Journal of Rock Mechanics and Geotechnical Engineering.

2012, 4 (2): 140–148

Comprehensive compressive-tensile strength criterion for intact rock

Hossein Bineshian1*, Abdolhadi Ghazvinian2, Zahra Bineshian3

1
University of Western Australia, Perth, Australia
2
Tarbiat Modares University, Tehran, Iran
3
Perlite Construction Company, Tehran, Iran
Received 12 December 2011; received in revised form 8 March 2012; accepted 25 March 2012

Abstract: This paper presents a strength criterion for intact rock, which can well describe triaxial test data under compressive
or tensile stress state. The proposed criterion is defined in terms of three parameters. One parameter expresses the apparent
unconfined compressive strength (AUCS), obtained from the Coulomb-Mohr criterion, as a regulated unconfined compressive
strength (RUCS). Two other parameters,  and , are material-dependent that can be determined by regression analysis. The
proposed criterion is compared with selected applicable strength criteria separately for compressive and tensile strengths.
Coefficient of determination and accordance coefficient are considered in comparisons between the proposed and selected
strength criteria.
Key words: strength criterion; failure mechanics; compression; tension

(1960), McClintock and Walsh (1962), Murrell (1963,

1 Introduction 1965), and Jaeger et al. (2007).
 Nonlinear characteristics of rock were described by
Strength criteria have a long history. The first linear von Karman (1911) on marble samples. Griggs (1936),
criterion was developed by Coulomb in 1773 (Jaeger et Mogi (1965, 1966a, 2006), Donath (1970) and
al., 2007). In 1900, Mohr extended the criterion and Hoshino et al. (1972) reported similar observations.
introduced the concept of the Mohr envelope Empirical fitting of curves to triaxial test data by
(Bineshian, 2000). The Coulomb-Mohr criterion is a Franklin and Hoek (1970), Franklin (1971), and Hoek
linear compressive-tensile criterion as follows: and Brown (1980a) provided additional evidence that
 1  C  B 3 (1) most triaxial behavior is nonlinear.
Over the past sixty years, many researchers have
where  1 and  3 are the major and minor principal
aimed at addressing limitations of various criteria.
stresses at failure, respectively; and C and B are the
Balmer (1952) developed a criterion, which can be
parameters of material. Parameter C in  1 -  3
used for compression but is not suitable for tension
coordinate system represents the apparent uniaxial
(Sheorey, 1997). This criterion was modified by
compressive strength (AUCS) and parameter B is the
Sheorey et al. (1989), however, it was not applicable to
slope of the Coulomb-Mohr linear envelope.
all rock types. Murrell (1963, 1965) developed a
In spite of its popularity, the Coulomb-Mohr
criterion, which was not reliable for tension although,
criterion cannot well describe the nonlinear
after being modified by Hoek (1968), Bieniawski
characteristics of rock strength under triaxial loading
(1974), Yudhbir and Prinzl (1983), Das and Sheorey
conditions. For brittle rocks, the criterion can
(1986) and Sheorey et al. (1986), its applicability to
overestimate the tensile strength.
compression was improved. These researchers
Griffith (1921, 1924) developed a nonlinear criterion
developed new criteria, based on Murrell criterion,
based on the initiation of crack propagation in brittle
which introduced new parameters and required
materials. This criterion was later modified by Brace
principal stresses to be normalized. Although the above
researchers modified the criterion to make it applicable
Doi: 10.3724/SP.J.1235.2012.00140
*
Corresponding author. Tel:+61-450108437; E-mail: Bineshian@gmail.com  to a wider range of rock types, the criterion was not
Hossein Bineshian et al. / J Rock Mech Geotech Eng. 2012, 4 (2): 140–148
applicable to tension (Bineshian, 2000). Despite the
fact that Bieniawski criterion (Bieniawski, 1974)
cannot predict tensile behavior of rock, it has been
widely used for evaluation of intact rock strength due
to its reliable prediction of compressive strength.
Bieniawski criterion is defined in terms of parameters
B and  as follows, and it is not applicable to a
tensile stress state (Bineshian and Bineshian, 2011a):
 None of the criteria above is comprehensively
1  
1 B 3  (2)
 c   c 
where  c is the unconfined compressive strength
(UCS) derived from test.
Other criteria, having forms different to those based
on Murrell criterion, are those by Fairhurst (1964),
Hobbs (1964), Mogi (1966a, 1966b), and Franklin
(1971). These criteria have limited applicability. For
example, Hobbs criterion is only applicable to coal.
Fairhurst criterion, which is another form of Griffith
criterion, cannot predict the tensile strength (Bineshian,
2000). Franklin criterion cannot predict the UCS.
Ohnaka (1973) developed a normalized form of Mogi
criterion, but it is limited to specific types of rocks.
Up to the 1980s, many criteria were proposed, all of
which were not applicable to either compression or
tension. However, a new empirical strength criterion
was proposed by Hoek and Brown (1980b, 1988):
 1  3 m 3
  1 (3)
 c  c  c
where  1 and  3 are the major and minor effective
principal stresses, respectively;  c is the effective
UCS; and m is the parameter of Hoek-Brown
criterion.
Hoek-Brown criterion was based on the Griffith
criterion (1921, 1924) as modified by McClintock and
Walsh (1962). At that time, the criterion was found to
be applicable to many rock types, and subsequent
modifications made it even more widely applicable.
Further developments have been suggested for the
criterion (e.g. Yoshida et al., 1990), however, most
have failed to deliver significant improvements in the
capability of the criterion to estimate strength
(Bineshian and Bineshian, 2011b).
A nonlinear criterion was also developed by
Johnston and Chiu (1984) and revised by Johnston
(1985). The characteristic of this criterion is that its
parameters depend on the UCS. From 1984 to 1988,
Rao (1984) and Ramamurthy et al. (1985, 1988)
suggested a nonlinear criterion through conducting
tests on 80 rock types. The criterion can be defined as
141
follows:

