Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Journal of Rock Mechanics and Geotechnical Engineering 7 (2015) 361e366

Contents lists available at ScienceDirect

Journal of Rock Mechanics and


Geotechnical Engineering
journal homepage: www.rockgeotech.org

Full length article

A theoretical derivation of the HoekeBrown failure criterion for rock


materials
Jianping Zuo a, b, *, Huihai Liu c, Hongtao Li a
a
State Key Laboratory of Coal Resources and Safe Mining, China University of Mining and Technology, Beijing 100083, China
b
Institute of Rock Mechanics and Fractals, School of Mechanics and Civil Engineering, China University of Mining and Technology, Beijing 100083, China
c
Aramco Research Center, Houston, TX 77084, USA

a r t i c l e i n f o a b s t r a c t

Article history: This study uses a three-dimensional crack model to theoretically derive the HoekeBrown rock failure
Received 7 March 2015 criterion based on the linear elastic fracture theory. Specifically, we argue that a failure characteristic
Received in revised form factor needs to exceed a critical value when macro-failure occurs. This factor is a product of the micro-
18 March 2015
failure orientation angle (characterizing the density and orientation of damaged micro-cracks) and the
Accepted 20 March 2015
Available online 7 April 2015
changing rate of the angle with respect to the major principal stress (characterizing the microscopic
stability of damaged cracks). We further demonstrate that the factor mathematically leads to the
empirical HoekeBrown rock failure criterion. Thus, the proposed factor is able to successfully relate the
Keywords:
Failure criterion
evolution of microscopic damaged crack characteristics to macro-failure. Based on this theoretical
Triaxial compression development, we also propose a quantitative relationship between the brittleeductile transition point
Micro-failure orientation and confining pressure, which is consistent with experimental observations.
Fracture mechanics Ó 2015 Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. Production and hosting by
Brittleeductile transition Elsevier B.V. All rights reserved.

1. Introduction predicts that uniaxial compressive strength sc is about eight times


the uniaxial tensile strength st, most experimental results show
The strength theory of brittle materials, such as rock and con- that values for sc/st are larger than 10, even up to 50.
crete, has been a subject of intensive study for many years, because Griffith’s criterion was modified by McLintock and Walsh (1962)
the failure of brittle material is of great concern in designing and to ðs1  s3 Þð1 þ mÞ1=2 þ mðs1 þ s3 Þ ¼ 4st , where m is the friction
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi
evaluating the safety and stability of engineering structures and coefficient. This criterion leads to sc =st ¼ 4=ð m2 þ 1 þ mÞ as a
geological engineering facilities. Consequently, the failure criterion uniaxial compression condition for s3 ¼ 0 and sc ¼ s1. In this case, the
for brittle material has become one of the most important subjects ratio sc/st is still smaller than the ratio commonly measured for rocks.
in many areas (Yu, 2004; Paterson and Wong, 2005; Jaeger et al., Murrell (1965) suggested that the biaxial criterion could be
2007; Zuo et al., 2008). generalized for triaxial stress as a paraboloid: ðs2  s3 Þ2 þ
Griffith (1921) first postulated that the fracture of brittle mate- ðs3  s1 Þ2 þ ðs1  s2 Þ2  24T0 ðs1 þ s2 þ s3 Þ ¼ 0. Under the uni-
rials occurs when the energy-release rate during crack growth axial compression conditions for s2 ¼ s3 ¼ 0 and sc ¼ s1, it can
exceeds the rate that energy is required. Griffith’s theory, which predict that sc/st ¼ 12. However, there seems to be no a priori
indicates that micro-cracks are a major cause of failure of materials reason for selecting this triaxial criterion except on empirical
(Griffith, 1921, 1924), has been extended for a biaxially stressed grounds. Based on a micro-mechanical analysis of sliding cracks,
body with randomly oriented cracks (Odé, 1960). A macro-failure Wiebols and Cook (1968) proposed a new criterion. However, their
criterion, generally known as the Griffith’s criterion, is expressed theory suggests that with respect to polyaxial compression, the
in terms of the biaxial principal stresses s1, s2 and the magnitude of intermediate principal stress has a pronounced and predictable
the uniaxial tensile strength T0, i.e. ðs1  s2 Þ2  8T0 ðs1 þ s2 Þ ¼ 0 effect, which may not always be considered in the existing theories.
for s1 >  3s2 and s2 ¼ T0 for s1 <  3s2. While Griffith’s criterion The classic HoekeBrown failure criterion, empirical in its
formulation and based on numerous experimental data, has been
widely used to predict the failure of rocks (Hoek and Brown, 1980a,
b). While the criterion has been modified or extended for a number
* Corresponding author. Tel.: þ86 10 62331358.
E-mail address: zjp@cumtb.edu.cn (J. Zuo). of fields, linkage between this well-known criterion and the micro-
Peer review under responsibility of Institute of Rock and Soil Mechanics, mechanical mechanisms of sliding cracks is still rare in the litera-
Chinese Academy of Sciences. ture (Hoek, 1983, 1998; Hoek and Brown, 1997), providing such a
1674-7755 Ó 2015 Institute of Rock and Soil Mechanics, Chinese Academy of Sci- linkage would have significant implications for improving our un-
ences. Production and hosting by Elsevier B.V. All rights reserved.
derstanding of rock damage/failure processes.
http://dx.doi.org/10.1016/j.jrmge.2015.03.008
362 J. Zuo et al. / Journal of Rock Mechanics and Geotechnical Engineering 7 (2015) 361e366

