Modeling Earth Systems and Environment

https://doi.org/10.1007/s40808-022-01499-z

ORIGINAL ARTICLE

An exact solution of three‑dimensional rock mass strength criterion

Vinay Kumar1 · Brijbhan Rao2 · Avijit Burman2 · Sudeep Kumar2 · Abidhan Bardhan2

Received: 2 July 2022 / Accepted: 13 August 2022

© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2022

Abstract
The present study proposed an exact solution of a three-dimensional strength criterion for rock material. The Hoek–
Brown (HB) yield criterion, in its original form, did not consider the effect of intermediate principal stress on the failure
response of rocks. In 2005, Priest combined the original Hoek–Brown strength equation with general three-dimensional
Drucker–Prager criterion to incorporate the effect of intermediate principal stress on the failure behavior(of rocks.
) The cri-
terion was formulated by considering effective minor principal stress 𝜎3′ and effective normal stress 𝜎zf′ as unknown
( )

parameters. To solve the three-dimensional yield criterion involving these unknowns, the error term 𝜀c has been minimized.
Detailed analytical expressions are derived and a MAPLE-18-based
( ) code has been developed for this purpose. It is observed
that the solution consists of three pairs of real roots 𝜎3′ , 𝜎zf′ among which two represent local minimum values and one pair
represents a saddle point. The solution is general in all sense and can cater all possible ranges of values of HB material
parameters such as mi , mb , s, GSI , and 𝜎ci values. In this study, a rigorous analytical solution is developed to minimize 𝜀c and
the three-dimensional yield criterion of Priest has been solved. The solution is coded in MAPLE-18. Three examples from
existing literatures have been solved to validate the proposed technique. It is observed that the solution of the error term
yields three real roots which are also possible solutions of the problem being investigated. The nature of the roots has been
further investigated and finally one of them is accepted as the practical solution.

Keywords 3D yield criterion for rocks · Hoek–Brown criterion · Drucker–Prager criterion

Introduction Hoek 1983; Hoek and Brown 1997; Marinos and Hoek
2001) applied failure criteria successfully for wide range of
The Hoek–Brown (HB) yield criterion is very popular for intact and fractured rock types. The HB rock mass yield cri-
estimation of rock strength of wide varieties (Hoek and terion has been expressed in terms of few parameters which
Brown 1980). Various researchers (Hoek et al. 2002a, b; are mi , mb , s, a, geological
( ) strength index (GSI) and uniaxial
compressive strength 𝜎ci . Here, mi is a parameter signifying
the initial state of intact rock mass; mb , s, and a are the HB
* Avijit Burman material parameters that depend on the proportion, struc-
avijit@nitp.ac.in ture and surface of the rock masses. Parameters mb , s , and
Vinay Kumar a are expressed as functions of GSI which represented the
vinaykumarnitp@gmail.com degree of interlocking of the rock mass. Recently, the modi-
Brijbhan Rao fied version of nonlinear HB criterion known as Generalized
brijbhanr.phd19.ce@nitp.ac.in Hoek–Brown (GHB) failure criterion had been presented by
Sudeep Kumar Hoek et al. (2002a, b), in which a disturbance factor D was
Sudeep.2k7@gmail.com introduced to define the degree of stress relaxation and exist-
Abidhan Bardhan ing damage. The expressions of HB material parameters,
abidhan@nitp.ac.in i.e., mi , mb , s , and a have been modified to incorporate the
1 effect of disturbance factor D . Generally, the nature of fail-
Civil Engineering Department, Government Engineering
College, Nawada, Bihar, India ure envelopes of all geomaterials are almost nonlinear (Agar
2 et al. 1987; Anyaegbunam 2015; Serrano and Olalla 2006).
Department of Civil Engineering, National Institute
of Technology, Patna, India The GHB failure criterion has been applied successfully

