You are on page 1of 10

SPE 36424

Determination of Pressure Dependent Leakoff and Its Effect on Fracture Geometry


R. D. Barree, SPE, Marathon Oil Company and H. Mukherjee, SPE, Dowell Schlumberger

Copyright 1996, Society of Petroleum Engineers, Inc.


relationship between the leakoff coefficient and the rate of
This paper was prepared for presentation at the 71st Annual Technical Conference and pressure decline. The assumptions in this work were
Exhibition held in Denver, CO, U.S.A., 6-9 October, 1996.
influenced by prior findings of Walsh, that natural fissures
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
dilate with increased treatment pressure resulting in an
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to exponential increase in permeability.3 Warpinski extended this
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at finding to relate net pressure to leakoff coefficient.4
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Permission to copy is restricted to an abstract of not more than 300
words. Illustrations may not be copied. The abstract should contain conspicuous Though the pressure dependent leakoff analysis resulted in
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435. a means to estimate the pressure dependent leakoff function,
the method could not be validated because of the lack of a
suitable fracture design model capable of using this function
Abstract to calculate fracture geometry. In this study the method is
The effects of natural fissure opening, or pressure dependent validated by a fully 3D fracture geometry model.5 A
leakoff, on the pressure behavior observed during fracturing diagnostic superposition derivative of the analytical G-
are significant. Previous work has suggested that this behavior function is shown to uniquely distinguish between pressure
can be identified from pressure diagnostic plots during dependent leakoff, fracture height changes, and tip extension
pumping, or from pressure falloff analysis. However, these and recession. Use of the superposition derivative to analyze
techniques lead to ambiguous conclusions regarding the fracture treating pressure behavior during pumping has been
magnitude, and even existence, of pressure dependent leakoff. presented by Ayoub, et al.6 The correct application of pressure
This paper presents a method of pressure falloff analysis dependent leakoff to fracture geometry calculations
which removes this ambiguity and allows an accurate substantially alters the resulting fracture size and fluid
determination of the magnitude of pressure dependent leakoff. efficiency. Accurate modeling of this effect can reduce the
The effects of fracture tip extension and recession, height occurrence of premature screenouts, which are common in the
recession, and transient flow in the fracture are also identified. case of pressure dependent leakoff, and improve the overall
The effect of pressure dependent fracture compliance, which treatment design.
has not been previously published, is also described.

Introduction
Post fracture pressure decline during shut-in is often affected Pressure Dependent Leakoff
by fracture height growth, extension or recession of the Naturally fractured reservoirs requiring hydraulic fracture
fracture tip after shut-in, and pressure dependence of leakoff. treatments exhibit pressure dependence of the fluid leakoff
These factors can alter the shape of the pressure decline curve. coefficient. In general, the induced hydraulic fracture
The G-function, which describes fracture pressure decline propagates perpendicular to the minimum horizontal in-situ
behavior, was analytically shown to be linear for constant stress. The transverse natural fissures are held closed by the
leakoff with a wall-building type of fluid.1 However, many action of the maximum horizontal stress. As the treating
factors which affect pressure decline behavior can result in pressure increases and approaches the maximum horizontal
non-linearities in the computed G-function. In 1990, stress, natural fissures dilate resulting in exponential increase
Mukherjee et al proposed a method of fracture pressure in the leakoff coefficient. Conventional fracture pressure
decline analysis for cases of pressure dependent leakoff, decline analysis7 fails to predict this phenomenon uniquely,
assuming the G-function to be piecewise linear during especially when fracture height growth or tip extension or
pressure decline in naturally fractured reservoirs.2 Methods recession occurs during fracture closure. For the improvement
were presented to derive a very simple exponential

References and illustrations at end of paper.