The major objective of this study is to make an effort to establish sn components on the disk-shaped crack can be given in terms of
the linkage mentioned above, based on a recent study in which Zuo the three principal stresses, and as a function of the direction
et al. (2008) have derived a theoretical nonlinear strength criterion cosines l, m, n:
for rock-like material based on fracture mechanics, a criterion with
a similar mathematical expression to the HoekeBrown criterion. sn ¼ l2 s1 þ m2 s2 þ n2 s3 (2)
Hoek and Martin (2014) recently compared our theoretical criterion
with MohreCoulomb and Griffith’s criterion in Fig. 3 of their paper.
This study is an effort to extend the work of Zuo et al. (2008) by s2n ¼ l2 s21 þ m2 s22 þ n2 s23  s2n (3)
including additional insights and analyses. The most important
issue in Zuo et al. (2008) is how to choose the failure characteristic 3. Propagation conditions for a disk-shaped crack
factor that relates the micro-mechanism to macro-failure, and our
study provides more physical justifications for the chosen factor. Although inelastic effects are involved at crack tips in various
Also, the study of Zuo et al. (2008) was based on a two-dimensional rocks (Hoagland et al., 1973), and even elastic behavior in the most
model for an initial sliding crack, while in this study we focus on highly stressed regions may be nonlinear, the practical analysis of
three-dimensional (3D) cases. In addition, we discuss how to the stress distribution in the neighborhood of the crack tip is
interpret the brittleeductile transition phenomenon in rocks. usually based on the classical linear theory of elasticity. Therefore,
according to linear elastic fracture mechanics (Anderson, 2005), the
2. Fundamental hypothesis and model crack mode in Fig. 1 is in-plane shear (Mode II), a mode corre-
sponding to in-plane shear loading that tends to slide one crack
There is abundant evidence for the existence of Griffith cracks or face with respect to the other. The stress intensity factor depends
other crack-like flaws in brittle materials (Griffith, 1921, 1924; Peng on both shear stress sn and friction stress sm, which is parallel to the
and Johnson, 1972; Schovanec, 1986; Mura, 1987; Vardoulakis and crack plane. According to Amontons’ law, sm ¼ msn, where m and sn
Papamichos, 1991). In addition, the low strength of some brittle are the friction coefficient for the pre-existing sliding crack surface
materials results mainly from the presence of these micro-cracks or and the normal stress acting on the disk plane, respectively. It is
flaws. In the present work, we assume that numerous Griffith obvious that sm is a key factor in suppressing crack propagation. The
micro-cracks are randomly distributed in rocks. A model of fric- effective shear stress se can be expressed as
tional sliding cracks has been discussed in the literature (Cook,
1965; Holcomb, 1978; Kachanov, 1982a, b; Li and Yang, 2006). se ¼ jsn j  msn (4)
Although the crack interactions can be taken into account by
modifications of a self-consistent method, in this paper, these
For the 3D problem, the effective shear stress se initiates fric-
micro-crack interactions are neglected, so that traction on crack
tional sliding. In addition, if se is sufficiently high to overcome the
faces can be calculated as being induced by external load. In this
frictional resistance along the initial crack, frictional slip results in
way, a simplified model for a sliding, disk-shaped, closed crack of
tensile stress concentrations at the two tips of the sliding crack,
diameter 2a in an infinite plate can be generated, as shown in Fig. 1.
which, in turn, may induce nucleation of “wing cracks” (Kachanov,
We adopt the convention common in geologic and rock mechanics
1982a). Prior to the onset of crack propagation, the maximum stress
literature that compressive stress is positive and tensile stress is
intensity factor KIImax of mode II at the initial crack tip approxi-
negative, and designate the major, intermediate, and minor prin-
mately is
cipal stresses as s1, s2, s3, respectively, where the three subscripts
represent the coordinate axes x1, x2, x3, respectively. The orientation pffiffiffi
4se a
of a disk-shaped crack can be defined by direction cosines l, m, n. If a KIImax ¼ pffiffiffiffi (5)
ð2  nÞ p
unique sphere is built up with the following equation:
where n is the Poisson’s ratio.
l2 þ m2 þ n2 ¼ 1 (1)
If the wing crack length is infinitesimal, the propagation con-
then the orientation of the crack plane is determined by the ditions for a disk-shaped crack under triaxial compression, based
normal direction of point M. The normal stress sn and shear stress on the mixed fracture criterion, can be expressed as (Kachanov,
1982b; Cotterell and Rice, 1980):

