close

Вход

Забыли?

вход по аккаунту

?

j.geothermics.2018.03.003

код для вставкиСкачать
Geothermics 74 (2018) 135–149
Contents lists available at ScienceDirect
Geothermics
journal homepage: www.elsevier.com/locate/geothermics
Field observations at the Fenton Hill enhanced geothermal system test site
support mixed-mechanism stimulation
T
⁎
Jack H. Norbecka, , Mark W. McClureb, Roland N. Hornea
a
b
Department of Energy Resources Engineering, Stanford University, Stanford, CA 94305, USA
McClure Geomechanics LLC, Palo Alto, CA 94305, USA
A R T I C LE I N FO
A B S T R A C T
Keywords:
Enhanced geothermal systems (EGS)
Shear stimulation
Hydraulic fracture propagation
Induced seismicity
Reservoir geomechanics
Renewable energy
The Fenton Hill enhanced geothermal system (EGS) test site was the first of its kind, and interpretations of field
observations from the project have influenced the past four decades of EGS development. In this study, we
hypothesized that stimulation (i.e., permeability enhancement) in the Fenton Hill reservoir occurred through a
mixed-mechanism process that involved propagation of hydraulic splay fractures encouraged by the stress
changes induced as natural fractures opened and failed in shear. We used a hydromechanical fractured reservoir
numerical model to validate the efficacy of the mixed-mechanism stimulation conceptual model. Our modeling
results were consistent with the observations recorded during the Fenton Hill field experiments in three distinct
ways: (1) a marked increase in injectivity occurred at a threshold injection pressure, (2) the near wellbore
injectivity enhancement following each stimulation treatment was reversible, and (3) seismicity propagated in a
direction that was inconsistent with the orientation of the maximum principal stress, despite injection having
occurred at pressures significantly above the fracturing pressure. The modeling results demonstrate that several
independent hydromechanical observations could be replicated by the mixed-mechanism stimulation conceptual
model. In contrast, the observations could not be explained by a pure mode-I hydraulic fracture propagation nor
by pure shear stimulation. Distinct fracture sets are activated through the mixed-mechanism stimulation process;
the natural fractures provide most of the heat transfer surface area, and the tensile splay fractures form the bulk
of the fluid storage volume. Future EGS projects could take advantage of mixed-mechanism stimulation to design
wellbore completion and reservoir engineering and strategies to increase effective transmissivity, improve heat
mining efficiency, and extend useful reservoir lifetime.
1. Introduction
The Fenton Hill enhanced geothermal system (EGS) test site, located
in New Mexico, USA, was the first EGS project in the world (Brown
et al., 2012). Major accomplishments of the project include successful
hydraulic stimulation of two deep wells creating a flow connection,
fluid circulation through the geothermal reservoir for several months,
and generation of electricity. The original design of the geothermal heat
exchange system was similar to the conceptual model studied by
Gringarten et al. (1975). In this idealized stimulation strategy, a set of
vertical hydraulic fractures would connect two deviated wells. Initial
attempts at hydraulic stimulation were unsuccessful at connecting the
wells, however, and microearthquake event locations indicated that an
unanticipated region of the reservoir had been affected. A hydraulic
connection was made possible by redrilling each of the wells through
the cloud of microseismicity and performing several additional stimulation treatments (Brown et al., 2012). To explain the unexpected
⁎
behavior, the scientists and engineers involved in the project developed
two competing hypotheses of the hydraulic stimulation mechanism
(Brown et al., 2012; Duchane, 1991). Observations that injection
pressures exceeded the magnitude of the least principal stress and
pressure-rollover behavior suggested that planar hydraulic fractures
were forming in the reservoir. In contrast, the development of a broad
cloud of microseismicity that migrated predominantly in a direction
that was inconsistent with the expected orientation of a planar hydraulic fracture for the in situ stress state suggested that perhaps permeability enhancement was caused by shear slip on preexisting fractures.
In the present study, we investigated how fracture pressurization,
poroelastic stress, and thermal stress affected the stimulation process
and the evolution of microseismicity that was observed during various
injection experiments carried out during the 1980s. We analyzed multiple data sets to develop a conceptual model of the Fenton Hill EGS
fractured reservoir system, focusing on four stimulation treatments in
Corresponding author at: Earthquake Science Center, U.S. Geological Survey, Menlo Park, CA 94025, USA.
E-mail address: jnorbeck@stanford.edu (J.H. Norbeck).
https://doi.org/10.1016/j.geothermics.2018.03.003
Received 19 December 2017; Received in revised form 19 February 2018; Accepted 5 March 2018
Available online 19 March 2018
0375-6505/ © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
2015, 2016, 2017). Fig. 2 shows the event locations recorded during
Expt. 2032 (MHF) in plan and cross-sectional views. Events migrated
away from the well during injection. In plan view, the microseismic
cloud tended to migrate in an overall NNW-SSE direction. This unanticipated observation helped to form the basis for our conceptual
reservoir model. Fig. 3 illustrates the rate of migration of the seismicity
away from the wellbore. During the injection treatment, events tended
to occur across the entire stimulated region. Upon shut-in after 60 h of
injection, the events occurred predominantly at the edges of the stimulated region. Following shut-in, the event rate decayed steadily over
the period of about one day.
Well EE-2 leading up to and including the massive hydraulic fracture
(MHF) experiment (Expts. 2018, 2020, MHF prepump, and Expt. 2032).
Our preliminary investigations were first presented by Norbeck et al.
(2016a,c).
We hypothesized that the dominant process contributing to enhanced reservoir permeability during hydraulic stimulation could be
characterized as a mixed-mechanism stimulation process. Mixed-mechanism stimulation has been described previously by Weng et al.
(2011), McClure (2012), McClure and Horne (2013a, 2014), Jeffrey
et al. (2015), and Zhang and Jeffrey (2016). In the mixed-mechanism
stimulation conceptual model, a combination of several hydromechanical processes can influence permeability evolution. In particular, opening-mode deformation and shear-induced dilation can influence a fracture's hydraulic aperture. Furthermore, both opening and
sliding deformations can promote the formation of splay fractures that
initiate and propagate off of natural fractures (Norbeck and Shelly,
2018; Pollard and Fletcher, 2005; Ye and Ghassemi, 2018; Zhang and
Jeffrey, 2016). For application to geothermal reservoir stimulation, the
mixed-mechanism process has the potential to generate a fracture
network that could be beneficial for thermal recovery due to its relatively complex geometry compared to pure hydraulic fracturing or pure
shear stimulation (McClure and Horne, 2014; Jeffrey et al., 2015).
In this work, we present evidence that a mixed-mechanism stimulation process occurred in the Fenton Hill geothermal reservoir by using
a numerical reservoir model to replicate several distinct behavioral
characteristics observed during stimulation at Fenton Hill. Because
subsurface reservoir engineering data at this site are insufficient to
describe the system with certainty, it is possible to develop multiple
hypotheses that can explain the observations. In Section 5, we analyze
several alternative conceptual models that cannot be ruled out completely, but that we believe are unlikely.
The remainder of this paper is organized as follows. In Section 2, we
present our conceptual model of the Fenton Hill reservoir and describe
the field data and observations used to constrain the model. The hydromechanical numerical model used in this study is described in
Section 3. Results from our mixed-mechanism simulations are presented
in Section 4. In Section 5, we propose several alternative hypotheses
and discuss why we believe they are unlikely. In Section 6, we discuss
the implications of the mixed-mechanism process for reservoir engineering design of future geothermal projects. Our concluding remarks
are listed in Section 7.
2.1. State of stress
Our interpretation for the state of stress in the Phase II reservoir was
based on wellbore stress measurements (Barton et al., 1988), earthquake focal mechanisms (House et al., 1985), minifrac tests (Brown,
1989; Kelkar et al., 1986), and observations during step-rate injection
tests (Brown et al., 2012; Matsunaga et al., 1983). Varying estimates of
the fracture gradient are available in the literature. Kelkar et al. (1986)
summarized a large number of minifrac tests, illustrated in Fig. 4, to
estimate that the minimum principal stress gradient was 19 MPa/km,
implying that the minimum horizontal stress was σh = 68.4 MPa at
3.6 km depth. However, Kelkar et al. (1986) noted that tests shallower
than 3.3 km depth indicated a much lower fracture gradient. Based on
these observations, Brown (1989) proposed that the minimum principal
stress gradient was 13 MPa/km, implying a minimum horizontal stress
of σh = 46.8 MPa at depth. Brown (1989) hypothesized that due to the
high tensile strength of granite, hydraulic fractures were unable to form
at the wellbore, so the fracturing pressure observed during injection
tests corresponded to the pressure required to exceed the normal stress
on preexisting fractures intersecting the well. If these fractures were
oblique to the principal stresses, then their opening pressure would
have been greater than the minimum principal stress. Similar behavior
has been observed during other field-scale stimulation treatments
(Baisch et al., 2015) and pressurization experiments in wellbores with
preexisting fractures (Baumgartner and Zoback, 1989; Rutqvist and
Stephansson, 1996). Therefore, Brown (1989) proposed that the apparent increase in fracturing pressure at 3.3 km was caused by a discontinuity in natural fracture orientation rather than stress, and that the
tests shallower than 3.3 km reflected a more accurate measure of the
minimum principal stress. Focal mechanism analysis indicated both
strike-slip and normal faulting mechanisms, suggesting that the maximum horizontal stress was roughly equal to the vertical stress (Barton
et al., 1985; House et al., 1985; Fehler, 1989; Phillips et al., 1997).