2 DETERMINATION OF PRESSURE DEPENDENT LEAKOFF AND ITS EFFECT ON FRACTURE GEOMETRY SPE 36424
of fracture design, pressure dependent leakoff must be closure pressure of 4000 psi). The pressure vs. G-function
uniquely diagnosed and quantitatively determined. This paper curve does not show any definite period of straight-line
presents an analysis technique which provides a unique behavior. This is clearly indicated by the continuously
evaluation of pressure dependent leakoff behavior, as well as changing pressure derivative. However, the superposition
a means to diagnose other non-linear pressure effects. The derivative does approach straight-line behavior after a G
following case studies, based on hypothetical numerical function value of about 0.4. The early time superposition
simulation examples, present methods of uniquely interpreting derivative data falls below the extrapolation of the apparent
the pressure behavior observed during post-treatment decline. straight-line. This early-time curvature in both derivatives is
characteristic of fracture extension after shut-in with
associated transient flow in the fracture.
Case Studies
Five hypothetical case studies are used to illustrate the
behavior of pressure, the derivative of pressure with respect to
G-function (dP/dG), and the superposition derivative, defined
as (GdP/dG) during shut-in. In classical G-function analysis Case 2: Constant fracture geometry with no pressure
the pressure vs. G-function plot should be linear over the dependent leakoff. This case illustrates the most basic
range of pressure decline corresponding to constant fracture behavior which can be observed during fracture closure. The
geometry and constant leakoff coefficient. Therefore, the normal leakoff data set was used to illustrate the type of
pressure derivative should be constant. Non-linearity in the pressure and derivative behavior which may be observed
pressure curve, and variability in the pressure derivative, during closure with no pressure dependent leakoff behavior
signal changing leakoff or fracture geometry. The case studies and constant fracture geometry after shut-in. Figure 3 shows
presented here illustrate many causes of this non-linear the pressure and derivative curves versus G-function during
behavior. Unique characteristic shapes of the pressure closure. The P vs. G-function data show some early time
derivative and superposition derivative are illustrated for each curvature, but gradually approach linearity for large values of
example. Using these case studies the pressure decline G-function (greater than 1) as indicated by the flattening of
behavior observed during shut-in can be better understood. the pressure derivative curve. The superposition derivative
For consistency in analysis of the pressure decline results curve shows perfectly linear behavior from a G-function value
presented in this paper, the same basic fracture and reservoir of 0.2 onward, indicating constant leakoff coefficient,
properties are used for all case studies. The data are grouped although the pressure derivative curve still indicates
in two categories: One for “normal” leakoff rates and one for considerable variability in leakoff behavior.
“low” leakoff. Data for each group are presented in Table 1. Since the model results show no significant changes in
A constant fracture height of 70 feet is used in all geometry during the early shut-in time, it appears that the non-
simulations except the case of fracture growth into high stress linearity in the pressure derivative is caused by transient flow
confining zones. In each case a fluid volume of 17,000 gallons effects in the fracture. Figure 4 shows the model calculated
is pumped at a rate of 20 bpm. The fluid has properties similar pressure distribution in the fracture at shut-in, 22 minutes after
to a 35 #/Mgal crosslinked guar-borate frac gel with n’=0.31. shut-in, and 40 minutes after shut-in. Several interesting
No proppant is injected in these case studies so that features are apparent in the pressure behavior. At shut-in the
