SPE 166332-MS

Characterizing Hydraulic Fracturing with a Tendency for Shear Stimulation

Test
Mark McClure, SPE, University of Texas at Austin, Roland Horne, SPE, Stanford University

Copyright 2013, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Annual Technical Conference and Exhibition held in New Orleans, Louisiana, USA, 30 September–2 October 2013.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by t he 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 mus t contain conspicuous acknowledgment of SPE copyright.

Abstract
The classical concept of hydraulic fracturing is that a single, planar, opening mode fracture forms. In recent years, there has
been a growing consensus that in many formations, natural fractures play an important role during stimulation. There is not
universal agreement on the mechanisms by which natural fractures affect stimulation, and these mechanisms vary depending
on formation properties. One potentially important mechanism is shear stimulation, where an increase in fluid pressure
induces slip and permeability enhancement on preexisting fractures. We propose a Tendency for Shear Stimulation (TSS) test
as a direct, relatively unambiguous method for determining the degree to which shear stimulation contributes to stimulation in
a particular formation. In the TSS test, fluid is injected at a bottomhole pressure that is intentionally maintained below the
minimum principal stress, ideally at a constant pressure. Under these conditions, shear stimulation is the only possible
mechanism for permeability enhancement (except perhaps thermally induced tensile fracturing). Standard pressure transient
tests could be performed before and after the TSS test to estimate formation permeability. The flow rate rate transient during
injection may also be interpreted to identify shear stimulation. Numerical simulations of shear stimulation were performed
with a discrete fracture network model that couples fluid flow with the stresses induced by fracture deformation. These
simulations were used to qualitatively investigate how shear stimulation and fracture connectivity affect the results of a TSS
test. The simulations neglected matrix flow and were two-dimensional, which made it impossible to forward simulate the
transients that would be expected in practical application. Modeling improvements will make this possible in future work.
Two specific field projects are discussed as examples of a TSS test, the Enhanced Geothermal System (EGS) projects at Desert
Peak and Soultz-sous-Forêts. At Soultz, the formation had a high TSS, and at Desert Peak, formation TSS was minimal.

Introduction
Classically, hydraulic fracturing has been conceptualized as creating a single, planar, opening mode tensile fracture. But in
low matrix permeability applications such as oil or gas production from shale or geothermal production from granite, the
process of hydraulic stimulation has been conceptualized as creating a complex network of newly forming fractures and/or
natural fractures that slip and open in response to injection (Fisher et al., 2004; Bowker, 2007; Gale et al., 2007; Cipolla et al.,
2008; King, 2010; Pine and Batchelor, 1984; Murphy and Fehler, 1986; Brown, 1989; Ito, 2003; Ito and Hayashi, 2003; Evans,
Moriya, et al., 2005; Ledésert et al., 2010). The precise geometry of these networks is a major uncertainty. The networks
cannot easily be observed directly in the subsurface; it is difficult to know how laboratory experiments relate to the reservoir
scale, and microseismic interpretations with respect to network geometry are nonunique.

In shale, it is widely believed that new fractures form and propagate through the formation, but apparently there is
disagreement about the role of preexisting fractures and how they contribute to production.

One potentially important process is termination of propagating natural fractures against preexisting fractures. This has been
observed in laboratory experiments (Blanton, 1982; Renshaw and Pollard, 1995; Zhou et al., 2008; Gu et al., 2011), mine-back
experiments (Warpinski and Teufel, 1987; Warpinski et al., 1993; Mahrer, 1999; Jeffrey et al., 2009), and computational
investigations (Dahi-Taleghani and Olson, 2009; Gu and Weng, 2010; Fu et al., 2012). If termination occurs, then it may be
difficult for a single, continuous, large fracture to propagate across the formation, and pathways for flow through the reservoir
may occur through both new and preexisting fractures (a process we refer to as Mixed-Mechanism Stimulation, MMS). This
process could play an important role in generating stimulated fracture surface area and therefore increasing recovery. MMS
2 SPE 166332-MS

has been modeled at the field scale by several authors (Damjanac et al., 2010; Weng et al., 2011; Wu et al., 2012; Section
2.3.3.2 of McClure, 2012).

In contrast to the MMS concept, some authors have modeled stimulation in shale with a single, large, continuous tensile
fracture per stage (Warpinski et al., 2001; Palmer et al., 2007; Rogers et al., 2010; Nagel et al., 2011; Roussel and Sharma,
2011). With this approach, a problem arises because high production rates from low permeability shale formations apparently
require more stimulated fracture surface area than can be explained by a single, linear fracture per stage (Mayerhofer et al.,
2010; Fan et al., 2010; Cipolla et al., 2010). This problem can be resolved by postulating that there are secondary fractures
surrounding the large primary fracture that contribute significantly to production. Some modelers are agnostic about the
nature of the secondary fractures (Warpinski et al., 2001), and others suggest the secondary fractures may be newly forming
tensile fractures (Roussel and Sharma, 2011). Other authors have modeled secondary fractures as shear stimulating (or
possibly opening) preexisting fractures (Palmer et al., 2007; Rogers et al., 2010; Nagel et al., 2011). We refer to this latter
mechanism as Primary Fracturing with Shear Stimulation Leakoff (PFSSL). Other combinations of mechanisms could be
Primary Fracturing with Mixed-Mechanism Leakoff or Primary Fracturing with Secondary Tensile Fracturing.

Projects where hydraulic stimulation is performed for geothermal energy production are often referred to as "Enhanced
Geothermal Systems," or EGS. The original concept of EGS, which remains the most common, was to perform hydraulic
fracturing in deep wells drilled into the crystalline basement (3 - 5 km depth). The potential size of the geothermal resource is
very large, because temperatures at this depth are high enough for geothermal energy production across a significant
percentage of the Earth's surface. However, the natural permeability in the crystalline basement is typically very low, and so
hydraulic stimulation is needed to improve economic performance (Tester, 2007). EGS is most often performed by injecting
liquid water into a long openhole section of a wellbore (vertical or gently inclined) without using proppant.

Most authors believe that during EGS stimulation in crystalline formations, the dominant mechanism of stimulation is induced
slip on preexisting fractures (which we refer to as Pure Shear Stimulation, PSS). Early EGS experience indicated that flow
from the wellbores was localized at preexisting fractures and that microseismic hypocenter locations were not consistent with
the vertical, penny-shaped tensile crack model of hydraulic fracturing (Murphy and Fehler, 1986; Pine and Batchelor, 1984).
Some early authors postulated a "dendritic" network model of stimulation (Murphy and Fehler, 1986), and authors eventually
converged on the shear stimulation idea, at the exclusion of new propagating fractures (Pine and Batchelor, 1984; Lanyon et
al., 1993; Willis-Richard et al., 1996; Ito and Hayashi, 2003; Tester, 2007). Since then, most EGS modeling has used this
conceptual model. The modeling workflow is to stochastically generate a realization of the preexisting fracture network and
then simulate flow and shear stimulation in the network (Kohl and Hopkirk, 1995; Bruel, 1995; Bruel, 2007; Kohl and Mégel,
2007; Rahman et al., 2002; Jing et al., 2000; Cladouhos et al., 2011; Zhou and Ghassemi, 2011; Rachez and Gentier, 2010;
Riahi and Damjanac, 2013).