The state of stress used in our hydromechanical reservoir simulations can be summarized in Figs. 4 and 5. We assumed that the “low
stress” profile was an accurate reflection of the state of stress at Fenton
Hill. The orientation of the maximum horizontal stress was N30°E based
on interpretations of wellbore breakouts from acoustic borehole televiewer logs (Barton et al., 1988). Brown (1989) stated that the reservoir
fluid pressure was about 5 MPa subhydrostatic, which has been taken
into consideration in the stress state representation shown in Fig. 5.
2. Background on Fenton Hill and model constraints
The EGS experiments at Fenton Hill involved many field tests in
both the shallow Phase I and deeper Phase II reservoirs. The primary
hydraulic stimulation experiments in the Phase II reservoir took place
in Wells EE-2 and EE-3A, which each had openhole intervals at a depth
of roughly 3.6 km. The most significant stimulation treatment was
performed in Well EE-2 during December 1983 (Expt. 2032, also called
the Massive Hydraulic Fracture (MHF) treatment), in which roughly
21,000 m3 of water was injected over 60 h at a maximum flow rate of
106 kg/s and maximum wellhead pressure of 49 MPa (Brown et al.,
2012). We considered the following stimulation treatment experiments
in Well EE-2:
2.2. Geologic structure
• Expt. 2018 (July 1982)
• Expt. 2020 (October 1982)
• Expt. 2032 prepump (December 1983)
• Expt. 2032 massive hydraulic fracture (December 1983)
Microseismic events observed during hydraulic fracturing treatments are often interpreted as shear slip events on natural fractures that
surround the main hydraulic fracture (Warpinski, 2009; Warpinski
et al., 2012). At Fenton Hill, if this was the appropriate mechanism,
then the microseismic cloud would be expected to migrate in the direction of maximum horizontal stress (N30°E), but this was not the case.
Irrespective of the assumption about the low- or high-stress profile,
injection pressures during the hydraulic stimulation treatments often
exceeded the magnitude of the minimum horizontal stress significantly
(see Figs. 1 and 4), which would suggest that hydraulic fractures were
The injection rate and injection pressure data recorded during these
experiments are shown in Fig. 1 (the data were reformatted based on
the data reported by Brown et al. (2012)).
Los Alamos National Laboratory provided microseismic event locations and timing recorded during Expts. 2032 (MHF) (White et al.,
136
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
Fig. 1. Injection rate and injection bottomhole pressure measurements during the four main stimulation treatments performed on well EE-2. Data has been reformatted from Brown et al.
(2012).
3. Hydromechanical numerical model
indeed forming in the reservoir. In addition, a repeated observation
during multiple step-rate tests was that near-wellbore injectivity increased substantially after exceeding a pressure corresponding to a
bottomhole pressure of about 74 MPa (Brown et al., 2012). A conventional minifrac analysis would interpret this value as the fracture
opening or fracture propagation pressure. Finally, a key observation
was that wellbore temperature logs indicated three or four distinct
feedzones during Expt. 2018, suggesting that natural fractures (as opposed to a single hydraulic fracture) were taking flow in the nearwellbore region and that flow was localized into a small number of
highly permeable pathways with vertical separation on the order of
30 m (Figure 6-6 from Brown et al. (2012)).
The orientation of the primary fracture set was determined based on
three independent data sets that each support one another. First, the
“fracture opening pressure” of 74 MPa was interpreted as the fluid
pressure required to overcome the normal stress acting on natural
fractures intersecting the well. Given the stress state measured at
Fenton Hill, the fracture orientation corresponding to a normal stress of
74 MPa (equivalent to an effective stress of 43 MPa at ambient reservoir
conditions) is roughly N23°W (see Fig. 5). Second, analysis of the focal
mechanisms for several of the largest observed microearthquakes
yielded solutions with a nodal plane suggesting the presence of nearvertical fracture or fault structures oriented N30°W (House et al., 1985).
Finally, the overall migration of the seismicity was in the NNW-SSE
direction (see Fig. 2). We generated a stochastic realization of a fractured reservoir based on this conceptual model for the simulations
performed in this study (see Fig. 6).
Geothermal reservoirs are often considered to be highly fractured,
and numerical reservoir simulation with discrete fracture modeling
approaches are commonly used to study complex and coupled physical
processes, including thermal, chemical, and hydromechanical effects
(Garipov et al., 2016; Ghassemi and Zhou, 2011; Karimi-Fard et al.,
2004; Karvounis and Jenny, 2016; Lee et al., 2000; McClure and Horne,
2013b; Norbeck and Horne, 2016a,b,c; Norbeck et al., 2016b;
Salimzadeh et al., 2018a,b; Sonnenthal et al., 2005). The numerical
simulations were performed using a coupled fracture mechanics and
fluid flow model within an embedded fracture modeling framework.
Detailed descriptions of the numerical formulation are provided by
McClure (2012), Norbeck et al. (2016b), Norbeck and Horne
(2016a,b,c), and in Chapter 2 of Norbeck (2016).
Mixed-mechanism stimulation is controlled by a nonlinear coupling
between the mechanical deformation of fractures and changes in their
hydraulic properties. We allowed for the possibility of shear-enhanced
dilation to increase fracture transmissivity. In addition, hydraulic splay
fractures were able to form off the tips of natural fractures in a mixed
mode-I/mode-II failure process (Pollard and Fletcher, 2005). Poroelastic stresses can be generated as fluid pressure in the rock surrounding the fractures changes due to leakoff or production (Segall,
1989). Thermal stresses can be generated as the rock cools during injection (Mossop, 2001; Rana, 1984). In some cases, porothermoelastic
effects can influence the behavior of microseismicity during fluid injection (Norbeck and Horne, 2015, 2016a,b). In this study, we
137
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
Fig. 2. Microearthquake locations recorded during the Expt. 2032 (MHF) stimulation treatment. The events are colored by event timing. The white square represents the location of the
openhole section of Well EE-2. The well was shut-in after roughly 60 h of injection. Seismic data courtesy of Los Alamos National Laboratory. (For interpretation of the references to color
in this figure legend, the reader is referred to the web version of the article.)
considered porothermoelastic effects related to fluid leakoff and cooling
of the rock volume surrounding the fractures. The poroelastic and
thermal stresses affected the local state of stress throughout the reservoir, and their effect on fracture deformation was accounted for
(Norbeck and Horne, 2016a,b,c).
∂
mf
∼+∼
∇ ·(ρλek f ∇p f ) + m
Ω =
(ρE ).
∂t
3.1. Governing equations
In the embedded fracture modeling approach, the mass and energy
balance equations are described separately for the matrix and fracture
domains (referenced with superscripts m and f, respectively). For a
porous medium saturated with single-phase fluid, the mass balance
equations can be written, for flow in the matrix domain, as:
∂
fm
∼+∼
∇ ·(ρλk m∇pm ) + m
Ω =
(ρϕm),
∂t
(2)
Here, p is fluid pressure in the matrix domain, p is fluid pressure in the
fracture domain, ρ is fluid density, λ is inverse of fluid viscosity, k is
permeability, ϕ is porosity, e is fracture hydraulic aperture, E is fracture
∼ is a normalized mass source term related to wells.
void aperture, and m
∼
The Ω terms are the embedded fracture source terms that account for
mass transfer between the fractures and the surrounding rock. In Eq.
(2), the fracture hydraulic properties were allowed to vary with mechanical opening of the fractures and with shear slip as described in
Section 3.2.
Fluid was assumed to be in thermal equilibrium with the reservoir
rock. Energy conservation can be described, in the matrix domain, as
(Charoenwongsa et al., 2010; Jaeger et al., 2007):
m
f
∼fm
∇ ·(κ m∇T m) − ∇ ·[ρφ cpφ v m (T m − T0m)] + ρφ cpφ q͠ (T w − T0m) + Π
(1)
=
and, for flow in the fracture domain, as:
138
∂
∂t
{[ϕmρφ c vφ + (1 − ϕm) ρr cr ](T m − T0m)}
(3)
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
Fig. 3. Location of each microseismic event represented as distance relative to the injection well. The lack of seismicity in the first 8 h is indicative of reservoir behavior and
can be attributed to the Kaiser effect. The lack of seismicity during the period between
1.25 and 1.75 days was due to a data acquisition problem. Seismic data courtesy of Los
Alamos National Laboratory.
Fig. 6. Illustration of the natural fracture network geometry used in the simulations of the
mixed-mechanism hypothesis. The white arrows indicate the orientation of the principal
stress tensor. The white square shows the location of Well EE-2.
and, in the fractured domain, as:
∼mf
∇ ·(κ f ∇T f ) − ∇ ·[ρφ cpφ v f (T f − T0f )] + ρφ cpφ q͠ (T w − T0f ) + Π
=
Fig. 4. The black diamonds represent the stress measurements reported originally by
Kelkar et al. (1986). The solid black line and dashed black line represent the stress profile
with depth of the minimum principal stress based on the low stress and high stress
models, respectively. The solid red line represents the estimated profile of the maximum
and intermediate principal stresses. The solid blue line represents the estimated fluid
pressure profile. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of the article.)
∂
∂t
{[ϕ f ρφ c vφ + (1 − ϕ f ) ρr cr ](T f − T0f )}
(4)
Tw
Here, T is rock temperature, T0 is a reference temperature,
is the
temperature of injected fluid, κ is the diagonal thermal conductivity of
rock, cpφ is the constant pressure heat capacity of the fluid, c vφ is the
constant volume heat capacity of the fluid, cr is the heat capacity of
∼
rock, ρr is the density of rock, v is the Darcy fluid velocity, and Π is
embedded fracture heat transfer term.
