SPE-189826-MS

DFIT Analysis in Low Leakoff Formations: A Duvernay Case Study

Behnam Zanganeh, Mason K. MacKay, and Christopher R. Clarkson, University of Calgary; Jack R. Jones, NSI
Fracturing LLC

Copyright 2018, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Canada Unconventional Resources Conference held in Calgary, Alberta, Canada, 13-14 March 2018.

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 have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
Diagnostic Fracture Injection Test (DFIT) responses in some shale reservoirs, such as the Duvernay shale
in western Canada, are not consistent with those interpreted through traditional analysis methods. Indeed,
interpretation with traditional techniques may result in significantly incorrect estimates of closure pressure,
pore pressure and formation permeability. The goal of this paper is to explain the observed DFIT behaviours
for selected Duvernay shale wells in terms of low leakoff of fracturing fluid to the formation, activation of
pre-existing fractures, and tip extension during the test.
DFIT data in the Duvernay shale are analyzed using pressure transient analysis methods. Two scenarios
are presented to explain the overall falloff behavior; moving-hinge closure with tip extension, and activation
of secondary natural fractures. The validity of each scenario is examined using rigorous coupled flow-
geomechanical simulation, geological information and geomechanical settings in the Duvernay Formation.
Due to extremely low leakoff, the main mechanism affecting pressure falloff during the DFIT is pressure
dissipation through the primary fracture created during injection. This results in significant tip extension or
activation of secondary fractures. The fluctuations and spikes observed on G-function or pressure derivative
plots are explained in the context of these scenarios. The leakoff rate varies with the pressure change, and
the enhanced fracture surface area, during tip extension. Therefore, the assumption of Carter leakoff, and
the traditional closure picks based on a straight-line tangent to the semi-log derivative on a G-function plot
or 3/2 slope on Bourdet-derivative plot are not valid. Due to very low matrix permeability and the additional
fracture length created through tip extension, it is unlikely that formation radial flow is established during
the test, compromising the ability to obtain a valid pore pressure or formation permeability.

Introduction
A Diagnostic Fracture Injection Test (DFIT), also known as a minifrac or fracture calibration test, has
become the standard pre-stimulation method for determining fracturing treatment properties, effective
reservoir permeability and initial reservoir pressure in unconventional reservoirs. A DFIT is an injection-
falloff test conducted before the main hydraulic fracturing treatment. The DFIT procedure involves
breakdown of the rock and creation of a short fracture during the injection of high pressure fluid (usually
2 SPE-189826-MS

water), and recording of the pressure response for long enough that the created fracture closes, and the
transient associated with this closure event is dissipated before the onset of reservoir flow regimes.
In the following, a brief overview of the Duvernay Formation geological and geomechanical
characteristics is provided, with a focus on those characteristics favoring episodic fracture growth. A
hydraulic fracturing treatment example with microseismic observations is presented to illustrate this
concept. A Theory and Methods section is then provided, including a brief review of DFIT interpretation
methods. The appropriateness of interpreting Duvernay DFITs with the conventional interpretation methods
is then examined. Two simulation models are built to generate synthetic DFIT responses. Finally, in the
Results and Discussion section, field data are interpreted in the context of simulation results.

