You are on page 1of 13

Geothermics 54 (2015) 96108

Contents lists available at ScienceDirect

Geothermics
journal homepage: www.elsevier.com/locate/geothermics

Experimental and numerical studies on the mechanical behaviour of


Australian Strathbogie granite at high temperatures: An application to
geothermal energy
Shishi Shao a , P.G. Ranjith a, , P.L.P. Wasantha a , B.K. Chen b
a
b

Deep Earth Energy Research Lab, Department of Civil Engineering, Monash University, Building 60, Melbourne, Victoria 3800, Australia
Department of Mechanical & Aerospace Engineering, Monash University, Building 36, Melbourne, Victoria 3800, Australia

a r t i c l e

i n f o

Article history:
Received 1 October 2012
Accepted 25 November 2014
Available online 13 January 2015
Keywords:
Geothermal
Granite
Brittleplastic
High temperature

a b s t r a c t
The effect of temperature on the mechanical behaviour of Strathbogie granite (ne-grained) was studied under unconned stress conditions. Fracturing behaviour of test specimens was studied using an
acoustic emission (AE) detection system and some crack propagation was also performed using electron microscopy scanning (SEM). The stressstrain curves showed plastic and post-peak behaviour for
temperatures above 800 C and the brittleplastic transition was observed to occur between 600 and
800 C for the uniaxially tested Strathbogie granite at a strain rate of 0.1 mm/min and room humidity.
Specimens were heated at a rate of 5 C/min with a 1 h holding period before testing. The AE results
showed that the increasing temperature reduces the stress thresholds for crack initiation and crack damage and extends the duration of stable crack propagation. Prevalence of ductile properties with increasing
temperature was also observed from AE results. The stressstrain and AE results reveal that the failure
modes of Strathbogie granite specimens changed from brittle fracturing to quasi-brittle shear fracturing
and eventually to ductile failure with increasing temperature. Temperature was observed to inuence the
colour of granite, and the initial white/grey colour changed to an oxidated reddish colour with increasing
temperature. The stressstrain data of tested specimens were incorporated into a nite element model
(ABAQUS 6.7.1), so that both plastic and ductile behaviour of the Strathbogie granite could be predicted
over a wide range of temperatures.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
Geothermal heat is now a recommended renewable energy
resource on the time-scales of both technological and societal systems, with cost, reliability and environmental advantages
over conventional energy resources (Rybach, 2003; Gallup, 2009;
Axelsson, 2010). Exploration of geothermal resources has posed
new challenges for engineers and geologists to counter rock engineering problems at high temperatures. Laboratory testing is an
important aspect of rock mechanics, which provides essential input
data for the design of engineering structures in the Earths crust
and mantle subjected to tectonic forces. Since the 1970s, a large
number of laboratory studies have been carried out to investigate the effect of temperature on the physical and mechanical

Corresponding author at: Deep Earth Energy Research Lab, Civil Engineering
Department, Clayton Campus, Monash University, VIC 3800, Australia.
Tel.: +61 3 99054982; fax: +61 3 99054944.
E-mail address: ranjith.pg@monash.edu (P.G. Ranjith).
http://dx.doi.org/10.1016/j.geothermics.2014.11.005
0375-6505/ 2014 Elsevier Ltd. All rights reserved.

properties of rocks in engineering applications, such as deep


mining, underground chambers for nuclear disposal storage (at
temperatures which generally vary from 100 to 300 C and will
increase over the storage period) and high-temperature thermal
cracking of rocks during mechanical drilling (at temperatures as
high as 1000 C), as well as the design of enhanced geothermal systems (EGSs) (at temperatures about 200 C) (Francois, 1980; Bauer
et al., 1981; Paquet et al., 1981; Heuze, 1983; Hirth and Tullis,
1989). The mechanical behaviour of rock can be signicantly inuenced by elevated temperatures in such underground projects. In
addition, different types of underground engineering applications
encounter different temperatures, which vary from room temperature to extremely high temperatures. Therefore, understanding the
effect of temperature on the physical and mechanical properties of
rock is of great importance for the design of rock structures and
safety assessment in underground rock engineering.
During the 1970s1980s, investigations mainly focused on
understanding the natural processes in the Earths crust, such
as rock deformation (faulting, folding and shearing), geothermal
activity and magmatic intrusions. Most of these studies reported