KIImax  KIIc ¼ kKIc (6)

where KIc and KIIc are the fracture toughness of Mode I and Mode II,
respectively; k is a proportionality factor, which may be determined
by a mixed fracture criterion of linear elastic fracture mechanics.
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
For example, k is 3ð1  nÞ=ð2 þ 2n  n2 Þ for the maximum strain
energy criterion for plain stress (Sih and Macdonald, 1974; Sih,
1974); k is 0.886 for the maximum stress criterion (Erdogan and
Sih, 1963); and k is 1 for the maximum energy release rate crite-
rion (Nuismer, 1975).
Because the wing crack is caused by the local tensile stress at the
disk-shaped crack tip, the fracture toughness KIc can be calculated
as
pffiffiffiffiffiffiffiffiffi
KIc ¼ 2st a=p (7)

Fig. 1. A disk-shaped crack of diameter 2a in an infinite plate under triaxial where st is the local tensile strength at the tip of the disk-shaped
compression. crack. Substituting Eqs. (5) and (7) into Eq. (6), we then get:
J. Zuo et al. / Journal of Rock Mechanics and Geotechnical Engineering 7 (2015) 361e366 363

 n
se  1  kst (8) cos q sin qðs1  s3 Þ  mðs1 cos2 q þ s3 sin2 qÞ  bst (12)
2
Let the dimensionless parameters s1 and s3 stand for s1/st and
Using b to denote ð1  n=2Þk, Eq. (8) can then be rewritten as
s3/st, respectively. With the corresponding trigonometric trans-
se  bst (9) formation, Eq. (12) can be rewritten as

ðms3 þ bÞtan2 q  ðs1  s3 Þtan q þ ðms3 þ bÞ  0 (13)


Then, we have
4. The attitude angle parameter f to characterize micro-
failure orientation tan q1  tan q  tan q2 (14)