Geological Overview
Duvernay Formation
The Duvernay Formation is a Devonian aged shale exploited for hydrocarbons within the Western Canadian
Sedimentary Basin (WCSB). Depositional conditions within shales results in marked vertical heterogeneity,
which leads to mechanical complexity (Harris et al. 2011). Within the Duvernay Formation, at least ten
microfacies are identified, each with their own natural fracture fabrics and elastic properties (Knapp et al.
2017). This vertical mechanical heterogeneity favors the creation of bed-contained, sub-vertical fracture
systems (Cooke and Underwood 2010). Natural fractures within the Duvernay are observed and inferred
over a variety of scales from microfractures (Ghanizadeh et al. 2015) to larger fractures (Fox 2015) and
even fault systems (Chopra et al. 2017).
The in-situ stress conditions are interpreted in terms of the orientation and magnitude of the effective
stress. Through basin scale modelling work, the stress regime is thought to be in a strike slip regime in
much of the basin with thrust fault regimes encountered proximal to the Rocky Mountain deformation zone
(Reiter and Heidbach 2014). The regional stress in much of the sedimentary basin trends with an azimuth
of 047 as indicated through borehole breakout directions and other geologic indicators (Reiter et al. 2014).
The Duvernay Formation has a characteristically high organic content with TOC values ranging from 0.1
up to 11 percent (Rivard et al. 2014) which suggests significant potential for hydrocarbon generation. In fact,
it is the source rock for most of the hydrocarbons trapped within the upper Devonian section of the WCSB
(Creaney and Allan 1990). Hydrocarbon generation can raise fluid pressures and may be the underlying
cause of significant overpressure conditions observed within the Duvernay (Fox and Soltanzadeh 2015,
Davis and Karlen 2014). The highly overpressured conditions create a geomechanical sensitivity that brings
discontinuities to a state of incipient failure. This phenomenon is directly observed through reported cases
of induced seismicity arising from treatment operations within the Duvernay Formation (Bao and Eaton
2016; Schultz et al. 2017).
The Duvernay Formation is primarily composed of mudstones and thus is associated with low matrix
permeability in the range of 3.7 × 10−7 to 1.2 mD (Ghanizadeh et al. 2015). This low permeability
suggests that significant leak-off into the matrix is unlikely over the timescale of a DFIT. Instead, fracture
fluid volume may be accommodated through crack tip extension of the hydraulic fractures as well as
accommodation through secondary natural fractures. This process of rupture of the rock and subsequent
flow of fluid into the fracture is theorised to occur in temporally intermittent periods. Evidence of sporadic
spatial-temporal evolution of fluid flow is supported through various case studies of microseismicity in
shales. For example, Goertz-Allman et al. (2017) provide a case study where microseismicity occurs due to
punctual activation of faults which controlled fluid movement in the subsurface.

Evidence for Episodic Fracture Growth in the Duvernay

Evidence of temporally-intermittent fracture behaviour is gleaned from microseismic observations during
the main hydraulic treatment of a nearby well within the Duvernay Formation. The microseismic temporal
SPE-189826-MS 3

evolution shows that the hydraulic fracture does not grow in a simple, steady and continuous way, but rather
the fracture expands to new areas of the reservoir over time. This occurs episodically through growth phases
followed by relative calm. In Fig. 1b, four major microseismic clusters are identified during a hydraulic
fracturing treatment stage, each following a growth front envelope. While there is a trend to follow these
pressure diffusion fronts (Shapiro and Dinske 2009), microseismicity may jump ahead of the theoretical
curve or fall behind, indicating that fracture growth is not continuous and steady. Furthermore, multiple
microseismic clusters form within a single treatment stage, suggesting that rock deformation processes
follow an episodic behavior as well. Fig. 1b also illustrates how the hydraulic fracture accesses different
parts of the reservoir throughout the treatment. The first hydraulic growth cluster produces a large half
length towards the northwest. Subsequently, active deformation occurs closer to the wellbore before finally
moving towards the southeast. Even though continual pumping occurred, the fracture did not continually
grow in length, but rather accessed different areas of the reservoir at different times.

Figure 1—Microseismic clusters showing episodic fracture propagation. (a) Treatment parameters showing
consistent injection (top), yet nonconsistent growth phases (bottom); (b) microseismic events plotted
in plan view and colored by time illustrating episodic spatial temporal growth throughout the reservoir.

This intermittent behavior may be a response to the mechanical layering found within the Duvernay
Formation. Brenner and Gudmundsson (2004) described how hydrofractures become arrested at mechanical
contrasts between layers especially when propagating from a stiff layer into a soft layer. This arresting
behavior would mean that as a fracture grows, it may encounter a difficult layer to propagate into; thus, a
new area of the reservoir will be accessed instead. Additionally, the aperture of the hydraulic fracture varies
depending on the Young's modulus of the surrounding rock matrix and the fluid pressure within the fracture.
Therefore, the hydraulic fracture likely has a non-uniform aperture distribution resulting from the complex
mechanical heterogeneity. The heterogeneous mechanical contrast suggests that a hydraulic fracture will
undergo a period of dilation and extension as it extends in length. This transition between dilation and
rupture likely contributes to episodic fracture growth. Furthermore, soft elastic rock will be able to store
elastic potential from dilation of the fracture during injection. This mechanical potential energy can be a
driving mechanism for growth after the injection has ceased if the pressure was not able to leak off due to low
matrix permeability. This effect has been observed through the generation of Krauklis waves, where fracture
wall deformation coupled to fluid flow creates resonances within the hydraulically connected fluid network
(Krauklis 1962). In this case, the dilation and subsequent collapse of fracture aperture is coupled to fluid
movement within the fracture (Liang 2017). The time scale that this process occurs on is partially dependent
on the elasticity of the surrounding rock mass, where low shear modulus values of the rock mass produces
4 SPE-189826-MS