1   3  
 B  c  (4)
3  3 
This criterion is only applicable to positive
confining pressures (  3  0 ), and it has limitations at
low confining pressures (Bineshian, 2000; Bineshian
and Bineshian, 2012a).
applicable to all rock types and all loading conditions.
Therefore, an alternative criterion, which can address
these limitations, needs to be developed.
The current study is based on the results obtained
from a comprehensive theoretical and experimental
study on the failure characteristics of 1 743 samples of
rock, coal and concrete. An alternative criterion is
proposed, which is applicable to compressive and
tensile stress states. This criterion was proposed in
1999–2000 as a part of first author’s master thesis
available at the Library of Tarbiat Modares University
as the properties of the Ministry of Science, Research
and Technology of Iran. Since then, it has been
developed and evaluated for more types of intact rocks,
rock masses, concretes and soils in Iran and Australia.
This criterion enables a Mohr envelope to be fitted
closely to actual triaxial data from more than 29 rock
types. The criterion is compared with the original
Coulomb-Mohr criterion (Jaeger et al., 2007),
Bieniawski criterion (Bieniawski, 1974; Bineshian and
Bineshian, 2011a), Hoek-Brown criterion (Hoek and
Brown, 1980; Bineshian and Bineshian, 2011b) and
Ramamurthy et al. criterion (Ramamurthy et al., 1985;
Bineshian and Bineshian, 2012a), and it is shown to
provide a good alternative to these criteria. The
parameters of the proposed criterion for igneous,
sedimentary and metamorphic rocks and also a special
consideration of inherent anisotropy of coal will be
presented in another paper.
2 Proposed strength criterion
The proposed strength criterion is defined as
    3 
1   3   c  c  (5)
  c   3 
where  and  are rock constants;  c is the
regulated unconfined compressive strength (RUCS),
 c   C , where  is a factor that defines RUCS in
terms of AUCS.
The parameter C is determined as the major
principal stress at the point where the Coulomb-Mohr
142
envelope in  3 -  1 coordinate system intersects with
 1 axis. This parameter is not necessarily the UCS of
rock, because if more than one uniaxial compression
test is carried out on the samples, then several possibly
different values of UCS will be obtained. When only
one uniaxial compression test or point load test (PLT)
has been carried out, its test result may not be
representative. Therefore, it should be regulated based
on the natural trend of triaxial test data. We call this
regulated value the RUCS.
The RUCS is the regulated value for UCS. It means
that the observed value of the uniaxial compression
test is corrected by the regulation factor (  ). Thus, to
calculate the best regulation factor,  should be
added as a parameter to the strength criterion (Eq. (5)).
Then the criterion should be fitted to triaxial test data
by regression analysis and, therefore,  that
regulates the AUCS will be obtained. AUCS in this
study is calculated by the Coulomb-Mohr criterion.
Hence  c can be calculated in following three steps:
(1) Calculate AUCS by fitting the Coulomb-Mohr
criterion to triaxial test data.
(2) Calculate the regulation factor  by fitting the
proposed criterion (Eq. (5)) to triaxial test data.
(3) Calculate RUCS by multiplying AUCS by  .
The value of RUCS is determined by setting  3  0
in Eq. (5):
 3  0  1   c   C (6)
By setting  1  0 in Eq. (5), the value of  3 is equal
to the uniaxial tensile strength (  t ):
  1   2  2  4  1
1  0   3   t  C
2
(7)
The originality of the proposed strength criterion
goes back to 2000 (Bineshian, 2000). In the process of
creating the form of the criterion (Eq. (5)), the authors
generated a simple mathematical function, which can
present both linear and nonlinear envelopes on the
triaxial test data in accordance with the input  3 -  1
data pairs to cover both linearity and nonlinearity of
intact geomaterials’ mechanical properties. Unlike
many empirical criteria from 1952 to the present (i.e.
criteria of Ballmer, Fairhurst, Hobbs, Murrell, Mogi,
Hoek, Franklin, Bieniawski, Hoek-Brown, Johnston,
Ramamurthy et al., Sheorey et al., Yoshida et al.,
Carter et al., etc.), this criterion is not expressed
through a power function, as which would limit the
applicability of the criterion (Bineshian, 2000;
Bineshian and Bineshian, 2011a, 2011b, 2012a). The
proposed functional dependence for the criterion was
Hossein Bineshian et al. / J Rock Mech Geotech Eng. 2012, 4 (2): 140–148
based on 2 076 linear and nonlinear regression
analyses (Bineshian, 2000), and was evaluated during
2000 to 2009 for intact rocks, rock masses, concretes
and soils (Bineshian, 2000, 2006, 2008; Bineshian and
Bineshian, 2011a, 2011b, 2012a). The proposed
criterion is a compressive-tensile strength criterion,
which fits well the compressive and tensile strength
data. It is an empirical-theoretical strength criterion,
because RUCS is calculated using the Coulomb-Mohr
criterion (a theoretical strength criterion) and the
function has been formed by empirical study. The
constants of the proposed criterion can be determined
easily by performing the regression analysis of the data
groups containing the triaxial compression and tension
test data. The general method to determine the
constants of the proposed criterion together with the
selected criteria is presented in sub-section 4.1 of this
paper.
3 Required triaxial test data and
data evaluation procedure
A total of 2 149 data pairs obtained from triaxial
compression tests on 392 samples of 34 rock types
were collected from literatures (Schwartz, 1964;
Horino and Ellikson, 1970; Ouyang and Elsworth,
1991; Vutukuri and Hossaini, 1992; Hossaini, 1993;
Sheorey, 1997; Mahab Ghodss Consulting Engineers
Co., 1999, 2003; Bineshian, 2000, 2006, 2008;
Bineshian and Bineshian, 2011a, 2011b, 2012a). The
following conditions were applied to eliminating
unsuitable data:
(1) Data pairs must be in the form of major and
minor principal stresses ( 3 ,  1 ) at failure.
(2) There have to be 5 or more ( 3 ,  1 ) data pairs
in each data group (Bineshian, 2000).
(3) All data groups must contain the value of UCS
(Bineshian, 1999).
(4) All data pairs in each data group must contain
the principal stress combination above the Mogi’s
transition limit (1966a, 1966b) from brittle to ductile
failure (Sheorey, 1997), i.e.
 1  4.40 3 (8)
(5) Each data group must contain the results of
triaxial compression tests, in which the confining
pressure is greater than zero (Bineshian, 2000;
Bineshian and Bineshian, 2011a).
(6) Each data group must contain a range of
confining pressures with at least one test at a confining
pressure greater than a half of UCS (Hossaini, 1999).
Hossein Bineshian et al. / J Rock Mech Geotech Eng. 2012, 4 (2): 140–148 143