Assume there is a sphere with a unit length radius, with the where
coordinate origin O as the center of the sphere. Based on the geo- qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
metric symmetry of sphere, one can concentrate on the eight equal ðs1  s3 ÞH ðs1  s3 Þ2  4ðms1 þ bÞðms3 þ bÞ
tanqq1 ¼ (15)
parts in the sphere, as shown in Fig. 2. Since there are numerous 2 2ðms1 þ bÞ
Griffith micro-cracks distributed randomly in rock materials, we
assume that all these cracks will converge. Assuming the move- An attitude angle can be defined, i.e. f ¼ q2  q1, then we have
ment of one endpoint of any disk-shaped crack to the coordinate qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
origin O, there develops an intersection point between the normal ðs1  s3 Þ2  4ðms1 þ bÞðms3 þ bÞ
line of any cracks between the normal line of any cracks and the tan f ¼ (16)
mðs1 þ s3 Þ þ 2b
surface of the upper hemisphere. For n rock samples under triaxial
compression, all intersection points can make up an intersection Eq. (16) gives the relationship between attitude angle f and
region that depends on the three principal stresses. Although evi- external load. In addition, the range of the attitude angle f is from
dence has shown that s2 does influence the failure of rock (Al-Ajmi 0 to p/2.
and Zimmerman, 2005), the shape of the intersection region is
rather complicated under true triaxial compression (s1 > s2 > s3).
5. The failure characteristic factor for rocks and derivation of
Therefore, we discuss only the special case of s1 > s2 ¼ s3. Ac-
the HoekeBrown criterion
cording to the axial symmetry, these intersection points will make
up a band region located on two latitude lines AB and CD (the
Rock failure processes can be approximately described as fol-
shadow region in Fig. 2). An angle q is used to denote the “micro-
lows. There are various pre-existing micro-cracks in rock materials
failure orientation angle” under triaxial compression. The angle q
that are potentially of vital significance to their brittle behavior.
corresponds to the angle a in Zuo et al. (2008), and they have the
When the external load achieves a critical value, pre-existing (and
same physical meaning, i.e. “micro-failure orientation angle”.
some new) micro-cracks will initiate and propagate. These micro-
When s2 ¼ s3, Eq. (2) becomes the following form derived from Eq.
cracks will tend not to propagate in their own planes, but rather
(1) and cosq ¼ l:
swing into an orientation more nearly parallel to the orientation of
the major principal stress s1 (Diederichs, 2003). With a further
sn ¼ s1 cos2 q þ s3 sin2 q (10)
increase in external load, some subcritical micro-cracks will
Then, Eq. (3) can be further simplified as initiate, propagate, and tend to become unstable. When numerous
micro-failures approach a critical condition, a macro-failure will
s2n ¼ l2 ð1  l2 Þðs1  s3 Þ2 or sn ¼ cos q sin qðs1  s3 Þ (11) occur, and the rock will then be broken. By contrast, micro-failures
are usually constrained by the confining pressure s3; thus, the
From Eqs. (4) and (9)e(11), we can get
higher the confining pressure, the smaller the micro-failure.
In Section 3, we obtained an expression for the portion of micro-
cracks that are subjected to micro-damage. The key issue is how to
relate this damage to macro-failures. The latter is determined by
two microscopic variables. The first variable is the attitude angle, f.
Obviously, a larger f corresponds to a larger relative portion of
damaged internal micro-cracks, and thus to a larger possibility for
macro-failure.
The second variable is jvf=vs1 j, the change rate in f with respect
to the dimensionless axial stress. This variable reflects the stability
of the rock (for the given f) under further stress increase. A larger
jvf=vs1 j means a larger degree of instability in the rock with further
stress increase (or a larger possibility for macro-failure). There is
considerable laboratory evidence that near macro-failure, the
number of observed damaged/activated micro-cracks increases
quickly with stress (Paterson and Wong, 2005; Jaeger et al., 2007).
Therefore, the failure characteristic factor (F) can be expressed as
 
fvf
F ¼   (17)
vs 1

Fig. 2. The band region of intersecting point between the normal line of any micro- The parameter F must exceed a critical value for macro-failure to
cracks and unit sphere. occur. Note that Eq. (17), while the simplest mathematical
364 J. Zuo et al. / Journal of Rock Mechanics and Geotechnical Engineering 7 (2015) 361e366

expression for relating microscopic variables f and jvf=vs1 j to factor could be used as a damage variable in damage mechanics
macro-failure, is justified by its consistency with our above dis- studies, because it is more or less equivalent to the HoekeBrown
cussion. (We acknowledge that the relationship between micro- criterion, as we have demonstrated here. Note that how to choose
scopic variables and macro-failure cannot be rigorously and the damage variable is at present a key issue in the damage me-
uniquely determined at this stage. Nevertheless, Eq. (17) can be chanics community.
considered a reasonable first step.) In many cases, we expect f to be
small, given the fact that observed micro-damage often occurs 6. The micro-mechanisms of brittleeductile transitions
along narrow bands in rock samples. Thus, f can be approximately
replaced by sinf in Eq. (17). Eq. (17) then becomes It is usually accepted that material failure is brittle type if no
  apparent plastic deformation takes place before fracturing. In
vcos f
F ¼   (18) ductile failure, by contrast, extensive plastic deformation takes
vs1  place. Experimental results indicate that while most rocks exhibit
brittleness under atmospheric pressure, ductility in rock can also be
From Eq. (16), we have
achieved in the laboratory with the aid of sufficiently high
  confining pressures and temperatures (Paterson and Wong, 2005;
m s þ b=m
cos f ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 þ 2 3 (19) Jaeger et al., 2007). In other words, there exists a critical condi-
1 þ m2 s1  s3
tion for the transition from brittleness to ductility. In this paper, we
For s3 ¼ 0 (or s3 ¼ 0), the above equation becomes concern with the effects of confining pressure. The brittleeductile
transition occurs when the external confining pressure reaches a
 