larger time period fluid movement events (Tary et al. 2014). This is because the rock can accommodate
more elastic deformation and thus larger apertures may be reached before the fracture constricts.
Abundant vertical natural fractures are observed in image logs taken from a horizontal leg within the
Duvernay Formation (Fig. 2a). These natural fractures occur in specific orientations called sets. The first set
is an extension fracture and follows the orientation of maximum horizontal stress. The second set is a shear
fracture and occurs at an angle to the main set. The third set is a cross joint perpendicular to the maximum
horizontal stress. To calculate the resolved normal and shear stress on the natural fractures, an understanding
of the in-situ effective stress field is required. In this case, we use a vertical stress based on 24 KPa/m and
a pressure gradient of 20 KPa/m located at 3 km depth. We choose the minimum horizontal stress to be 0.8
times the vertical while the maximum horizontal stress is 1.5 times the vertical stress. Using tensor rotations,
the normal and shear stresses are resolved for each fracture set and shown on a Mohr's circle (Fig. 2b).

Figure 2—(a) Stereonet representation of natural fracture orientations observed from image logs in the horizontal leg
within the Duvernay Formation. Fracture planes are shown as great circles while the poles to the fractures are plotted
as points. (b) Mohr's circle representation of normal and shear stresses resolved onto fractures under the estimated
in-situ stress conditions. A Mohr-Coulomb envelope with no cohesion and 20 degree friction angle is plotted to show
how close to failure the fracture system is. (c) Natural fracture system within the Duvernay Formation as exposed in
outcrop. Fluid alteration (steel blue) follows the fracture network with some leakoff occurring into the rock matrix.

As fluid pressure is increased during injection, the normal effective stress in the fracture system decreases
within the fluid front. The extension fracture will accept fluid most easily as it has the least resolved normal
stress; however, when the linking cross joint or shear joint are encountered, they will transmit fluid from one
fracture to another. Before they can transmit the fluid, there must be a buildup of fluid pressure because they
require a higher normal stress to open. Once the fluid builds up to sufficient level to overcome the normal
stress, fluid may be transmitted to another extension fracture which in turn will release the pressure to a
lower state. Thus, flow through the fracture system itself leads to episodic buildup and release of pressure
as different fracture systems are accessed. Fig. 2c illustrates this effect in a Duvernay Formation equivalent
outcrop exposure at Roche Miette in Jasper National Park, Alberta. The natural fracture system is comprised
of multiple connecting fracture sets in which fluid flow pathways are established. The low permeability of
SPE-189826-MS 5

the matrix concentrates fluid into the fracture system as observed from the mineralogical alteration front
following the natural fracture system.
From the above summary, it is concluded that the Duvernay possesses several properties that may
complicate DFIT analysis: low matrix permeability, limiting leakoff from fractures; an extensive natural
fracture system; and fine-scale mechanical layering, the latter two properties increasing the propensity
for intermittent fracture growth that could affect the DFIT signature. Direct observations of this behavior
are illustrated through episodic spatial-temporal clustering behavior of microseismicity during a hydraulic
fracture treatment.

Theory and Methods

DFIT Interpretation Methods
DFIT analysis was pioneered by Nolte (1979), who introduced the Gfunction, based on material balance
and the Carter leakoff model (Howard and Fast 1957) for before closure analysis. Gu et al. (1993) and Nolte
et al. (1997) demonstrated the possibility of observing afterclosure reservoir linear and pseudo-radial flow
and their application in estimation of fracture geometry and reservoir properties.
Barree and Mukherjee (1996) and Barree (1998) recommended using the G-function and its diagnostic
derivatives to identify fracture closure. Barree (1998) and Barree et al. (2009) also presented
signatures for non-ideal leakoff behaviors such as tip extension, pressure-dependent leakoff (PDL) and
height recession/transverse storage based on G-function, square-root of shut-in time, and log-log plot of
pressure change with shut-in time plots. Fig. 3 summarizes normal leakoff and nonideal leakoff signatures
on G-function combination plots based on the Barree at al. (2009) interpretation methodology.

Figure 3—(a) Normal leakoff; (b) PDL; (c) tip extension; (d) storage/height
recession signatures on G-function combination plots (Barree et al. 2009).
6 SPE-189826-MS

Mohamed et al. (2011) and Marongiu-Porcu et al. (2011) presented a model to predict the falloff pressure
trend of an idealized DFIT (normal leakoff) using standard pressure transient (PTA) log-log diagnostic plots.
A log-log diagnostic plot of the basic falloff response shape predicted by their model is shown in Fig. 4.
They identified normal leakoff as a 3/2 slope on the Bourdet-derivative (with respect to superposition time),
and picked fracture closure as the deviation from the 3/2 slope.