S. Shao et al. / Geothermics 54 (2015) 96108

the inuence of temperature on the mechanical behaviour of


granite (Heuze, 1983; Wang et al., 2002; Dwivedi et al., 2008;
Xu et al., 2008a,b). For example, a review study (Heuze, 1983)
reported that some mechanical, physical and thermal properties of
granitic rocks, including deformation modulus, Poissons ratio, tensile strength, compressive strength, cohesion and internal friction
angle and viscosity, all vary considerably with increasing temperatures. Although some other scholars have also studied the
physical and mechanical properties of rocks under high temperatures, the vast majority of previous experimental studies have
been performed by heating the specimens to predetermined temperature levels, but testing them at room temperature (Xu et al.,
2008a,b, 2009; Zhang et al., 2009). Since rocks in natural geothermal
reservoirs are subjected to continuous heating conditions, these
preheating test conditions may not exactly reproduce the in situ
temperature conditions. In addition, when pre-heated rock specimens are cooling down to room temperature, micro-structural
changes and irreversible thermally induced micro-cracking can
take place. Therefore, the results of experimental studies performed
on pre-heated specimens at room temperature are insufcient to
represent the essential characteristics of rocks at high temperature
in geothermal applications. The aim of this paper is to investigate
the mechanical behaviour of granite, which is a common rock type
in the Earths crust, at high temperatures (with continuous heating) under unconned stress conditions using both experimental
and numerical studies.
A review of the most pertinent studies in the literature regarding
the basic mineralogy of granite, brittleplastic transition of granite
and fracturing behaviour of rock is presented in the following subsections, followed by the results and discussion of the experimental
work and numerical simulation undertaken for the present paper.
1.1. Basic mineralogy of granitic rocks
Granites are generally medium- to coarse-grained igneous crystalline rocks that form by crystallization of certain slow-cooling
magma. The main minerals that form granite are quartz, plagioclase feldspars and alkali feldspar, and some amount of biotite,
muscovite and/or hornblende (Farndon, 2010). Granite is rich in
elements with heat-producing radioactive isotopes (K, Th, U), and
is thus commonly associated with temperature anomalies and elevated geothermal gradients within the crust. This feature makes it
a suitable geothermal reservoir rock. It also has extremely low permeability and high strength, which also make it a good potential
storage site for nuclear waste.
1.2. Brittleplastic transition of granitic rocks
Brittle to plastic transition in response to increasing temperature has been studied for different types of rocks. The experimental
results of Tullis and Yund (1987) demonstrated a transition from
dominantly micro-cracking to dominantly dislocation at approximately 300400 C for quartz and 550650 C for feldspar. Hueckel
et al. (1994) reported that the conning pressure at brittle to plastic
transition is generally reduced by elevated temperatures. According to their study, when westerly granite is subjected to triaxial
conditions, the conning pressure at brittle to semi-brittle transition drops from 2000 MPa at 360 C to about 500 MPa at 800 C, and
the compressive strength also drops as the temperature increases,
with a dramatic drop at 668 C.
Xu et al. (2009) had used different temperatures ranging from
room temperature to 1200 C for their testing. The results showed
that the phase-changing behaviour of brittleplastic transition
appears around 800 C and the mechanical properties of the granite
samples did not signicantly vary before that (Fig. 1). This transition temperature is higher than that of westerly granite, which was

97

Fig. 1. Stressstrain curves of granite after high temperature (Xu et al., 2008a,b).

found to be 668 C by Hueckel et al. (1994). Therefore, it is clear


that the brittleplastic transition temperature can be signicantly
different, even for the same type of rock. This is understandable,
considering the variation of rock mineral composition and microstructural properties for the same rock type obtained from different
sources.
1.3. Fracture development behaviour in rock
1.3.1. Stages of fracture development
Based on the characteristics of the axial stressaxial strain and
axial stresslateral strain curves of uniaxial compression tests,
Hoek and Bieniawski (1965) found that the crack propagation process of brittle materials consists of three main stages: (1) crack
closure followed by an elastic region; (2) crack initiation followed
by a stable crack propagation region; (3) crack damage followed
by unstable crack propagation until ultimate failure. In previous
studies, various methods such as stressstrain analysis, scanning
electron microscopy, photoelasticity, and acoustic emission have
been used to identify these crack development stages (Bieniawski,
1967; Eberhardt et al., 1998; Ranjith et al., 2008). When the material is subjected to compressive loading, closure of pre-existing
micro-cracks that are inclined to the applied loading direction takes
place (Bieniawski, 1967). Linear elastic deformation occurs after the
majority of pre-existing cracks have closed. This marks the beginning of the stable crack propagation region and the stress at the
transition stage is known as the crack initiation stress threshold
( ci ). In the stable crack propagation region the material deforms
elastically. As the loading is further increased, unstable crack propagation begins and the stress at the transition from stable crack
propagation to unstable crack propagation is referred to as the crack
damage stress threshold ( cd ) (Eberhardt et al., 1998).
1.3.2. Acoustic emission (AE) detection method to study
fracturing development
Acoustic emission (AE) detection is a non-destructive evaluation
method to study the crack propagation of a brittle material subjected to a stress eld (Ranjith et al., 2008). Brittle material suddenly
releases strain energy when a crack develops, which creates an elastic stress wave travelling from the location of energy release to the
samples surface. AE is used to detect and measure the transient
wave that is generated from the discrete acoustic waves produced
by each micro-crack, which can produce event data to interpret
the crack propagation of the material. AE detection is able to monitor micro-crack slip and formation relative to the stressstrain
response.

98

S. Shao et al. / Geothermics 54 (2015) 96108

Fig. 2. Summation of counts vs. axial stress of single fractured rock (Ranjith et al., 2004).

A number of previous studies used the acoustic emission technique to study crack growth in brittle material (Shield, 1997;
Eberhardt et al., 1998; Rudajev et al., 2000; Cai et al., 2007). Eberhardt used a combination of AE technology and stressstrain curve
measurement data in their experimental study (Eberhardt et al.,
1999). Chang and Lee (2004) studied the cracking and damage
mechanisms of Korean Hwang-deung granite and Yeo-san marble under triaxial compression using acoustic emissions. These
researchers have divided the process of crack development into ve
different stages: crack closure, crack initiation, secondary cracking,
crack coalescence and crack damage, which mirror the divisions
established by Hoek and Bieniawski (1965) based on stressstrain
behaviour. A similar type of division of stress threshold has been
carried out by Ranjith et al. (2008) in their experimental studies
on fractured rock under uniaxial compression. They used the characteristics of the curve of cumulative AE count versus axial stress
to dene the stress thresholds for different fracture development
stages. The crack initiation stress threshold ( ci ) has been dened
as the point where the curve of cumulative AE counts marks the
initial lift-off (Fig. 2). The region, which sustains a linear increase
of cumulative AE counts with increasing stress, is termed stable
crack propagation. When the curve starts exhibiting an exponential growth, the corresponding stress is called the crack damage
stress threshold ( cd ) and the region between crack damage and
failure is dened as unstable crack growth.

Fig. 3. Grain size distribution of the testing material.

Cylindrical core specimens 22.5 mm in diameter were prepared


following the International Society for Rock Mechanics (ISRM)
recommendations for uniaxial compressive strength testing. Core
samples with visible cracks were discarded. Core specimens were
then cut to limit their length to 45 mm. The two ends of the
specimens were ground in order to produce two parallel surfaces
perpendicular to the long dimension of the cylindrical specimen.