13
Vol.:(0123456789)
 Modeling Earth Systems and Environment

for isotropic rock mass, intact and jointed rock mass (Hoek on the intermediate effective principal stress. Researchers
1990; Serrano and Olalla 1994; Shen et al. 2012; Sofianos applied similar concept on 2D HB failure criteria to study
2003; Sofianos and Halakatevakis 2002; Zhang and Zhu the influence of 𝜎2 ′ on application of strength of rock mate-
2007) to capture the nonlinear stress–strain response of the rials to develop 3D failure criteria (Pan and Hudson 1988;
rock mass under external loading. Hoek and Brown (1997) Priest 2005; Zhang and Zhu 2007). Other researchers (Priest
expressed failure criterion in which parameter s was equal 2005; XD and JA 1988) considered circular cross sections
to zero when GSI < 25. However, in the failure criterion of in the deviatoric stress plane for 3D failure criterion. Priest
Hoek et al. (2002a, b), the value of parameter s was modi- (2012) compared five 3D failure criteria and recommended
fied as a very low positive quantity at small values of GSI . that further research and rock testing was required before
The generalized Hoek–Brown (GHB) criterion has gen- any of the 3D HB failure criterion can be applied with
erated a great deal of scholarly interest in recent years. For confidence.
instance, Deng (2020) used GHB to examine the stability The generalization proposed by Priest (2005) was based
of rock slopes while taking into account the interaction on the incorporation of DP yield function with HB strength
between strain softening and shear dilatancy. In addition, Liu criterion, which indeed considered the effect of the 𝜎2 ′, but
et al. (2020) explored the stability of slopes by employing ignored the influence of the Lode angle 𝜃 on the deviatoric
GHB instead of Mohr–Coulomb and the finite element limit stress at yield (being a function of deviatoric stress invari-
equilibrium approach. Similar to this, Wei et al. (2020) pre- ants J1′ and J2′ only). In the DP model, the lack of the third
dicted the ideal set of conditions causing rock failure using invariant (i.e., Lode angle θ) in the yield surface formulation
the strength reduction approach and GHB. Karrech et al. is responsible for the well-known limitation of having quite
(2021) employed the generalized Hoek–Brown Criterion to different friction angles in axis-symmetric compression and
do a limit analysis on a three-dimensional rock slope. extension conditions. If the DP constants are calibrated by
A major shortcoming of GHB failure criterion (Hoek matching the deviator stress at yield in compression, the
et al. 2002a, b) was that it was unable to consider the effect resulting friction angle in extension will be largely over-
of the intermediate principal stress 𝜎2 ′. Earlier experiments estimated. The issue of extending the original HB model
suggested that 𝜎2 ′ had a superficial effect on the strength to a full three dimensions involving stress invariants was
of rock mass (Al-Ajmi and Zimmerman 2005; Chang and elegantly presented by Benz et al. (2008) which did not suf-
Haimson 2000; Colmenares and Zoback 2002). Originally fer from the problem of over predicting the friction angle.
developed HB failure criterion ignored the effect of the The HB criterion is able to simulate the nonlinear nature of
𝜎2 ′ for estimating the strength of the rock materials. Priest stress characteristics of rock. An intrinsic material strength
(2005) used general 3D Drucker–Prager (DP) yield criterion factorization scheme was first proposed by Matsouka
to consider the effect of 𝜎2 ′ along with originally developed and Nakai (1982) to formulate an enhanced HB criterion
HB strength criterion. based on spatial mobilized plane (SMP). The modified
Haimson (2006) studied the behavior of rocks under the Hoek–Brown–Matsouka–Nakai criterion, as proposed by
action of intermediate principal stress for past many years Benz et al.(2008), also ensured that the apparent friction
and concluded that it was important to consider the effect angles in compression and extension remain same.
of 𝜎2 ′ during analysis of failure of any rock material. The Priest (2005) extended 2D GHB yield criterion to three
authors also stated that the yield criteria which did not con- dimensions by combining it with DP yield criterion. The HB
sider the effect of 𝜎2 ′ were unable to properly represent the yield criterion has corners in pi-plane (i.e., plane of princi-
failure behavior of rocks. A new criterion was proposed by pal stresses). These corners occur where either major and
Jiang et al. (2011) who considered poly-axial compression intermediate principal stress or minor and intermediate prin-
of rocks. Jiang and Zhao (2015) proposed a new 3D yield cipal stresses have the same value.( In case of tri-axial tests
criterion with convex yield surface for rocks which inher- on rocks, the later condition, i.e., 𝜎z ≥ 𝜎x = 𝜎y is usually
)