complications associated with closure on proppant are slope of the pressure-distance surface is significant, indicating
eliminated. A shut-in period of 40 minutes is modeled at the substantial flow and associated friction in the fracture. As
end of pumping for each case. closure begins, the pressure slope decreases substantially. In
the limit of leakoff rate approaching zero the pressure slope
should also approach zero. As closure continues the pressure
Case 1: Fracture extension after shut-in with no pressure slope stabilizes at a value characteristic of the total leakoff
dependent behavior. The first case study uses the low leakoff rate, which is related to the fluid velocity in the fracture.
data to illustrate the effects on pressure falloff behavior of Fracture tip recession is also apparent after the fluid pressure
fracture extension after shut-in. Figure 1 shows the created at the fracture tip declines to closure pressure (after 22
length, width, and height as a function of pumping time, along minutes following shut-in). The model indicates no fracture
with the actual pump rate during the treatment. Shut-in occurs width in the area of the fracture with fluid pressures less than
at approximately 25 minutes, as shown by the immediate the closure pressure of 4000 psi.
decrease in fracture width. Because of the stored energy in the In this case the superposition derivative appears to give the
fracture, the frac length continues to extend for about 10 best qualitative indication of stabilized leakoff behavior. The
minutes after shut-in. During this time the length increases pressure derivative appears to be related to the slope of the
from approximately 850 feet to 940 feet. pressure-distance function in the fracture, so that fluid leakoff
Figure 2 shows the pressure decline, pressure derivative, coefficient can be overestimated if significant friction (or flow
and superposition derivative observed during closure with velocity) exists in the fracture. This effect may be exacerbated
fracture extension. At the end of 40 minutes of shut-in the by non-Newtonian (shear-thinning) fluid
bottomhole pressure declines to only 4500 psi (with a fracture
4 DETERMINATION OF PRESSURE DEPENDENT LEAKOFF AND ITS EFFECT ON FRACTURE GEOMETRY SPE 36424
and the P vs.G-function slope is not constant. The pressure observed discontinuity may be partially caused by numerical
derivative indicates substantial change in leakoff which is discretization in the simulation. For smooth and continuous
supported by the non-linearity of the superposition derivative closure, the derivatives should exhibit continuous curvature as
curve. the fracture height recedes. The fracture height recession is
At the fissure opening pressure of 4400 psi the P-G- accompanied by several combined effects on the fracture
function curve shows a significant break and rapidly closure rate. First, as the overall fracture height decreases, the
approaches linearity. The pressure derivative undergoes a step fracture compliance will decrease. However, closure of the
increase then becomes nearly constant. Similarly, the higher stress bounding layers reduces the energy available to
superposition derivative increases sharply then assumes a drive fluid loss in the fracture. Finally, as the fracture closes in
linear trend. For these data the decrease in fracture compliance the bounding zones, the effective area of the open fracture
coincident with fissure closure dominates the system. The approaches the total permeable area available for fluid loss.
“stiffer” system behaves as if the overall leakoff coefficient The combination of these effects appears to result in an
were higher, based on the rate of pressure decline, although overall increase in apparent fluid loss coefficient, as shown by
the actual leakoff rate is considerably less after fissure closure. the increase in the pressure derivative and the increasing slope
of the superposition derivative function.

