t
Width:
1/ 9
q 3i 2 1 / 9
w 1.95 t
E
Net Pressure:
pn ,w 2.51 E 2 1 / 3 1 / 3
t
Looking at the net pressure equations above, we can see that while the PKN (or PK) model
predicts an increasing treating pressure curve, the other two models predict decreasing
pressure profiles. In addition, the PKN model implies that the net pressure is higher if the
injection rate is larger. The other two models predict a net pressure varying with time
independently of the injection rate, and therefore they are of limited practical use for
pressure related analysis. In general, we cannot assume that leakoff is negligible and the
above solutions in terms of time are not valid. Nevertheless, we can combine a particular
width equation with material balance relations and obtain a closed design model.
Other Processes Controlling Fracture Extension
Our present understanding of fracture propagation is that in most cases, the simple two‐
dimensional models described above predict faster fracture propagation than actually
occurs in the formation. In other words, the tip propagation is usually retarded. This means
higher‐than‐zero net pressure at the tip, because there is intensive energy dissipation in the
near‐tip area. Several attempts have been made to incorporate this tip phenomenon into
fracture propagation models. One reasonable approach is to introduce an apparent fracture
toughness that increases with the size of the fracture. Other approaches include a
controlling relationship for the propagation velocity, uf, incorporating some additional
mechanical property of the formation (dilatancy factor, continuum damage mechanics
parameter, etc).
In principle, the lateral and vertical propagation of the fracture is subjected to the same
mechanical laws. The substantial difference is that the fracture tip meets the same
32
lOMoARcPSD|3236323
minimum stress during lateral propagation, while the vertical tip crosses several layers with
different material properties and stress state.
The equilibrium height concept of Simonson, et al. (1978) provides a simple and reasonable
method of calculating the height of the fracture if there is a sharp stress contrast between
the target layer and the over‐ and under‐burden strata. If the minimum horizontal stress is
considerably larger in the over‐ and under‐burden layers (i.e., by several hundred psi), we
may assume that the fracture height is determined by the requirement of reaching the
critical stress intensity factor at both the top and bottom tips. This requirement of
equilibrium poses two constraints, and so the two penetrations can be obtained solving a
system of two nonlinear equations. The solution can be plotted as a height‐map, indicating
what fracture height will be reached at a given treating pressure (Figure 6)
Figure 6, Height map.
The dashed line is a second (unstable) solution to the system of equations). Height‐maps are
advantageous for selecting the fracture heights to be used in simple two‐dimensional design
models. They also help us determine a treatment pressure limit (if, for instance, we must
avoid fracturing into a water zone).
33
lOMoARcPSD|3236323
Exercises
Q1.
a. Calculate a well’s skin effect due to radial damage if the permeability impairment is k/ks =
5 fold, the wellbore radius is rw= 0.328 ft and the penetration distance is 0.5 ft.
b. Assume that pseudo‐steady state flow conditions and a drainage radius of re = 2980 ft
apply to this well. What portion of the pressure drawdown is lost in the skin zone?
c. Assume that the well has been matrix acidized and the original permeability has been
restored in the skin zone. By what factor will the production rate increase assuming the
pressure drawdown is the same before and after the treatment? By what factor will the
Productivity Index increase?
d. Assume that this well has been fracture treated and a negative pseudo skin factor has
been created: sf = ‐5. By what factor will the Productivity Index with respect to the damaged
well?
Sol ‐1:
a. The damage radius is the sum of the wellbore radius and the penetration distance, or rs =
0.828:
k r
s 1 ln s
ks rw
Since we deal only with ratios, we do not have to change units:
0.828
s (5 1) ln 3.7
0.328
b. The fraction of pressure drawdown in the skin zone is given by:
s
re
0.75 ln s
rw
Since we deal only with ratios, we do not have to convert units:
3.7
0.31
2980
0.75 ln 3.7
0.328
Therefore, 31 percent of the pressure drawdown is not utilized because of the near‐
wellbore damage.
c. We can assume that the skin after the acidizing treatment becomes zero. Then the
increase in production assuming the same pressure drawdown is:
re
0.75 ln s
rw
re
0.75 ln
rw
Since we deal only with ratios, we do not have to convert units:
34
lOMoARcPSD|3236323
2980
0.75 ln 3.7
0.328 1.44
2980
0.75 ln
0.328
The Productivity Index increase will be the same: 44 percent.
d. The ratio of Productivity Indices after and before the treatment is
re
0.75 ln s
rw
re
0.75 ln s f
rw
Since we deal only with ratios, we do not have to convert units:
2980
0.75 ln 3.7
0.328 3.6
2980
0.75 ln 5
0.328
The Productivity Index will increase by a factor of 3.6.