McClure and Horne (2013) argued that the Pure Shear Stimulation is not as universal in crystalline EGS as the conventional
wisdom suggests, and that in many or most projects, Mixed-Mechanism Stimulation occurs, with both new and preexisting
fractures playing an important role. McClure and Horne (2013) pointed out that many geological conditions must be present
for Pure Shear Stimulation to be possible, and these conditions cannot always be expected to be satisfied. Furthermore, in
most EGS projects, the bottomhole pressure has exceeded the minimum principal stress, for example, the projects at Hijiori,
Fjallbacka, Le Mayet, Rosemanowes, Fenton Hill (Willis-Richards et al, 1995), Ogachi (assuming vertical stress gradient of 25
MPa/km) (Kaieda et al., 2010), Desert Peak (Chabora et al., 2012), and Groβ Schonebeck (Zimmermann et al., 2008).

Overall, shear stimulation is suspected of being important for hydraulic stimulation in both shale and geothermal resources, but
the role of shear stimulation is not fully understood. If shear stimulation contributes significantly to production in shale, this
would have important implications for stimulation design and formation assessment. TSS should depend on various
geological parameters, and geologists need to know whether to incorporate these variables into their resource assessments.
Modelers need to know whether to incorporate shear stimulation into their simulations. Generally, concepts about shear
stimulation shape the physical intuition of engineers making design decisions.

In geothermal, most EGS modeling neglects propagation of new fractures, and recent EGS projects have been focused on
trying to intersect large faults because they are seen as being most favorable for shear stimulation. If tensile fracturing
(perhaps Mixed-Mechanism Stimulation) was reconsidered as a viable mechanism, it would open up new possibilities for
overall system design (such as using proppant).

The tendency for shear stimulation (TSS) must depend on many geological factors. At a minimum, shear stimulation (either
PSS or PFSSL) cannot occur without self-propping fractures that are well oriented in the local stress state to slip in response to
fluid increase. For Pure Shear Stimulation to occur, the formation must be capable of accepting fluid rapidly enough to
prevent excessive pressure buildup, which would result in propagation of new fractures (McClure and Horne, 2013). The
importance (or lack therof) of shear stimulation in hydraulic stimulation must depend on the geological parameters, but these
SPE 166332-MS 3

relationships are not well defined, and it is not clear how the importance of shear stimulation for a formation could be
measured or quantified.

Because the tendency for shear stimulation to occur depends on geological setting, it could be considered a formation property.
TSS could be estimated through indirect analyses such as Coulomb stress analyses of the local fracture network (Evans, 2005;
Ito and Hayashi, 2003) and rock mechanics testing of core to evaluate the tendency for self-propping (Lutz et al., 2010). But
ultimately, these techniques require interpretation, there are important factors that could be difficult to characterize, and it is
unclear how different factors could be integrated into a single assessment. It may be possible to develop methodologies to
assess TSS, but these efforts will be hampered without a way to calibrate them with direct measurements of TSS.

We propose that a "Tendency to Shear Stimulation Test" (TSS Test) could be used in the field to evaluate the importance of
shear stimulation practically and unambiguously. In a TSS Test, fluid would be injected into an openhole section of the
wellbore at a fluid pressure near to, but less than, the minimum principal stress. Under these conditions, shear stimulation
would be the only possible mechanism of stimulation. The formation permeability could be measured with pressure transient
techniques before and after the stimulation treatment, and the change in permeability could be used to assess TSS. The
behavior of the injection rate over time during the TSS test could be used to diagnose TSS and possibly some additional
information about the properties of the formation.

To investigate the properties of TSS Tests, three simulations were performed using CFRAC, a two-dimensional discrete
fracture network simulator that fully couples fluid flow and the stresses induced by fracture deformation (McClure, 2012).
The simulations were configured to loosely resemble the reservoir parameters at the Desert Peak EGS project. At the Desert
Peak project, fluid injection was intentionally performed with a bottomhole pressure less than the minimum principal stress
(Chabora et al., 2012; Benato et al., 2013), which makes it a prototype example of a TSS test.

Several simplifications are made by CFRAC (most importantly, it neglects matrix flow and is two-dimensional), which limit
its ability to be used for forward modeling of reservoir transients (either wellbore pressure or production rate). Therefore, the
simulations in this paper are intended to be interpreted qualitatively. In future work, we plan to repeat these simulations with a
three-dimensional model that includes matrix flow to do realistic forward models of the pressure and rate transients before,
during, and after the TSS test. CFRAC is used rather than a more conventional code because the character of the overall shear
stimulation process is significantly affected by (1) idiosyncracies of flow in fracture networks and (2) stresses induced by
deformation.

Methodology

Computational Model
The full details of CFRAC are summarized in Chapter 2 of McClure (2012), and only a brief overview is given in this section.
The model assumes single-phase liquid water (no proppant), isothermal, Darcy flow in the fractures, and no flow in the matrix
around the fractures. Gravity was neglected, but this had a minimal effect on the simulations because fluid density variations
were small. Stresses induced by fracture deformation are calculated with the Shou and Crouch (1995) Displacement
Discontinuity (boundary element) method using quadratic basis functions assuming homogeneous, isotropic, linear elastic
deformation. A code called Hmmvp (Bradley, 2012) is used to very accurately and efficiently approximate the matrices of
interaction coefficients arising from the Displacement Discontinuity method, which increases efficiency significantly. When
creating the matrix approximations for this study, a relative error, εtol, equal to 10-6 was used (defined in Section 2.3.5 of
McClure, 2012).

Stresses induced by normal displacement of closed fractures are neglected. For a crack, these displacements are due to
fracture stiffness and are quite small. For a thicker, fault-zone like feature, displacements may be bigger, but they are spread
over a larger volume, reducing stress and strain (see Section 2.2.3.3 in McClure, 2012 for more discussion).

The simulations are two-dimensional, and in this paper should be interpreted as showing normal faults viewed from the side,
looking in the direction of the maximum horizontal stress. In this paper, the vertical direction is referred to as the y-axis
direction and the horizontal direction is referred to as the x-axis direction. The Shou and Crouch (1995) method assumes plane
strain deformation, implying infinite fracture size in the out-of-plane dimension, but in the study described in this paper, the
Olson (2004) adjustment was used to approximate the effect of a finite sized out-of-plane dimension (given by the variable h).