2. Experimental work
2.1. Specimen preparation
The tested rock specimens were produced from one macroscopically homogeneous Strathbogie granite block. The Strathbogie
granite is a high-level, discordant, composite granitoid intrusion in
south-eastern Australia. It is ne-grained and white grey and dark
brown in colour with a tested porosity of 0.463%. It contains magmatic garnet, and cordierite and biotite predominate in the bothlith
(Neil Phillips et al., 1981). Grain size (which is measured from the
thin section of rock samples) ranges from 0 mm to 200 m, with
a few grains larger than 300 mm (Fig. 3). An optical microscopy
image of Strathbogie granite is presented in Fig. 4. At room temperature, Strathbogie granite showed a bulk density of 1805.7 kg/m3 ,
peak compressive strength of 215.97 MPa and elastic modulus of
8.97 GPa.

Fig. 4. Scanning electron microscopy image of Strathbogie granite.

S. Shao et al. / Geothermics 54 (2015) 96108

2.2. Testing methodology


Unconned compressive strength testing was carried out to
study the mechanical behaviour of dry granite samples under
elevated temperatures. Specimens were tested at nine different temperatures (i.e. 23 C, 100 C, 200 C, 400 C, 600 C, 800 C,
900 C, 1000 C and 1100 C).
Specimens were heated in a high-temperature furnace, using
a heating rate of 5 C/min to reach the appointed testing temperatures, with the exception of 23 C at which the specimens were
tested at room temperature. This heating rate was adopted in order
to avoid provoking thermal shock and the development of stress
fractures. The temperature was kept constant for two hours in the
furnace after it reached the assigned temperature, to ensure uniformity in temperature across the specimen.
The testing procedure was divided into two stages due to
the limitations of laboratory facilities. Tests for the temperatures
from 23 C to 600 C were carried out using a servo-controlled
Instron testing machine with a loading capacity of 100 kN, and
the experimental temperatures were achieved inside an environmental furnace with a square transparent window (Fig. 5a). The
specimens were loaded in compression with a constant displacement rate of 0.1 mm/min (we conducted testing only with this
strain rate and different strain rates may produce signicantly different behaviours). Mechanical behaviours related to axial load
and axial strains were measured in the process of axial compression until specimens failed after exposure to high temperatures.
In addition, AE data were recorded in the AE detection system
during all testing to analyze the fracturing behaviour of test specimens subjected to elevated temperatures. These AE measurements
were obtained using the attached piezoelectric transducers which
transducers identify AE events when they receive a signal with
amplication.
Other tests (i.e. tests for the temperatures from 800 C to
1100 C) were carried out using a Universal testing machine with
a loading capacity of 500 kN and high temperatures were achieved
inside a cylindrical furnace (Fig. 5b). The specimens were loaded
in compression with a constant displacement rate of 0.1 mm/min.
During the entire deformation process the temperatures were
maintained constant at their designated levels for all specimens.
2.3. Results and discussion of mechanical testing
The results of the experimental work are discussed under four
sub-topics; (1) compressive strength, elastic modulus and strain
at failure, (2) failure mechanisms, (3) mineralogical analysis using
X-ray diffraction testing and (4) fracturing behaviour using AE monitoring.
2.3.1. Compressive strength, elastic modulus and strain at failure
The axial stress-axial strain plots of representative tests at
each temperature considered are given in Fig. 6. Table 1 summarizes the important mechanical observations obtained from axial
stressaxial strain variations. According to Fig. 6, the stressstrain
curves of 23 C, 100 C, 200 C, 400 C, 600 C, 800 C, 900 C,
1000 C, 1100 C show two different behaviours. The stressstrain
curves of 23 C to 600 C display an initial elastic increase of stress
with increasing strain until peak strength is reached and then a
sudden drop of stress. On the other hand, the curves of temperatures over 800 C display a more gradual decrease in stress after
peak strength with increasing strain at failure. The 1100 C curve
presents a distinct difference from other curves, displaying a signicant strain increase with a little increase in stress. This indicates
that at or above 1100 C Strathbogie granite behaves as partial melting (granite melts at about 1200 C).

99

Table 1
Mechanical properties of Strathbogie granite at different test temperatures (averaged results).
Temperature,