(7) Data groups must be definable in terms of Nelder and Mead (Neter and Wasserman, 1977; Neter
regressive shaped curve (Bineshian and Bineshian, et al., 1988) was considered. However, no significant
2011b). difference between the results obtained using the
(8) Data pairs are considered invalid if  1 is less proposed method and those obtained using the above
than UCS (Bineshian, 2000). methods was noted.
(9) Data groups must have similar slenderness ratios 4.2 Qualification criteria
(Bineshian, 1999, 2006). The coefficients of determination ( R 2 ) were
After eliminating unsuitable data, a total of 1 743 calculated for nonlinear and linear strength criteria,
data pairs were remained from 250 data groups for 15 respectively. For nonlinear strength criterion, R 2
igneous rock types, 6 sedimentary rock types, 7 (0  R 2  1) can be expressed as
metamorphic rock types and 4 coal types. For the coal
R2  1 
[ 1 jexp  f ( 3jexp )]2 (14)
samples, different orientations of main cleats and

 12jexp   1 jexp / n 
2

bedding planes relative to the orientation of the major

For linear strength criterion, R 2 can be expressed
principal stress were considered. All laboratory tests
as
cited were carried out according to the International
     3jexp1jexp / n 
2
Society for Rock Mechanics (ISRM) suggested R2 
3jexp 1jexp

 2     / n 12jexp 
/ n   
 3jexp
2 2
methods (Brown, 1981).
3jexp    1jexp

4 Quantification of performance of (15)

proposed strength criterion
4.1 Linear and nonlinear regression
The following two equations are examples of linear
and nonlinear regression models, respectively, in terms
of independent variable X i and dependent variable
Yi :
Yi   0  1 X i   i (9)
Yi   0   1
 2 X i
 i
where  0 , 1 and  2 are the regression coefficients;
and  i is the error in the regression model. The
models are suitable for describing linear and nonlinear
relationships between principal stress data pairs.
A linear regression model was applied for the
Coulomb-Mohr criterion, and nonlinear models for the
proposed criterion and criteria of Bieniawski,
Hoek-Brown, and Ramamurthy et al.
For the Bieniawski criterion (Bineshian, 2000;
Bineshian and Bineshian, 2011a), there are
   
X i  lg  3  , Yi  lg  1  1 , B  10 0 ,   1 (11)
  c    c  rocks. In 52% of the total cases, the coefficients of
For Hoek-Brown criterion (Bineshian, 2000;
Bineshian and Bineshian, 2011b), there are
X i   3 , Yi  ( 1   3 )2 , m  1 (  0 )1 ,  c   0
(12)
For Ramamurthy et al. criterion (Bineshian, 2000;
Bineshian and Bineshian, 2012a), there are
   