Flow is not upscaled to an effective continuum model. Because flow in the matrix is neglected and the boundary element
method is used to calculate stresses induced by deformation, it is only necessary to discretize the fractures, not the matrix.
Implicit time-stepping is used. During every time-step, the fluid pressure and (if the element is opening and/or sliding)
opening and sliding displacements are calculated to satisfy simultaneously for all elements the unsteady-state mass balance
equation and appropriate stress conditions. Elements may be closed (walls in contact) or open (walls out of contact),
4 SPE 166332-MS

depending on whether the fluid pressure has reached the normal stress. If the walls are in contact, Coulomb's law with
constant coefficient of friction is used to determine if the fracture should slide, and if so, displacements are calculated so that
Coulomb's law is satisfied. If walls are out of contact, displacements are calculated so that the walls bear zero shear stress. A
radiation damping term (Rice, 1993; Segall, 2010) is included for shear stress to approximate the effect of inertia at high
slipping velocity (though high slipping velocity is uncommon if a constant coefficient of friction is used) and to prevent
sliding from happening instantaneously.

The Coulomb failure criterion with a radiation damping term is (Jaeger et al., 2007; Segall, 2010):

|   v |  f  n'  S 0 , ............................................................................................................................. ...... (1)

where τ is the shear stress, η is the radiation damping coefficient, v is the sliding velocity of the fracture, μf is the coefficient of
friction (assumed constant in this study), S0 is fracture cohesion, and σn’ is the effective normal stress, defined as (Jaeger et al.,
2007; Segall, 2010):

 n'   n  P , ............................................................................................................................. ..................... (2)

where compressive stresses are taken to be positive. For fractures with shear stress less than the frictional resistance to slip,
shear deformation is assumed to be negligible.

Nonlinear relationships are used between fracture stress, fluid pressure, opening displacement, sliding displacement, hydraulic
aperture, and void aperture (modified from Willis-Richards et al, 1996):

E0  Edil
E  DE ,eff tan , ..................................................................................... (3)
1  9 n /  n, Eref
'
1  9 n' /  n, Eref

where E0, σn,Eref, and φEdil are specified constants. DE,eff is defined as equal to D if D < DE,eff,max, and equal to DE,eff,max
otherwise. The constants are allowed to be different for hydraulic aperture, e, and void aperture E. Nonzero φEdil corresponds
to pore volume dilation with slip, and non-zero φedil corresponds to transmissivity enhancement with slip.

Hydraulic aperture is equal to void aperture between two smooth plates but is lower than void aperture for rough surfaces such
as rock fractures (Liu, 2005). A "fracture" in a Discrete Fracture Network (DFN) model may represent a crack, but it may also
represent a more complex feature such as a fault zone (Section 2.2.3.3 in McClure, 2012). In the latter case, the void aperture
may be much larger than the hydraulic aperture, which is why the model allows e and E to be different. An elevated void
aperture E may also loosely approximate the effect of fluid leakoff into the matrix, which is not currently included in the
model.

The cubic law for fracture transmissivity (the product of permeability and hydraulic aperture) is (Jaeger et al., 2007):

e3
T  ke  . ............................................................................................................................. ..................... (4)
12
In the simulations in this paper, it was assumed that no new fractures could form. Even though the injection pressure was
below the minimum principal stress, it was a simplification to neglect fracture propagation because concentrations of tensile
stress could locally develop from sliding fractures, causing tensile fractures. In geology, this type of fracture is called a splay
crack or wing crack (Mutlu and Pollard, 2008).

Details of the Simulations

Three simulations (A, B, and C) were performed to investigate the behavior of a TSS Test under different conditions.
Simulation A was the baseline simulation in which the conditions needed for shear stimulation were met: coupling of slip to
transmissivity enhancement, adequate initial transmissivity, and a percolating fracture network. In Simulation B, a non-
percolating fracture network was used. Shear stimulation was possible, but fluid was unable to propagate far from the
wellbore due to the lack of continuous pathways for flow. The total number of fractures intersecting the wellbore was roughly
the same in Simulations A and B, but Simulation B had a lower average fracture length. Simulation C used the same fracture
network as Simulation A, but there was no coupling between slip and transmissivity (φedil equal to zero), making shear
stimulation impossible.
SPE 166332-MS 5

In Simulations A and B, the initial fracture transmissivity was very low, causing the fracture transmissivity to increase very
substantially during slip, from around 10 -18 m3 to around 5×10-14 m3. The initial fracture transmissivity was set to be higher in
Simulation C than in Simulations A and B (around 10-16 m3) because otherwise very little flow would have occurred (since
there was no shear stimulation to increase transmissivity).

The simulations were designed to loosely resemble the shear stimulation performed in Well 27-15 at the Desert Peak EGS
project. The simulations represent a vertical cross-section viewed from the side looking in the direction of the maximum
principal stress. At Desert Peak, the openhole section was roughly 150 m long and centered at a depth of 990.5 m. According
to the stress profile of Hickman and Davatzes (2010), at the center of the openhole section, at a depth of 990.5 m, σhmin = 14.5
MPa, σv = 23.8 MPa, and P = 8.7 MPa. The gradients of effective σhmin and σv with depth (stress value minus in-situ fluid
pressure) were 0.005 MPa/m and 0.014 MPa/m, respectively. Fracture orientation data was not available for the Desert Peak
project. In the simulations, the initial stress was initialized with a vertical gradient in σxx and σyy of 0.005 MPa/m and 0.014
MPa/m, respectively, and with σxx and σyy equal to 14.5 MPa and 23.8 MPa at y = 0.

During the low rate stimulation of Well 27-15, injection was performed at a roughly constant wellhead pressure of 3.1 MPa
(roughly 13 MPa at 990.5 m). In the simulations, injection was performed at a constant pressure of 12.4 MPa. Because the
simulations were two-dimensional and neglected matrix flow, no attempt was made to match the injection rate trends from the
Desert Peak data.

The fracture networks in the simulations were not necessarily intended to appear geologically realistic. To enable shear
stimulation, the networks contained abundant fractures that were well-oriented to slip.

The simulations were performed for seven days. Table 1 gives the settings that were common to all the simulations. Table 2
gives the settings that were different between the simulations.

Results
Figure 1, Figure 2, and Figure 3 show the final pressure distribution at the end of simulations for Simulations A, B, and C.
Figure 4 and Figure 5 show the final pressure distribution from Simulation A with contours of induced change in Coulomb
Stress and shear stress (on a fracture oriented 30° clockwise/counterclockwise from the y-axis direction). Figure 6 shows the
normalized flow rate versus time for Simulations A, B, and C.

Figure 1: The final pressure distribution for Simulation A. The black line represents the wellbore. The blue/red lines are
natural fractures with color proportional to fluid pressure.
6 SPE 166332-MS

Figure 2: The final pressure distribution for Simulation B. The black line represents the wellbore. The blue/red lines are
natural fractures with color proportional to fluid pressure.

Figure 3: The final pressure distribution for Simulation C. The black line represents the wellbore. The blue/red lines are
natural fractures with color proportional to fluid pressure.
SPE 166332-MS 7

Figure 4: Gold/green contours in the background show the change in Coulomb stress at the end of Simulation A for fractures
oriented 30° clockwise/counterclockwise from the y-axis. The black line represents the wellbore, and the blue/red lines
represent natural fractures with color proportional to fluid pressure.