Figure 4—Log-log diagnostic plot for an idealized fracture-injection/falloff test (Marongiu-Porcu et al. 2011).

Bachman et al. (2012) presented a combination of conventional diagnostic PTA plots to identify various
flow regimes before and after fracture closure, summarized in Table 1. They demonstrated that a 3/2 on
Bourdet-derivative corresponds to a straight line trend on G-function derivatives, representing Carter leakoff
(Nolte flow).

Table 1—Identification of flow regimes using derivative slope on log-log plot

of pressure difference versus equivalent time functions (Bachman et al. 2012)

Log-Log Flow Regime!

Carter Linear Bilinear

Equivalent Time Early Late Early Late Early Late

Derivative Time Time Time Time Time Time
Function Slope Slope Slope Slope Slope Slope
------- ------- ------- ------- ------- -------
Ends Starts Ends Starts Ends Starts
tD=0.08 tD=5.0 tD=0.04 tD=4.0 tD=0.13 /n=8.0

terdΔP/dter 1/1 3/2 1/2 1/2 1/4 1/4

(Radial)

ΔtdΔP/dtec 1/1 1/1 1/2 0 1/4 -1/4

(Carter)

Δt0.5dΔP/dtel 1/1 3/2 1/2 1/2 1/4 1/4

(Linear)

Δt0.25dΔP/dteb 1/1 3/2 1/2 1/2 1/4 1/4

(Bilinear)
SPE-189826-MS 7

Recently, Zanganeh et al. (2018) explained the spectrum of flow patterns and signatures observed before
and after closure during DFITs. Their interpretation was based on rigorous numerical simulation which
included the dynamic nature of fracture geometry (i.e. fracture expansion, tip extension, hinge closure,
progressive closure), variable leakoff rate (as a function of fracture surface area, fracture pressure, and
falloff time) and relative magnitude of afterflow with respect to total leakoff rate. They showed that with
DFITs, unlike conventional buildup/falloff tests, afterflow caused by wellbore storage can last for a much
longer period of time, affecting pressure transient in various time periods during the pressure falloff.

Problems with Application of Conventional DFIT Analysis Methods to the Duvernay

Some DFIT responses in shale reservoirs, such as the Duvernay shale, are not consistent with those
interpreted through traditional analysis methods. Fig. 5 provides pressure profiles during injection for two
DFITs conducted in the Duvernay Formation. There is no clear breakdown, or a sharp pressure drop after
the breakdown, as expected in more conventional reservoirs. In fact, in some cases there is an increasing
pressure trend after the breakdown even though the injection rate is constant or decreasing (see Field
Example 2, Fig. 5b). Another common signature observed in Duvernay DFITs is a large pressure drop at the
time of shut-in, e.g. 20 MPa and 25 MPa for Field Example 1 and 2, respectively. This significant pressure
drop is likely caused by another mechanism other than friction in the wellbore or tortuosity.

Figure 5—Pressure profile during injection for (a) Field Example 1; (b) Field Example 2.

PTA and G-function diagnostic plots for Field Example 1 are provided in Fig. 6. The trend of
curve is not similar to any of the signatures proposed by Barree et al. (2009). Except for the early time
unit slope, no other straight line trend with a specific slope is observed on any of the derivative plots.
According to Zanganeh et al. (2018), the long transition after unit slope, with no sharp characteristic hump,
indicates that leakoff from fracture to surrounding formation is not dominant. This is probably due to very
low matrix permeability of the formation coupled with small fracture surface area of the created fracture.
Several fluctuations are observed on all of the derivative plots, each lasting for a considerable period of
time. In this case, there is no signature of fracture closure based on any of the identification methods.
8 SPE-189826-MS

Figure 6—G-function and PTA diagnostic plots for Field Example 1.

Based on the geological information, geomechanical settings and evidence for episodic hydraulic fracture
growth in the Duvernay Formation provided in the previous section, combined with the anomalous DFIT
observations just illustrated, two hypotheses are presented to explain the DFIT behavior in the Duvernay
formation; moving-hinge closure with tip extension, and propagation through activation of secondary
fractures. As will be discussed below, model-generated synthetic results are compared with field data, and
necessary modifications are made to calibrate the model. The observations from the calibrated model are
then used to explain field behaviors.