m b=m critical value. The state in which this change happens is called a
cos f ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 þ 2 (20)
1 þ m2 s1 brittleeductile transition state. In addition, the investigation of the
brittleeductile transition has become very important in many
Substituting Eq. (19) into Eq. (18) yields geologic situations, and may also have relevance to some engi-
  neering applications.
vcos f m 2s3 þ 2b=m
  Studies of brittleeductile transition phenomena have been re-
 vs  ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi (21)
1 1 þ m2 ðs1  s3 Þ2 ported by a number of researchers (Paterson and Wong, 2005;
Jaeger et al., 2007). However, most of these studies are from a
We can determine the critical F value (corresponding to macro-
macroscopic or phenomenological point of view, and take no ac-
failure) through a uniaxial compression experiment. Substituting
count of the micro-mechanisms, especially the micro-crack
the uniaxial compression conditions s1 ¼ sc =st ¼ sc and s3 ¼ 0
mechanisms under high confining pressure. For example, some
into Eq. (21), we have
investigators artificially define the failure of rock as brittle type (or
  ductile type) if the value of strain or confining pressure of failure is
vcos f m 2b
 
 vs  ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi (22) less than (or larger than) a specific value. For example, the strain
1 1 þ m2 ms2c value of 3%e5% corresponding to failure is often taken for defining
Consequently, the critical value for F is the right part of Eq. (22). the brittleeductile transition (Heard, 1960; Evans et al., 1990). Some
Substituting this value into Eq. (21), we have investigators have proposed that, as a general rule, the brittlee
ductile transition should be related to the strength of rocks (Mogi,
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
m 2 2005). For instance, in silicate rocks under compression, it occurs
s1 ¼ s3 þ sc s3 þ s2c (23)
b when the confining pressure is equal to roughly 1/3 the stress
difference at failure; in carbonate rocks, at about 1/4 the stress
Substituting the two parameters s1 ¼ s1 =st and s3 ¼ s3 =st difference. The ductility of rock in extension requires much higher
back into Eq. (23), we obtain confining pressures (Heard, 1960). However, in the case of porous
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sandstone, the transition to ductility occurs at an effective
m sc
s1 ¼ s3 þ sc s3 þ s2c (24) confining pressure of about 0.15 times the critical effective pressure
b st for the onset of grain crushing under hydrostatic pressure (Wong
et al., 1997). In other words, there is no strict or unified standard
where m is about 0.2e0.8 (Paterson and Wong, 2005). The param-
from which to choose the critical parameter for the brittleeductile
eters m, b, st, sc depend on rock properties, and all can be obtained
transition. Parameter selection varies with individual investigator
from experimental data.
or rock types, indicating the difficulty in making clear and strict
The original HoekeBrown empirical criterion is (Hoek and
definitions of the brittleeductile transition point on a macroscale.
Brown, 1980a, b):
Therefore, it is necessary to revisit the concepts of brittle, ductile,
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi and brittleeductile transition states for the micro/mesoscale. This
s1 ¼ s3 þ msc s3 þ ss2c (25) section provides some new information on the subject.
In light of the previous derivation of the strength criterion, we
where m and s are the material parameters for the rock. With s ¼ 1 define the brittleeductile transition as occurring in rock when the
for intact rock, Eq. (25) becomes confining pressure achieves a suitable value. That value is achieved
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi when the micro-failure in rock is completely restrained. This is a
s1 ¼ s3 þ msc s3 þ s2c (26) definition of the rock brittleeductile transition point at the micro/
mesoscale. Since there is clearly a causal relationship between
Thus, Eq. (24) is the same as Eq. (26) for m ¼ msc/(bst). In other micro-failure and rock dilatancy, this definition is entirely equiva-
words, we are able to derive the HoekeBrown criterion from micro- lent to the macro-definition.
mechanical considerations, based on the failure characteristic fac- According to this definition, the confining pressure value for the
tor defined in Eq. (17). This fact in itself strongly supports the brittleeductile transition point can be derived as follows. Previ-
reasonableness and usefulness of our failure characteristic factor. ously, we used the attitude angle f to characterize the micro-failure
While it is beyond the scope of this study, we speculate that this orientation, and the relationship between the angle f and external
J. Zuo et al. / Journal of Rock Mechanics and Geotechnical Engineering 7 (2015) 361e366 365