The numerical formulation for the porothermoelastic framework
was presented originally in Norbeck and Horne (2016a,b,c). Assuming
infinitesimal strains and using Hooke's law as the constitutive relationships between stress and strain, momentum balance can be expressed as (Jaeger et al., 2007):
G∇2 u + (G + Λ) ∇ (∇ ·u) = −F − α P∇ (pm − p0 ) − 3αT K ∇ (T m − T0),
(5)
where u are the material displacements caused by poroelastic and
thermoelastic deformation as well as mechanical deformation of fractures, G is shear modulus, Λ is Lamé's modulus, K is bulk modulus, αP is
Biot's coefficient, αT is the linear thermal expansion coefficient, and F
are body forces. In this work, we used a sequential solution strategy to
solve for stress changes related to fracture deformation (using a
boundary element method described by Bradley (2014)) and the poroelastic and thermal stresses (using a finite element method described
by Smith et al. (2014)).
Fig. 5. The Mohr circle representation of the state of stress at the Fenton Hill site at a
depth of 3.6 km. The red lines represent the shear failure envelope assuming a friction
coefficient f = 0.7. The blue diamond represents the fracture orientation corresponding to
the fracture opening pressure observed in the rate/pressure data, which was the average
orientation of the natural fracture network used in the simulations. (For interpretation of
the references to color in this figure legend, the reader is referred to the web version of
the article.)
139
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
addition to allowing for fracture propagation to occur, it was critical to
include this hydromechanical coupling to investigate the mixed-mechanism stimulation conceptual model.
In order to solve the fault deformation problem, mechanical equilibrium was enforced along the fault surfaces. Throughout a simulation,
the state of stress at each discretized fault element was checked and
updated continuously. Fault elements that bore a compressive effective
normal stress deformed in the mode-I direction subject to a nonlinear
joint stiffness constitutive law (Willis-Richards et al., 1996), and fracture aperture was therefore able to be calculated explicitly. Fracture
mechanics theory suggests that if the fluid pressure acting in the fault
overcomes the remote tectonic loading, then the opening-mode deformations induce stresses that serve to balance the overpressure. In
this case, opening-mode mechanical equilibrium can be expressed as:
σn = σnR + Δσn − p f = 0.
3.3. Hydraulic splay fracture propagation
The mixed-mechanism stimulation hypothesis relies on the interaction between preexisting natural fractures and propagating hydraulic
fractures. The numerical method used to perform the fracture propagation calculations has been described by McClure (2012, 2015). A set
of potentially forming hydraulic fractures were specified stochastically
at the beginning of the simulation. The state of stress was continually
evaluated for all fracture elements, and if the fluid pressure of a potentially forming splay fracture element reached the least principal
stress it was activated. Once an element became activated, the mode-I
stress intensity factor was calculated using an approach developed for
the displacement discontinuity method (Norbeck et al., 2016b; Olson,
2007):
(6)
σnR
In Eq. (6),
is the tectonic loading resolved in the fault-normal direction, Δσn is any change in normal stress due to mechanical deformation, poroelastic stress, or thermal stress, and pf is the fluid
pressure acting within the fracture.
To solve the sliding deformation problem, mechanical equilibrium
was enforced such that the shear stress causing sliding was balanced by
the frictional resistance to slip. Therefore, the sliding mechanical
equilibrium equation can be expressed as (Jaeger et al., 2007):
τR
+ Δτ = fσn + c.
1/2
Y
⎤ ⎛ π ⎞ Δe .
KI = 0.806 ⎡
2
⎢ 4(1 − ν ) ⎦
⎥⎝ a ⎠
⎣
Here, Y is Young's modulus, ν is Poisson's ratio, a is the half-length of
the crack tip element and Δe is the mode-I opening displacement discontinuity of the crack tip element. Once KI exceeded the mode-I
fracture toughness, KIC, the tip element was added into the flow and
mechanics system of equations. The splay fractures were assumed to
propagate in the plane perpendicular to the least principal stress.
It was necessary to consider hydraulic fracture propagation in the
model for several reasons. First, the fluid pressure during injection
exceeded the minimum principal stress, so it is plausible that hydraulic
fractures would have formed. Even if the high tensile strength of the
rock prevented hydraulic fractures from forming at the well, concentrations of stress created by natural fracture opening and sliding
would facilitate the formation of hydraulic fractures away from the well
(McClure and Horne, 2014; Zhang and Jeffrey, 2016). Because the fluid
pressure increased well-above the minimum principal stress, the hydraulic fractures could have opened mechanically to large apertures,
providing the bulk of the fluid storage during injection. However, it was
critical for the hydraulic fractures to terminate against the natural
fractures, because otherwise they would have propagated continuously
across the formation and the microseismic cloud would have been oriented primarily perpendicular to the minimum principal stress. In the
simulations, fracture termination was assumed to occur 50% of the time
when a hydraulic fracture reached a natural fracture. Physics-based
criteria exist that describe the interaction, termination, and crossing of
hydraulic and natural fractures (Renshaw and Pollard, 1995; Gu et al.,
2012), however we used a stochastic crossing criterion because of the
uncertainty associated with upscaling the models to the field (Fu et al.,
2016).
(7)
Here, τ is the shear stress resolved due to the tectonic loading, Δτ is the
stress change caused by mechanical deformation, f is the friction
coefficient and c represents fault cohesion. Eqs. 1 through 7 were solved
to determine the distributions of fluid pressure, temperature, and material displacements within the matrix and fractures.
R
3.2. Constitutive equations for hydromechanical deformation
The fracture mass balance equation (see Eq. (2)) can behave in a
highly nonlinear fashion due to changes in the hydraulic properties
caused by mechanical deformation and shear failure of fractures. The
fracture transmissivity, ϒ, was assumed to behave according to the
cubic law for flow between parallel surfaces (Snow, 1965; Witherspoon
et al., 1980):
Υ = ek f =
e3
.
12
(8)
We applied an empirical model introduced originally by Bandis et al.
(1983) and Barton et al. (1985) in order to describe the nonlinear relationship between opening deformation (mode-I) and effective stress
as well as the effect of shear-slip (mode-II) induced dilation. We applied
the form of the constitutive model proposed by Willis-Richards et al.
(1996):
e*
e (σn, p f , δ ) =
1+9
(
σn − p f
σ e
*
)
+ δ tan φe + eres.
(9)
In Eq. (9), σn is the total normal stress acting on the fracture, δ is cumulative shear slip, e* and σ*e are laboratory-derived properties that
describe the fracture stiffness, φe is the shear dilation angle, and eres is
the residual fracture aperture. An equivalent formulation is also used to
describe fracture void aperture, E, where the empirical constants are
allowed to be different.
Eq. (9) was used to evaluate fracture aperture for fracture elements
that bore a compressive effective normal stress (i.e., when σn > 0 ). As
soon as the fluid pressure in the fracture became equal to or exceeded
the normal stress, the opening displacement, Δe, was calculated according to fracture mechanics theory with a boundary element method.
Then, the total aperture was evaluated as:
e = e* + Δe + δ tan φe .
(11)
4. Modeling mixed-mechanism stimulation at Fenton Hill
The numerical simulations involved modeling four hydraulic stimulation treatments performed in Well EE-2 that occurred during 1982
and 1983. In an attempt to represent the operational activity accurately, the stimulation treatments were modeled coherently as a single
simulation. Specifically, we modeled Experiments 2018, 2020, 2032
(prepump), and 2032 (MHF). We based the injection well operational
parameters on the data described by Matsunaga et al. (1983) and Brown
et al. (2012) (see Fig. 1).
4.1. Initial conditions, boundary conditions, and model properties
(10)
In our model, we chose to use the Brown (1989) estimate for the
magnitude of the minimum principal stress. The state of stress was a
Eqs. (9) and (10) were designed to be continuous as the fracture transitioned from being closed to open (McClure and Horne, 2014). In
140
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
was specified as a constant-rate mass injection source based on simplified interpretations of the rates in Fig. 1.
A limitation of our simulations is that we modeled a two-dimensional domain with an assumed thickness of h = 200 m.
Microseismicity during the MHF treatment formed a cloud roughly
200 m wide, 1000 m long, and 1000 m in vertical extent. The stimulated
reservoir volume implied by the microseismicity represents a combination of the fracture density, the surface area and aperture of individual fractures, and the volumetric extent of seismicity. The total
volume of fluid injected during the treatment must balance with the
stimulated fracture volume (after accounting for the initial pore fluid
volume and leakoff into the surrounding rock). Because virtually no
constraints exist for the density, variability of surface area, or aperture
distribution of the natural fractures in the Fenton Hill reservoir, these
modeling details were specified at our discretion. Given this significant
level of uncertainty, no advantage would be gained by extending our
analysis to three dimensions.
Table 1
Initial conditions, boundary conditions, and model domain geometry.
Parameter
Value
Unit
Description
Depth
h
3.6
200
90
km
m
MPa
Depth at the center of the model domain
Vertical thickness of the model domain
Maximum horizontal stress
σhR
θ σH
46
MPa
Minimum horizontal stress
N30°E
–
p0
T0
Tw
31
230
100
MPa
°C
°C
Azimuth of σHR
Initial reservoir fluid pressure
Initial reservoir temperature
Temperature of injected fluid
σHR
Table 2
Fluid properties.