X i  lg  3  , Yi  lg  1  1 , B  10 0 ,   1 (13)
  c   3  and coal are presented in Table 3. Based on the results
The nonlinear regression method proposed by
where  1jexp and  3jexp are the observed values of
 1 and  3 for the jth data pair, respectively; n is the
number of ( 3 ,  1 ) data pairs. The closer the value
of R 2 approaches to 1, the better the fitting between
the criterion and the experimental data is.
The accordance coefficient ( 2 ) quantifies how a
strength criterion defines the actual data, i.e.
2 
 ( 1jexp   1jcal ) 2
( 2  0) (16)
(10)  ( 1jexp   1jexp ) 2
where  1jcal is the calculated value of  1 for the jth
data. The smaller the value of  2 is, the better the
strength criterion accords with the triaxial test data.
The accordance is considered good when  2  0 .
5 Evaluation of proposed strength
criterion
5.1 Evaluation using coefficient of determination
Table 1 compares the coefficients of determination
obtained from the application of the proposed criterion
and other four criteria previously described to intact
determination applicable to the proposed criterion were
close to 1. This result confirms the good correlation
between the estimated and the actual data. In only 25%
and 3% of the total cases, the values of R 2 applicable
to the Bieniawski and Coulomb-Mohr criteria were
also close to 1.
The comparison for coal (25 data groups) is
presented in Table 2, and summations for intact rocks
listed in Tables 1–3, and considering the correlation of
144 Hossein Bineshian et al. / J Rock Mech Geotech Eng. 2012, 4 (2): 140–148

triaxial compression and tension test data, the proposed

Table 1 Evaluation of correlation between strength criteria and criterion is found to compare favorably with the other
actual triaxial test data for intact rocks.
four criteria.
Correlation of the criteria with actual data (%)
R2
Figs. 1 and 2 compare four criteria with Betourney
Proposed Bieniawski Hoek-Brown Ramamurthy Coulomb-Mohr
et al.’s triaxial test data for gneiss obtained from
criterion criterion criterion et al. criterion criterion
Strathcona Mine, Canada (Sheorey, 1997).
1.00 52 25 11 13 3
≥0.99 80 45 31 37 14
Triaxial data
≥0.98 88 55 46 46 26
≥0.97 93 62 55 55 38
≥0.96 94 68 64 62 50
≥0.95 95 72 71 66 57
≥0.94 99 78 74 68 65
≥0.93 99 81 77 72 75
≥0.92 99 87 79 76 77
≥0.91 99 87 83 78 82
≥0.90 99 87 84 81 85
Fig. 1 Fitting the proposed, Coulomb-Mohr and Hoek-Brown
Table 2 Evaluation of correlation between strength criteria and criteria to Betourney et al.’s data for gneiss (Sheorey, 1997).
actual triaxial test data for coal.
Correlation of the criteria with actual data (%) Triaxial data
R2
Proposed Bieniawski Hoek-Brown Ramamurthy Coulomb-Mohr
criterion criterion criterion et al. criterion criterion
1.00 24 6 4 0 0
≥0.99 48 28 24 6 0
≥0.98 80 44 40 17 4
≥0.97 92 50 44 33 4
≥0.96 92 61 60 44 8
Fig. 2 Fitting Bieniawski criterion to Betourney et al.’s data for
≥0.95 96 72 64 50 12 gneiss (Sheorey, 1997).
≥0.94 96 89 76 56 24
≥0.93 96 89 80 61 24 The Ramamurthy et al. criterion fits poorly and
≥0.92 96 89 80 78 28
hence has not been included (Bineshian and Bineshian,
≥0.91 100 89 80 78 40
2012a). The proposed strength criterion and the
≥0.90 100 94 84 78 52
Bieniawski and Hoek-Brown criteria fit the triaxial test
data well. Note that the Bieniawski criterion was
Table 3 Evaluation of correlation between strength criteria and
actual triaxial test data for intact rocks and coal. regressed after the tensile data pairs were eliminated,
Correlation of the criteria with actual data (%) because it cannot provide strength envelope for tensile
R2
Proposed Bieniawski Hoek-Brown Ramamurthy Coulomb-Mohr
regime of stresses (Bineshian and Bineshian, 2011a),
criterion criterion criterion et al. criterion criterion therefore, it has been presented separately in Fig. 2.
1.00 47 22 9 11 2 Calculating parameters of strength criteria in this study
≥0.99 75 43 30 33 12
are listed in Table 4, in which all the test data are
referred to Sheorey (1997).
≥0.98 87 53 44 40 22
Figs. 1 and 2 show the fitting results under low
≥0.97 93 60 53 52 32
confining pressures. Medium level of confining
≥0.96 93 67 63 59 43
pressures is assessed as well. Fig. 3 shows the fitting
≥0.95 95 71 70 63 49 results of the criteria to Franklin and Hoek’s triaxial
≥0.94 98 78 73 66 58 test data for Pennant sandstone (Sheorey, 1997).
≥0.93 98 81 77 70 66 Because of presence of UCS data pair, Ramamurthy et
≥0.92 98 86 78 76 69
al. criterion could not be able to fit the triaxial test data.
In Fig. 3, the criteria have been assessed for medium
≥0.91 99 86 81 77 75
level of confining pressures. As can be seen in Fig. 3,
≥0.90 99 87 83 81 80
the fitting envelope of the proposed criterion is
Hossein Bineshian et al. / J Rock Mech Geotech Eng. 2012, 4 (2): 140–148 145

Table 4 Calculating parameters of strength criteria in this study.