load was derived. It has also been proved that the use of qffiffiffiffiffiffiffiffiffiffiffiffiffiffi 
jvcos f=vs1 j as a rock brittle-failure characteristic factor is in 2b
2b 1 þ m2 þ m  sc  rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2ffi (30)
accordance with the HoekeBrown empirical criterion.
m2 þ 1 þ m2  m  m
Obviously, when the confining pressure increases to a specific
value s*3 , the internal micro-failure is completely restrained when
This also means that Eq. (28) is in accord with Mogi’s criterion
the rock is broken. In this case, the corresponding value of f is 0.
under certain conditions. Fig. 3 also shows that when the value of sc
According to Eqs. (16) and (22), we have
is less than left part of Eq. (30), rock does not exhibit brittle
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 behavior under compression. In addition, when sc is larger than
  1 þ m2 cos f=m  1 right part of Eq. (30), higher confining pressure is needed for a
vcos f m m 2b
 
 vs  ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2s3 þ 2b=m
¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi brittleeductile transition.
1 1 þ m2 1 þ m2 ms2c Obviously, to ensure that the rock exhibits brittle behavior un-
(27) der triaxial pressure, we must have s*3 > 0. According to Eq. (28),
this requires
Substituting f ¼ 0 and s3 ¼ s*3 into Eq. (27), we have
" qffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 # qffiffiffiffiffiffiffiffiffiffiffiffiffiffi 
1 s2c sc  2b 1 þ m2 þ m
s*3 ¼ 1 þ m2  m  b (28) (31)
m 4b
Eq. (28) is the critical condition with respect to the brittlee Eq. (31) indicates that (the ratio of rock uniaxial compressive
ductile transition for rocks. This formula indicates that the greater strength sc to uniaxial tensile strength st) is an important param-
sc ¼ sc =st of the rock, the greater the confining pressure s*3 that eter for evaluating rock brittleness. The greater this ratio, the more
triggers the brittleeductile transition. This is consistent with the brittle the rock, and vice versa.
experimental results. In addition, the critical confining pressure of
the brittleeductile transition is also related to the friction coeffi- 7. Conclusions
cient m and the fracture parameter b of rocks.
An empirical failure criterion has also been proposed in Wong A 3D crack model is employed to theoretically derive the Hoeke
et al. (1997), indicating that for most rocks, the confining pres- Brown rock failure criterion based on the linear elastic fracture
sure must always be smaller than the uniaxial compressive theory. Specifically, we argue that our proposed failure character-
strength to maintain rock brittle behavior: istic factor needs to exceed a critical value when macro-failure
occurs. This factor is a product of a micro-failure orientation
angle (characterizing the density and orientation of damaged
s*3  sc (29)
micro-cracks) and the changing rate of the angle with respect to the
To compare two criteria given by Eqs. (28) and (29), we plot major principal stress (characterizing the microscopic stability of
them in Fig. 3, together with experimental data from page 45 in damaged cracks). We further demonstrate that the factor mathe-
Mogi (2005). The figure shows that only when sc satisfies the matically leads to the empirical HoekeBrown rock failure criterion.
following condition, Eq. (29) is physically valid: One important implication of the study is that the proposed factor