Peak strength,
MPa

Elastic
modulus, GPa

Failure strain, %

23
100
200
400
600
800
900
1000
1100

215.98
190.64
234.02
203.67
119.56
96.01
39.63
37.07
5.29

8.98
8.28
8.48
7.95
5.78
3.72
1.71
1.34
0.23

2.70
2.60
3.15
2.95
2.55
4.05
5.75
5.75
7.45

Variations of peak compressive strength, elastic modulus


and axial strain at failure (i.e. at peak compressive strength)
against the temperature are presented in Fig. 7(a)(c), respectively. From Fig. 7(a) and (b) it can be seen that the peak
compressive strength and elastic modulus decrease with increasing test temperature, with the exception of 200 C. In addition,
salient drops of both peak strength and elastic modulus can
be observed when the temperature increases from 600 C to
800 C. Strains at failure do not show a noteworthy temperature dependence until the temperature reaches 600 C,
but increase dramatically with increasing test temperature from
800 C (Fig. 7(c)). Overall, from the plotted variations of three
mechanical properties (compressive strength, elastic modulus and
strain to failure) against test temperature, a clear transition stage
of the mechanical response can be identied when the temperature increases from 600 C to 800 C for the tested granite. It is
believed that the brittleplastic transition for Strathbogie granite
(ne-grained) occurs within this temperature range (i.e. 600 and
800 C) under unconned stress conditions.
The average values of compressive strength, elastic modulus and
strain at failure were normalized to their respective values at room
temperature (i.e. 23 C) for a better comparison with similar results
reported in the literature. The calculated average values and normalized values of compressive strength ( c ), elastic modulus (E)
and strain at failure () for different temperatures (T) are shown in
Table 2.
Normalized values of compressive strength and elastic modulus,
observed in this study for Strathbogie granite, were plotted against
different temperatures, along with some results reported in the
literature for some other rock types (Figs. 8 and 9).
As a type of crystalline rock, granite displays signicant variability in characteristics, due to both its depositional history, such as
mineralogy, shape, size and size distribution, and post-depositional
experience, such as burial, thermal and uid interaction history
and cementation. The results of compressive testing from Xu et al.
(2009) and Neil Phillips et al. (1981) present study show an initial strength weakening and softening pattern and subsequent
strengthening and stiffening (Fig. 8). This pattern continues up to
200 C in this study, but in the study of Xu et al. (2009) it lasts up to
800 C. This may be due to the differences the mineral composition
and micro-structure of granite. In general, Strathbogie granite and
the UCS testing results of Bauer et al. (1981) and Xu et al. (2009)
show decreasing trends in both compressive strength and elastic
modulus with increasing temperature, which is more clearly visible for tests carried out at temperatures above 800 C (Figs. 8 and 9).
The transition in mechanical character denoted by the more steep
negative deection in the plots of Fig. 8 can be taken to mark a
transition in deformation mechanism with increasing temperature.
This deection in the curves of Fig. 8 (the elasticplastic transition) occurs at approximately the same temperature for the various
granites tested. However, the strength developing patterns dened

100

S. Shao et al. / Geothermics 54 (2015) 96108

Fig. 5. Two testing devices used for experiments (a) servo-controlled Instron machine with a loading capacity of 100 kN and environmental furnace; (b) universal testing
machine with loading capacity of 500 kN and cylindrical furnace.

Fig. 6. Axial stressaxial strain of Strathbogie granite over a range of temperatures.

by the curves of Fig. 9 are distinctly different, which illustrates a


variation in the mechanical response of the various granites tested
by heating.
2.3.2. Failure mechanisms
Failure mechanisms of tested samples evident from visual
inspection of post-failure samples can provide important information about the behaviour of the tested granite samples at
different temperatures. Different failure mechanisms include
inherent mechanical characteristics such as the amount of

energy released at the failure and microstructural changes during


deformation. Therefore, the understanding of failure mechanisms
of rock at different temperatures is very important for geothermal
applications, to enable practitioners to be aware of the expected
failure mechanism at different temperatures. Typical post-failure
images taken at the conclusion of each test are shown in Table 3.
The post-failure images show some distinct characteristics for
different temperatures. The failure patterns for test temperatures
from 23 C to 800 C are approximately diagonal and the specimens
were broken down to a number of little pieces showing more brittle

Table 2
Average and normalized values of compressive strength, elastic modulus and strain at failure.
T ( C)

Average  c (MPa)

Normalized  c

Average E (GPa)

Normalized E

Average

Normalized

23
100
200
400
600
800
900
1000
1100

215.97
190.63
234.02
203.66
119.56
96.01
38.53
37.07
5.29

1
0.883
1.084
0.943
0.554
0.445
0.178
0.172
0.024

8.97
8.28
8.48
7.94
5.77
3.72
1.71
1.34
0.23

1
0.923
0.945
0.885
0.644
0.415
0.190
0.150
0.026

0.027
0.026
0.032
0.030
0.026
0.040
0.058
0.057
0.075

1
0.966
1.172
1.103
0.952
1.500
2.142
2.134
2.775

S. Shao et al. / Geothermics 54 (2015) 96108

101

Fig. 7. (a) Peak strength of Strathbogie granite at different test temperatures; (b) elastic modulus of Strathbogie granite at different test temperatures; (c) failure strain at
peak load of Strathbogie granite at different test temperatures. The dashed line indicates the melting point (600 C) of quartz, K-feldspar, Na-plagioclase and micas.

characteristics. For the specimens tested at 900 C and 1000 C, failure occurred along only a few failure planes and closely resembled
shear failure. Specimens tested at 1100 C showed a very different failure mode, in which the failure took place in a more ductile
manner. The overall trend of the failure mechanisms clearly shows
that increasing temperature inuences the failure mechanisms of
Strathbogie granite in such a way that the ductile properties are
dominant at higher temperatures.
The micro-structural changes which occur with increasing
temperature in granite inuence the variation of failure mechanisms for different temperatures. Scanning electron microscopic
(SEM) images of granite thin sections are provided in Fig. 10, and
indicate that cracks developed and opened up with increasing

temperature. The mineral composition of rock inuences its


strength signicantly, since cracks are propagated by the weakest plane within the rock. The crack propagation of igneous rock
is determined not only by the micro-cracks through grain boundaries, but also the inter-granular cracks in some of the weaker
mineral constituents, such as feldspar and biotite grains (Brace
et al., 1972). It is known that at low temperature and low conning
pressure, the brittle fracture of polycrystalline rocks may correspond to general axial splitting by macroscopic cracks (pre-existing
joints and faults) extending in the direction of axial compression
(Hoek and Bieniawski, 1965; Mogi, 1966; Nemat-Nasser, 1985).
When the temperature increases, the crystal particles of rock fractures form new microscopic cracks (pre-existing grain boundaries)

Fig. 8. Normalized compressive strength vs. test temperature curves for the present study and previous studies.

102

S. Shao et al. / Geothermics 54 (2015) 96108

Fig. 9. Normalized elastic modulus vs. test temperature curves for the present study and previous studies.