Model Description and Setup

A customized fully-coupled stress-pore pressure simulator (Abaqus Analysis User's Guide 2016) is used
herein to generate synthetic DFIT responses. The customized model is capable of simulating all the physical
processes involved in a typical DFIT including: porous media deformation; fluid flow inside the reservoir;
hydraulic fracture initiation, propagation and closure (based on the cohesive zone method); compliance
change before and after closure; residual fracture aperture and conductivity; and fluid flow inside the fracture
and fluid interaction between the fracture and reservoir (leakoff).
The modeling procedure is described in detail in Zanganeh et al. (2017) and Zanganeh et al. (2018).
The cohesive zone method (CZM) is used to model hydraulic fracture initiation, propagation and closure.
With the CZM, the fracture is modeled as a gradual separation between two material (rock) surfaces. This
separation is modeled as a progressive degradation of cohesive strength along the cohesive layer, which
is a pre-defined surface embedded in the rock and follows a traction-separation law. Model 1 assumes
propagation of a single planar primary fracture (HF; Fig. 7a). However, in Model 2, there is a pre-existing
network of natural fractures (NF; Fig. 7b), similar to the system observed in the Duvernay outcrop (Fig.
2c). The cohesive elements are embedded in the formation rock - they are assumed to be non-existent in the
model until fracture initiation and propagation criteria are reached, at which time these elements act as the
potential pathway for fracture growth. In both simulation models, a matrix permeability of 25 nd was used.
SPE-189826-MS 9

Figure 7—Simplified schematics of (a) Model 1; (b) Model 2.

In the previous simulation models of Zanganeh et al. (2017 and 2018), tangential fluid flow inside the
fracture is calculated using Poiseuille's law:

(1)

where w is the fracture opening, μ is the fluid viscosity and Pf is the fluid pressure along the fracture length
(x direction). For an injection fluid with low viscosity (i.e. water), even a small fracture aperture of 0.1
mm, results in a fracture with large tangential permeability (conductivity) without any significant pressure
gradient inside the fracture. To achieve synthetic pressure responses similar to the presented field examples
in the Duvernay Formation, especially during pumping and early shut-in time, the β coefficient is introduced
into Poiseuille's formula (Eq. 1). After several sensitivity simulation runs, a β value of 1000 was selected
as the optimum value.

Results and Discussion

Model 1: Tip Extension
Fig. 8a provides the simulated pressure profile, G-function semi-log derivative and PTA plots during
pressure falloff for Model 1. The overall trends of the curves are similar to the Field Example 1 (Fig. 6).
At the time of breakdown, the created fracture has finite conductivity with a considerable pressure gradient
inside the fracture. This results in the sudden pressure drop at the time of shut-in.
10 SPE-189826-MS

Figure 8—(a) PTA diagnostic plots; (b) pressure profile and G dP/
dG curves. Tip extension phases are shown with the dotted squares.

As shown by Zanganeh et al. (2018), moving hinge closure occurs during the transition from afterflow
dominance (caused by wellbore storage) to leakoff dominance. Due to low matrix permeability and the
pressure gradient inside the fracture in the Model 1, coupled with the presence of afterflow, the pressure
front moves to the tip of fracture and causes tip extension of the primary fracture. Pressurization and
depressurization during tip extension can be repeated several times (Fig. 9) during the transition period
before leakoff dominates the falloff process. Tip extension phases are shown as fluctuations on pressure
derivative plots. It must be noted that in the simulation models, the magnitude and duration of these
fluctuations are controlled by the size of fracture elements which is uniform throughout the length of
fracture. However, in reality, the magnitude and duration of each tip extension phase can be different.

Figure 9—Plan view of a single wing of the fracture showing pressure gradient inside the
fracture and tip extension phases during falloff. The fracture aperture is magnified 1000 times.

In Model 1, leakoff dominance starts after 4.5 log-cycles and fracture closure does not occur.
SPE-189826-MS 11

Model 2: Pre-existing Fractures and Tip Extension

As mentioned in the model description and setup section, Model 2 uses a pre-existing network of natural
fractures (NF), similar to the system observed in the Duvernay outcrop. Pre-existing natural fractures can
be activated during injection and falloff and affect propagation direction and pressure response. Fig. 10a
illustrates the pressure profile during injection and early shut-in period. There is no significant pressure
drop at the time of breakdown due to the low conductivity of the primary fracture. As the primary fracture
hits the pre-existing fracture, propagation temporarily stops (upward trend on pressure profile) until the pre-
existing fracture is activated. Then, the propagation continues in the original direction perpendicular to the
minimum horizontal stress (Fig. 10b).

