Rock Mech Rock Eng (2012) 45:989–993
DOI 10.1007/s00603-012-0277-3
ISRM SUGGESTED METHOD
Three-Dimensional Failure Criteria Based
on the Hoek–Brown Criterion
Stephen Priest
Published online: 18 July 2012
Ó Springer-Verlag 2012
List of symbols true triaxial stress conditions. Although a number of three-
mb Hoek–Brown material constant dimensional failure criteria have been developed, such as
s Hoek–Brown material constant the Drucker and Prager (1952) criterion and Lade criterion
a Hoek–Brown material constant (Kim and Lade 1984), these criteria were not primarily
Co Uniaxial compressive strength developed for the application to rocks.
I10 First invariant of the effective stress tensor The widespread adoption of the empirical two-dimen-
r01 Major principal effective stress sional Hoek–Brown failure criterion (2DHB) (Hoek and
r02 Intermediate principal effective stress Brown 1997; Hoek et al. 2002) for rock engineering appli-
r03 Minor principal effective stress cations has prompted a number of researchers to develop
soct Octahedral shear stress three-dimensional versions, in which the predicted major
r01hb Major principal effective stress at failure for the 2D effective principal stress at failure is dependent on the
Hoek–Brown criterion intermediate effective principal stress, in addition to the
r03hb Minor principal effective stress at failure for the 2D parameters in the existing 2DHB failure criterion. Three-
Hoek–Brown criterion dimensional versions of the 2DHB failure criterion have been
a Simplified Priest material constant proposed by Pan and Hudson (1988), Priest (2005) and Zhang
b Simplified Priest material constant and Zhu (2007). Zhang (2008) presented a generalised ver-
w Simplified Priest material constant sion of the Zhang–Zhu criterion. Melkoumian et al. (2009)
presented an explicit version of the ‘comprehensive’ Priest
criterion. Conventionally, in the literature, each criterion has
been named after the author(s) who first described the crite-
rion; this convention will be adopted here. It is likely that
1 Description
additional new three-dimensional versions of the Hoek–
Brown criterion will be developed over the next few years.
There is a growing body of experimental evidence
(Takahashi and Koide 1989) to suggest that the interme-
diate principal stress has a substantial influence on the
2 Background
strength of rock materials. Widely adopted failure criteria,
such as the Coulomb and Hoek–Brown criteria, ignore the
The most recent generalised version of the 2DHB failure
influence of the intermediate principal stress and therefore
criterion is introduced by Eberhardt and Rahjoo (this vol-
may not provide a reliable prediction of rock strength under
ume). This version of the Hoek–Brown criterion is here
referred to as ‘generalised’ because the key parameters mb,
s and a can take any general values to allow the application
S. Priest (&)
to intact rock and to rock masses. The paper explains how
School of Civil, Environmental and Mining Engineering,
The University of Adelaide, Adelaide, SA, Australia the parameters mb, s and a for a fractured rock mass can be
e-mail: [Link]@[Link] estimated from empirical expressions.
123
990 S. Priest
For intact rock, the parameters mb, s and a are mi, 1.0 Pan and Hudson (1988) is the absence of the intermediate
and 0.5, respectively. A number of authors, including Hoek principal stress in the third term of Eq. (1). The generalised
and Brown (1997), provide tabulations of suggested values form of the Pan–Hudson criterion can be written as
of mi for a range of rock types. Alternatively, the param- 1
ð11aÞ 3soct a 3mb soct
eters mi, s and a can be determined from a series of con- s Co ¼ Co pffiffiffi þ pffiffiffi mb I10 ð4Þ
ventional triaxial tests on intact rock, as explained by 2 2 2
Eberhardt and Rahjoo (this volume). where the parameters are as defined earlier. Again, a
In the following section, three-dimensional versions of numerical strategy is required to determine the value of r01
the Hoek–Brown criterion have been expressed in terms of in Eq. (4). Although there is apparently only a minor dif-
the parameters mb, s and a, in order to provide a general- ference between the GZZ and GPH criteria, these criteria
ised formulation. However, since these criteria have not predict very different strength values.
been shown to be, nor indeed claimed to be, applicable to
fractured rock masses, the parameters mb, s and a should be 3.3 Generalised Priest (GP) Criterion
replaced by mi, 1.0 and 0.5, respectively, and the criteria
limited to the application to intact rock materials. A three-dimensional version of the Hoek–Brown yield
criterion was developed by Priest (2005) by combining
3 Formulation the two-dimensional Hoek and Brown (1997) and the
three-dimensional Drucker and Prager (1952) criteria.