Fig. 10. SEM images of negrained Strathbogie granite: (a) thin section at room temperature; (b) thin section heated to 400 C; (c) thin section heated to 800 C.

between mineral grains, as a result of differential thermal expansion between grains with different thermo-elastic moduli and
thermal conductivities (Dmitriyev et al., 1972; Heard and Page,
1982; Kranz, 1983).
Micro-cracking patterns and the distribution of inter- and intragranular micro-cracks in granites have been studied previously
(Kudo et al., 1992; Nasseri et al., 2005) and after being affected by
high temperature (Nasseri et al., 2007). Mineralogy is known to play
an important role in granite thermal degradation. As David et al.
(2012) reported, the crack density of La Peyratte granite is increased
from 0.2 at room temperature to 4.4 at 600 C temperature. They

further stated that differences of thermal expansion between different minerals cause crack nucleation, and the dramatic increase
of crack density of La Peyratte granite between 500 C and 600 C
is a result of transition of quartz, which occurs at 576 C. Even
with a temperature increase below the transition threshold, the
high and anisotropic thermal expansion leads to the fact that with
the variation of quartz content a different deterioration is expected
upon the same temperature changes. Especially in crystalline rock,
the amount of quartz has a signicant effect on thermally induced
micro-cracks because of the complexity of its thermal expansivity. It is reasonable to state that when the temperature increases

S. Shao et al. / Geothermics 54 (2015) 96108

103

Table 3
Post-failure images of tested specimens.
Gf5 at 23 C

Gf7 at 100 C

Gf8 at 200 C

Gf12 at 400 C

Gf13 at 600 C

Gf19 at 800 C

Gf23 at 900 C

Gf20 at 1000 C

Gf24 at 1100 C

Fig. 11. Cumulative AE events vs. axial stress at room temperature 23 C (Gf6).

Fig. 12. Cumulative AE events vs. axial stress at temperature 100 C (Gf7).

to 600 C or higher, the inuence of temperature becomes greater


on the mechanical behaviour of granite, due to the signicant thermal response at higher temperatures.In addition, a clear change
of colour of the Strathbogie granite specimens can be observed
with increasing temperature (Table 3). The approximate colours
observed for different temperatures are shown in Table 4. In general, Strathbogie granite specimens appear white and green with
dark brown at 23 C, and the colour turns to white and reddish from
400 C and subsequently the colour deepens to pink at 1100 C.

Fig. 13. Cumulative AE events vs. axial stress at temperature 200 C (Gf10).

2.3.3. Analysis of fracturing behaviour using AE monitoring


As discussed in Section 1.3.2, the stress thresholds of various crack propagation stages can be determined by monitoring
AE counts. Typical variations of cumulative AE counts against
the axial stress for samples tested for temperatures from 23 C
to 600 C are shown in Figs. 1115 (we had difculties in
operating the AE device subjected to very high temperatures
such that no AE data are available for temperatures beyond
600 C).
2.3.3.1. Crack initiation stress threshold ( ci ). The cumulative AE
counts in Figs. 1115, for temperatures between 23 C and 400 C,
are characterized by a sudden increase of AE activities followed by
a few AE events occurring with little strength change. The stress
threshold for crack initiation ( ci ) for these cases was identied as
the point where the AE event counts rst increase above the background level of events (Figs. 1115). AE activities at 600 C showed

Fig. 14. Cumulative AE events vs. axial stress at temperature 400 C (Gf12).

a smooth increase of cumulative AE counts with increasing stress


after the beginning of crack initiation. This trend is analogous to the
trend observed by Ranjith et al. (2004), which is shown in Fig. 2. At
600 C, the crack initiation stress threshold is marked at the point
where the curve initially lifts off (Ranjith et al., 2012).

Table 4
Different colours observed at different temperatures for the tested specimens.
T ( C)

23

100

200

400

600

800

900

1000

1100

Colour

White and grey

White and grey

White and grey

White and reddish

White and reddish

White and reddish

Reddish

Reddish

Pink

104

S. Shao et al. / Geothermics 54 (2015) 96108

Fig. 15. Cumulative AE events vs. axial stress at temperature 600 C (Gf14).
Table 5
Stress thresholds for different fracturing stages.
T ( C)

 ci , MPa

 ci2 , MPa

 cd , MPa

 peak , MPa

23
100
200
200
400
400
600
600

200
135.8
209
232.4
190
169.5
31.6
35.900

202
140.8
214.7
235
194
182.8

203
165.5
217.8
238.7
197
190
96.8
115.8

214.467
181.989
228.074
239.975
203.198
204.128
117.321
121.804

Fig. 16. Crack initiation and crack damage stress thresholds and peak compressive
strength for different temperatures.

Table 6
Normalized values of crack initiation and crack damage for different temperatures.
Temperature ( C)

23

100

200

200

400

400

600

600

 ci / peak
 cd / peak

0.93
0.95

0.75
0.91

0.92
0.95

0.97
0.99

0.94
0.97

0.83
0.93

0.27
0.82

0.29
0.95

2.3.3.2. Secondary cracking stress threshold ( ci2 ). Tests performed


for temperatures between room temperature and 400 C showed
some stages with sudden AE count rises after the crack initiation.
Previous studies have dened the secondary cracking as the point
when the cumulative AE counts curve starts to diverge followed
by crack initiation, and the corresponding stress as  ci . Secondary
cracking was readily apparent for temperatures below 400 C, but
results were not available for 600 C. From Figs. 1014, it is clear
that at the test temperature of 600 C, the stable crack propagation
region is much longer than that at other lower temperatures. This
could be an indication of the changes in mechanical response to
heating changes for Strathbogie granite.
2.3.3.3. Crack damage stress threshold ( cd ). For test temperatures
below 400 C,  cd was identied as the point where AE counts begin
to show an exponential growth. At 600 C,  cd was identied as the
point where the cumulative AE curve switches from linear increase
to an exponential growth. The unstable crack propagation starts
from  cd and continues until the failure of specimens.
Stress thresholds for different crack development stages identied from plots of cumulative AE counts against axial stress are
summarized in Table 5. Average values of crack initiation and crack
damage stress thresholds for different test temperatures are graphically presented in Fig. 16.
As Fig. 16 depicts, crack initiation and damage occur at lower
levels of stress when temperature increases and the strength range
of stable crack propagation expands with increasing temperature.
Crack initiation stress and crack damage stress were then normalized to the peak strength (Table 6), in order to understand the
proximity of crack development stages to the ultimate failure of
specimens for different temperatures. Fig. 17 reveals that crack initiation normalized stress decreases with increasing temperature,
except for the case of 100 C. This reects that the temperature