Q 2: a. Assume an undamaged well of radius rw = 0.322 ft, formation pay thickness h = 50 ft,
drainage radius re = 2100 ft and reservoir permeability k = 0.5 md. By what factor will the
Productivity Index theoretically increase if 500 ft3 of proppant can be placed into the
formation to give a proppant pack permeability of kf = 60,000 md?
b. Find the optimal fracture half‐length and width needed to realize the Productivity Index
increase calculated in part a.
c. Repeat all the calculations from parts a and b, again assuming an undamaged well of
radius rw = 0.322 ft, formation pay thickness h = 50 ft and drainage radius re = 2100 ft, but
now assume that the reservoir permeability k = 20 md. Again, 500 ft3 of proppant can be
placed into the formation to give a proppant pack permeability of kf = 60,000 md
d. By what factor will the productivity index be increased for the well and treatment in part
c, if the 500 ft3 of proppant is not placed optimally, but using an arbitrary requirement of,
say, CfD = 30.
Sol 2:
a. Since 500 ft3 of proppant can be placed into the formation, the volume of one wing will
be 250 ft3. The maximum Productivity Index improvement is given by:
r
0.75 ln e
rw
re2 hk
0.869 0.5 ln
Vf k f
Since we deal only with ratios, we do not have to convert units:
35
lOMoARcPSD|3236323
2100
0.75 ln
0.322 4 .3
2100 2 50 0.5
0.869 0.5 ln
250 60000
Therefore, if the 500 ft3 of proppant is placed optimally into the formation, the Productivity
Index increases by 330 %.
b. Since the optimum CfD is 1.6,
1/ 2
1.6V f k f
1/ 2
Vf k f
x f and w
hk
1.6hk f
Since we deal only with ratios of permeabilities, their units should not be converted. The
equations remain consistent if we use cubic feet for volume and feet for all the length
variables.
250 60000
1/ 2
xf 612 ft
1.6 50 0.5
and
1.6 250 0.5
1/ 2
c. The Productivity Index increase
re
0.75 ln
rw
re2 hk
0.869 0.5 ln
Vf k f
is now only
2100
0.75 ln
0.322 =2.16
2100 2 50 20
0.869 0.5 ln
250 60000
But this will be a much larger production rate increase in absolute terms, because the
original production rate of the well is much larger than in the low permeability case.
The optimal fracture half‐length and width are now:
250 60000
1/ 2
xf 97 ft
1.6 50 20
36
lOMoARcPSD|3236323
and
1.6 250 20
1/ 2
d. The increase in the Productivity Index is
re
0.75 ln
rw
r 2 hk
0.869 0.5 ln e 0.5 ln C fD f (C fD )
Vf k f
where
1.65 0.328u 0.116u 2
f (C fD ) and u ln C fD
1 0.18u 0.064u 2 0.005u 3
We first calculate f(30)
u ln C fD ln 30 3.401
1.65 0.328 3.401 0.116 3.4012
f (C fD )
1 0.18 3.401 0.064 3.4012 0.005 3.4013
Substituting the obtained value into our basic equation, we obtain
2100
0.75 ln
0.322 1.77
2100 50 20
2
0.869 0.5 ln 0.5 ln 30 0.736
250 60000
Therefore, the arbitrary placement of the proppant yields a 77% increase of production rate,
while the optimal placement of the same proppant volume would result in 116 % increase
(assuming unchanged pressure drawdown).
We note that the arbitrary requirement, CfD = 30 , would require a fracture 22 ft in length
and 2.7 inches in width, which is difficult to create even in a soft formation.