ited all the features of HB criterion while simulating stress found to prevail. Therefore, the instantaneous cohesion (c′)
behavior of rocks under tri-axial compression. The crite- and instantaneous angle of internal friction (𝜙′) to be esti-
rion also incorporated the effect of 𝜎2 ′ . The criterion was mated for HB criterion would be same as those calculated
calibrated using tri-axial compression test data on rocks. by the DP criterion (Drucker and Prager 1952). Numerous
Applications of these criteria to 2D and 3D problems of data of tri-axial tests carried out on rocks suggested that the
rock mechanics have been earlier reported (Jiang et al.2011; yield surface should be nonlinear (as described by HB yield
Priest 2012, 2005). criterion) and stress/failure points should also lie inside a
It was required to calibrate the developed 3D strength cri- tapering cone. These facts inspired the development of 3D
teria with the laboratory test results. Many experiments were strength criterion for rocks by Priest (2005).
performed related to 3D problem, in which the predicted Priest (2005) defined an error term and suggested an
major effective principal stress (𝜎1′ ) at failure was dependent iterative trial-and-error based numerical strategy to

13
Modeling Earth Systems and Environment

minimize it for solving the 3D yield criterion for rocks and input parameters that depend on the proportion, structure
applied the technique to two example problems to illustrate and surface of the rock masses (Hoek et al.1995). The HB
how the solution of the 3D yield criterion may be obtained. input parameters, i.e., mb , s , and a are functions of GSI of
Priest (2005) was unable to develop an analytical solution the rock mass. GSI describes the degree of interlocking of
for the 3D yield criterion for rock mass till date and encour- the blocks in the rock mass. The relations between these
aged other researchers to take up the problem of finding new parameters are as follows:
solutions. Later, Melkoumian et al. (2009) reported an ana-
GSI − 100
( )
lytical solution of the 3D rock mass yield criterion originally mb = mi 𝑒𝑥𝑝 (2)
28 − 14D
developed by Priest (2005) and recommended its use for the
purpose of stress modeling. In the present work, a rigorous
GSI − 100
( )
analytical approach to minimize the error term defined by s = 𝑒𝑥𝑝 (3)
Priest (2005) is presented to solve the 3D strength criterion 9 − 3D
of rocks. Detailed analytical expressions are derived and a ( ) ( )
MAPLE-18-based code was developed for this purpose. It − GSI − 20
e −e (4)
15 3