Strength C
Test data R2    B  m
criterion (MPa)
Proposed 0.997 20.06 10.25 1.24 — — — —

Betourney Bieniawski 0.966 — — — — 3.51 0.75 –

et al.’s data Hoek-Brown 0.966 — — — — — — 20.48

Coulomb-Mohr 0.909 — — — 263.49 12.95 — —

Proposed 0.988 11.95 3.79 0.89 — — — —

Bieniawski 0.985 — — — — 3.65 0.74 —

Fig. 4 Fitting the proposed and all selected criteria to Mogi’s
Franklin and
data for Haruyama peridotite (Sheorey, 1997).
Hoek’s data Hoek-Brown 0.979 — — — — — — 13.52

Coulomb-Mohr 0.967 — — — 220.76 4.94 — —

Proposed 0.996 4.79 0.45 0.79 — — — —

Bieniawski 0.991 — — — — 3.77 0.96 —

Hoek-Brown 0.972 — — — — — — 10.12

Mogi’s data
Coulomb-Mohr 0.989 — — — 304.94 3.85 — —

Ramamurthy
0.99 — — — — 3.77 0.04 —
et al.
Gustkiewicz’s Proposed 0.997 8.06 0.52 0.69 — — — — Fig. 5 Fitting the proposed and Coulomb-Mohr criteria to
Gustkiewicz’s data for Rudna sandstone (Sheorey, 1997).
data Coulomb-Mohr 0.964 — — — 101.68 4.35 — —

Proposed 0.993 7.43 0.87 1.13 — — — — group of Haruyama peridotite (Sheorey, 1997), which
Brace’s data Hoek-Brown 0.963 — — — — — — 11.3 contains high confining pressure’s data pairs with
Coulomb-Mohr 0.98 — — — 385.81 5.51 — — R 2 =0.997. It shows better correlation with triaxial
Proposed 0.999 11.43 1.27 0.7 — — — — test data in comparison to the selected criteria. Fig. 5
Hobbs’s data shows the fitting results for the proposed criterion and
Coulomb-Mohr 0.911 — — — 15.26 4.04 — —
Coulomb-Mohr criterion to Gustkiewicz’s data for
Johnston’s Proposed 0.996 13.99 7.08 1.25 — — — —
Rudna sandstone (Sheorey, 1997) under high confining
data Coulomb-Mohr 0.893 — — — 100.93 6.99 — — pressure as well. But other selected criteria in this case,
because of presence of tensile data, could not be fitted,
even Hoek-Brown criterion that generally presents
strength envelope for tensile regime of stresses. The
proposed criterion for this data group shows a
satisfactory fitting with R 2 =0.997. The parameters
for the proposed and selected strength criteria are listed
in Table 4.
Fig. 6 shows fitting results of the proposed,
Coulomb-Mohr and Hoek-Brown criteria to Brace’s
Fig. 3 Fitting the proposed and selected criteria to Franklin and data for Frederick diabase (Sheorey, 1997). This data
Hoek’s data for Pennant sandstone (Sheorey, 1997).
group contains tensile data pairs under very high
satisfactory and better than other selected criteria. The confining pressures. The maximum confining pressure
parameters for the proposed and selected strength in this data group is almost 318 MPa. Bieniawski
criteria are listed in Table 4. criterion and Ramamurthy et al. criterion could not be
Figs. 4 and 5 show the fitting results for two groups fitted because of presence of tensile data pairs. As can be
of triaxial test data under high confining pressures. In seen in Fig. 6, the proposed criterion fits well the data
Fig. 4, all selected and proposed criteria could be fitted. groups under high confining pressures as well, but
This data group does not contain any UCS data pairs, Hoek-Brown criterion in the range of high confining
and after fitting the Coulomb-Mohr criterion on it, C is pressures does not present a suitable fit on data pairs
used for the selected criteria as UCS. As can be seen in under very high confining pressures. Calculating
Fig. 4, the proposed criterion well fits Mogi’s data parameters for this case are available in Table 4.
146 Hossein Bineshian et al. / J Rock Mech Geotech Eng. 2012, 4 (2): 140–148

Fig. 7 shows good accordance for the proposed

criterion when applied to Hobbs’s data for Linby coal
loaded parallel to the bedding planes and vertical to the
main cleat (Sheorey, 1997). It can be seen from Fig. 7
that the coefficient of determination is almost unity,
which indicates good correlation.

Fig. 6 Fitting the proposed, Coulomb-Mohr and Hoek-Brown

criteria to Brace’s data for Frederick diabase (Sheorey, 1997).