Case 5: Fracture height recession in bounding layers. The


final case study illustrates the effect of fracture closure in high
stress bounding layers on observed pressure behavior. For this Summary of Case Study Results
case the “normal” leakoff data shown in Table 1 are used. A The examples shown here illustrate several possible types of
confining stress contrast of 1000 psi is assumed between the pressure behavior which may be observed during fracture
target and bounding intervals. A target zone height of 60 feet closure. If only the pressure versus G-function plot is utilized
is used in this study. No pressure dependent leakoff or for analysis ambiguous results can be obtained. In fact, it is
compliance is assumed. difficult to determine whether early time non-linearity in the
Figure 7 is a plot of fracture geometry including length, pressure curve is caused by pressure dependent leakoff,
width, and height during the simulation. Pump rate is also transient flow, fracture tip extension or recession, or changes
shown to indicate the point of shut-in. Note that the fracture in fracture height. The use of the pressure and superposition
height grows from an initial value of 60 feet to approximately derivatives, along with the pressure vs. G-function data,
140 feet during the job. After shut-in the extra energy uniquely clarifies all these ambiguities. All three diagnostic
provided by the high stress boundary zones causes some fluid plots are required for unambiguous interpretation of post-
to be forced out of the upper and lower tips. This results in fracture pressure decline data.
some fracture extension after shut-in. The fracture half length A complete understanding of the fracture closure behavior
grows from about 600 feet to 620 feet during the first three must involve an overall analysis of a complete data set,
minutes after shut-in. With continued closure, the fracture including pressure behavior during pumping and shut-in,
height recedes to about 100 feet after 40 minutes of shut-in. along with any available stress and mechanical property data
However, the fracture is far from closed at this time, with a and post-frac temperature and tracer logs. A successful
remaining width of about 0.13 inches at the wellbore. analysis must consistently honor all available data.
Figure 8 is a plot of the pressure and derivative data during Pressure dependent leakoff behavior is not discontinuous,
closure for this example. The pressure curve shows the typical but represents a continuum of leakoff behavior during
early time curvature, then approaches what could be pumping and shut-in. Its importance at various times during
considered a straight-line section. At a pressure of about 4650 the treatment must be measured against other factors including
psi the pressure plot begins to curve significantly downward. overall leakoff rate, pump rate, fracture growth rate (in both
The pressure derivative shows this behavior more distinctly. height and length), and many other parameters.
After a rapid initial decline the pressure derivative For example, it has been suggested that pressure
becomes nearly constant until a G-function value of dependent leakoff can be characterized by a zero slope on the
approximately 0.9 when it steps upward. Following the step log-log net pressure diagnostic plot during pumping.4
change the value of the derivative remains nearly constant However, simulation results show that a zero slope can only
until a second step-change occurs at G-function value of about be achieved if the overall leakoff rate approaches the total
1.2. The superposition derivative indicates linear behavior pump rate. Otherwise, fracture extension continues and net
from G-function values of 0.1 to slightly less than 0.9 then pressure continues to rise, although at a slower rate.
shows both a step change in value and an increase in slope. To illustrate this observation the data from the pressure
Early-time straight line behavior in the superposition dependent leakoff case study presented above (Case 3) was
derivative (for G-function values between 0.1 and 0.9), along used to generate net pressure diagnostic plots at three different
with the stabilizing pressure derivative, indicate nearly pump rates. The pump rate used in the case study was 20 bpm
constant leakoff behavior and fracture geometry during this throughout most of the job. Rates of 10 bpm and
period. The step changes in both derivatives coincide with the
beginning of closure in the high stress bounding layers. The
SPE 36424 R. D. BARREE AND H. MUKHERJEE 5
30 bpm were also used. Figure 9 shows the net pressure Implications of Pressure Dependent Leakoff on
plots for these three cases. The top line is for 30 bpm injection Fracture Geometry
rate, and exhibits the greatest net pressure slope. Decreasing The effect of the correct use of this pressure dependent leakoff
the pump rate to 10 bpm causes the slope to decrease, but it function on created fracture geometry is significant. Pressure
still remains positive even though the fracture pressure is dependent leakoff behavior must be accounted for correctly in
above the fissure opening pressure throughout most of the job. design simulators. Although the elevated leakoff coefficient
The positive slope still exists because the maximum leakoff appears to be a transient effect when viewed from the
rate is 8.5 bpm compared to the injection rate of 10 bpm. standpoint of post-frac pressure decline, leakoff can be
Additional simulation cases (not shown here) have indicated elevated throughout most of the pumping time of the
that the combination of increased leakoff and decreased treatment. Because of this, leakoff coefficients calculated from
compliance above the fissure opening pressure further the stabilized part of the pressure decline curve can
decreases the observed net-pressure slope. significantly underestimate actual leakoff conditions. This can