Figure 5: Gold/green contours in the background show the change in the absolute value of shear stress at the end of Simulation
A for fractures oriented 30° clockwise/counterclockwise from the y-axis. The black line represents the wellbore, and the
blue/red lines represent natural fractures with color proportional to fluid pressure.
8 SPE 166332-MS

Figure 6: Normalized flow rate versus time for Simulations A (red), B (blue), and C (green). Flow rates for each simulation
are normalized by dividing by their flow rate after one hour, 0.164 kg/s, 0.114 kg/s, and 1.077 kg/s, respectively.

Discussion

Overall Behavior
In Simulation A, the final fracture network was more spatially dispersed and sparsely connected than in Simulation B (Figure
1 and Figure 3). This result is typical for simulations of shear stimulation that couple flow, stresses induced by shear, and
transmissivity coupling to shear (McClure and Horne, 2010; Section 3.4.2.2 of McClure, 2012). Sparsely connected networks
form from shear stimulation because of a process that we refer to as Crack-like Shear Stimulation (CSS) (Section 3.4.2.2 in
McClure, 2012; McClure and Horne, 2011). Figure 6 illustrates the CSS mechanism.

Crack-like Shear Stimulation arises because of the interaction between induced stresses, transmissivity enhancement, and fluid
flow. Along a particular fault, a concentration of shear stress develops at the boundary between where slip has occurred and
slip has not occurred (effectively, this is a shear crack tip). The concentration of shear stress causes slip and transmissivity
enhancement to propagate ahead of the fluid pressure front. This allows the fluid pressure front to propagate along the fracture
at a rate related to the stimulated, not the initial, fracture transmissivity.

Figure 7 shows various fracture properties along a fracture in a network that is experiencing shear stimulation. Fluid is
generally flowing from right to left. The fluid pressure front is at around 80 m, and the front of shear displacement and
transmissivity enhancement is at around 40 m. The transmissivity at the fluid pressure front is much higher than the initial
transmissivity (which was extremely low in this particular simulation and is seen at the far left side of the fracture).
CSS can help stimulation propagate along a fracture, but cannot cause the first patch of slip to occur on a fracture. Initiation of
slip on a fracture requires fluid to flow into it at the initial transmissivity. If the initial transmissivity is much lower than the
stimulated transmissivity, slip will tend to propagate along fractures much more rapidly than slip can initiate on fractures. This
creates an episodic behavior where individual fractures are stimulated in relatively short durations of time, separated by
periods of relatively slow propagation of stimulation (Section 3.4.2.2 of McClure, 2012).
SPE 166332-MS 9

Figure 7: Illustration of crack-like shear stimulation. Reproduced from Figure 3-33 in McClure (2012). The distribution of
transmissivity, shear stress, frictional strength, pressure, and shear displacement is shown along a particular fracture in a
fracture network simulation involving shear stimulation. The fluid is moving from right to left.

Changes in Coulomb Stress, ΔCS, can be calculated to estimate whether induced stresses, Δτ and Δσn', on a fracture will
encourage discourage slip, where positive changes encourage slip:

| | . ............................................................................................................................. ........... (4)

Figure 4 shows contours of induced changes in Coulomb Stress for μ equal to 0.6 on a fracture oriented 30° clockwise or
counterclockwise from the y-axis direction (equivalent to a fault dipping at 60°), which are roughly the orientations of the
fractures in Simulations A, B, and C.

The stresses in Figure 4 were calculated assuming that there were no fluid pressure changes induced in response to the induced
stresses. But actually, there can be a poroelastic response to changes in normal stress, Δσn. This occurs because void aperture,
E, is proportional to effective normal stress (Equation 3), and the water contained in the fractures (in the model) is only
slightly compressible. The fluid pressure changes in response to the induced normal stresses in order to keep the effective
normal stress nearly constant. With liquid water filling the fractures, as would occur in EGS applications, the poroelastic
response would almost completely counteract the induced normal stress. Over time, the fluid pressure would reequilibrate
with the surroundings, and induced change in effective normal stress would become equal to the induced change in normal
stress. In fact, this behavior has been proposed as one explanation for the time delay in earthquake aftershocks (King, 2007).
If the fractures were full of gas, as in a gas shale reservoir, the much greater compressibility of the infillling fluid would make
the poroelastic response on fluid pressure less significant. The poroelastic response is determined by the relative
compressibility of fluid density and the void aperture. Figure 5 shows contours of induced changes in the absolute value of
10 SPE 166332-MS

shear stress, which is equal to the change in Coulomb stress in the limiting case of the fully undrained response for an
incompressible fluid (which reasonably approximates water).

Figure 4 and Figure 5 show that when the fractures slip, they reduce Coulomb and shear stress along their sides, creating a
stress shadow that inhibits their neighbors from slipping. The Coulomb and shear stress changes are positive ahead of the
crack tips. These stresses encourage the further propagation of slip along the fractures, as described earlier. Shear fractures
should not be expected to propagate further in their own plane (Segall and Pollard, 1983), because the strength of intact rock is
much greater than the strength of preexisting flaws.

The result of these processes is that stimulation tends to localize into a smaller number of fractures, but propagate further from
the wellbore, than would be expected from a continuum model, or a model that behaves similarly to a continuum model, such
as Simulation C. The tendency for localization to occur depends on the average fracture length because with shorter fractures,
it is necessary for flow to initiate on new fractures more frequently. In Simulation A, individual fractures were quite long and
the difference between the initial and stimulated transmissivity was quite high, both factors that encouraged localization of
stimulation.

In Simulation B, shear stimulation occurred readily, as in Simulation A. However, stimulation remained confined close to the
wellbore because the fracture network was not percolating (Figure 2). If injection had been performed at constant rate, rather
than constant pressure, the fluid pressure would have eventually been forced to rise, eventually leading to propagation of new
fractures through the formation and the generation of continuous pathways for flow. With constant pressure injection, the
injection rate tended to zero as the fluid pressure become nearly equal to the injection pressure everywhere in the fractures
connected to the wellbore.

Flow Rate
In Figure 6, flow rate during the simulations is shown. Injection rates are normalized by the flow rate at one hour, which was
0.11 kg/s in Simulation A, 0.17 kg/s in Simulation B, and 1.08 kg/s in Simulation C. The qualitative behavior of flow rate
with time was different for the three models. In practice, it may be possible to use flow rate behavior over time to diagnose the
effectiveness of shear stimulation during a TSS Test. However, a fully realistic simulation of these transients would require a
three-dimensional simulation that included flow in the matrix.

In Simulation C (no coupling of slip and transmissivity), the injection rate had a negative one-half slope on the log-log plot
(Figure 6), indicating decay with the inverse of the square root of time. This behavior is consistent with the one-dimensional
solution to the diffusivity equation for a constant pressure boundary adjacent to an infinite half-space (Bird et al., 2006). This
is the same behavior that would have been expected in a continuum pressure transient model, in analogy to the matrix linear,
ellipsoidal, and late radial flow regime sequence expected from a hydraulically fractured well (Chapter 11 of Kamal, 2009).