3.1 Generalised Zhang–Zhu (GZZ) Criterion The nomenclature ‘Priest criterion’ has been adopted
following Zhang (2008). The term ‘comprehensive’
The Zhang–Zhu criterion was first presented by Zhang and three-dimensional Hoek–Brown criterion was adopted by
Zhu (2007). A generalised version of this criterion, based Priest (2005) to distinguish this failure criterion from the
on the generalised Hoek–Brown criterion, was presented ‘simplified’ version described below. The term ‘com-
by Zhang (2008) as follows: prehensive’ is somewhat misleading, since this criterion
1 0 is no more comprehensive than the other criteria outlined
0
ð11aÞ 3soct a 3mb soct mb 3I1 r2 above. This criterion will therefore be referred to as the
s Co ¼ Co pffiffiffi þ pffiffiffi ð1Þ
2 2 2 2 generalised Priest criterion (Priest 2009). The formula-
tion presented by Priest (2005) required a numerical
where r03 is the minor effective principal stress at failure,
solution strategy. Melkoumian et al. (2009) addressed
r02 is the intermediate effective principal stress at failure, r01 this problem by developing an explicit version of this
is the major effective principal stress at failure, and the three-dimensional Hoek–Brown criterion involving the
other Hoek–Brown parameters are as defined earlier. 2DHB minimum effective stress at failure r03hb , as sum-
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
0 2 2 2 marised below:
r1 r02 þ r02 r03 þ r03 r01
soct ¼ ð2Þ mb r02 þ r03
3 C ¼sþ ð5Þ
2 Co
and I10 is given by
E ¼ 2 C a Co ð6Þ
r01 þ r02 þ r03
I10 ¼ ð3Þ F ¼ 3 þ 2 a C a1 mb ð7Þ
3
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2ffi
In Eq. (1), E E2 F r02 r03
0 r02 þ r03
r03hb ¼ þ ð8Þ
mb 3I1 r02 mb r03 þ r01 2 2F
¼ :
2 2 Equation (8) gives two values for r03hb , one of which can
Unfortunately, this failure criterion cannot easily be be negative and the other positive. In a compressive stress
formulated to express r01 explicitly in terms of the input regime, r03hb will be positive, so Melkoumian et al. (2009)
data. It is, however, a relatively straightforward matter to recommended that the greater or positive root in Eq. (8)
apply a numerical strategy to determine the value of r01 that should be adopted.
satisfies Eqs. (1)–(3). a
mb r03hb
P ¼ Co þs ð9Þ
3.2 Generalised Pan–Hudson (GPH) Criterion Co
Finally,
Zhang and Zhu (2007) demonstrated that the only differ-
r01 ¼ 3r03hb þ P r02 þ r03 : ð10Þ
ence between their yield criterion and the one proposed by
123
Three-Dimensional Failure Criteria 991
Major effective principal stress at failure (MPa)
3.4 Simplified Priest (SP) Criterion 1400
1300 Minor effective principal stress = 60 MPa GP
Priest (2005) proposed a ‘simplified’ three-dimensional 1200
1100 GZZ
version of the Hoek–Brown criterion, which has the merit 1000 SP
Test data
of providing an easily computed estimate for the three- 900 GPH
dimensional effective failure stress r01 . 800
700
r01 ¼ r01hb þ 2r03hb r02 þ r03 ð11Þ 600 2DHB
500
where, as before, r03hb is the minimum 2DHB effective 400
300
stress at failure, and r01hb is the maximum 2DHB effective
200
stress at failure, calculated from Eq. (2), and 100
0
r03hb ¼ wr02 þ ð1 wÞr03 ð12Þ 0 50 100 150 200 250 300 350 400 450 500
Intermediate effective principal stress (MPa)
where w is a weighting factor in the range 0–1, which
governs the relative influence of r02 and r03 on the strength Fig. 1 Four three-dimensional Hoek–Brown failure criteria and also
the two-dimensional Hoek–Brown failure criterion for KTB amphib-
of the rock. Priest (2005) suggested that for a wide range of olite, mi = 35.4, rci = 159.1 MPa
rock types, w can be estimated from the following simple
power law.
shows that the 2DHB failure criterion is, as expected,
w a r0b
3 ð13Þ insensitive to the intermediate principal stress. All criteria,
except the GPH, diverge from the common point where
Preliminary studies by Priest (2005) suggest that, as a r02 ¼ r03 = 60 MPa and r01 = 662.9 MPa. This somewhat
first approximation, a = b = 0.15. anomalous behaviour of the GPH criterion merits further
investigation. The generalised Priest criterion (GP) is the
most sensitive to the influence of the intermediate principal
4 Experimental Data on Rock stress, predicting substantially higher values of r01 than the
other criteria and the test data. The simplified Priest cri-
True triaxial rock test data published by Chang and terion (SP), adopting Eq. (13) to calculate the weighting
Haimson (2000) for the KTB amphibolite and by Haimson factor w, is the least sensitive, with the GZZ lying between
and Chang (2000) for Westerly granite were selected to these two. These latter two criteria appear to model the test
compare the predictions of the four three-dimensional data reasonably well. This same general pattern is repeated
Hoek–Brown failure criteria. Data published in these for the test data at other values of intermediate principal
papers include uniaxial and ‘conventional’ triaxial test data stress, presented by Chang and Haimson (2000), with the
for these rocks (where r02 ¼ r03 ), so it was possible to simplified Priest (SP) and the GZZ criteria offering the best
determine the experimental values of the Hoek–Brown models for the test data. The test data do, however, indicate
parameter mi and the uniaxial compressive strength of the that the sensitivity of r01 to r02 reduces at higher values of
intact rock material Co, on the assumption that the Hoek– r03 .