These results illustrate the difficulty in diagnosing fissure also be illustrated using the Case 3 data.
opening from pumping pressure alone. Accurate As an example, if the pressure versus G-function data
characterization of pressure dependent behavior requires a shown for Case 3 are analyzed using the 75% rule discussed
detailed analysis of the pressure falloff and its derivatives above, without special regard to pressure dependent leakoff
during closure. behavior, actual leakoff conditions during pumping are
underestimated. For these data the ISIP value is 4730 psi and
closure pressure is 4000 psi. By application of the 75% rule
Analysis of Pressure Dependent Leakoff Behavior the value of the pressure derivative at 4548 psi is
At pressures above the fissure opening pressure the overall approximately 730. This yields a total leakoff coefficient (CL)
leakoff coefficient is significantly increased above the of 0.0016 by Equation 2. While this is larger than the base
normally assumed value for matrix fluid loss. Mukherjee et al leakoff coefficient (Co=0.0012) used in the study, it is
have suggested that the magnitude of the leakoff coefficient at considerably less than the leakoff coefficient which applies
any point above fissure opening pressure, relative to the during pumping, when treating pressures are above the fissure
coefficient under closed-fissure (or constant leakoff) opening pressure.
conditions, is proportional to the ratio of the pressure The effect of the underestimation of leakoff on fracture
derivative at the two conditions. geometry is shown by Figure 11 which is a contour plot of
Using the data from Case 3, the ratio of leakoff fracture width at the end of pumping for two cases. The upper
coefficients (Cp/Co) was calculated for all pressures above the plot shows the created geometry using the pressure dependent
fissure opening pressure. The results are shown in Figure 10. leakoff function applied for Case 3. The lower plot shows the
The heavy line is the estimated leakoff ratio obtained from the resulting geometry using a constant leakoff coefficient
ratio of the pressure derivative to the derivative after fissure estimated by the 75% rule, without pressure dependent
closure at each pressure. The actual leakoff ratio used in the behavior. Failure to account for pressure dependent leakoff
simulation of Case 3 is given by Equation 3, and is plotted as behavior results in a gross overestimation of fracture length.
the thin line in Figure 10. In general the agreement is very The errors induced by ignoring pressure dependent leakoff
good. Other cases (not shown) have been run with lower are more easily shown by comparing fluid efficiency at the
overall leakoff values and different levels of pressure end of pumping for the two cases. With pressure dependent
dependence. These studies indicate that the leakoff ratio can leakoff the actual efficiency at shut-in is 0.276. With the
be accurately estimated using the derivative ratio constant leakoff coefficient computed from the 75% rule, the
approximation, and that the accuracy of the approximation computed efficiency is 0.486. Interestingly, Equation 1 yields
increases with lower leakoff or more Newtonian fluid very similar efficiency estimates for both cases (η=0.465 for
behavior. pressure dependent leakoff and η=0.480 for the constant
It must be noted that the method only applies in cases of leakoff case). These results agree with the actual leakoff for
actual pressure dependent leakoff. If early time derivative the 75% rule case (constant leakoff), but are significantly in
curvature results from other factors such as fracture tip error for the pressure dependent case. The reason for the
extension or recession, or height variation, the method should difference is that the rate of pressure decline during shut-in is
not be applied. However, the method does allow estimation of primarily controlled by the matrix fluid loss rate, once the
the magnitude of pressure dependent leakoff for use in pressure drops below fissure opening (or closing) pressure.
fracture treatment design calculations, when pressure However, during the treatment the pumping pressure remains
dependent behavior is indicated by the behavior of all three above fissure opening pressure throughout the job. Therefore,
diagnostic curves. the leakoff coefficient which controls efficiency at shut-in is
different than that which controls closure after shut-in (at
pressures below the fissure opening pressure). Obviously,
decisions regarding the volume of pad fluid and proppant
SPE 36424 R. D. BARREE AND H. MUKHERJEE 7
6. Pressure dependent leakoff significantly affects fracture 6. Ayoub, J. A., Brown, J. E., Barree, R. D. ,and Elphick, J. J.:
geometry and fluid requirements. It must be accounted for ”Diagnosis and Evaluation of Fracturing Treatments,” SPE
accurately in fracture design calculations. Production Engineering, (February 1992) 39-46.
7. Nolte, K.G.: “A General Analysis of Fracturing Pressure
Decline With Application to Three Models,” SPE-FE
(Dec.1986) 571-583.
Nomenclature
a = pressure dependent compliance factor
b = pressure dependent leakoff factor Table 1: Basic data for Case Studies
ct = total formation compressibility, Lt2/m,psi-1
Property Normal Leakoff Low Leakoff
C = fluid loss coefficient, L/t0.5, ft/min0.5
Frac Height, ft 70 70
E = Young’s Modulus, m/Lt2, psi
G = dimensionless time function Modulus, psi 4.2e6 4.2e6
h = gross fracture height, L, ft Poissons’ Ratio 0.25 0.25
k = permeability, L2, md Permeability, md 0.60 0.10
P = pressure, m/Lt2, psi Porosity, fraction 0.18 0.10
rp = ratio of permeable height to gross height Cw, ft/√min 0.0012 0.0003
s = formation skin factor Pore Pressure, psi 2000 2000
tp = producing time, t, min Min. Stress, psi 2000 2000
φ = porosity, fraction Max. Stress, psi 2400 2400
η = fracture fluid efficiency, fraction
µ = dynamic viscosity, m/Lt, cp
1000 0.35