Parameter
Value
κφ
cφ
λ−1
ρφ,0
βφ
Unit
Description
−1
0.6
4200
0.15 × 10−3
930
4.4 × 10−4
−1
W m °C
J kg−1 °C−1
Pa s
kg m−3
MPa−1
Thermal conductivity of fluid
Heat capacity of fluid
Viscosity of fluid
Density of fluid
Compressibility of fluid
4.2. Hydraulic stimulation of Well EE-2
We aimed to replicate the reversible injectivity behavior that occurred during multiple stimulation treatments in Well EE-2. In Fig. 1,
the change in injectivity can be characterized by an initial rapid pressurization at low injection rates up until a critical pressure level, followed by only modest increases in pressure even for relatively large
changes in injection rate when pressure is above the critical threshold.
This pressure-rollover behavior was observed in each of the four repeated stimulation treatments, suggesting that any near-wellbore permeability enhancement achieved during each treatment was reversible.
Our simulation results matched the injectivity transition accurately
for each stimulation treatment (see Fig. 7). In the model, the stressdependent coupling between fracture opening and fracture transmissivity (see Eqs. (9) and (10)) was responsible for the pressure-rollover
behavior. As the injection pressure approached the magnitude of the
normal stress acting on the fractures that intersected the well, fracture
transmissivity was able to increase dramatically. While the fractures
had high transmissivity, they were able to accommodate large changes
in flow rate with negligible resistance (Fig. 7).
We reiterate that the injectivity at low flow rates occurred consistently even after several repeated stimulation treatments in Well EE2. A shear dilation angle of φ = 0 was necessary to replicate this behavior. Therefore, our simulation results suggest that shear stimulation
was ineffective at enhancing the permeability of natural fractures at
Fenton Hill (see Section 5.1 for a more detailed analysis of the effects of
shear stimulation).
The MHF treatment was performed at extremely high net-pressures
for 60 h. The assumed stress state indicated a fracturing pressure of
roughly 46 MPa (bottomhole pressure), and injection pressures were
over 70 MPa for most of the treatment. Traditional mode-I fracture
propagation analyses based on the theory of linear elastic fracture
mechanics would predict unrealistically large hydraulic fractures for
these operational conditions. High net-pressures support the fracture
termination and branching assumption that formed the basis of our
conceptual and numerical models. Our analysis supports the hypothesis
that stimulation occurred as a mixed-mechanism combination of mechanical fracture opening (as the fluid pressure in natural fractures
approached their normal stress) and creation of hydraulic splay fractures.
Table 3
Rock properties.
Parameter
Value
Unit
Description
−1
−1
κr
αP
αT
2.4
0.4
8 × 10−6
Wm
–
°C−1
km
ϕm
KIC
cr
G
ν
ρr
βr
1 × 10−19
0.05
1.5
800
15
0.25
2650
4.4 × 10−4
m2
–
MPa m1/2
J kg−1 °C−1
GPa
–
kg m−3
MPa−1
°C
Thermal conductivity of rock
Biot's coefficient of rock
Linear thermal expansion coefficient of
rock
Permeability of matrix rock
Porosity of matrix rock
Mode-I fracture toughness
Heat capacity of rock
Shear modulus of rock
Poisson's ratio of rock
Density of rock
Pore compressibility of rock
Table 4
Fracture properties.
Parameter
Value
Unit
Description
e*
E*
σ*e
σ*E
eres
Eres
φe
φE
f
s
0.0004
0.0004
8
8
0
0
0
0
0.7
0.5
m
m
MPa
MPa
m
m
deg.
deg.
–
MPa
Reference fracture hydraulic aperture
Reference fracture void aperture
Stiffness parameter
Stiffness parameter
Residual fracture hydraulic aperture
Residual fracture void aperture
Shear dilation angle
Shear dilation angle
Coefficient of friction
Fracture cohesion
transitional strike-slip/normal faulting stress regime where the state of
stress at a depth of 3.6 km was σHR ≈ σVR = 90 MPa, σhR = 46 MPa, and
p0 = 31 MPa. The initial reservoir temperature at the injection interval
was T0 = 230°C. The host rock was assumed to have a very low permeability of km = 1 ×10−19 m2 (100 nd) and a porosity of ϕm = 0.05.
Elastic properties of the matrix rock were assumed to be typical of
granite (Jaeger et al., 2007). The fractures had a constant coefficient of
friction of f = 0.7. The initial conditions, boundary conditions, and
model domain geometry are listed in Table 1. The fluid and rock
properties are listed in Tables 2 and 3, respectively. The fracture
properties and constitutive parameters that related fracture deformation and flow are listed in Table 4. The boundary condition at the well
4.3. Stimulated fracture network growth and microseismicity
The growth of the stimulated reservoir volume resulting from the
mixed-mechanism process is illustrated in Fig. 8. As preexisting natural
fractures were subjected to pressure perturbations, shear slip occurred
resulting in complex stress redistributions that concentrated stresses
near fracture tips. Tensile stress concentrations encouraged hydraulic
141
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
Fig. 7. Simulated injection rate and injection bottomhole pressure during the four main stimulation treatments performed on well EE-2.
splay fractures to nucleate. The continuous fluid injection source allowed the splay fractures to continue to propagate in the direction of σH
(N30°E). Newly formed fractures tended to terminate against natural
fractures, resulting in an overall north–south migration of the stimulated region. Hydraulic fracture termination, which was a model assumption, was requisite to control the spatial migration of the stimulated zone.
The microseismicity associated with the stimulation is shown in
Fig. 9. The location of the microseismic events were calculated by assuming that the rate of sliding deformation of the fractures was related
directly to the generation of seismicity. The dimensions of the cloud of
seismicity (roughly 1 km in the north–south direction and 500 m in the
east-west direction) was in agreement with the field observations
(compare to Fig. 2). Seismicity was observed to propagate relatively
uniformly away from the injection well.
Comparing the trends the migration of seismicity away from the
well, shown in Figs. 3 and 10, we observe several qualitative similarities. The lack of seismicity in the first several hours of injection, most
likely attributable to the Kaiser effect resulting from prior stimulation
treatments, was captured well by the model. In the model results, bursts
of seismicity occurred as hydraulic fractures terminated against natural
fractures, leading to rapid sliding along newly connected natural fractures. This is a plausible mechanism for the bursts in seismicity observed in the field experiments. The rate at which seismicity propagated
away from the well was similar to the observations. In the model, the
rate of extension of the stimulated zone was influenced by the density of
natural fractures and their initial transmissivity. Following shut-in of
the injection well, the modeled seismicity occurred only at the edges of
the stimulated region, similar to the behavior observed in the field and
at other EGS sites (Majer et al., 2007; Mukuhira et al., 2007). This
behavior supports the observation discussed by McClure and Horne
(2011) and McClure (2015) that pressure is able to rise near the edges
of the stimulated region even while pressure is falling in the nearwellbore region after shut-in. The mixed-mechanism behavior caused a
nonuniform distribution of fracture transmissivity throughout the reservoir.
5. Alternative conceptual models of stimulation process
We explored two alternative hypotheses of the stimulation process
at Fenton Hill: (1) Was shear stimulation the primary mechanism for
permeability enhancement? (2) Is it possible that a pure opening mode
fracture propagated away from the wellbore and arrested against a
permeable fault structure such that leakoff into the fault zone caused
the misoriented seismicity? In Section 5.1, we present the results from
additional numerical simulations performed to test the pure shear stimulation hypothesis (Fig. 11(a) shows the natural fracture network
used in the simulations). In Section 5.2, we present an analysis of how
leakoff into a permeable fault can influence the arrest duration of a
propagating hydraulic fracture (see Fig. 11(b)). We found that both of
these scenarios are unlikely and prefer the mixed-mechanism stimulation conceptual model to explain the hydromechanical and geophysical
observations at the Fenton Hill site.
5.1. Pure shear stimulation
To investigate whether shear stimulation may have been a plausible
mechanism for permeability enhancement at Fenton Hill, we performed
142
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
Fig. 8. Illustration of the stimulated fracture network at the end of each of the four main stimulation treatments performed on Well EE-2. The black lines are the preexisting natural
fractures, and the red lines are the newly formed hydraulic splay fractures. As a result of the mixed-mechanism process, the interaction between the natural fractures and propagating splay
fractures caused an overall north–south migration of the stimulated zone. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the
article.)
143
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
intersecting the well the fracture transmissivity began to increase due to
shear dilation. Accordingly, the well was able to accept fluid at a bottomhole pressure of approximately 45 MPa, significantly lower than
recorded in the actual data. The maximum pressure reached during the
simulation of Expt. 2018 was only 55 MPa. In the next treatment (Expt.
2020) the well began to accept fluid at about 39 MPa and pressure
began to rise steadily, in contrast to the sharp pressure-rollover behavior observed in the mixed-mechanism simulations and the recorded
data. The influence of the irreversible injectivity gains caused by shear
stimulation were again evident in the simulations of Expt. 2032 (prepump) and Expt. 2032 (MHF). Our simulation results demonstrate that
pure shear stimulation is largely inconsistent with the hydromechanical
observations at Fenton Hill.