Brown parameters s and a are 1.0 and 0.5, respectively, for Figure 2 shows the four three-dimensional Hoek–Brown
the intact rock specimens. Simple curve fitting against the failure criteria and also the 2DHB failure criterion for a
data presented for these two rock types gave the following minor principal effective stress r03 of 20 MPa and an
best estimates for the key parameters: KTB amphibolite intermediate effective principal stress at failure r02 ranging
mi = 35.4, Co = 159.1 MPa; Westerly granite mi = 40.5, from 20 to 202 MPa, for Westerly granite. This figure
Co = 191.0 MPa. Although the test data can be compared again shows that the 2DHB failure criterion is, as expected,
with the predictions of the yield criteria in a number of insensitive to the intermediate principal stress. Again, all
different ways, including, for example, plots of failure criteria, except the GPH, diverge from the common point
envelopes in the deviatoric plane, the primary focus here where r02 ¼ r03 = 20 MPa and r01 = 457.1 MPa. Again,
will be to examine how well the failure criteria model the the generalised Priest criterion (GP) is the most sensitive to
influence of the intermediate principal stress r02 . the influence of the intermediate principal stress, predicting
Figure 1 shows the four three-dimensional Hoek–Brown substantially higher values of r01 than the other criteria and
failure criteria and also the 2DHB failure criterion for a the test data. The simplified Priest criterion (SP), adopting
minor principal effective stress r03 of 60 MPa and an Eq. (13) to calculate the weighting factor w, is the least
intermediate effective principal stress at failure r02 ranging sensitive, with the GZZ again lying between these two.
from 60 to 450 MPa for KTB amphibolite. This figure These latter two criteria appear to model the test data
123
992 S. Priest
None of the criteria examined, with the exception of the
Major effective principal stress at failure (MPa)
1100
Minor effective principal stress = 20 MPa GP simplified Priest criterion, require additional input param-
1000
900
eters beyond r02 and the parameters required for the 2DHB
800 GZZ
criterion. It is, of course, possible to obtain a close fit to
700 SP
almost any experimental data by incorporating additional
600 parameters (or ‘fudge factors’) into the formulation of a
GPH
500 criterion. Adoption of a criterion with one or more addi-
2DHB
400 tional parameters would necessitate the determination of
300 Test data these parameters for the particular rock type from a series
200 of true triaxial tests. Such testing facilities are not generally
100 available to rock mechanics practitioners, so existing and
0
0 20 40 60 80 100 120 140 160 180 200 220 240
future three-dimensional Hoek–Brown failure criteria with
Intermediate effective principal stress (MPa) additional parameters are likely to be of limited practical
use.
Fig. 2 Four three-dimensional Hoek–Brown failure criteria and also A potential advantage of three-dimensional failure cri-
the two-dimensional Hoek–Brown failure criterion for Westerly
granite, mi = 40.5, rci = 191.0 MPa
teria based on the Hoek–Brown criterion is that, theoreti-
cally, it would be possible to adopt values of mb, s and
a that reflect the properties of a fractured rock mass.
Consideration of the strength of fractured rock masses is,
reasonably well. This same general pattern is repeated for
however, beyond the scope of these suggested methods.
the test data at other values of intermediate principal stress,
presented by Haimson and Chang (2000), with the sim-
plified Priest and the GZZ criteria offering the best models
6 Recommendations
for the test data. As for the KTB amphibolite, the test data
for Westerly granite indicate that the sensitivity of r01 to r02
A significant obstacle to recommending which, if any, of
reduces at higher values of r03 .
the above three-dimensional Hoek–Brown failure criteria
should be applied to rock materials and rock masses is the
relative paucity of rock strength test data for specimens
5 Advantages and Limitations
loaded under uniaxial, conventional triaxial and true tri-
axial conditions for a range of rock types.