Fig. 17. Normalized stress for crack initiation and crack damage at different temperatures.

inuences the crack initiation stress thresholds of Strathbogie granite and increasing temperature allows crack initiation to happen
earlier. The early occurrence of crack initiation as temperature is
increased indicates the domination of ductile properties at elevated
temperature. This correlation is consistent with the prevalence of
ductile failure at higher temperatures (Table 3).
However, the normalized stress values for crack damage do not
show any important scatter with increasing temperature for Strathbogie granite, suggesting that crack damage occurs very close to
failure during uniaxial loading, but independent of the temperature.
3. Finite element modelling
Conducting a rigorous laboratory testing programme is usually time-consuming, labour-intensive and expensive. Validating a
numerical model using few experimental results and then extending the models for many other scenarios is a far more efcient
approach, and the availability of sophisticated and user-friendly
software packages has made this task much easier. This approach
is extremely useful, since results can be obtained even under
test conditions that the experimental approach fails to simulate.
In addition, some critical insights, such as stress/strain distribution within model specimens, such as the distribution of plastic
strain, can be comprehensively obtained from numerical simulations. Therefore, the experimental results of this study were rst
used to validate a nite element numerical model and then some

S. Shao et al. / Geothermics 54 (2015) 96108

105

Table 7
Temperature-dependent mechanical properties of Strathbogie granite.
T ( C)

 c (MPa)

E (GPa)

23
100
200
400
600
800
900
1000
1100

215.97
190.63
234.02
203.66
119.56
96.01
38.53
37.07
5.29

0.027
0.026
0.032
0.030
0.026
0.040
0.058
0.057
0.075

8.97
8.28
8.48
7.94
5.77
3.72
1.71
1.34
0.23

Table 8
Temperature-dependent plastic properties of Strathbogie granite.
T ( C)

Yield stress (MPa)

800
800
900
900
1000
1000
1100
1100

82.51

Ultimate yield stress (MPa)


101.76

26
38
25
31
7.11
7.65

Plastic strain
0
0.0125
0
0.0149
0
0.0140
0
0.0100

temperatures above 800 C, an elasticplastic model was assumed.


The temperature-dependent plastic properties (rst yield stress
and ultimate yield stress of the material), assigned for the material,
are shown in Table 8.
To perform the nite element modelling, several assumptions
were made as follows:

Fig. 18. Geometry of model specimen (a) specimen and the platens of the testing apparatus (the shaded area is the portion of the specimen that was modelled
by ABAQUS) and (b) the two-dimensional axis-symmetric model of the uniaxial
compression test.

results were obtained from models under temperatures other than


the temperatures used in the experimental work.
3.1. Model geometry and material properties
A nite element model of the compression tests carried out on
Strathbogie granite specimens at elevated temperature was performed using ABAQUS 6.7.1. A two-dimensional (axis-symmetric)
model of the compression test was set up in two parts. As Fig. 18(a)
shows, only a quarter of the specimen and half of the top platen was
modelled. The geometry of the model and boundary conditions are
shown in Fig. 20(b).
Temperature-dependent constitutive equations for Strathbogie
granite based on compressive tests performed over a range of temperatures were used in this nite element model. The stressstrain
behaviour under unconned uniaxial compression was found to be
strongly dependent on temperature and the brittleplastic transition was found to take place between 600 C and 800 C.
The temperature-dependent mechanical properties assigned to
Strathbogie granite are shown in Table 7 and they were incorporated into the nite element model via an in-built interpolation
routine. For the temperature range of 25 C to 600 C, the material
properties were assumed to be elastic up to the failure strain. For

(1) Uniform temperature across the entire granite was predened


before the start of each analysis.
(2) The platen was assumed to be steel with elastic modulus of
210 GPa and Poissons ratio of 0.3 and the Poissons ratio of the
Strathbogie granite was assumed to be 0.25.
(3) The interaction between the platen (master) and granite
(slave) assumes a friction coefcient of 0.1. Vertically downward displacements were applied to the bottom surface of the
platen and enabled the granite to be loaded in compression.
3.2. Results and discussion of simulation
The results of the numerical simulation for different temperatures along, with the corresponding experimental results are
shown in Fig. 19. Numerical simulation was not performed for
1100 C temperature as the material behaved similar to molten lava
in the experimental study for that temperature. The experimental stressstrain results obtained from compression tests appear
to show two sections of linear elastic behaviour through all tested
temperatures: (1) from 23 to 600 C, with a low modulus at the initial stage of loading up to a strain of about 0.0025, followed by a
relatively higher linear modulus up to the failure strain; (2) from
800 to 1000 C, with a low modulus at the initial stage of loading
up to a strain of 0.010.015 followed by a relatively higher linear
modulus up to the failure strain. The constitutive equation is based
on the temperature-dependent modulus after a small correction
is applied to the initial strain, which allows the model to predict
the axial stress up to the failure strain of the specimen with an
acceptable measure of success.
The FEM model is able to reproduce the results of the experiments for test temperatures below 600 C (elastic behaviour), and
for the test temperatures of 800, 900 and 1000 C (plastic and
ductile behaviour), as shown in Fig. 19. After validation and calibration of the nite element model, it can be used to predict
the stressstrain response of Strathbogie granite under unconned

106

S. Shao et al. / Geothermics 54 (2015) 96108

Fig. 19. Experimental results vs. values predicted by FEM (red squares) for different test temperatures.

stress conditions at any given temperature between 23 and 1000 C.