In Simulations A and B, shear stimulation occurred, arresting the decay in injection rate. The shape of the normalized rate
curve for these two simulations started identically. The actual rate was initially slightly higher in Simulation B because a few
more fractures were intersecting the wellbore than in Simulation A. In Simulation B, the boundary of the "reservoir" (the
fractures connected to the wellbore through a percolating flow pathway) began to be felt after a few hours, causing a
downward deviation in the flow rate curve. After around 100 hours, the rate dropped with a slope much more rapidly than the
-1/2 slope observed in Simulation C.

In practice, comparison of the observed flow rate versus time curve to the flow rate versus time curve expected from a
standard continuum model (assuming constant transmissivity) could be used to diagnose the presence of shear stimulation. In
two-dimensions, the continuum result was linear (for early time), and in three-dimensions, the continuum result would be
radial. The well test solution for radial constant pressure injection was given by Ehlig-Economides (1979). In practice, this
solution would probably need to include dual-porosity behavior. Shear stimulation would be indicated by an upward deviation
in rate above the expected result. An anomolously rapid decay in rate could be an indication of a poorly connected network.
These trends in rate could be caused by other reservoir effects, such as boundary effects, and so it would be useful to perform
an injectivity test some period after the TSS test (using a constant pressure low enough to avoid shear stimulation) and
compare between the two transients.

Thermal Stresses
Thermal stresses have the potential to be important if long-term injection of colder fluid is performed. Thermal contraction
could cause the formation and propagation of tensile fractures through the formation. If tensile fractures form, it could
increase the injectivity of the well, arresting the expected decline in injection rate, and creating an effect that could be
mistaken for shear stimulation. Presumably, this increase in permeability would be reversible. This behavior needs to be
investigated in future work.
SPE 166332-MS 11

Combining TSS Tests with Standard Well Tests

The simplest way to interpret a TSS Test would be to perform standard pressure transient tests, such as injection/falloff tests,
before and after the TSS Test. Shear stimulation creates a permanent increase in formation permeability, and this could be
observed by comparing the permeabilities estimated from these transients. The falloff that occurs at the end of the TSS Test
probably could not be used because the redistribution of fluid pressure after injection could cause shear stimulation to occur at
the periphery of the stimulated region, creating nonlinearity that is not included in standard well test models.

The injection/falloff test performed prior to the TSS Test would need to be performed with a bottomhole pressure low enough
to avoid causing shear stimulation to occur. The injection/falloff test performed after the TSS Test would not be expected to
cause further shear stimulation as long as pressures and volumes were kept below the values used during the TSS Test because
fractures that have already slipped in response to a particular fluid pressure will not slip again when returned to that same fluid
pressure.

A production/buildup transient test could be useful after the TSS Test because the production of fluid from the formation
would tend to reheat the region of the formation that was cooled by injection, which would reduce any potential impact from
thermal stresses.

Desert Peak EGS Project

At the low rate stimulation at the Desert Peak project (described in the Details of the Simulations section above), the injection
rate was roughly constant at about 0.45 l/s for around five days, and then showed a gradual increase to around 1.89 l/s over the
subsequent week. During this period, injection pressure was mostly between 2.76 and 3.1 MPa (Chabora et al., 2012).
Assuming an injection pressure of 3.0 MPa, the well demonstrated an increasing injectivity from around 0.15 (l/s)/MPa to 0.63
(l/s)/MPa. As stated above, no attempt was made to match these trends in the modeling because the simplified dimensionality
of the simulations makes it impossible to use them to match real data.

A period of time after the low rate stimulation, an injectivity test was performed that estimated the injectivity at 0.37 (l/s)/MPa,
indicating a somewhat lower injectivity than before. The partial reversibility of the injectivity increase may indicate that
thermal fracturing, rather than shear stimulation, was responsible for the increase in injectivity during the test. Thermal
fracturing would be expected to be reversible because reheating of the formation would be expected after shut-in, which would
reclose some of the fractures unloaded by thermal stresses.

Overall, the results at the Desert Peak project indicate that the formation had minimal TSS. Low rate injection increased the
injectivity by perhaps a factor of four, but this increase was from a low initial injectivity, and was not fully retained during a
subsequent injectivity test. The low TSS was recognized by the project managers, who subsequently injected at high rate with
the intention of performing tensile fracturing (Chabora et al., 2012).

Soultz-sous-Forêts EGS Project

The EGS project at Soultz-sous-Forêts involves several wells drilled and stimulated to depths of 3.5 and 5 km in granite. All
wells were stimulated by injecting water with no proppant into long openhole wellbore sections. During the June 2000
stimulation of the well GPK2, fluid was injected at rates up to 50 l/s (Weidler et al., 2002) without the fluid pressure reaching
the minimum principal stress (Valley and Evans, 2007). Hydraulic tests prior to the stimulation estimated an injectivity
around 0.2 (l/s)/MPa. Hydraulic tests immediately after the test did not give conclusive estimates of injectivity (Weidler,
2000), but a test in 2003 (with no intervening stimulations having been performed) showed a peak overpressure of 5 MPa after
several days of injection at 15 l/s, equivalent to an injectivity of 3.0 (l/s)/MPa (Hettkamp et al., 2004). Wellbore observations
in GPK2 were not possible due to a wellbore obstruction, but in neighboring GPK3, production logs confirmed that flow from
the wellbore was concentrated at preexisting, large scale fault zones, and critical stress analysis and caliper logs confirmed that
these faults had failed in shear due to injection (Evans, 2005; Evans et al., 2005).

Overall, the results at the Soultz project indicate that the formation had a high TSS. Injection at pressure less than the
minimum principal stress resulted in an enduring, large increase in the well injectivity, and wellbore observations confirmed
that this increase was due to induced slip on the preexisting faults.

Conclusions
Shear stimulation is a potentially important process in hydraulic stimulation of shale formations and in EGS. However in
practice, the role of shear stimulation is inferred and subject to interpretation. The tendency for shear stimulation should be
highly dependent on local geological factors such as fracture orientation and ability to self-prop, and could be considered a
formation property.

Understanding the importance of shear stimulation will have important implications for modeling, stimulation design, and
resource assessment in low matrix permeability shale and geothermal resources. The Tendency for Shear Stimulation Test is a
12 SPE 166332-MS

simple, unambiguous way to directly measure the effect of shear stimulation in-situ. Using modeling, we showed qualitatively
how a TSS Test might be interpreted, showed how geological parameters could affect the results, and described how shear
stimulation can affect the process of fluid propagation in a fracture network. We reviewed two projects as examples of TSS
Tests, the EGS projects at Desert Peak and Soultz, and showed how their results clearly demonstrated the difference between
high and low TSS.

TSS Tests could be interpreted by performing pressure transient tests before and after the TSS Test and comparing the
interpreted formation permeability. It may also be possible to interpret TSS tests directly by analyzing the trends in rate over
time, but developing this technique in future work will require three-dimensional simulations that include matrix flow.