Subscripts 900
Length
c = closure 800
0.3

f = fluid 0.25
p = pressure dependent 700
Width
Height, Length, and Rate

o = original 600
0.2

L = overall leakoff

Width
500

0.15
400

Acknowledgments 300
0.1

The authors would like to thank the management of Marathon 200

Oil Company and Dowell Schlumberger for their support and 100 Height
0.05

permission to publish this paper. We also thank Dr. W. C. Chu 0


Rate
0
for his transient test analysis expertise. 0 10 20 30 40 50 60 70

Pumping Time, minutes

Figure 1: Fracture geometry for Case 1 showing fracture


References extension after shut-in.

1. Nolte, K.G., Mack, M.G., and Lie, W.L.: “A Systematic


Method of Applying Fracturing Pressure Decline : Part I,” SPE 500 4950

25845, Proc., SPE Rocky Mountain Regional/Low Permeability 450 4900


Reservoirs Symposium, Denver (1993) 31-50. dP/dG
2. Mukherjee, H., Larkin, S., and Kordziel, W.: “Extension of 400
4850

Fractured Pressure Decline Curve Analysis to Fissured 350


4800
Formations,” SPE 21872, Proc., SPE Rocky Mountain
Bottomhole Pressure, psi
Pressure Derivatives

300
Regional/Low Permeability Reservoirs Symposium, Denver 4750

(1991 ) 671-685. 250

4700
3. Walsh, J.B.: “Effect of Pore Pressure and Confining Pressure on 200
GdP/dG
Fracture Permeability,” Int. J. Rock Mechanics, Min., Sci., & 4650
150
Geomech. Abstract, Vol. 18, pp. 429-435, 1981.
4600
4. Warpinski, N.R.: “Hydraulic Fracturing in Tight Fissured 100

Media,” JPT (Feb. 1991) 146-152. 50 P 4550

5. Barree, R.D.: “A Practical Numerical Simulator for Three


Dimensional Fracture Propagation in Heterogeneous Media,” 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
4500

SPE 12273, presented at the 1983 SPE Symposium on G Function (α=1.0)

Reservoir Simulation, San Francisco, CA, Nov. 15-18.


Figure 2: G-function plot of pressure falloff data for Case 1 -
fracture extension after shut-in.
8 DETERMINATION OF PRESSURE DEPENDENT LEAKOFF AND ITS EFFECT ON FRACTURE GEOMETRY SPE 36424
1000 4900 500 4700

900 450
4800
GdP/dG 4600

800 400

4700
4500
700 350

Bottomhole Pressure, psi


Pressure Derivatives

Pressure Derivatives
dP/dG

Bottomhole Pressure, psi


4600
600 300
P 4400

500 4500 250

4300
400
dP/dG 4400
200

300 150
P 4200
4300

200 100

4100
4200
100
GdP/dG 50

0 4100 0 4000
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

G Function (α =1.0) G Function (α=1.0)

Figure 3: G-function analysis for Case 2 - normal leakoff with Figure 6: G-function analysis for Case 4 - pressure dependent
no pressure dependence. leakoff and fracture compliance.