For example, although compression tests were not performed at
temperatures of 500 and 700 C, the stressstrain response can be
predicted using the model. The FEM results for these two temperatures are shown in Fig. 20(a) and (b), respectively. For comparison
with the predictions of nite element modelling, experimentally
observed curves for 400 and 600 C and 600 and 800 C test

temperatures are shown in Fig. 20(a) and (b), respectively. The


results show that the nite element model has excellent capability
for predicting the behaviour of temperature-dependent materials.
The evolution of plastic strains with increasing compression
can be visualized in nite element models. These plastic strain
distributions dictate the failure mechanisms for different test temperatures. The distribution of plastic strain in the specimen that was

S. Shao et al. / Geothermics 54 (2015) 96108

107

Fig. 20. Predicted compressive strength of Strathbogie granite using FEM (red squares) for (a) 500 C and (b) 700 C.

Fig. 21. The distribution of plastic strain in the specimen deformed at 800 C (a) strain = 0.0251, (b) strain = 0.0343 and (c) strain = 0.0457.

deformed at 800 C at the onset of plastic strain, partway through


the plastic deformation and just before nal failure, is shown in
Fig. 21. Plastic strain initially develops at the corner and gradually
extends through the specimen at an inclined angle of about 45 C
before failing roughly at that angle. Experiments also conrm that
failure of compressed cylindrical Strathbogie granite specimens
occurs in this fashion.
4. Conclusions
A series of unconned compressive strength tests was carried out on Strathbogie granite (ne grain) specimens at various
temperatures between 23 C and 1100 C. Specimens were heated
at a rate of 5 C/min and deformed at a displacement rate of
0.1 mm/min. A nite element model for predicting the thermomechanical behaviour of Strathbogie granite was analyzed using
ABAQUS 6.7.1 to cover the range of test temperatures. Both
experimental and numerical investigations revealed signicant
behaviours of granite at higher temperatures as follows.
(1) When the testing temperature was increased from room temperature to 200 C, both compressive strength and elastic
modulus were found to increase. Further increase of temperature caused a decrease in compressive strength and elastic
modulus. Above 800 C, the stressstrain curves showed a signicant increase in strain with very little change in stress,
and the compressive strength decreases dramatically at temperatures greater than 800 C. The brittleplastic transition

temperature of this type of granite under unconned conditions


was found to be between 600 and 800 C.
(2) Analysis of the changes of stress thresholds for various crack propagation stages at different temperatures
using acoustic emission monitoring showed a considerable dependence on the testing temperature. Increasing
temperature reduced the stress thresholds for crack initiation and crack damage and extended the stable crack
propagation stage. In addition, the increasing temperature
was observed to produce crack initiation earlier, indicating that ductile/quasi-brittle properties dominate at higher
temperatures.
(3) Finite element modelling showed stressstrain behaviours
that reproduced the experimentally observed stressstrain
variations. Numerical modelling work was extended to
study the mechanical behaviour of Strathbogie granite under
temperatures other than those considered in the experimental work. Moreover, the plastic strain distributions within
numerical specimens were compared with the observed failure patterns of the experiments and close similarity was
observed.
This paper reports an initial study of the behaviour of Strathbogie granite under unconned stress conditions and elevated
temperatures. The results of more detailed investigations under
very high conning pressures and very high temperatures (using
the experiments that the authors are currently conducting) will be
published later.

108

S. Shao et al. / Geothermics 54 (2015) 96108

Acknowledgement
The authors acknowledge the use of facilities within the Monash
Centre for Electron Microscopy.
References
Axelsson, G., 2010. Sustainable geothermal utilization case histories; denitions;
research issues and modelling. Geothermics 39 (4), 283291.
Bauer, S.J., Friedman, M., Handin, J., 1981. Effects of water saturation on strength and
ductility of three igneous rocks at effective pressure to 50 MPa and temperatures
to partial melting. In: Proceedings of the 22nd Symposium on Rock Mechanics,
Cambridge, MA, pp. 7984.
Bieniawski, Z.T., 1967. Mechanism of brittle fracture of rock. Part II: Experimental studies. Int. J. Rock Mech. Min. Sci. 4 (4), 407408, IN413IN414, 409418,
IN415IN418, 419423.
Brace, W.F., Silver, E., Hadley, K., Goetze, C., 1972. Cracks and pores: a closer look.
Science 178 (4057), 162164.
Cai, M., Kaiser, P.K., Morioka, H., Minami, M., Maejima, T., Tasaka, Y., Kurose, H.,
2007. FLAC/PFC coupled numerical simulation of AE in large-scale underground
excavations. Int. J. Rock Mech. Min. Sci. 44 (4), 550564.
Chang, S.H., Lee, C.I., 2004. Estimation of cracking and damage mechanisms in rock
under triaxial compression by moment tensor analysis of acoustic emission. Int.
J. Rock Mech. Min. Sci. 41 (7), 10691086.
David, E.C.B.N., Schubnel, A., Zimmerman, R.W., 2012. Sliding crack model for nonlinearity and hysteresis in the uniaxial stressstrain curve of rock. Int. J. Rock
Mech. Min. Sci. 52, 917.
Dmitriyev, A.P.K.L.S., Protasov, Y.I., Yamshchikov, V.S., 1972. Physical Properties of
Rocks at High Temperatures. National Aeronautics and Space Administration,
NASA Technical Translation N 72-26286.
Dwivedi, R.D., Goel, R.K., Prasad, V.V.R., Sinha, A., 2008. Thermo-mechanical properties of Indian and other granites. Int. J. Rock Mech. Min. Sci. 45 (3), 303315.
Eberhardt, E., Stead, D., Stimpson, B., Read, R.S., 1998. Identifying crack initiation
and propagation thresholds in brittle rock. Can. Geotech. J. 35 (2), 222233.
Eberhardt, E., Stimpson, B., Stead, D., 1999. Effects of grain size on the initiation and
propagation thresholds of stress-induced brittle fractures. Rock Mech. Rock Eng.
32 (2), 8199.
Farndon, J., 2010. The Complete Guide to Rocks & Minerals. Joanna Lorenz, Hermes
House, London.
Francois, J.P.A.P., 1980. Experimental deformation of partially melted granite rocks
at 600900 C and 250 MPa conning pressure. Tectonophysics 68, 131146.
Gallup, D.L., 2009. Production engineering in geothermal technology: a review.
Geothermics 38 (3), 326334.
Heard, H.C., Page, L., 1982. Elastic moduli, thermal expansion, and inferred permeability of two granites to 350 C and 55 megapascals. J. Geophys. Res. 87 (B11),
93409348.
Heuze, F.E., 1983. High-temperature mechanical, physical and thermal properties
of granitic rocks a review. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 20 (1),
310.
Hirth, G., Tullis, J., 1989. The effects of pressure and porosity on the micromechanics of the brittleductile transition in quartzite. J. Geophys. Res. 94 (B12),
17,82517,838.
Hoek, E., Bieniawski, Z.T., 1965. Brittle fracture propagation in rock under compression. Int. J. Fract. 1 (3), 137155.