Acknowledgements
Thank you to the Precourt Institute for Energy for supporting this research from 2009-2012.

Nomenclature
CS = Coulomb stress, MPa
D = cumulative fracture sliding displacement, m
DE,eff, De,eff = Effective cumulative fracture sliding displacement in Equation 3, m
DE,eff,max,, De,eff,max = Maximum effective cumulative sliding displacement in Equation 3, m
e = hydraulic aperture (m)
E = void aperture (m)
E0, e0 = reference aperture in Equation 3, m
h = out-of-plane dimension in the two-dimensional simulations, m
k = permeability, m2
P = fluid pressure, MPa
S0 = fracture cohesion, MPa
T = transmissivity, m3
v = fracture sliding velocity, m/s
η = radiation damping coefficient, MPa/(m/s)
μf = coefficient of friction, unitless
σn = normal stress, MPa
σn’ = effective normal stress, MPa
σn,Eref, σn,eref = 90% fracture closure stress in Equation 3, MPa
τ = shear stress, MPa
φEdil, φedil = shear dilation angle, °
SPE 166332-MS 13

Tables
Table 1: Simulation settings used in Simulations A, B, and C.

De,eff,max 0.005 m η 3 MPa/(m/s)

E0 .005 m

G (shear modulus) 15 GPa ηtarg (for time stepping, 0.5 MPa

Section 2.2.3.9 in
McClure, 2012)

h 100 m μf 0.6

Pinit (initial fluid pressure) 8.7 MPa σn,Eref 20 MPa

S0 0.5 MPa σn,eref 20 MPa

mechtol (error tolerance .003 MPa σxx (initial x-direction 14.5 MPa
used in solving shear principal stress at y = 0)
stress equations, Section
2.2.3.5 in McClure, 2012)

itertol mechtol (error 0.01 MPa σyy (initial y-direction 23.8 MPa
tolerance used in iterative principal stress at y = 0)
coupling scheme, Section
2.2.3.1 in McClure, 2012)

φEdil 0° σxx,trend (trend in initial σxx 0.005 MPa/m

with distance in y-axis
direction)

υp (Poisson's ratio) 0.25 σyy,trend (trend initial σxx 0.014 MPa/m

with distance in y-axis
direction)

Table 2: Simulation settings specific to Simulations A, B, and C

A B C

φedil 2.5° 2.5° 0°

e0 .00001 m .00001 m .00005 m

14 SPE 166332-MS

References
Benato, S., D. M. Reeves, R. Parashar et al. 2013. Computational investigation of hydro-mechanical effects on transmissivity
enhancement during the initial injection phases at the Desert Peak EGS Project, NV. Paper presented at the Thirty-
Eighth Workshop on Geothermal Reservoir Engineering, Stanford, CA.
Bird, R. Byron, Warren E. Stewart, Edwin N. Lightfoot. 2006. Transport Phenomena, 2nd edition, John Wiley & Sons, Inc.
Blanton, Thomas. 1982. An experimental study of interaction between hydraulically induced and pre-existing fractures. Paper
SPE 10847 presented at the SPE Unconventional Gas Recovery Symposium, Pittsburgh, Pennsylvania.
Bowker, Kent A. 2007. Barnett Shale gas production, Fort Worth Basin: issues and discussion. AAPG Bulletin 91 (4): 523-
533. 10.1306/06190606018.
Bradley, Andrew M. 2012. H-matrix and block error tolerances. arXiv:1110.2807v2.
Brown, D. W. 1989. The potential for large errors in the inferred minimum earth stress when using incomplete hydraulic
fracturing results. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts 26 (6):
573-577. 10.1016/0148-9062(89)91437-X.
Bruel, D. 1995. Heat extraction modelling from forced fluid flow through stimulated fractured rock masses: application to the
Rosemanowes Hot Dry Rock reservoir. Geothermics 24 (3): 361-374. 10.1016/0375-6505(95)00014-H.
Bruel, Dominique. 2007. Using the migration of the induced seismicity as a constraint for fractured Hot Dry Rock reservoir
modelling. International Journal of Rock Mechanics and Mining Sciences 44 (8): 1106-1117.
10.1016/j.ijrmms.2007.07.001.
Chabora, Ethan, Ezra Zemach, Paul Spielman et al. 2012. Hydraulic stimulation of Well 27-15, Desert Peak Geothermal Field,
Nevada, USA. Paper presented at the Thirty-Seventh Workshop on Geothermal Reservoir Engineering, Stanford
University, Stanford, CA.
Cipolla, Craig, Norman Warpinski, Michael Mayerhofer et al. 2008. The relationship between fracture complexity, reservoir
properties, and fracture treatment design. Paper SPE 115769 presented at the SPE Annual Technical Conference and
Exhibition, Denver, Colorado, USA.
Cipolla, C. L., E. P. Lolon, J. C. Erdle et al. 2010. Reservoir modeling in shale-gas reservoirs. SPE Reservoir Evaluation &
Engineering 13 (4): 638-653. 10.2118/125530-PA.
Cladouhos, Trenton T., Matthew Clyne, Maisie Nichols et al. 2011. Newberry Volcano EGS demonstration stimulation
modeling. Geothermal Resources Council Transactions 35: 317-322.
Dahi-Taleghani, Arash, Jon Olson. 2009. Modeling simultaneous growth of multiple hydraulic fractures and their interaction
with natural fractures. Paper SPE 119739 presented at the SPE Hydraulic Fracturing Technology Conference, The
Woodlands, Texas.
Damjanac, Branko, Ivan Gil, Matt Pierce et al. 2010. A new approach to hydraulic fracturing modeling in naturally fractured
reservoirs. Paper presented at the 44th U.S. Rock Mechanics Symposium and 5th U.S.-Canada Rock Mechanics
Symposium, Salt Lake City, Utah.
Ehlig-Economides, Christine. 1979. Well Test Analysis for Wells Produced at a Constant Pressure. PhD Thesis, Stanford
University.
Evans, Keith F. 2005. Permeability creation and damage due to massive fluid injections into granite at 3.5 km at Soultz: 2.
Critical stress and fracture strength. Journal of Geophysical Research 110 (B4). 10.1029/2004JB003169.
Evans, Keith F., Albert Genter, Judith Sausse. 2005. Permeability creation and damage due to massive fluid injections into
granite at 3.5 km at Soultz: 1. Borehole observations. Journal of Geophysical Research 110 (B4).
10.1029/2004JB003168.
Fan, Li, John Thompson, John Robinson. 2010. Understanding gas production mechanism and effectiveness of well
stimulation in the Haynesville Shale through reservoir simulation. Paper SPE 136696 presented at the Canadian
Unconventional Resources and International Petroleum Conference, Calgary, Alberta, Canada.
Fisher, M. K., J. R. Heinze, C. D. Harris et al. 2004. Optimizing horizontal completion techniques in the Barnett Shale using
microseismic fracture mapping. Paper SPE 90051 presented at the SPE Annual Technical Conference and Exhibition,
Houston, Texas.
Fu, Pengcheng, Scott M. Johnson, Charles R. Carrigan. 2012. An explicitly coupled hydro-geomechanical model for
simulating hydraulic fracturing in arbitrary discrete fracture networks. International Journal for Numerical and
Analytical Methods in Geomechanics. 10.1002/nag.2135.
Gale, Julia F. W., Robert M. Reed, Jon Holder. 2007. Natural fractures in the Barnett Shale and their importance for hydraulic
fracture treatments. AAPG Bulletin 91 (4): 603-622. 10.1306/11010606061.
Gu, H., X. Weng. 2010. Criterion for fractures crossing frictional interfaces at non-orthogonal angles. Paper presented at the
44th U.S. Rock Mechanics Symposium and 5th U.S.-Canada Rock Mechanics Symposium, Salt Lake City, Utah.
Gu, Hongren, Xiaowei Weng, Jeffrey Lund et al. 2011. Hydraulic fracture crossing natural fracture at non-orthogonal angles, a
criterion, its validation and applications. Paper SPE 139984 presented at the SPE Hydraulic Fracturing Technology
Conference, The Woodlands, TX.
Hettkamp, T., J. Baumgärtner, R. Baria et al. 2004. Electricity production from hot rocks. Paper presented at the Twenty-Ninth
Workshop on Geothermal Reservoir Engineering, Stanford University.
SPE 166332-MS 15