Figure 10—(a) Pressure profile during injection and early shut-in time for Model 2; (b) Propagation as the
primary fracture hits and activates a preexisting fracture (fracture apertures are magnified 1000 times).

The falloff behavior is similar to Model 1. Given the low matrix permeability and leakoff, reactivation
of secondary fractures can happen during the falloff period, too. The response on pressure derivative plots
will be similar to tip extension phases as illustrated in Model 1.

Field Example 1
The observed trends of Field Example 1 in Fig. 6 can be explained based on the simulation Model 1. The
absence of a clear breakdown and the significant pressure drop at the time of shut-in (Fig. 5a) is caused by
low conductivity and the large pressure gradient inside the created fracture.
The PTA plots for this dataset are revisited in Fig. 11. The long transition after the early time unit slope,
with no sharp characteristic hump, indicates that leakoff from the fracture to surrounding formation is
not dominant. This is due to low matrix permeability of the formation coupled with low conductivity and
pressure gradient inside the fracture. Three major extension cycles are observed during the falloff period
demonstrated by the fluctuations and inflection points. The pressure changes at the end of fall period are
very small as the leakoff is low and no additional extension occurs. As a result, the end of falloff period
is highly affected by ambient temperature changes. No signature of fracture closure is observed after 200
hours. Therefore, no reliable estimate of closure pressure and formation permeability can be made in this
example.
12 SPE-189826-MS

Figure 11—PTA plots for Field Example 1 showing the interpretation based on tip extension cycles.

Field Example 2
In this DFIT, a total volume of 6.8 m3 of fresh water was injected during 3 minutes of pumping. The pressure
values were recorded using surface gauges. The increasing trend after breakdown while the injection rate
is kept constant (Fig. 5b) indicates the activation of secondary fractures in addition to the primary fracture
during the injection. Again, there is a long transition of about 2 log-cycles after the early unit slope. There are
two major opening cycles during this transition, caused by activation of secondary fractures or extension of
primary fracture during falloff, which creates additional fracture surface area for fluid leakoff from fracture
system to the formation. As a result, leakoff dominance starts after about 2.2 hours. The slope of and

plots during leakoff dominance is smaller than 3/2 and 1/2, respectively; and there is no evidence of
Carter leakoff. The closure is picked at the inflection point on curve that corresponds to change of

slope on PPD and curves, with closure pressure of 32.1 MPa and closure time of 51.4 hours. Formation
linear or radial flow are not observed after 300 hours of shut-in time, and no reliable estimate of formation
permeability or initial pressure can be made.

Figure 12—PTA plots for Field Example 2.

SPE-189826-MS 13

Conclusions
DFIT responses in the Duvernay Formation for the cases studied are controlled by the following
mechanisms:

• Episodic fracture growth during injection and falloff. The hydraulic fracture does not grow in a
simple, steady and continuous way. Growth occurs episodically through growth phases followed
by relative calm.
• Low permeability of the matrix. This property delays the transition to leakoff dominance in the
falloff period, and results in long fracture closure times. Also, low leakoff favors the episodic
fracture growth (moving hinge-closure and tip extension) as it provides a mechanism to release
pressure and create additional fracture surface area.
• Long lasting afterflow (caused by wellbore storage) and transition period due to limited leakoff.

• Low conductivity fractures. The fractures, both primary and pre-existing, have a low conductivity
resulting in pressure gradient within the fractures. This explains the unclear breakdown point and
pressure responses after the breakdown.
The above mechanisms result in complex leakoff and pressure behavior. Therefore, the assumption of
Carter leakoff, and the traditional fracture closure picks based on a straight-line tangent to the semi-log
derivative on a G-function semi-log derivative plot or 3/2 slope on Bourdet-derivative plot are not reliable.
Finally, it is very unlikely for the formation radial flow to occur during short falloff periods of days or weeks.

Nomenclature

Field Variables
GG-function time, dimensionless
mSlope of straight line, dimensionless
PPressure, Pa
Pf
Fracture pressure (fluid pressure in the cohesive element), Pa
qfrac Gap flow rate within the cohesive element, m3/sec
Shmax Maximum horizontal stress, Pa
Shmin Minimum horizontal stress, Pa
ta Agarwal's time, dimensionless
tp Pumping time, sec
w Fracture opening (aperture), m

Greek Variables
β Coefficient in the Poiseuille's formula, dimensionless
Δt Shut-in time, sec
μ Fracturing fluid viscosity, cp