β
μ + +μ −μ −μ

σ =σc

σc
σ = +μ −μ −β
μ β

o σc
β +μ +μ

Fig. 3. The relationship between the confining pressure s*3 at brittlenesseductility transition and the value of sc/st.
366 J. Zuo et al. / Journal of Rock Mechanics and Geotechnical Engineering 7 (2015) 361e366

is able to relate the evolution of microscopic damaged-crack char- Hoek E. Reliability of HoekeBrown estimates of rock mass properties and their
impact on design. International Journal of Rock Mechanics and Mining Sciences
acteristics to macro-failure.
1998;35(1):63e8.
Based on our theoretical development, this paper also gives a Holcomb DJ. A quantitative model of dilatancy in dry rock and its application to
clear interpretation of the brittleness and ductility of rock on the Westerly granite. Journal of Geophysical Research 1978;83(B10):4941e50.
microscale. A quantitative relationship between the brittleeductile Jaeger JC, Cook NGW, Zimmerman RW. Fundamentals of rock mechanics. 4th ed.
Oxford, UK: Blackwell; 2007.
transition point and confining pressure is derived and shown to be Kachanov ML. A microcrack model of rock inelasticity, part I: frictional sliding on
consistent with experimental observations. microcracks. Mechanics of Materials 1982a;1(1):19e27.
Kachanov ML. A microcrack model of rock inelasticity, part II: propagation of
microcracks. Mechanics of Materials 1982b;1(1):29e41.
Conflict of interest Li YP, Yang CH. On sliding crack model for brittle solids. International Journal of
Fracture 2006;142(3e4):323e30.
The authors wish to confirm that there are no known conflicts of McLintock FA, Walsh JB. Friction on Griffith cracks in rocks under pressure. In:
Proceedings of the 4th U.S. National Congress of Applied Mechanics, Berkeley,
interest associated with this publication and there has been no California, vol. II; 1962. p. 1015e21.
significant financial support for this work that could have influ- Mogi K. Experimental rock mechanics. London, UK: Taylor & Francis; 2005.
enced its outcome. Mura T. Micromechanics of defects in solids. Dordrecht, Netherlands: Kluwer Ac-
ademic Publishers; 1987.
Murrell SAF. The effect of triaxial stress systems on the strength of rocks at at-
Acknowledgments mospheric temperatures. Geophysical Journal of the Royal Astronomical Society
1965;10(3):231e81.
Nuismer RJ. An energy release rate criterion for mixed mode fracture. International
This work is supported by the National Natural Science Foun-
Journal of Fracture 1975;11(2):245e50.
dation of China (No. 51374215), Fok Ying Tung Education Founda- Odé H. Faulting as a velocity discontinuity in plastic deformation. In: Griggs D,
tion (No. 142018), Beijing Major Scientific and Technological Handin J, editors. Rock deformation. Boulder, Colorado, USA: Geological Society
Achievements into Ground Cultivation Project and the 111 Project of America; 1960. p. 293e321.
Paterson MS, Wong TF. Experimental rock deformation e the brittle field. Berlin,
(No. B14006). The second author (Huihai Liu) thanks the manage- Germany: Springer; 2005.
ment of Aramco Research Center for approving publication of the Peng S, Johnson AM. Crack growth and faulting in cylindrical specimens of
work. Also, the authors would like to acknowledge the anonymous Chelmsford granite. International Journal of Rock Mechanics and Mining Sci-
ences & Geomechanics Abstracts 1972;9(1):37e86.
reviewers for their valuable comments and suggestions, which Schovanec LA. Griffith crack problem for an inhomogeneous elastic material. Acta
helped to significantly improve the quality of this paper. Mechanica 1986;58(1e2):67e80.
Sih GC, Macdonald B. Fracture mechanics applied to engineering problems-strain
energy density fracture criterion. Engineering Fracture Mechanics 1974;6(2):
References 361e86.
Sih GC. Strain-energy-density factor applied to mixed mode crack problems. In-
Al-Ajmi AM, Zimmerman RW. Relation between the Mogi and the Coulomb failure ternational Journal of Fracture 1974;10(3):305e21.
criteria. International Journal of Rock Mechanics and Mining Sciences Vardoulakis I, Papamichos E. Surface instabilities in elastic anisotropic media with
2005;42(3):431e9. surface-parallel Griffith cracks. International Journal of Rock Mechanics and