From Figs. 1–6, it can be concluded that the

proposed criterion fits the triaxial test data containing
low to very high confining pressures’ data pairs with a
suitable value of R 2 . The Hoek-Brown criterion does Minor principal stress at failure (MPa)
not present proper fitting to the data groups, which Fig. 7 Accordance of the proposed criterion to Hobbs’s data for
contains data pairs with high and very high confining Linby coal (Sheorey, 1997).
pressures. Bieniawski strength criterion cannot be
fitted to the data groups with tensile data pairs. 5.3 Evaluation from tensile strength aspect
Ramamurthy et al. criterion cannot present any The ability of the proposed criterion to define tensile
strength envelope on data groups that contain tensile strength ( t ) characteristics was evaluated by
and UCS data pairs. Coulomb-Mohr strength criterion reference to the 22 data groups of coal tested by Hobbs
is fitted to all data groups, but its correlation with the (Sheorey, 1997). This ability was compared to those
data pairs is not acceptable for intact rocks, while its applicable to the Hoek-Brown and Coulomb-Mohr
correlation with high and very high confining criteria.
pressures’ data pairs is still better than Hoek-Brown The estimated tensile strengths ( t ) were
criterion. compared with the values determined by Brazilian
5.2 Evaluation using accordance coefficient testing (Sheorey, 1997; Bineshian, 2000). The test data
The compatibility of the proposed criterion with are referred to as values of uniaxial tensile strength
triaxial test data was evaluated in terms of the (UTS). The comparison results are summarized in
accordance coefficient for 12 data groups (containing Table 6. It is indicated that, in 82% and 100% of the
136 pairs of triaxial test data) obtained from different comparisons, the proposed criterion yielded estimates
types of limestone. of  t better than those estimated with the
The results are summarized in Table 5. In 42% of Hoek-Brown and Coulomb-Mohr criteria, respectively.
the cases, the proposed criterion has an accordance The proposed criterion estimated  t approximately
coefficient lower than 0.01, which is indicative of 1.2 times the values determined by UTS testing
satisfactory compatibility. This proposed criterion has whereas the Hoek-Brown and Coulomb-Mohr criteria
better compatibility than those applicable to the other estimated  t 2.2 and 5.1 times the applicable values,
criteria. respectively (Bineshian and Bineshian, 2011b). These
results are confirmed by the accordance coefficients
Table 5 Evaluation of accordance of strength criteria with actual
determined for these criteria in tension, as shown in
triaxial test data for different types of limestone.
Table 7.
Accordance of the criteria with actual data (%)
2 The research also highlighted the fact that the
Proposed Bieniawski Hoek-Brown Ramamurthy Coulomb-Mohr
criterion criterion criterion et al. criterion criterion
Hoek-Brown criterion should not be applied when a
≤0.001 9 0 0 0 0
data group has more than one pair of tensile strength
data (Bineshian and Bineshian, 2011b). The proposed
≤0.005 25 9 0 0 0
criterion can however be applied in such circumstances
≤0.01 42 17 0 17 0
(Bineshian and Bineshian, 2012b). Fig. 8 shows the
≤0.05 75 75 67 50 67
proposed criterion envelope on Johnston’s data for
≤0.15 92 83 83 83 75 Westerly granite at 600 C (Sheorey, 1997).
Hossein Bineshian et al. / J Rock Mech Geotech Eng. 2012, 4 (2): 140–148 147

function of the unconfined compressive strength (  c )

Table 6 Comparison of  t with UTS of coal tested by Hobbs obtained by Coulomb-Mohr criterion. This parameter
(Sheorey, 1997; Bineshian, 2000). adjusts the value of  c to reflect the trend of the
 t (MPa) triaxial test data.
Data UTS
Hoek-Brown Coulomb-Mohr
A rigorous program of comparing estimated to
group (MPa) Proposed criterion
criterion criterion actual test data has suggested that the criterion can
1 0.83 1.061 0.903 2.851 provide linear and nonlinear envelopes capable of
2 0.76 1.086 1.89 5.172 describing most triaxial test data. The proposed
3 3.41 3.689 9.992 14.546 criterion can be applied to compressive and tensile
4 1.79 2.079 2.99 6.456 strength data and provide estimation of these values
5 2.07 2.392 3.535 7.386 that are similar to those determined in laboratory. The
6 2.45 2.69 3.514 7.159 alternative criteria considered in this research may be
7 0.69 0.701 1.099 1.882 less able to do so.
8 2.69 3.565 2.103 4.785
Acknowledgment
9 0.69 1.327 0.639 2.96
10 0.55 1.18 0.383 2.746 We would like to show our special gratitude and
11 0.62 0.664 3.489 5.643 appreciation to Professor Arcady V. Dyskin, Head of
12 0.76 0.793 3.178 7.004 Rock Mechanics Group of the University of Western
13 1.58 1.746 0.958 8.01 Australia, for his worthwhile technical comments in
14 0.89 1.215 1.464 3.781
early 2010 and Dr. Anthony Meyers, ISRM’s Vice
15 1.1 1.713 2.96 7.281
President for Australasia, for his very helpful technical
16 1.1 1.177 2.603 5.937
editing of the paper in 2011. We would also like to
17 1.1 1.544 3.076 6.941
thank Tarbiat Modares University for supporting the
18 0.69 0.899 1.332 2.989
19 0.55 0.951 1.345 3.897
initial part of this research, Mahab Consulting
20 0.76 0.875 3.29 5.87 Engineers for making the results of the triaxial tests
21 1.1 1.644 2.112 5.317 available to us, and Perlite Construction Company for
providing the cost of soil and concrete testing.
Table 7 Calculated accordance coefficients for strength criteria
with pairs of tensile data.
References
Strength criterion Accordance coefficient, 2