The generalised Priest criterion seems to overestimate the
It is recommended that a substantial amount of further
experimentally determined true triaxial rock strength for
research and rock testing should be conducted before any
KTB amphibolite and Westerly granite by around 10–30 %
of the three-dimensional Hoek–Brown failure criteria can
for r02 in the approximate range 2r03 to 4r03 . This overes-
be applied with confidence. This testing, which should
timate rises to more than 50 % at higher values of r02 .
cover a wide range of rock types and rock strengths, should
Clearly, the generalised Priest criterion (GP) should be
follow the testing strategy adopted by Chang and Haimson
used with some caution at this stage, particularly at higher
(2000) for the KTB amphibolite and by Haimson and
levels of intermediate principal stress.
Chang (2000) for Westerly granite, as follows:
The simplified Priest criterion (SP), adopting Eq. (13) to
calculate the weighting factor w, and the GZZ criterion (a) A series of conventional uniaxial and triaxial tests
both provide a reasonably good model of the experimen- should be conducted on intact rock specimens to
tally determined true triaxial rock strength for KTB determine the uniaxial compressive strength and the
amphibolite and Westerly granite. The generalised and Hoek–Brown parameter mi for the rock material,
simplified Priest criteria (GP, SP) do, however, have the following the relevant ISRM Suggested Methods. It is
benefit of being amenable to direct explicit evaluation and also recommended that this series of conventional
so are more suitable for incorporation into numerical uniaxial and triaxial tests should be repeated in the
modelling software. The simplified Priest criterion sub- true triaxial testing apparatus to assess if there is any
stantially underestimates the experimentally determined specimen geometry or testing machine influence on
true triaxial rock strength for KTB amphibolite and Wes- the strength results.
terly granite when the minor principal stress is zero. Under (b) A series of true triaxial tests should be conducted on
these conditions the weighting factor w in Eq. (13) is zero, specimens of the same intact rock, covering a range
which creates a negative slope for the graph of r01 versus r02 of minor and intermediate effective principal stresses.
for this failure criterion. If it is assumed that the geological strength index
123
Three-Dimensional Failure Criteria 993
(GSI) is 100 for intact rock, it will then be possible to strength and deformability in Westerly granite. Int J Rock Mech
assess the predictions of the published three-dimen- Min Sci 37:285–296
Hoek E, Brown ET (1997) Practical estimates of rock mass strength.
sional Hoek–Brown failure criteria over a range of Int J Rock Mech Min Sci Geomech Abstr 34:1165–1186
rock types and stress levels. Hoek E, Carranza-Torres S, Corkum B (2002) Hoek–Brown failure
criterion—2002 version. Rockscience. [Link]
Evaluation of the three-dimensional Hoek–Brown fail- com/highlights
ure criteria for fractured rock masses presents a substantial Kim MK, Lade PV (1984) Modelling rock strength in three
challenge. The sampling and testing of undisturbed speci- dimensions. Int J Rock Mech Min Sci Geomech Abstr 21:21–33
Melkoumian N, Priest SD, Hunt SP (2009) Further development of
mens of fractured rock of a size sufficient to represent in
the three-dimensional Hoek–Brown yield criterion. Rock Mech
situ rock mass conditions presents a significant technical Rock Eng 42:835–847
and financial difficulty. Furthermore, true triaxial testing Pan XD, Hudson JA (1988) A simplified three-dimensional Hoek–
equipment is currently only capable of testing relatively Brown yield criterion. In: Romana M (ed) Rock mechanics and
power plants. Balkema, Rotterdam, pp 95–103
small specimens of intact rock. One promising strategy
Priest SD (2005) Determination of shear strength and three-dimen-
might be in situ pressuremeter tests in boreholes coupled sional yield strength for the Hoek–Brown criterion. Rock Mech.
with testing of recovered core and/or chips and detailed Rock Eng 38:299–327
downhole surveys. Priest SD (2009) Comparisons between selected three-dimensional
yield criteria applied to rock. Rock Mech Rock Eng 43:379–389
Takahashi M, Koide H (1989) Effect of the intermediate principal
stress on strength and deformation behavior of sedimentary rocks
at the depth shallower than 2000 m. In: Maury V, Fourmaintraux
References D (eds) Rock at great depth, vol 1. Balkema, Rotterdam,
pp 19–26
Chang C, Haimson BC (2000) True triaxial strength and deforma- Zhang L (2008) A generalized three-dimensional Hoek–Brown
bility of the German Continental Deep Drilling Program (KTB) strength criterion. Rock Mech Rock Eng 41:893–915
deep hole amphibolite. J Geophys Res 105:18999–19014 Zhang L, Zhu H (2007) Three-dimensional Hoek–Brown strength
Drucker D, Prager W (1952) Soil mechanics and plastic analysis or criterion for rocks. J Geotech Geoenviron Eng ASCE
limit design. Q Appl Math 10:157–169 133:1128–1135
Haimson BC, Chang C (2000) A new true triaxial cell for testing
mechanical properties of rock, and its use to determine rock
123