5.2. Hydraulic fracture arresting against large faults
In rare instances, hydraulic fractures have been observed to intersect with and terminate against relatively large fault structures (Yang
and Zoback, 2014). It is possible that the misoriented seismicity was
caused by leakoff from a hydraulic fracture into a permeable, preexisting fault zone. In this conceptual model, the fault zone permeability must have been sufficient to accommodate the high rates of
leakoff necessary to arrest the hydraulic fracture for over 60 h (otherwise seismicity would have propagated in the direction of σH). Moreover, because fluid pressures exceeded the fracturing pressure for the
entire duration of the MHF treatment and given that seismicity migrated symmetrically away from the wellbore, it would have been necessary that two faults existed to prevent a two-wing fracture from
propagating in both directions (see Fig. 11(b)).
Fig. 9. Distribution of microseismicity resulting from the mixed-mechanism stimulation
process during the simulation of Expt. 2032 (MHF).
5.2.1. Model constraints
If kilometer-scale fault structures existed at the Fenton Hill site, they
must have been relatively isolated and not connected to the far field
based on the following lines of evidence:
• The geothermal gradient is conductive, not advective, indicating
•
that no large-scale hydrothermal activity occurs at the site (Brown
et al., 2012). Therefore, faults in the Fenton Hill system must not be
well-connected.
During long-term circulation tests, the difference between injected
and produced fluid volumes was minimal (Brown et al., 2012). This
suggests that fluid losses into the matrix rock or to the far-field was
not significant, and therefore the fault structures must have finite
extent.
The seismicity in Fig. 2 can be characterized as a broad cloud of
roughly 1 km (north–south) by 1 km (vertical) by 300 m (east-west)
dipping steeply to the east (Phillips et al., 1997). The errors in event
location were estimated to be 10–30 m (Fehler et al., 1987; Fehler,
1989; Phillips et al., 1997). The earthquake magnitudes were observed
to range between −3 ≤ M ≤ 0 (Murphy and Fehler, 1986). This suggests that, under the assumptions of this alternative conceptual model,
the entire fault zone was unable to slip coherently (Fehler, 1989;
Phillips et al., 1997).
Assuming that fluid flow along the fault zone caused the microearthquakes, the seismicity can provide constraints on the hydraulic
properties of the fault. It has been argued that the hydraulic diffusivity
of a porous medium, D, can be estimated by fitting a curve that is
proportional to D t to the leading edge of seismicity propagation in a
space–time plot (Rothert and Shapiro, 2007; Shapiro et al., 2005).
Based on this approach, Shapiro et al. (2005) analyzed the Fenton Hill
seismicity and estimated the hydraulic diffusivity to be D = 0.15 m2/s.
The fault rock compressibility and porosity was assumed to be similar
to the values for granitic basement rock cited by Townend and Zoback
(2000). The void aperture was assumed to be E = 0.1 m based on a 5 m
wide fault damage zone with an average porosity of 0.02. Water
Fig. 10. Migration of microseismicity away from the wellbore during the simulation of
Expt. 2032 (MHF).
additional simulations that assumed strong shear dilation effects. All
model parameters were the same as in the mixed-mechanism simulations, except the shear dilation angle was set to a nonzero value of
φe = φE = 0.5°. The primary fracture set had the same orientation as in
the mixed-mechanism simulations. No new hydraulic splay fractures
were allowed to form, but a secondary fracture set was specified with
an average orientation perpendicular to the least principal stress (similar to the hypothesis of Brown et al. (2012)). The natural fracture
network geometry is illustrated in Fig. 11(a).
As in the simulations of the mixed-mechanism process, we modeled
the four main stimulation treatments performed on Well EE-2. A comparison of the rate/pressure response between the two simulations is
shown in Fig. 12. In the first treatment (Expt. 2018) the pressure initially rose quickly indicating low initial near-well injectivity. As the
fluid pressure reached a level sufficient to cause slip on the fracture
144
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
Fig. 11. Illustration of the alternative conceptual models. (a) The natural fracture network geometry used in the simulations of the pure shear stimulation hypothesis. (b) A two-wing
hydraulic fracture propagating from the well and arresting against large, permeable fault structures.
shown to be D = ϒλ/(Eβ) (Norbeck and Horne, 2016a,b,c). Then, the
mass leakoff rate can be described in terms of the fault and fluid
properties as:
compressibility and viscosity were the same as in the mixed-mechanism
simulations.
Using the diffusivity and storativity values, we were able to estimate
the fault transmissivity as ϒ = 10−15 m3. We also varied the fault
transmissivity by an order of magnitude about this base case estimate
(D was held constant by varying S along with ϒ). The pressure drop
driving leakoff was calculated as Δp = σh − p0 = 35 MPa based on the
assumption that the pressure in the hydraulic fracture was equal to the
least principal stress (the high-stress profile shown in Figs. 4 and 5 was
used in this analysis, under the interpretation that the pressure-rollover
behavior was in response to hydraulic fracture opening at the wellbore).
mL = 2ϒhρλ Δp (π D t)−1/2 .
The Carter leakoff rate (Eq. (14)) can be used to estimate the timedependent leakoff rate from a hydraulic fracture into a fault zone.
5.2.3. Results of fracture arrest analysis
Throughout the majority of the MHF treatment the injection rate
was roughly 100 kg/s (see Fig. 1). For a two-wing fracture forming off
the well, the critical leakoff rate necessary to arrest each fracture wing
would have been 50 kg/s. In stark contrast to typical hydraulic fracture
treatments performed today, high-rate, high-pressure injection occurred into a single stage for over 60 h.
The time-dependent leakoff rate is shown in Fig. 13. In order to
maintain arrest of the hydraulic fracture, the leakoff rate must remain
larger than the pumping rate. Due to the qL ∼ t−1/2 scaling, when the
fault transmissivity is increased by an order of magnitude, the arrest
duration is increased by two orders of magnitude. For the base case
estimate ϒ = 10−15 m3, we found that the arrest duration would have
been less than 10 min. For the case with the highest fault transmissivity
we found that the arrest duration was approximately 5 h, significantly
shorter than the full pumping time of 61 h. Fault transmissivity is
known to vary over several orders of magnitude, so it may be possible
that an extremely permeable fault could have accommodated the
leakoff rate necessary to arrest the hydraulic fracture for the full 61 h.
However, the existence of large-scale permeable fault structures is at
odds with most other observations at the site. For example, the absence
of advective temperature gradients suggests the lack of permeable faults
in the reservoir. In addition, Wells EE-2 and EE-3 were both sidetracked
following the MHF treatment to intersect the zone of microseismicity,
but did not experience any significant lost circulation events that would
suggest the presence of permeable faults. Similarly, the sidetracked
wells both had to be stimulated at pressures above the least principal
5.2.2. Carter leakoff into fault zones
In the Carter leakoff model (Economides and Nolte, 2000), the
leakoff rate is largest initially and decays over time as pressure in the
porous medium near the hydraulic fracture increases. The fluid leakoff
velocity is:
vL = CL t −1/2,
(12)
where CL is the Carter leakoff coefficient and t is the time elapsed since
the hydraulic fracture intersected the fault. The fault has a vertical
height h and a hydraulic aperture e. The total volumetric leakoff rate
(accounting for leakoff into the fault from both sides of the hydraulic
fracture) is qL = 2CLAt−1/2, where the cross-sectional area of the fault is
A = eH. The mass leakoff rate is mL = ρqL.
The Carter leakoff coefficient for diffusion-driven leakoff is:
1/2
k f λϕ f β ⎞
CL = ⎛
⎝ π ⎠
⎜
⎟
Δp ,
(13)
where ϕ is the fault porosity, k is the fault permeability, β is the total
compressibility of the fault pore volume and fluid within the fault zone,
λ is inverse fluid viscosity, and Δp = pf − p0 is the pressure drop driving
leakoff. The hydraulic diffusivity for fluid flow along a fault zone can be
f
(14)
f
145
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
Fig. 12. Comparison of bottomhole pressure profiles between the mixed-mechanism and pure shear stimulation treatments. The mixed-mechanism case exhibited a reversible injectivity
change that was consistent with the observations, whereas the shear stimulation case exhibited permanent injectivity enhancement after the first treatment.
stress before a hydraulic connection was achieved. Therefore, it is unlikely that the misoriented seismicity was caused by leakoff into a fault
that arrested a propagating hydraulic fracture.
6. Discussion
The shear stimulation conceptual model was invented largely based
on interpretations of the Fenton Hill data (Tester et al., 1989, 2006),
and has guided the design of nearly all historic EGS projects (McClure
and Horne, 2014). Several EGS projects have shown clear evidence for
effective shear stimulation, for example at Soultz-sous-Forêts (Evans
et al., 2005; Genter et al., 2000) and the Habanero well at Cooper Basin
(Wyborn et al., 2005; Baisch et al., 2006, 2009). Based on the results of
our numerical simulations, we argue that the mixed-mechanism stimulation conceptual model is more consistent with the variety of hydromechanical and geophysical observations at the Fenton Hill site.
The modeling results suggest that the Fenton Hill reservoir experienced nonlinear permeability changes that depended on the evolution
of the state of stress throughout the reservoir. Fractures were able to
dilate as the effective normal stress acting on the fractures decreased
during injection. This type of deformation depended on the orientation
of the fractures and was reversible. In addition, hydraulic splay fractures were able to propagate through the reservoir while injecting at
pressures above the least principal stress. The splay fractures created
Fig. 13. Carter leakoff rate from a hydraulic fracture into a permeable fault structure (see
Eq. (14)). The solid colored lines represent the time-dependent leakoff rate for several
different values of fault transmissivity, ϒ. For each case, the hydraulic diffusivity, D, was
held constant. The horizontal dashed black line represents the minimum leakoff rate necessary to arrest the hydraulic fracture. The vertical solid black line represents the 61 h
pumping duration of Expt. 2032. (For interpretation of the references to color in this
figure legend, the reader is referred to the web version of the article.)