Hueckel, T., Peano, A., Pellegrini, R., 1994. A thermo-plastic constitutive law for
brittleplastic behavior of rocks at high temperatures. Pure Appl. Geophys. 143
(1), 483510.
Kranz, R.L., 1983. Microcracks in rocks: a review. Tectonophysics 100 (13), 449480.
Kudo, Y., Sano, O., Murashige, N., Mizuta, Y., Nakagawa, K., 1992. Stress-induced crack
path in Aji granite under tensile stress. Pure Appl. Geophys. 138 (4), 641655.
Mogi, K., 1966. Pressure dependence of rock strength and transition from brittle
fracture to ductile ow. Bull. Earthq. Res. Inst. 44, 215232.
Nasseri, M.H.B., Mohanty, B., Robin, P.Y.F., 2005. Characterization of ve microstructures and fracture toughness in ve granitic rocks. Int. J. Rock Mech. Min.
Sci. 42, 450460.
Nasseri, M.H.B., Schubnel, A., Young, R.P., 2007. Coupled evolutions of fracture toughness and elastic wave velocities at high crack density in thermally treated
Westerly granite. Int. J. Rock Mech. Min. Sci. 44 (4), 601616.
Neil Phillips, G., Wall, V.J., Clemens, J.D., 1981. Petrologoyf Thes Thathbogbiea
Tholitha: Cordierite-Bearginrga Nite. Can. Miner. 19, 4763.
Nemat-Nasser, H.H.A.S., 1985. Compression-induced microcrack growth in brittle
solids: axial splitting and shear failure. J. Geophys. Res. 90, 31053125.
Paquet, J., Francois, P., Nedelec, A., 1981. Effect of partial melting on rock deformation: experimental and natural evidences on rocks of granitic compositions.
Tectonophysics 78 (14), 545565.
Ranjith, P.G., Fourar, M., Pong, S.F., Chian, W., Haque, A., 2004. Characterisation of
fractured rocks under uniaxial loading states. Int. J. Rock Mech. Min. Sci. 41
(Suppl. 1), 1A 08 01-06.
Ranjith, P.G., Jasinge, D., Song, J.Y., Choi, S.K., 2008. A study of the effect of displacement rate and moisture content on the mechanical properties of concrete: use
of acoustic emission. Mech. Mater. 40 (6), 453469.
Ranjith, P.G., Viete, D., Chen, B., Perera, M.S.A., 2012. Transformation plasticity and
the effect of temperature on the mechanical behaviour of Hawkesbury sandstone at atmospheric pressure. Eng. Geol. 151 (29), 120127, http://dx.doi.
org/10.1016/j.enggeo.2012.09.007.
Rudajev, V., Vilhelm, J., Lokajcek, T., 2000. Laboratory studies of acoustic emission
prior to uniaxial compressive rock failure. Int. J. Rock Mech. Min. Sci. 37 (4),
699704.
Rybach, L., 2003. Geothermal energy: sustainability and the environment. Geothermics 32 (46), 463470.
Shield, C.K., 1997. Comparison of acoustic emission activity in reinforced and
prestressed concrete beams under bending. Construct. Build. Mater. 11 (3),
189194.
Tullis, J., Yund, R.A., 1987. Transition from cataclastic ow to dislocation creep of
feldspar: mechanisms and microstructures. Geology 15 (7), 606609.
Wang, Y., Zhang, H., Huang, X., Qiu, Y., 2002. Experimental study on whole
stressstrain course of marble under the action of high temperature. Yanshilixue
Yu Gongcheng Xuebao 21 (Suppl.), 23452349.
Xu, X.-L., Kang, Z.-X., Ji, M., Ge, W.-X., Chen, J., 2009. Research of microcosmic mechanism of brittleplastic transition for granite under high temperature. Procedia
Earth Planet. Sci. 1 (1), 432437.
Xu, X.-L., Gao, F., Shen, X.-M., Xie, H.-P., 2008a. Mechanical characteristics and microcosmic mechanisms of granite under temperature loads. J. China Univ. Min.
Technol. 18 (3), 413417.
Xu, X.L., Gao, F., Gao, Y.N., Xie, H.P., 2008b. Effect of high temperatures on the
mechanical characteristics and crystal structure of granite. Zhongguo Kuangye
Daxue Xuebao/J. China Univ. Min. Technol. 37 (3), 402406.
Zhang, L.Y., Mao, X.B., Lu, A., 2009. Experimental study on the mechanical properties of rocks at high temperature. Sci. China Ser. E: Technol. Sci. 52 (3),
641646.