Hickman, Stephen H., Nicholas C. Davatzes. 2010. In-situ stress and fracture characterization for planning of an EGS
stimulation in the Desert Peak Geothermal Field, Nevada. Paper presented at the Thirty-Fifth Workshop on
Geothermal Reservoir Engineering, Stanford University, Stanford, CA.
Ito, Hisatoshi. 2003. Inferred role of natural fractures, veins, and breccias in development of the artificial geothermal reservoir
at the Ogachi Hot Dry Rock site, Japan. Journal of Geophysical Research 108 (B9). 10.1029/2001JB001671.
Ito, T., K. Hayashi. 2003. Role of stress-controlled flow pathways in HDR geothermal reservoirs. Pure and Applied
Geophysics 160 (5): 1103-1124. 10.1007/PL00012563.
Jaeger, J. C., Neville G. W. Cook, Robert Wayne Zimmerman. 2007. Fundamentals of Rock Mechanics, 4th edition. Malden,
MA, Blackwell Pub.
Jeffrey, Robert, Xi Zhang, Marc Thiercelin. 2009. Hydraulic fracture offsetting in naturally fractured reservoirs: quantifying a
long-recognized process. Paper SPE 119351 presented at the SPE Hydraulic Fracturing Technology Conference, The
Woodlands, Texas.
Jing, Z., J. Willis-Richards, K. Watanabe et al. 2000. A three-dimensional stochastic rock mechanics model of engineered
geothermal systems in fractured crystalline rock. Journal of Geophysical Research 105 (B10): 23663-23679.
10.1029/2000JB900202.
Kaieda, H., S. Sasaki, D. Wyborn. 2010. Comparison of characteristics of micro-earthquakes observed during hydraulic
stimulation operations in Ogachi, Hijiori, and Cooper Basin HDR projects. Paper presented at the World Geothermal
Congress, Bali, Indonesia.
Kamal, Medhat M., ed. 2009. Transient Well Testing, Vol. Monograph Series Vol. 23, Society of Petroleum Engineers.
King, George. 2010. Thirty years of gas shale fracturing: what have we learned? Paper SPE 133456 presented at the SPE
Annual Technical Conference and Exhibition, Florence, Italy.
King, G. C. P. 2007. 4.08 - Fault interaction, earthquake stress changes, and the evolution of seismicity. In Treatise on
Geophysics, ed. Schubert Gerald, 225-255. Amsterdam, Elsevier.
Kohl, T., R. J. Hopkirk. 1995. "FRACTURE" - a simulation code for forced fluid flow and transport in fractured, porous rock.
Geothermics 24 (3): 333-343. 10.1016/0375-6505(95)00012-F.
Kohl, T., T. Mégel. 2007. Predictive modeling of reservoir response to hydraulic stimulations at the European EGS site Soultz-
sous-Forêts. International Journal of Rock Mechanics and Mining Sciences 44 (8): 1118-1131.
10.1016/j.ijrmms.2007.07.022.
Lanyon, G.W., A.S. Batchelor, P. Ledingham. 1993. Results from a Discrete Fracture Network model of a Hot Dry Rock
system. Paper presented at the Eighteenth Workshop on Geothermal Reservoir Engineering, Stanford University.
Ledésert, Béatrice, Ronan Hebert, Albert Genter et al. 2010. Fractures, hydrothermal alterations and permeability in the Soultz
Enhanced Geothermal System. Comptes Rendus Geoscience 342 (7–8): 607-615. 10.1016/j.crte.2009.09.011.
Liu, Enru. 2005. Effects of fracture aperture and roughness on hydraulic and mechanical properties of rocks: implication of
seismic characterization of fractured reservoirs. Journal of Geophysics and Engineering 2 (1): 38-47. 10.1088/1742-
2132/2/1/006.
Lutz, S. J., S. Hickman, N. Davatzes et al. 2010. Rock mechanical testing and petrological analysis in support of well
stimulation activities at the Desert Peak Geothermal Field, Nevada. Paper presented at the Thirty-Fifth Workshop on
Geothermal Reservoir Engineering, Stanford University.
Mahrer, Kenneth D. 1999. A review and perspective on far-field hydraulic fracture geometry studies. Journal of Petroleum
Science and Engineering 24 (1): 13-28. 10.1016/S0920-4105(99)00020-0.
Mayerhofer, Michael, Elyezer Lolon, Norman Warpinski et al. 2010. What Is Stimulated Reservoir Volume? SPE Production
& Operations 25 (1). 10.2118/119890-PA.
McClure, M. W. 2012. Modeling and characterization of hydraulic stimulation and induced seismicity in geothermal and shale
gas reservoirs. PhD Thesis, Stanford University, Stanford, California.
McClure, M. W., R. N. Horne. 2010. Discrete fracture modeling of hydraulic stimulation in Enhanced Geothermal Systems.
Paper presented at the Thirty-Fifth Workshop on Geothermal Reservoir Engineering, Stanford University.
McClure, M. W., Roland N. Horne. 2011. Investigation of injection-induced seismicity using a coupled fluid flow and
rate/state friction model. Geophysics 76 (6): WC181-WC198. 10.1190/geo2011-0064.1.
McClure, M. W., R. N. Horne. 2013. Is Pure Shear Stimulation always the mechanism of stimulation in EGS? Paper presented
at the Thirty-Eighth Workshop on Geothermal Reservoir Engineering, Stanford, CA.
Murphy, H. D., M. C. Fehler. 1986. Hydraulic fracturing of jointed formations. Paper SPE 14088 presented at the International
Meeting on Petroleum Engineering, Beijing, China.
Mutlu, O., D. D. Pollard. 2008. On the patterns of wing cracks along an outcrop scale flaw: a numerical modeling approach
using complementarity. Journal of Geophysical Research 113 (B6). 10.1029/2007JB005284.
Nagel, Neal, Ivan Gil, Marisela Sanchez-Nagel et al. 2011. Simulating hydraulic fracturing in real fractured rocks -
overcoming the limits of pseudo3D models. Paper SPE 140480 presented at the SPE Hydraulic Fracturing
Technology Conference, The Woodlands, Texas, USA.
Olson, Jon E. 2004. Predicting fracture swarms — the influence of subcritical crack growth and the crack-tip process zone on
joint spacing in rock. Geological Society, London, Special Publications 231 (1): 73-88.
10.1144/GSL.SP.2004.231.01.05.
16 SPE 166332-MS