37
lOMoARcPSD|3236323
Q‐3: An oil well of radius rw = 0.328 ft drains an area of radius re = 3000 ft. The pay
thickness h is 45 ft, and formation permeability k is 0.5 md. The formation fluid is oil with a
viscosity of 0.8 cp and a formation volume factor B = 1.2 RB/STB.
a) What is the production rate if the well is undamaged and the pressure drawdown
is 1500 psi (assume pseudosteady‐state flow conditions)?
b) What is the theoretically maximum production rate if 100,000 lbm proppant,
having specific gravity = 2.65 (with respect to water), proppant pack porosity p =
0.38 and proppant pack permeability kf =60,000 md, can be placed into the pay
layer? What is the optimal half‐length xf? What is the optimal areal proppant
concentration?
c) What would be the theoretically maximum incremental production rate if an
additional 100,000 lbm proppant could be placed into the formation? What would
be the optimal half‐length? What will be the optimal areal proppant concentration?
Sol 3:
a) What is the production rate if the well is undamaged and the pressure drawdown is 1500
psi (assume pseudo steady‐state flow conditions)?
In a consistent system of units the production rate from the undamaged well would be
2khp 1
qo
B re
0.75 ln
rw
Using field units, this equation has the form
khp 1
qo
141.2 B re
0.75 ln
rw
and hence
0.5 45 1500 1
qo 29.74 STB/d
141.2 1.2 0.8 0.75 ln 3000
0.328
b. The volume of the 100,000 lbm proppant is
100,000
975 ft 3
(1 0.38) 62.4 2.65
where 62.4 lbm/ft3 is the reference water density. Therefore, the one‐wing propped volume
is Vf = 487.5 ft3.
The theoretically maximum production rate in consistent units is
2khp 1
q1
B 2
r hk
0.869 0.5 ln e
Vf k f
which in oil‐field units takes the form:
38
lOMoARcPSD|3236323
khp 1
q1
141.2 B r hk
2
0.869 0.5 ln e
Vf k f
Substituting the known variables, the result is
0.5 45 1500 1
q1
141.2 1.2 0.8 3000 2 45 2
0.869 0.5 ln
487.5 60000
Therefore, placing the first 100,000 lbm proppant increases the production rate by 106.4
STB/D. The corresponding fracture‐half length in consistent units is
1/ 2
Vf k f
x f
1.6hk
and hence,
487.5 60000
1/ 2
xf1 901.4 ft
1.6 45 0.5
The optimal areal proppant concentration is 50,000 lbm / (901.4 ft ×45 ft ) = 1.23 lbm/ft2 and
the optimal width is 0.144 in.
c. The additional 100,000 lbm proppant would increase the one‐wing propped volume to Vf
= 975 ft3.
The theoretically maximum production rate is then,
0.5 45 1500 1
q1 167.1 STB/d
141.2 1.2 0.8 3000 2 45 2
0.869 0.5 ln
975 60000
Therefore, placing the second 100,000 lbm proppant increases the production rate by an
additional 31.5 STB/D.
The corresponding fracture‐half length is:
975 60000
1/ 2
xf 2 1275 ft
1.6 45 0.5
and the optimal areal proppant concentration in this case is: s
100,000 lbm / ( 1275×45 ft2) = 1.74 lbm/ft2. s
39
lOMoARcPSD|3236323
Q4: Assuming a poroelastic constant of = 0.7, estimate the absolute minimum horizontal
stress at 10,000 ft depth if the Poisson ratio is = 0.2 and the pore pressure is po = 5000 psi.
What will be the effect of depleting the pore pressure to 2000 psi?
Sol4: The absolute vertical stress is calculated using the approximate overburden gradient
(g) =1.1 psi/ft . In this case v = 11,000 psi.
Applying Equation
h ( v p) p
1
we obtain
0.2
h (11,000 0.7 5000) 0.7 5000 5375 psi
1 0.2
After depletion to 2000 psi the absolute horizontal stress will be
0.2
h2 (11,000 0.7 2000) 0.7 2000 3800 psi
1 0.2
Therefore, depleting the reservoir by 3000 psi will cause 1575 psi decrease in the minimum
horizontal stress (a 0.1575 psi/ft decrease in the fracture gradient).