146
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
Fig. 14. Growth of the stimulated fracture network during the simulation of Expt. 2032 (MHF). The relative contribution of the hydraulic splay fractures and natural fractures is
highlighted. The top frame shows the evolution of the connected fracture surface area and the bottom frame shows the evolution of the fluid storage volume of the stimulated fracture
network.
new flow pathways which caused an irreversible, reservoir-scale change
in permeability. The overall stimulated network therefore consisted of
both natural and hydraulic fractures.
As opposed to shear stimulation and pure opening-mode hydraulic
fracturing, the mixed-mechanism process activates multiple fracture
sets that may result in a relatively complex network of flow pathways.
In the simulations, we observed that each of the fracture sets contributed to the reservoir growth in unique ways that would have a
significant impact on EGS reservoir engineering design. In Fig. 14, we
show the evolution of the stimulated fracture network during Expt.
2032. The fracture surface area of the stimulated zone was dominated
by the preexisting natural fractures. In contrast, the newly formed splay
fractures formed the bulk of the storage volume. Because the splay
fractures were oriented perpendicular to the least principal stress,
therefore bearing a relatively low normal stress, they were able to dilate
significantly compared to the natural fractures. Heat recovery from
geothermal reservoirs is known to be influenced significantly by the
surface area available for heat transfer between the working fluid
flowing though the fractures and the surrounding rock, and the fluid
storage volume impacts the residence time of fluids circulating through
the reservoir (Grant and Bixley, 2011). By recognizing the relative influence of various fracture sets on the flow and heat recovery processes
during active circulation, it may be possible to optimize future EGS
designs to leverage mixed-mechanism behavior.
It is likely that although fluid pressures exceeded the magnitude of
the minimum principal stress, hydraulic fractures did not initiate at the
wellbore. This type of behavior may encourage mixed-mechanism stimulation. Baumgartner and Zoback (1989) observed similar behavior in
a shallow vertical wellbore completed in gneiss, and attributed the inability to form a hydraulic fracture to the high tensile strength of the
rock. At the same site, Baumgartner and Zoback (1989) reported that a
subhorizontal plane of weakness was able to be opened without the
formation of a hydraulic fracture (the least principal stress was vertical
at that site). The Fenton Hill reservoir was a fractured granitic
reservoir. Similar behavior has been interpreted at other EGS projects in
granite reservoirs including Hijiori, Ogachi, Le Mayet de Montange, and
Rosemanowes, and Cooper Basin (Baisch et al., 2015; Cornet and
Morin, 1997; Holl and Barton, 2015; McClure and Horne, 2014; Pine
and Batchelor, 1984; Sasaki, 1998). At Cooper Basin, Australia, the
stimulation of well Jolokia 1 behaved in a manner contradictory to
most other stimulation treatments at the site (Baisch et al., 2015; Holl
and Barton, 2015). At Jolokia 1, fluid pressures exceeded the least
principal stress by at least 10–20 MPa, but the well only achieved injection rates on the order of 1 L/s (only 380 m3 of fluid were injected
over the 8 day injection period), suggesting that significant hydraulic
fracturing did not occur (Baisch et al., 2015). At the Pohang EGS site in
South Korea, it has been reported that stimulation treatments caused
shear stimulation in one well and hydraulic fracturing at another well
located in close proximity and at the same depth (Park et al., 2017). The
Pohang observations suggest that it is possible to achieve shear stimulation and propagation of new fractures in the same reservoir, providing
possible evidence for mixed-mechanism processes in geothermal reservoirs. It is not apparent whether this behavior occurs in all types of
formations.
7. Conclusions
In this work, we performed an investigation of the Fenton Hill EGS
test site to develop an improved understanding of the geologic structure
and hydromechanical behavior of fractured geothermal reservoirs.
Using interpretations of the data sets recorded during several field experiments at Fenton Hill, we designed a conceptual model of the geologic structure and stimulation mechanism at the site. We hypothesized
that stimulation (i.e., permeability enhancement) occurred though a
mixed-mechanism process controlled largely by propagating splay
fractures that formed due to stress concentrations created as pressurized
natural fractures slipped and opened. We applied a numerical model
that coupled fluid flow, heat transfer, elasticity, shear failure, and
147
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
fracture propagation in order to validate our hypothesis.
During the stimulation phase, our simulation results were consistent
with observations recorded during the field experiments in three distinct ways:
Bull. Seismol. Soc. Am. 96 (6), 2242–2256. http://dx.doi.org/10.1785/0120050255.
Bandis, S., Lumsden, A., Barton, N., 1983. Fundamentals of rock joint deformation. Int. J.
Rock Mech. Min. Sci. Geomech. Abstr. 20 (6), 249–268. http://dx.doi.org/10.1016/
0148-9062(83)90595-8.
Barton, C., Zoback, M., Burns, K., 1988. In-situ stress orientations and magnitude at the
Fenton Hill geothermal site, New Mexico, determined from wellbore breakouts.
Geophys. Res. Lett. 15 (5), 467–470. http://dx.doi.org/10.1029/GL015i005p00467.
Barton, N., Bandis, S., Bakhtar, K., 1985. Strength, deformation and conductivity coupling
of rock joints. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 22 (3), 121–140. http://
dx.doi.org/10.1016/0148-9062(85)93227-9.
Baumgartner, J., Zoback, M., 1989. Interpretation of hydraulic fracturing pressure–time
records using interactive analysis methods. Int. J. Rock Mech. Min. Sci. Geomech.
Abstr. 26 (6), 461–469. http://dx.doi.org/10.1016/0148-9062(89)91422-8.
Bradley, A., 2014. Software for efficient static dislocation-traction calculations in fault
simulators. Seismol. Res. Lett. 85 (6), 1–8. http://dx.doi.org/10.1785/0220140092.
Brown, D., 1989. The potential for large errors in the inferred minimum earth stress when
using incomplete hydraulic fracturing results. Int. J. Rock Mech. Min. Sci. Geomech.
Abstr. 26 (6), 573–577. http://dx.doi.org/10.1016/0148-9062(89)91437-X.
Brown, D., Duchane, D., Heiken, G., Hriscu, V., 2012. Mining the Earth's Heat: Hot Dry
Rock Geothermal Energy. Springer.
Charoenwongsa, S., Kazemi, H., Miskimins, J., Fakcharoenphol, P., 2010. A fully-coupled
geomechanics and flow model for hydraulic fracturing and reservoir engineering
applications. In: Canadian Unconventional Resources & International Petroleum
Conference, vol. 6. Calgary, Alberta, Canada.
Cornet, F.H., Morin, R.H., 1997. Evaluation of hydromechanical coupling in a granite
rock mass from a high-volume high-pressure injection experiment: Le Mayet de
Montagne, France. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 34 (3–4). http://dx.
doi.org/10.1016/S1365-1609(97)00185-8. 207.e1–207.e14.
Duchane, D., 1991. Hot Dry Rock Geothermal Energy Development Program. Tech. Rep.
HDR-90-Annual Report. Los Alamos National Laboratory.
Economides, M., Nolte, K., 2000. Reservoir Stimulation, 3rd edition. Wiley.
Evans, K., Moriya, J., Niitsuma, H., Jones, R., Phillips, W., Genter, A., Sausse, J., Jung, R.,
Baria, R., 2005. Microseismicity and permeability enhancement of hydrogeologic
structures during massive fluid injections into granite at 3 km depth at the Soultz
HDR site. Geophys. J. Int. 160 (1), 388–412. http://dx.doi.org/10.1111/j.1365246X.2004.02474.x.
Fehler, M., 1989. Stress control of seismicity patterns observed during hydraulic fracturing experiments at the Fenton Hill hot dry rock geothermal energy site, New
Mexico. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 26 (3–4), 211–219. http://dx.
doi.org/10.1016/0148-9062(89)91971-2.
Fehler, M., House, L., Kaieda, H., 1987. Determining planes along which earthquakes
occur: method and application to earthquakes accompanying hydraulic fracturing. J.
Geophys. Res. 92 (B9), 9407–9414. http://dx.doi.org/10.1029/JB092iB09p09407.
Fu, W., Ames, B., Bunger, A., Savitski, A., 2016. Impact of partially cemented and nonpersistent natural fractures on hydraulic fracture propagation. Rock Mech. Rock Eng.
49 (11), 4519–4526. http://dx.doi.org/10.1007/s00603-016-1103-0.
Garipov, T., Karimi-Fard, M., Tchelepi, H., 2016. Discrete fracture model for coupled flow
and geomechanics. Comput. Geosci. 20, 149–160. http://dx.doi.org/10.1007/
s10596-015-9554-z.
Genter, A., Traineau, H., Ledésert, B., Bourgine, B., Gentier, S., 2000. Over 10 years of
geological investigations within the HDR Soultz project, France. In: World
Geothermal Congress. Kyushu - Tohoku, Japan.
Ghassemi, A., Zhou, X., 2011. A three-dimensional thermo-poroelastic model for fracture
response to injection/extraction in enhanced geothermal systems. Geothermics 40
(1), 39–49. http://dx.doi.org/10.1016/j.geothermics.2010.12.001.
Grant, M., Bixley, P., 2011. Geothermal Reservoir Engineering, 2nd edition. Academic
Press.