References
Abaqus Analysis User's Guide 2016. Dassault Systemes Simulia Corp., Providence, RI.
Agarwal, R. G. 1980. A New Method to Account for Producing Time Effects when Drawdown Type Curves are Used to
Analyze Pressure Buildup and Other Test Data. Presented at the SPE Annual Technical Conference and Exhibition,
Dallas, Texas, 21-24 September. SPE-9289-MS. http://dx.doi.org/10.2118/9289MS
Bachman, R. C., Walters, D. A., Hawkes, R. A. et al 2012. Reappraisal of the G Time Concept in Mini-Frac Analysis.
Presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, 8-10 October. SPE-160169-
MS. http://dx.doi.org/10.2118/160169MS
14 SPE-189826-MS

Bao, X. and Eaton, D. W. 2016. Fault Activation by Hydraulic Fracturing in Western Canada. Science 354:1406-1409.
https://doi.org/10.1126/science.aag2583
Barree, R. D. and Mukherjee, H. 1996. Determination of Pressure Dependent Leakoff and Its Effect on Fracture Geometry.
Presented at the SPE Annual Technical Conference and Exhibition, Denver, Colorado, 6-9 October. SPE-36424-MS.
http://dx.doi.org/10.2118/36424MS.
Barree, R. D. 1998. Applications of Pre-Frac Injection/Falloff Tests in Fissured Reservoirs-Field Examples. Presented
at the SPE Rocky Mountain Regional/Low Permeability Reservoirs Symposium, Denver, Colorado, 5-8 April.
SPE-39932-MS. http://dx.doi.org/10.2118/39932MS.
Barree, R. D., Baree, V. L., and Craig, D. P. 2009. Holistic Fracture Diagnostics: Consistent Interpretation of Prefrac
Injection Tests Using Multiple Analysis Methods. SPE Production & Operations 24(3): 396-406. SPE-107877-PA.
http://dx.doi.org/10.2118/107877PA
Brenner, S. L. and Gudmundsson, A. 2004. Arrest and Aperture Variation of Hydrofractures in Layered Reservoirs.
Geological Society, London, Special Publications 231(1): 117-128. https://doi.org/10.1144/GSL.SP.2004.231.01.08
Chopra, S., Sharma, R. K., Ray, A. K. et al 2017. Seismic Reservoir Characterization of Duvernay Shale With
Quantitative Interpretation and Induced Seismicity Considerations—A case study. Interpretation 5(2): T185-T197.
https://doi.org/10.1190/INT20160130.1
Cooke, M. L. and Underwood, C. A. 2001. Fracture Termination and Step-Over at Bedding Interfaces Due to Frictional Slip
and Interface Opening. Journal of Structural Geology 23: 223–238. https://doi.org/10.1016/S01918141(00)000924
Creaney, S and Allan, J. 1990. Hydrocarbon Generation and Migration in the Western Canada Sedimentary Basin.
Geological Society, London, Special Publications 50: 189–203.
Davis, M. and Karlen, G. 2014. A Regional Assessment of the Duvernay Formation: A World-Class Liquids-Rich Shale
Play: Geoconvention, Calgary, Alberta, Canada.
Fox, A., Soltanzadeh, M., Watson, N. et al 2015. The Link Between Lithology and Rock Fractures in the Duvernay. 2015
Gussow Conference: Fine-Grained Rocks: Resources to Reserves, Banff, Alberta, Canada, October 13-15.
Fox, A. D. and Soltanzadeh, M. 2015, A Regional Geomechanical Study of the Duvernay Formation. Geoconvention,
Calgary, Alberta, Canada.
Ghanizadeh, A., Bhowmik, S., Haeri-Ardakani, O. et al 2015. A Comparison of Shale Permeability Coefficients
Derived Using Multiple Non-Steady-State Measurement Techniques: Examples from the Duvernay Formation, Alberta
(Canada). Fuel 140: 371–387. https://doi.org/10.1016/j.fuel.2014.09.073
Goertz-Allmann, B. P., Gibbons, S. J., Oye, V. et al 2017. Characterization of Induced Seismicity Patterns
Derived from Internal Structure in Event Clusters. J. Geophys. Res. Solid Earth 122: 3875–3894. http://
dx.doi.org/10.1002/2016JB013731
Gu, H., Elbel, J. L., Nolte, K. G. et al 1993. Formation Permeability Determination Using Impulse-Fracture Injection.
Presented at the SPE Production Operations Symposium, Oklahoma City, Oklahoma, 21-23 March. SPE-25425-MS.
https://doi.org/10.2118/25425MS
Harris, N. B., Miskimins, J. L. and Mnich, C. A. 2011. Mechanical Anisotropy in the Woodford Shale, Permian Basin:
Origin, Magnitude, and Scale. The Leading Edge 30(3): 284-291. https://doi.org/10.1190/1.3567259
Howard, C. C., and Fast, C. R. 1957. Optimum Fluid Characteristics for Fracture Extension. API Drilling and Production
Practice 24(1): 261-270. API-57-261.
Knapp, L. J., McMillan, J. M and Harris, N. B. 2017. A Depositional Model for Organic-Rich Duvernay Formation
Mudstones. Sedimentary Geology 347: 160–82. https://doi.org/10.1016/j.sedgeo.2016.11.012
Krauklis, P. V. 1962. On Some Low-Frequency Vibrations of A Liquid Layer in An Elastic Medium. J. App. Mathematics
and Mechanics 26(6): 1685-1692. https://doi.org/10.1016/00218928(62)902034
Liang, C., O'Reilly, O., Dunham, E. M. et al 2017. Hydraulic Fracture Diagnostics from Krauklis-Wave Resonance and
Tube-Wave Reflections. Geophysics 82(3):D171-D186. https://doi.org/10.1190/geo20160480.1
Marongiu-Porcu, M., Ehlig-Economides, C. A. and Economides, M. J. 2011. Global Model for Fracture Falloff Analysis.
Presented at the SPE North American Unconventional Gas Conference and Exhibition, The Woodlands, Texas, 14-16
June. SPE-144028-MS. http://dx.doi.org/10.2118/144028MS
Mohamed, I. M., Nasralla, R. A., Sayad, M. A. et al 2011. Evaluation of After-closure Analysis Techniques for Tight
and Shale Gas Formations. Presented at the SPE Hydraulic Fracturing Technology Conference and Exhibition, The
Woodlands, Texas, 24-26 January. SPE-140136-MS. http://dx.doi.org/10.2118/140136MS
Nolte, K. G. 1979. Determination of Fracture Parameters From Fracturing Pressure Decline. Presented at the
SPE Annual Technical Conference and Exhibition, Las Vegas, Nevada, 23-26 September. SPE-8341-MS. http://
dx.doi.org/10.2118/8341MS
Nolte, K. G., Maniere, J. L., and Owens, K. A. 1997. After-Closure Analysis of Fracture Calibration Tests. Presented
at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, 5-8 October. SPE-38676-MS. https://
doi.org/10.2118/38676MS
SPE-189826-MS 15