Proposed 0.26 Balmer G. A general analytical solution for Mohr envelope. American
Society of Testing Materials, 1952, 52: 1 260–1 271.
Hoek-Brown 6.536
Bieniawski Z T. Estimating the strength of rock materials. Journal of South
Coulomb-Mohr 40.638 African Institute of Mining and Metallurgy, 1974, 74 (8): 312–320.
Bineshian H, Bineshian Z. Prediction of geomaterials’ compressive strength
on the basis of defining new values for Bieniawski failure criterion’s
parameters. In: Proceedings of International Conference on Advances in
Geotechnical Engineering. Perth: [s.n.], 2011a: 227–234.
Bineshian H, Bineshian Z. Modification to the Hoek-Brown strength
criterion for intact geomaterials based on new values for the parameter.
In: Proceedings of the 13th International Conference of the International
Association for Computer Methods and Advances in Geomechanics.
Melbourne: [s.n.], 2011b: 365–370.
Bineshian H, Bineshian Z. Prediction of triaxial compressive strength of
geomaterials based on a failure criterion. In: Proceedings of the 31st
International Conference on Ocean, Offshore and Arctic Engineering.
Fig. 8 Compatibility of the proposed criterion to Johnston’s data
Rio De Janeiro: [s.n.], 2012a (in press).
for Westerly granite (Sheorey, 1997).
Bineshian H, Bineshian Z. Applicability of a new strength criterion in
comparison to failure criteria. In: Proceedings of Ground Engineering in
6 Conclusions a Changing World, ANZ 2012. Melbourne: [s.n.], 2012b (in press).
Bineshian H. Geomechanical properties of Towhid project. 1st ed. Tehran:
The proposed strength criterion is defined in terms Perlite Co., 2006.
of three parameters (,  and ). Parameter  is a Bineshian H. The results of triaxial compressive tests on injected materials of
148
Towhid project soil. Tehran: Perlite Co., 2008.
Bineshian H. Failure mechanics of rock. Tehran: Engineering Department,
Tarbiat Modares University, 1999.
Bineshian H. Verification of the applicability of rock failure criteria in
relation to the limestone of Iran and suggestion of a suitable failure
criterion. M.S. Thesis. Tehran: Tarbiat Modares University, 2000.
Brace W F. An extension of the Griffith theory of fracture to rocks. Journal of
Geophysical Research, 1960, 65 (10): 3 477–3 480.
Brown E T. Rock characterization testing and monitoring: ISRM suggested
methods. [S.l.]: Pergamon Press, 1981.
Das M N, Sheorey P R. Triaxial strength behavior of some Indian coals.
Journal of Mines, Metals and Fuels, 1986, 34 (3): 118–122.
Donath F A. Rock deformation apparatus and experiments for dynamic
structural geology. Journal of Geological Education, 1970, 18: 3–12.
Fairhurst C. On the validity of the Brazilian test for brittle materials.
International Journal of Rock Mechanics and Mining Sciences and
Geomechanics Abstracts, 1964, 1 (4): 535–546.
Franklin J A, Hoek E. Development in triaxial testing technique. Rock
Mechanics, 1970, 2 (4): 223–228.
Franklin J A. Triaxial strength of rock materials. International Journal of
Rock Mechanics and Mining Sciences and Geomechanics Abstracts,
1971, 3 (2): 86–98.
Griffith A A. The phenomena of rupture and flow in solids. Philosophical
Transactions of the Royal Society of London, 1921, 221: 163–198.
Griffith A A. Theory of rupture. In: Proceedings of the 1st International
Congress on Applied Mechanics. Delft: [s.n.], 1924: 55–63.
Griggs D T. Deformation of rocks under high confining pressure.
I—experiments at room temperature. Journal of Geology, 1936, 44 (5):
541–577.
Hobbs D W. The strength and the stress-strain characteristics of coal in
triaxial compression. Journal of Geology, 1964, 72 (2): 214–231.
Hoek E. Rock mechanics in engineering practice. London: Wiley, 1968.
Hoek E, Brown E T. Underground excavations in rock. London: Institution of
Mining and Metallurgy, 1980a.
Hoek E, Brown E T. Empirical strength criterion for rock masses. Journal of
Geotechnical Engineering Division, ASCE, 1980b, 106 (9): 1 013–1 035.
Hoek E, Brown E T. The Hoek-Brown failure criterion—a 1988 update. In:
Proceedings of the 15th Canadian Rock Mechanics Symposium.
Toronto: [s.n.], 1988: 31–38.
Horino F G, Ellikson M L. A method for estimating the strength of rock
containing planes of weakness. Washington, D.C.: Bureau of Mines, US
Deptartment of Interior, 1970.
Hoshino K, Koide H, Inami K, Iwamura S, Mitsui S. Mechanical properties
of Japanese Tertiary sedimentary rocks under high confining pressure.