Gringarten, A., Witherspoon, P., Ohnishi, Y., 1975. Theory of heat extraction from fractured hot dry rock. J. Geophys. Res. 80 (8), 1120–1124. http://dx.doi.org/10.1029/
JB080i008p01120.
Gu, H., Weng, X., Lund, J., Mack, M., Ganguly, U., Suarez-Rivera, R., 2012. Hydraulic
fracture crossing natural fracture at nonorthogonal angles: a criterion and its validation. SPE J. 27 (1), 20–26. http://dx.doi.org/10.2118/139984-PA.
Holl, H.-G., Barton, C., 2015. Habanero field – structure and state of stress. In: World
Geothermal Congress. Melbourne, Australia. pp. 1–8.
House, L., Keppler, H., Kaieda, H., 1985. Seismic studies of a massive hydraulic fracturing
experiment. In: Geothermal Resources Council Transactions, vol. 9. Davis, CA, USA.
Jaeger, J., Cook, N., Zimmerman, R., 2007. Fundamentals of Rock Mechanics, 4th edition.
Blackwell Publishing Ltd., Oxford.
Jeffrey, R., Zhang, X., Jung, R., 2015. Reassessing stimulation for EGS reservoirs. In:
Proceedings World Geothermal Congress. Melbourne, Australia.
Karimi-Fard, M., Durlofsky, L., Aziz, K., 2004. An efficient discrete-fracture model applicable for general-purpose reservoir simulators. SPE J. 9 (2), 227–236. http://dx.
doi.org/10.2118/88812-PA.
Karvounis, D., Jenny, P., 2016. Adaptive hierarchical fracture model for enhanced geothermal systems. SIAM 14 (1), 207–231. http://dx.doi.org/10.1137/140983987.
Kelkar, S., Murphy, H., Dash, Z., 1986. Earth stress measurements in deep granitic rock.
In: 27th US Symposium on Rock Mechanics. Tuscaloosa, Alabama, USA.
Lee, S., Jensen, C., Lough, M., 2000. Efficient finite-difference model for flow in a reservoir with multiple length-scale variations. SPE J. 5 (3), 268–275. http://dx.doi.
org/10.2118/65095-PA.
Majer, E., Baria, R., Stark, M., Oates, S., Bommer, J., Smith, B., Asanuma, H., 2007.
Induced seismicity associated with Enhanced Geothermal Systems. Geothermics 36,
185–222. http://dx.doi.org/10.1016/j.geothermics.2007.03.003.
Matsunaga, I., Kadowake, M., Murphy, H., 1983. Current Summary of Hydraulic
Fracturing Experiments in Phase II Reservoir. Tech. Rep. Los Alamos National
1. Extremely low injectivity at Well EE-2 was observed until reaching
an injection bottomhole pressure of roughly 70 MPa, whereupon
injectivity increased significantly.
2. Low injectivity (at low pressure) at Well EE-2 was observed consistently throughout multiple stimulation treatments. That is, near
wellbore permeability enhancement following each stimulation
treatment was reversible.
3. An overall north–south migration of microseismicity was observed.
This was despite injection pressures exceeding the fracturing pressure, which would tend to cause hydraulic fractures to propagate in
the direction of the maximum principal stress (N30°E at Fenton
Hill).
The results of our numerical experiments demonstrate that several
independent hydromechanical observations could be explained reasonably by the mixed-mechanism stimulation conceptual model. We
explored two alternative hypotheses for stimulation mechanism, but
found that neither the pure shear stimulation nor pure opening-mode
conceptual models were able to replicate the field observations at the
Fenton Hill EGS site with sufficient accuracy.
If mixed-mechanism stimulation can occur in geothermal reservoirs,
then future EGS projects could be designed to take advantage of the
interaction between distinct fracture sets activated during the stimulation process. Given the sensitivity of fracture flow properties to local
variability in stress and fluid pressure, the hydromechanical response of
hydraulic fractures and preexisting natural fractures may vary significantly. Our modeling results showed that the hydraulic splay fractures controlled the fluid storage volume in the stimulated fracture
network, whereas the heat transfer surface area was dominated by the
natural fractures. Reservoir engineering designs could leverage mixedmechanism stimulation to optimize for heat mining efficiency, delayed
thermal breakthrough, and extended reservoir lifetime. Moreover, if
hydraulic fracture propagation in geothermal reservoirs can play a
more significant role than previously thought, it may be possible to
make use of completions engineering technology from the oil and gas
industry to improve EGS stimulation treatment design.
Acknowledgments
The project was funded by the U.S. Department of Energy
Geothermal Technologies Office geothermal code comparison study
under grant number DE-EE0006646. The numerical simulations were
performed at the Stanford Center for Computational Earth and
Environmental Science using their high performance computing resources. The authors are grateful to M. White and the rest of the geothermal code comparison study participants for many stimulating discussions, as well as to J.O. Kaven, J. Taron, P. Fu and two anonymous
reviewers for their thoughtful feedback that improved this manuscript.
The reservoir simulation software used in this study, Complex
Fracturing ReseArch Code (CFRAC), can be licensed for research purposes by contacting J. Norbeck.
References
Baisch, S., Rothert, E., Stang, H., Voros, R., Koch, C., McMahon, A., 2015. Continued
geothermal reservoir stimulation experiments in the Cooper Basin (Australia). Bull.
Seismol. Soc. Am. 105 (1), 1–12. http://dx.doi.org/10.1785/0120140208.
Baisch, S., Vörös, R., Weidler, R., Wyborn, D., 2009. Investigation of fault mechanisms
during geothermal reservoir stimulation experiments in the Cooper Basin, Australia.
Bull. Seismol. Soc. Am. 99 (1), 148–158. http://dx.doi.org/10.1785/0120080055.
Baisch, S., Weidler, R., Vörös, R., Wyborn, D., de Graaf, L., 2006. Induced seismicity
during the stimulation of a geothermal HFR reservoir in the Cooper Basin, Australia.
148
Geothermics 74 (2018) 135–149
J.H. Norbeck et al.
microseismicity. J. Geophys. Res.: Solid Earth 112 (B04309), 1–16. http://dx.doi.
org/10.1029/2005JB003959.
Rutqvist, J., Stephansson, O., 1996. A cyclic hydraulic jacking test to determine the in situ
stress normal to a fracture. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 33 (7),
695–711. http://dx.doi.org/10.1016/0148-9062(96)00013-7.
Salimzadeh, S., Nick, H., Zimmerman, R., 2018a. Thermoporoelastic effects during heat
extraction from low-permeability reservoirs. Energy 142, 546–558. http://dx.doi.
org/10.1016/j.energy.2017.10.059.
Salimzadeh, S., Paluszny, A., Nick, H., Zimmerman, R., 2018b. A three-dimensional
coupled thermo-hydro-mechanical model for deformable fractured geothermal systems. Geothermics 71, 212–224. http://dx.doi.org/10.1016/j.geothermics.2017.09.
012.
Sasaki, S., 1998. Characteristics of microseismic events induced during hydraulic fracturing experiments at the Hijiori hot dry rock geothermal energy site, Yamagata,
Japan. Tectonophysics 289 (1–3), 171–188. http://dx.doi.org/10.1016/S00401951(97)00314-4.
Segall, P., 1989. Earthquakes triggered by fluid extraction. Geology 17 (10), 942–946.
http://dx.doi.org/10.1130/0091-7613.
Shapiro, S.A., Rentsch, S., Rothert, E., 2005. Characterization of hydraulic properties of
rocks using probability of fluid-induced microearthquakes. Geophysics 70 (2),
F27–F33. http://dx.doi.org/10.1190/1.1897030.
Smith, I., Griffiths, D., Margetts, L., 2014. Programming the Finite Element Method, 5th
edition. John Wiley & Sons, Ltd., Chichester.
Snow, D., 1965. A Parallel Plate Model of Fractured Permeable Media (Ph.D. thesis).
University of California, Berkeley.
Sonnenthal, E., Ito, A., Spycher, N., Yui, M., Apps, J., Sugita, Y., Conrad, M., Kawakami,
S., 2005. Approaches to modeling coupled thermal, hydrological, and chemical
processes in the drift scale heater test at Yucca Mountain. Int. J. Rock Mech. Min. Sci.
42, 698–719.
Tester, J., Anderson, B., Batchelor, B., Blackwell, A., DiPippo, R., Drake, E., Petty, S.,
2006. The Future of Geothermal Energy: Impact of Enhanced Geothermal Systems
(EGS) on the United States in the 21st Century. Tech. Rep. Massachusetts Institute of
Technology.
Tester, J., Brown, D., Potter, R., 1989. Hot Dry Rock Geothermal Energy – A New Energy
Agenda for the 21st Century. Tech. Rep. LA-11514-MS. Los Alamos National
Laboratory.
Townend, J., Zoback, M.D., 2000. How faulting keeps the crust strong. Geology 28 (5),
399–402. http://dx.doi.org/10.1130/0091-7613(2000)28<399:HFKTCS>2.0.CO;2.
Warpinski, N., 2009. Microseismic monitoring: inside and out. J. Pet. Technol. 61 (11),
80–85. http://dx.doi.org/10.2118/118537-JPT.
Warpinski, N., Du, J., Zimmer, U., 2012. Measurements of hydraulic-fracture-induced
seismicity in gas shales. SPE Prod. Oper. 27 (3), 240–252. http://dx.doi.org/10.
2118/151597-PA.