Anderson TL. Fracture mechanics: fundamentals and applications. Boca Raton, FL, Mining Sciences & Geomechanics Abstracts 1991;28(2e3):163e73.
USA: Taylor & Francis; 2005. Wiebols GA, Cook NGW. An energy criterion for the strength of rock in polyaxial
Cook NGW. The failure of rock. International Journal of Rock Mechanics and Mining compression. International Journal of Rock Mechanics and Mining Sciences &
Sciences 1965;2(4):389e403. Geomechanics Abstracts 1968;5(6):529e49.
Cotterell B, Rice JR. Slightly curved or kinked cracks. International Journal of Frac- Wong TF, David C, Zhu W. The transition from brittle faulting to cataclastic flow in
ture 1980;16(2):155e69. porous sandstones: mechanical deformation. Journal of Geophysical Research
Diederichs MS. Rock fracture and collapse under low confinement conditions. Rock 1997;102(B2):3009e25.
Mechanics and Rock Engineering 2003;36(5):339e81. Yu MH. Unified strength theory and its applications. Berlin, Germany: Springer;
Erdogan F, Sih GC. On the crack extension in plates under plane loading and 2004.
transverse shear. Journal of Fluids Engineering 1963;85(4):519e25. Zuo JP, Li HT, Xie HP, Ju Y, Peng SP. A nonlinear strength criterion for rocklike
Evans B, Frederich JT, Wong TF. The brittleeductile transition in rocks: recent materials based on fracture mechanics. International Journal of Rock Mechanics
experimental and theoretical progress. In: Duba AG, Durham WB, Handin JW, and Mining Sciences 2008;45(4):594e9.
Wang HF, editors. The brittleeductile transition in rocks. Washington DC, USA:
American Geophysical Union; 1990. p. 1e20.
Griffith AA. The phenomena of rupture and flow in solids. Philosophical Trans-
actions of the Royal Society A: Mathematical, Physical and Engineering Sciences Jianping Zuo obtained his M.Sc. and Ph.D. from China
1921;221:163e98. University of Mining and Technology (Beijing), China. He is
Griffith AA. The theory of rupture. In: Biezeno CG, Burgers JM, editors. Proceedings professor of Engineering Mechanics at the China Univer-
of the 1st International Congress for Applied Mechanics. Delft, Netherlands: sity of Mining and Technology (Beijing) where he has been
Waltman; 1924. p. 54e63. Head of the Department of Engineering Mechanics, School
Heard HC. Transition from brittle fracture to ductile flow in Solenhofen limestone as of Mechanics and Civil Engineering. He has been involved
a function of temperature, confining pressure, and interstitial fluid pressure. In: in rock mechanics and mining engineering research,
Griggs D, Handin J, editors. Rock deformation. Boulder, Colorado, USA: consulting and teaching for more than 10 years. He is in
Geological Society of America; 1960. p. 193e226. charge of and/or participates in 12 scientific research
Hoagland RG, Hahn GT, Rosenfeld AR. Influence of microstructure on fracture projects, which were funded by National Natural Science
propagation in rock. Rock Mechanics 1973;5(2):77e106. Foundation of China, National Basic Research Program of
Hoek E, Brown ET. Empirical strength criterion for rock mass. Journal of Geotech- China (973 Program), Beijing Major Scientific and Tech-
nical and Geoenvironmental Engineering 1980a;106(9):1013e35. nological Achievements into Ground Cultivation Project,
Hoek E, Brown ET. Underground excavations in rock. London, UK: Institution of the 111 Project and Coal Mines Corporations. He is the
Mining and Metallurgy; 1980b. author or co-author of more than 90 scientific papers. He has received 3 national pat-
Hoek E. Strength of jointed rock masses. Géotechnique 1983;23(3):187e223. ents, and 2 Natural Science and Technology Progress Awards. In recent years, he has
Hoek E, Brown ET. Practical estimates of rock mass strength. International Journal of been awarded National Excellent Doctoral Dissertation Award (2009), New Century
Rock Mechanics and Mining Sciences 1997;34(8):1165e86. Excellent Talents of the Ministry of Education (2009), Coal Youth Science and Technol-
Hoek E, Martin CD. Fracture initiation and propagation in intact rock e a review. ogy Award in China (2010), Beijing Outstanding Young Talent (2014), Fok Excellent
Journal of Rock Mechanics and Geotechnical Engineering 2014;6(4):287e300. Young Teachers Award (2014).

You might also like