Geological Survey of Japan Report, 1972, 244: 1–200.
Hossaini S M F. Essential hints in selection of laboratorial data to evaluate
failure criteria. Fanni Department Journal, 1999, 32 (1): 89–94.
Hossaini S M F. Some aspects of the strength characteristics of intact and
jointed rocks. Ph.D. Thesis. Sydney: The University of New South
Wales, 1993.
Jaeger J C, Cook N G W, Zimmerman R W. Fundamentals of rock mechanics.
Victoria: Blackwell Publishing Ltd., 2007.
Johnston I W, Chiu H K. Strength of weathered Melbourne mudstone.
Journal of Geotechnical Engineering Division, ASCE, 1984, 110 (G57):
875–898.
Johnston I W. Strength of intact geomechanical materials. Journal of
Geotechnical Engineering Division, ASCE, 1985, 111 (6): 730–749.
Mahab Ghodss Consulting Engineering Co. Rock mechanics report for
Karun III Dam—2. Izeh: Dam Deptartment, 1999.
Mahab Ghodss Consulting Engineering Co. Rock mechanics report for
Hossein Bineshian et al. / J Rock Mech Geotech Eng. 2012, 4 (2): 140–148
Seimareh Dam—1. Seimareh: Dam Deptartment, 2003.
McClintock F A, Walsh J. Friction on Griffith cracks in rocks under pressure.
In: Proceedings of the 4th US National Congress on Applied Mechanics.
Berkeley: [s.n.], 1962: 1 015–1 021.
Mogi K. Deformation and fracture of rocks under confining pressure.
II—elasticity and plasticity of some rocks. Bull. Earthquake Res. Inst.,
1965, 43: 349–379.
Mogi K. Some precise measurements of fracture strength of rocks under
uniform compressive stress. Rock Mechanics and Engineering Geology,
1966a, 4 (1): 43–55.
Mogi K. Pressure dependence of rock strength and transition from brittle
fracture to ductile flow. Bull. Earthquake Res. Inst., 1966b, 44:
215–232.
Mogi K. Experimental rock mechanics. Leiden: Taylor and Francis Group,
2006.
Murrell S A F. A criterion for brittle fracture of rocks and concrete under
triaxial stress and the effect of pore pressure on the criterion. Oxford:
Pergamon Press, 1963.
Murrell S A F. The effect of triaxial stress systems on the strength of rocks at
atmospheric temperatures. Geophysical Journal, 1965, 10 (3): 231–281.
Neter J, Wasserman W. Applied linear statistical models. Boston: Allyn and
Bacon Inc., 1977.
Neter J, Wasserman W, Whitmore G A. Applied statistics. Boston: Allyn and
Bacon Inc., 1988.
Ohnaka M. The quantitative effect of hydrostatic confining pressure on the
compressive strength of crystalline rocks. Journal of the Physics of the
Earth, 1973, 21 (2): 125–140.
Ouyang Z, Elsworth D. A phenomenological failure criterion for brittle rock.
Rock Mechanics and Rock Engineering, 1991, 24 (3): 133–153.
Ramamurthy T, Rao G V, Rao K S. A linearnonlinear strength criterion for
rocks. In: Proceedings of the 5th Australia-New Zealand Conference on
Geomechanics. Sydney, Australia: Institution of Engineering, 1988:
247–252.
Ramamurthy T, Rao G V, Rao K S. A strength criterion for rocks. In:
Proceedings of Indian Geotechnical Conference. Roorkee: [s.n.], 1985:
59–64.
Rao K S. Strength and deformation behavior of sandstone. Ph.D. Thesis.
Delhi: Indian Institute of Technology, 1984.
Schwartz A E. Failure of rock in the triaxial shear test. In: Proceedings of the
6th US Symposium on Rock Mechanics. Rolla: University of Missouri,
1964: 64–109.
Sheorey P R, Biswas A K, Choubey V D. An empirical failure criterion for
rocks and jointed rock masses. Engineering Geology, 1989, 26 (2):
141–159.
Sheorey P R, Das M N, Bordia S K, Singh B. Pillar strength approaches
based on a new failure criterion for coal seams. International Journal of
Mining and Geological Engineering, 1986, 4 (4): 273–290.
Sheorey P R. Empirical rock failure criteria. Rotterdam: A.A. Balkema, 1997.
von Karman T. Festigkeitsversuche unter allseitigem druck. Zeitschrift
Verein Deutscher Ingenieure, 1911, 55 (42): 1 749–1 757 (in German).
Vutukuri V S, Hossaini S M F. Assessment of applicability of strength
criteria for rock and rock mass to coal pillars. In: Proceedings of the
11th International Conference on Ground Control in Mining. New South
Wales, Australia: University of Wollongong, 1992: 1–8.
Yoshida N, Morgenstern N R, Chau D H. A failure criterion for stiff soils and
rocks exhibiting softening. Canadian Geotechnical Journal, 1990, 27 (2):
195–202.
Yudhbir L W, Prinzl F. An empirical failure criterion for rock masses. In:
Proceedings of the 5th ISRM International Congress on Rock
Mechanics. Melbourne: A. A. Balkema, 1983: B1–B9.