Weng, X., Kresse, O., Cohen, C., Wu, R., Gu, H., 2011. Modeling of hydraulic-fracturenetwork propagation in a naturally fractured formation. SPE Prod. Oper. 26 (4),
368–380. http://dx.doi.org/10.2118/140253-PA.
White, M., Fu, P., McClure, M., Danko, G., Elsworth, D., Sonnenthal, E., Kelkar, S.,
Podgorney, R., 2017. A suite of benchmark and challenge problems for enhanced
geothermal systems. Geomecha. Geophys. Geo-Energy Geo-Resour. http://dx.doi.
org/10.1007/s40948-017-0076-0.
White, M., Podgorney, R., Kelkar, S., McClure, M., Danko, G., Ghassemi, A., Fu, P.,
Bahrami, D., Barbier, C., Cheng, Q., Chiu, K., Detournay, C., Elsworth, D., Fang, Y.,
Furtney, J., Gan, Q., Gao, Q., Guo, B., Yao, H., Horne, R., Huang, K., Im, K., Norbeck,
J., Rutqvist, J., Safari, R., Sesetty, V., Sonnenthal, E., Tao, Q., White, S., Wong, Z.,
2016. Benchmark Problems of the Geothermal Technologies Office Code Comparison
Study. Tech. Rep. PNNL-26016. Pacific Northwest National Laboratory.
White, S., Purohit, S., Boyd, L., 2015. Using GTO-Velo to facilitate communication and
sharing of simulation results in support of the Geothermal Technologies Office code
comparison study. In: Fortieth Workshop on Geothermal Reservoir Engineering.
Stanford, CA, USA.
Willis-Richards, J., Watanabe, K., Takahashi, H., 1996. Progress toward a stochastic rock
mechanics model of engineered geothermal systems. J. Geophys. Res.: Solid Earth
101 (B8), 17481–17496. http://dx.doi.org/10.1029/96JB00882; doi:10.1029/9.
Witherspoon, P., Wang, J., Iwai, K., Gale, J., 1980. Validity of Cubic Law for fluid flow in
a deformable rock fracture. Water Resour. Res. 16 (6), 1016–1024. http://dx.doi.org/
10.1029/WR016i006p01016.
Wyborn, D., de Graaf, L., Davidson, S., Hann, S., 2005. Development of Australia's first
hot fractured rock (HFR) underground heat exchanger, Cooper Basin, South
Australia. In: World Geothermal Congress. Antalya, Turkey.
Yang, Y., Zoback, M., 2014. The role of preexisting fractures and faults during multistage
hydraulic fracturing in the Bakken Formation. Interpretation 2 (3), SG25–SG39.
http://dx.doi.org/10.1190/INT-2013-0158.1.
Ye, Z., Ghassemi, A., 2018. Experimental study on injection-induced fracture propagation
and coalescence for EGS stimulation. In: 43rd Workshop on Geothermal Reservoir
Engineering. Stanford, CA, USA. pp. 1–9.
Zhang, X., Jeffrey, R., 2016. Fluid-driven nucleation and propagation of splay fractures
from a permeable fault. J. Geophys. Res.: Solid Earth 121, 5257–5277. http://dx.doi.
org/10.1002/2016JB013304.
Laboratory.
McClure, M., 2012. Modeling and Characterization of Hydraulic Stimulation and Induced
Seismicity in Geothermal and Shale Gas Reservoirs (Ph.D. thesis). Stanford
University.
McClure, M., 2015. Generation of large postinjection-induced seismic events by backflow
from dead-end faults and fractures. Geophys. Res. Lett. 42, 6647–6654. http://dx.doi.
org/10.1002/2015GL065028.
McClure, M., Horne, R., 2011. Investigation of injection-induced seismicity using a coupled fluid flow and rate/state friction model. Geophysics 76 (6), WC181–WC198.
http://dx.doi.org/10.1190/geo2011-0064.1.
McClure, M., Horne, R., 2013a. Is pure shear stimulation always the mechanism of stimulation in EGS? In: Thirty-Eighth Workshop on Geothermal Reservoir Engineering.
Stanford, CA, USA.
McClure, M.W., Horne, R.N., 2013b. Discrete Fracture Network Modeling of Hydraulic
Stimulation: Coupling Flow and Geomechanics. Springer Briefs in Earth Sciences.
McClure, M., Horne, R., 2014. An investigation of stimulation mechanisms in Enhanced
Geothermal Systems. Int. J. Rock Mech. Min. Sci. 72, 242–260. http://dx.doi.org/10.
1016/j.ijrmms.2014.07.011.
Mossop, A., 2001. Seismicity, Subsidence and Strain at The Geysers Geothermal Field
(Ph.D. thesis). Stanford University.
Mukuhira, Y., Moriya, H., Ito, T., Asanuma, H., Häring, M., 2007. Pore pressure migration
during hydraulic stimulation due to permeability enhancement by low-pressure
subcritical fracture slip. Geophys. Res. Lett. 44, 1–10. http://dx.doi.org/10.1002/
2017GL072809.
Murphy, H., Fehler, M., 1986. Hydraulic fracturing of jointed formations. In: SPE Paper
14088 presented at the International Meeting on Petroleum Engineering. Beijing,
China. pp. 489–496.
Norbeck, J., 2016. Hydromechanical and Frictional Faulting Behavior of Fluid-InjectionInduced Earthquakes (Ph.D. thesis). Stanford University, Stanford, CA, USA.
Norbeck, J., Horne, R., 2015. Injection-triggered seismicity: an investigation of porothermoelastic effects using a rate-and-state earthquake model. In: Fortieth
Workshop on Geothermal Reservoir Engineering. Stanford, CA, USA.
Norbeck, J., Horne, R., 2016a. A numerical method for fractured reservoir poromechanics
using a mixed-continuum embedded fracture model. In: Geothermal Resources
Council Transactions. Sacramento, CA, USA.
Norbeck, J., Horne, R., 2016b. Evidence for a transient hydromechanical and frictional
faulting response during the 2011 Mw 5.6 Prague, Oklahoma earthquake sequence. J.
Geophys. Res.: Solid Earth 121 (12), 8688–8705. http://dx.doi.org/10.1002/
2016JB013148.
Norbeck, J., Horne, R., 2016c. Physical mechanisms related to microseismic-depletiondelineation field tests with application to reservoir surveillance. SPE J. 21 (4),
1279–1288. http://dx.doi.org/10.2118/178926-PA.
Norbeck, J., McClure, M., Horne, R., 2016a. Analysis of hydromechanical reservoir response during fluid circulation at the Fenton Hill Enhanced Geothermal System test
site. In: Proceedings of 38th New Zealand Geothermal Workshop. Auckland, New
Zealand.
Norbeck, J., McClure, M., Lo, J., Horne, R., 2016b. An embedded fracture modeling
framework for simulation of hydraulic fracturing and shear stimulation. Comput.
Geosci. 20 (1), 1–18. http://dx.doi.org/10.1007/s10596-015-9543-2.
Norbeck, J., McClure, M., Horne, R., 2016c. Revisiting stimulation mechanism at Fenton
Hill and an investigation of the influence of fault heterogeneity on the
Gutenberg–Richter b-value for rate-and-state earthquake simulations. In: 41st
Workshop on Geothermal Reservoir Engineering. Stanford, CA, USA. pp. 1–22.
Norbeck, J., Shelly, D., 2018. Exploring the role of mixed-mechanism fracturing and fluidfaulting interactions during the 2014 Long Valley Caldera, California, earthquake
swarm. In: 43rd Workshop on Geothermal Reservoir Engineering. Stanford, CA, USA.
pp. 1–16.
Olson, J.E., 2007. Fracture aperture, length and pattern geometry development under
biaxial loading: a numerical study with applications to natural, cross-jointed systems.
In: Lewis, H., Couples, G. (Eds.), The Relationship Between Damage and Localization,
vol. 289. Geological Society, London, pp. 123–142.
Park, S., Xie, L., Kim, K.-I., Kwon, S., Min, K.-B., Choi, J., Yoon, W.-S., Song, Y., 2017. First
hydraulic stimulation in fractured geothermal reservoir in Pohang PX-2 well.
Procedia Eng. 191, 829–837. http://dx.doi.org/10.1016/j.proeng.2017.05.250.
Phillips, W.S., House, L.S., Fehler, M.C., 1997. Detailed joint structure in a geothermal
reservoir from studies of induced microearthquake clusters. J. Geophys. Res. 102
(B6), 11745–11763. http://dx.doi.org/10.1029/97JB00762.
Pine, R.J., Batchelor, A.S., 1984. Downward migration of shearing in jointed rock during
hydraulic injections. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 21 (5), 249–263.
http://dx.doi.org/10.1016/0148-9062(84)92681-0.
Pollard, D., Fletcher, R., 2005. Fundamentals of Structural Geology. Cambridge
University Press, Cambridge.
Rana, R., 1984. Exploratory Study on the Influence of Thermal Stressing on the Strength
and Porosity of Granite (Ph.D. thesis). Stanford University.
Renshaw, C., Pollard, D., 1995. An experimentally verified criterion for propagation
across unbounded frictional interfaces in brittle, linear elastic materials. Int. J. Rock
Mech. Min. Sci. Geomech. Abstr. 32 (3), 237–249. http://dx.doi.org/10.1016/01489062(94)00037-4.
Rothert, E., Shapiro, S., 2007. Statistics of fracture strength and fluid-induced
149
Документ
Категория
Без категории
Просмотров
0
Размер файла
6 078 Кб
Теги
geothermics, 003, 2018
1/--страниц
Пожаловаться на содержимое документа