Reiter, K., O. Heidbach, D. Schmitt, K. et al 2014. A Revised Crustal Stress Orientation Database for Canada.
Tectonophysics 636: 111–124. https://doi.org/10.1016/j.tecto.2014.08.006
Rivard, C., Lavoie, D., Lefebvre, R. et al 2014. An Overview of Canadian Shale Gas Production and Environmental
Concerns. International Journal of Coal Geology 126: 64-76. https://doi.org/10.1016/j.coal.2013.12.004
Schultz, R., R. Wang, Y. J. Gu, et al 2017. A Seismological Overview of the Induced Earthquakes in the Duvernay Play
near Fox Creek, Alberta. J. Geophys. Res. Solid Earth 122: 492-505. http://dx.doi.org/10.1002/2016JB013570
Shapiro, S. A. and Dinske, C. 2009. Fluid Induced Seismicity: Pressure Diffusion and Hydraulic Fracturing. Geophysical
Prospecting 57(2): 301-310. http://dx.doi.org/10.1111/j.13652478.2008.00770.x
Tary, J., Baan, M. and Eaton, D. 2014. Interpretation of Resonance Frequencies Recorded During Hydraulic Fracturing
Treatments. J. Geophys. Res. Solid Earth 119: 1295-1315. http://dx.doi.org/10.1002/2013JB010904
Zanganeh, B., Clarkson C. R., and Hawkes, R. V. 2017. Reinterpretation of Fracture Closure Dynamics During Diagnostic
Fracture Injection Tests. Presented at the SPE Western Regional Meeting, Bakersfield, California, 23-27 April.
SPE-185649-MS. https://doi.org/10.2118/185649MS
Zanganeh, B., Clarkson C. R., and Jones, J. R. 2018. Reinterpretation of Flow Patterns During DFITs Based on Dynamic
Fracture Geometry, Leakoff and Afterflow. Presented at the SPE Hydraulic Fracturing Technology Conference, The
Woodlands, Texas, 23-25 January. SPE-189840-MS. https://doi.org/10.2118/189840MS