Palmer, Ian, Zissis Moschovidis, John Cameron. 2007. Modeling shear failure and stimulation of the Barnett Shale after
hydraulic fracturing. Paper SPE 106113 presented at the SPE Hydraulic Fracturing Technology Conference, College
Station, Texas, U.S.A.
Pine, R. J., A. S. Batchelor. 1984. Downward migration of shearing in jointed rock during hydraulic injections. International
Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts 21 (5): 249-263. 10.1016/0148-
9062(84)92681-0.
Rachez, X., S. Gentier. 2010. 3D-hydromechanical behavior of a stimulated fractured rock mass. Paper presented at the World
Geothermal Congress, Bali, Indonesia.
Rahman, M. K., M. M. Hossain, S. S. Rahman. 2002. A shear-dilation-based model for evaluation of hydraulically stimulated
naturally fractured reservoirs. International Journal for Numerical and Analytical Methods in Geomechanics 26 (5):
469-497. 10.1002/nag.208.
Renshaw, C. E., D. D. Pollard. 1995. An experimentally verified criterion for propagation across unbounded frictional
interfaces in brittle, linear elastic materials. International Journal of Rock Mechanics and Mining Sciences &
Geomechanics Abstracts 32 (3): 237-249. 10.1016/0148-9062(94)00037-4.
Riahi, Azadeh, Branko Damjanac. 2013. Numerical study of hydro-shearing in geothermal reservoirs with a pre-existing
discrete fracture network. Paper presented at the Thirty-Eighth Workshop on Geothermal Reservoir Engineering,
Stanford, CA.
Rice, James R. 1993. Spatio-temporal complexity of slip on a fault. Journal of Geophysical Research 98 (B6): 9885-9907.
10.1029/93JB00191.
Rogers, Stephen, Davide Elmo, Rory Dunphy et al. 2010. Understanding hydraulic fracture geometry and interactions in the
Horn River Basin through DFN and numerical modeling. Paper SPE 137488 presented at the Canadian
Unconventional Resources and International Petroleum Conference, Calgary, Alberta, Canada.
Roussel, Nicolas, Mukul Sharma. 2011. Strategies to minimize frac spacing and stimulate natural fractures in horizontal
completions. Paper SPE 146104 presented at the SPE Annual Technical Conference and Exhibition, Denver,
Colorado, USA.
Segall, Paul. 2010. Earthquake and Volcano Deformation. Princeton, N.J., Princeton University Press.
Segall, Paul, David D. Pollard. 1983. Nucleation and growth of strike slip faults in granite. Journal of Geophysical Research
88 (B1): 555-568. 10.1029/JB088iB01p00555.
Shou, K. J., S. L. Crouch. 1995. A higher order Displacement Discontinuity Method for analysis of crack problems.
International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts 32 (1): 49-55.
10.1016/0148-9062(94)00016-V.
Tester, J., ed. 2007. The future of geothermal energy: impact of Enhanced Geothermal Systems (EGS) on the United States in
the 21st century, Massachusetts Institute of Technology.
Warpinski, N. R., J. C. Lorenz, P. T. Branagan et al. 1993. Examination of a cored hydraulic fracture in a deep gas well, SPE
22876. SPE Production & Facilities 8 (3). 10.2118/22876-PA.
Warpinski, N. R., L. W. Teufel. 1987. Influence of geologic discontinuities on hydraulic fracture propagation. Journal of
Petroleum Technology 39 (2). 10.2118/13224-PA.
Warpinski, N. R., S.L. Wolhart, C. A. Wright. 2001. Analysis and prediction of microseismicity induced by hydraulic
fracturing. Paper SPE 71649 presented at the SPE Annual Technical Conference and Exhibition, New Orleans,
Louisiana.
Weidler, R. 2000. Hydraulic stimulation of the 5 km deep well GPK-2, Bureau de recherches géologiques et minières
(BRGM).
Weidler, R., A. Gerard, A. Baria et al. 2002. Hydraulic and micro-seismic results of a massive stimulation test at the 5 km
depth at the European Hot-Dry Rock test site Soultz, France. Paper presented at the Twenty-Seventh Workshop on
Geothermal Reservoir Engineering, Stanford University.
Weng, Xiaowei, Olga Kresse, Charles-Edouard Cohen et al. 2011. Modeling of hydraulic-fracture-network propagation in a
naturally fractured formation. SPE Production & Operations 26 (4). 10.2118/140253-PA.
Willis-Richards, J., A. S. P. Green, A. J. Jupe. 1995. A comparison of HDR geothermal sites. Paper presented at the World
Geothermal Congress, Florence, Italy.
Willis-Richards, J., K. Watanabe, H. Takahashi. 1996. Progress toward a stochastic rock mechanics model of engineered
geothermal systems. Journal of Geophysical Research 101 (B8): 17481-17496. 10.1029/96JB00882.
Wu, R., O. Kresse, X. Weng et al. 2012. Modeling of interaction of hydraulic fractures in complex fracture networks. Paper
presented at the SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, USA.
Zhou, Jian, Mian Chen, Yan Jin et al. 2008. Analysis of fracture propagation behavior and fracture geometry using a tri-axial
fracturing system in naturally fractured reservoirs. International Journal of Rock Mechanics and Mining Sciences 45
(7): 1143-1152. 10.1016/j.ijrmms.2008.01.001.
Zhou, Xiaoxian, Ahmad Ghassemi. 2011. Three-dimensional poroelastic analysis of a pressurized natural fracture.
International Journal of Rock Mechanics and Mining Sciences 48 (4): 527-534. 10.1016/j.ijrmms.2011.02.002.
Zimmermann, G., A. Reinicke, W. Brandt, G. Blöcher, H. Milsch, H.-G. Holl, I. Moeck, T. Schulte, A. Saadat, and E.
Huenges 2008. Results of stimulation treatments at the geothermal research wells in Groβ Schönebeck/Germany.
SPE 166332-MS 17

Paper presented at the Thirty-Third Workshop on Geothermal Reservoir Engineering, Stanford University.