700 0.35

4900
Length
600 0.3
4800

4700 t= 0, shut-in 500 0.25

Height, Length, and Rate


4600
Width
Pressure, psi

400 0.2
4500

Width
4400
300 0.15
t=22 minutes
4300

200 0.1
4200
t=40 minutes Height
4100
100 0.05

4000
0 10 20 30 40 50 60 70 Rate
0 0
Distance, feet
0 10 20 30 40 50 60 70

Figure 4: Pressure distribution with fracture length during Pumping Time, minutes

closure for Case 2. Figure 7: Created fracture geometry for Case 5 - growth into
high stress boundaries.

1000 4700

1000 4850
900
4600
900 4800
800

800
4500 4750
700

Bottomhole Pressure, psi


Pressure Derivatives

700
Bottomhole Pressure, psi

P
Pressure Derivatives

4700
600
4400 600 P
4650
500
500 GdP/dG
4300 4600
400
dP/dG 400

4550
300 300
4200
dP/dG 4500
200 200

4100 4450
GdP/dG
100
100

0 4400
0 4000 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
G Function (α=1.0)
G Function (α=1.0)
Figure 8: G-function analysis for Case 5 showing fracture
Figure 5: G-function analysis for Case 3 - pressure dependent
height recession during closure.
leakoff behavior.
SPE 36424 R. D. BARREE AND H. MUKHERJEE 9
1000
4000

30 bpm
20 bpm
10 bpm
3000
Bottomhole Net Pressure, psi

Pressure, psia
2000

1000

0
100
100 1000 10000 100000 0 2 4 6 8 10 12 14 16 18 20
Pumped Volume, gallons Time, hours

Figure 9: Net treating pressure diagnostic plots for three pump Figure 12: Observed pressure during a brine injection test.
rates with pressure dependent leakoff.

10000
100

∆p

1000
Calculated and Observed Ct

∆ p or ∆ p', psi ∆ p'

10

100

10
0.001 0.01 0.1 1 10
1
∆t, hours
4400 4450 4500 4550 4600 4650 4700

Bottomhole Pressure, psi


Figure 13: Log-log pressure diagnostic plot of transient fallof
Figure 10: Pressure dependent leakoff calculated from falloff data from brine injection test.
data compared to actual values.
100
4000

80
0 .1
H eig ht, fe et

60 0

0 .2 3000
40
Pressure, psia

20

2000
0
0 200 400 600

F r a c tu r e H a lf - L e n g th , f e e t
100
1000

80
H eig ht, fe et

0.1

60
0.2 0 0
0 2000 4000 6000 8000
40
Superposition Function

20
Figure 14: Rate superposition analysis of the pseudo radial
0
0 200 400 600
flow portion of the pressure falloff data from brine injection
F ra c tu re H a lf-L e n g th , fe e t
test.
Figure 11: Fracture width contours resulting from pressure
dependent leakoff (top) and 75% rule estimate of CL (bottom).
10 DETERMINATION OF PRESSURE DEPENDENT LEAKOFF AND ITS EFFECT ON FRACTURE GEOMETRY SPE 36424
4000 2000

1800
3500

1600
3000
1400
P

dP/dG and GdP/dG


2500
1200
Pressure, psi

2000 dP/dG 1000

800
1500

600
1000
400

500 GdP/dG 200

0 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
G function

Figure 15: G-function analysis of brine injection falloff data


showing characteristics of pressure dependent leakoff.

You might also like