Q 5: a. A fracture’s area evolves according to a Power Law model with exponent 2/3 (KGD
model; opening time distribution factor is = 1.478). The leakoff coefficient is CL = 0.001
ft/min0.5 and the pumping time is 40 min. Calculate the width lost because of leakoff.
What will be the total lost width if there is an additional spurt loss, the spurt loss coefficient
being Sp = 0.01 gal/ft2 ?
Sol 5: a. The width lost due to leakoff is
wL 2C L k t e
Therefore,
wL 2 0.001 1.478 40 0.0187 ft 0.224in 5.7mm
b. The spurt loss width is twice the spurt loss coefficient, or 20.01 gal/ft2 =0.032 in. (0.81
mm). The total lost width is 0.256 in. (6.5 mm).
40
lOMoARcPSD|3236323
Q6: Assume that a vertical fracture has an aspect ratio one, that is, 2xf = hf . What will be the
ratio of average widths calculated from the PKN and the KGD width equations, if other
parameters are the same?
Sol 6:
From Quantitative Description of Fracture Growth: Equations 5 and 6, the PKN average
width is
qi x f
1/ 4
0.628 3.27
E
and from Equations 6and 7, the KGD average width is
1/ 4
qi x 2f
0.785 3.22
E h
f
The ratio of the two average widths is
0.628 3.27x f
1/ 4
1/ 4
x 2f
0.785 3.22
h
f
0.628 3.27x f
1/ 4
= 1/ 4
0.785 3.22 x f 0.5
2x f
1/ 4
hf
0.628 3.27
=0.97
0.785 3.22 0.5 1
1/ 4
In other words, when the aspect ratio is one, the two width equations give (almost) the
same average widths.
41
lOMoARcPSD|3236323
Q 7: Estimate the maximum possible net pressure from the PKN, KGD and Radial width
equations if slurry is injected for te = 40 min with a two‐wing injection rate of (2qi) = 20
Bbl/min. Assume the following data are available:
Young's modulus, E = 2.0×105 psi;
Poisson ratio, = 0.2;
Fracture height = 40 ft; and,
Average equivalent fluid viscosity, = 180 cp.
Assume no‐leakoff to obtain maximum pressure.
Solution ‐7
The plane strain modulus is E' = 2.0×105/(1‐0.22) psi = 2.08×105 psi, and the one‐wing
injection rate is qi = 10 bbl/min.
For this problem, it is advantageous to use a strictly coherent system of units. The SI is
practically the only full and coherent system (moreover, it is supported by law). Therefore,
we use SI.
First we convert all the input data.
E' = 1.44×109 Pa
= 0.2
hf = 12.2 m
= 0.180 Pas
qi = 0.0265 m3/s
te = 2400 s
The PKN net pressure is
1/ 5
E 4 qi 2
pn , PKN 1.39 t 1/ 5
h 6 e
f
1/ 5
(1.44 10 9 ) 4 0.180 0.0265 2
1.39 (2400)1 / 5
12.2 6
1.16 10 Pa6
42
lOMoARcPSD|3236323
Similarly, for the KGD model we obtain
pn , KGD 1.09 E 2
1/ 3
te
1 / 3
1.09 (1.44 10 9 ) 0.180 1/ 3
(2400) 1 / 3
5.86 10 4 Pa
p n , KGD 8.5 psi
and for the radial model:
pn , Rad 2.51E 2 t 1 / 3
1/ 3
1.33 10 5 Pa
pn , Rad 19.4 psi
Comparing the net pressures, it should be obvious that the KGD and Radial models are not
suitable for analyzing treating pressures, because their net pressure predictions are not
realistic.
Reference for Additional Reading
1. “Petroleum Production Systems”, by Michael J. Economides, etl, 1994, Prentice Hall
Petroleum Engineering Series, Chapter 16‐17
2. “Reservoir Stimulation”, by Michael J. Economides, and Kenneth G. Nolte, 3rd
Edition, Wiley publisher, Chapter 1, 3, 4‐6
3. “Modern Fracturing – Enhancing Natural Gas Production”, by Michael J Economides,
Tony Martin, 2007, E. T. Publishing, Chapter 4
43