is observed
( that
) the solution consists of three pairs of real
a = 0.5 +
6
roots 𝜎3 , 𝜎zf among which two represent local minimum
′ ′
where mi is the HB constant for intact rock. Hoek et al.
values and one pair represents a saddle point. The solution
(2002a, b) introduced a disturbance factor D to account
is general in all sense and can cater all possible ranges of
for the disturbance caused by stress relaxation and dam-
values of HB material parameters such as mi , mb , s, GSI , and
age (Brady and Brown 2004; Hoek 1988). Generally, it is
𝜎ci values. The developed method has been validated by
assumed that there is no influence of the 𝜎2 ′ on the yield
comparing its results with already published results. Two
strength and the minor principal stress acts as the confin-
problems earlier solved by Priest (2005) and one problem
ing stress in both the x and y directions. However, Paterson
from the work of Melkoumian et al. (2009) have been
(1978) stated that a reasonable assumption should be made
selected for verification of the proposed solution in this
about the 𝜎2 ′ in 3D which exerts some influence on the yield
study. The developed solution was found to be robust from
strength. Figure 1 shows that Mohr–Coulomb as well as HB
computational point of view. The outline of the developed
criteria in two dimensions, plotted as irregular hexagons
MAPLE-18 code is also presented in this study.
in the pi-plane designated as curves B and A, respectively
(Fig. 1). Priest (2005) stated that the corners of the hexa-
gons represent either of the pair of principal stresses, i.e.,
Methodology the minor and intermediate, or the major and intermediate
principal stress. At the ‘sharp’ corners, the minor and inter-
The GHB criterion is developed based on the results of
mediate principal stresses have the same value. The ‘blunt’
numerous tri-axial tests carried on rocks. It is expressed as
corners are those where the major and intermediate principal
a function of the major and minor principal stresses at the
stresses have the same value. The condition 𝜎z ≥ 𝜎x = 𝜎y is
time of failure and 𝜎ci of rocks. In many situations, it is found
maintained at the ‘sharp’ corners while performing tri-axial
that the influence of 𝜎2 ′ is substantially larger than the minor
intermediate principal stress (𝜎3 ′) which is usually ignored
while assessing the strength of the rock mass. The condition
of intermediate stress can develop in the excavated walls
of drilled boreholes during petroleum or gas exploration.
Rapid development of plastic strain results into yielding and
failure of the rock material adjacent to the boreholes leading
to boreholes instability problems. Near any open surface
of rock mass, “yielding” leads to rock mass disintegration,
deterioration and cracking resulting into ultimate failure.
The expression for GHB criterion is as follows:
}a
mb 𝜎3�
{( )
� �
𝜎1 = 𝜎3 + 𝜎ci +s (1)
𝜎ci

where 𝜎1′ and 𝜎3′ are maximum and minimum effective prin-
cipal stresses at failure and 𝜎ci is the uniaxial compressive Fig. 1  Pi-plane plot illustrating the (A) DP, (B) Coulomb, (C) Com-
strength of the rock material. Here, mb,s , and a are the HB prehensive 3D HB, and (D) 2D HB yield criteria

13
 Modeling Earth Systems and Environment

test on rock specimens. To compute the external and cir- are modified as Ae , Be for the external or ‘circumscribed’
cumscribed DP model, Priest (2005) applied the idea of model and Ai , Bi for the internal or ‘inscribed’ models (refer
considering Coulomb cohesion ( c′ ) and angle of internal to Fig. 1).
friction (𝜙′) for the rock material and evaluated them from
the external or circumscribed DP model, i.e., the one that √
touched the three sharp corners of Coulomb irregular hexa- 2 3𝑠𝑖𝑛𝜙� c� Ae
Ae = and B e = (11)
gon’s surface. This way of linking Coulomb and DP model 3 − 𝑠𝑖𝑛𝜙� 𝑡𝑎𝑛𝜙�
meant that the circular yield surfaces on the pi-plane would
fall on the surface of the straight-sided cone. However, the
√
3𝑠𝑖𝑛𝜙� c� Ai
results of tri-axial tests on rocks also suggested that the yield Ai = � and Bi = (12)
� 𝑡𝑎𝑛𝜙�
surface should be nonlinear and therefore, the DP yield sur- 3 + 𝑠𝑖𝑛2 𝜙
face should resemble the shape of a tapering cone. Priest
(2005) proposed 3D strength criterion for rock mass which The Mohr–Coulomb shears strength parameters c′ and
′ aredetermined using GHB criterion under specified nor-
combined nonlinear HB yield criteria with DP criterion 𝜙
(refer to curve D in Fig. 1) and incorporated the effect of mal stress. In 3D, a clearly defined plane of shear failure
intermediate principal stress. Brief description of 3D rock cannot be specified for any normal stress. In these circum-
mass yield criterion proposed Priest (2005) is given below. stances, HB and DP criteria are combined to predict equal
A cubic element of rock mass subjected to any loading strength at the apex of the HB criterion based on uniaxial
conditions and general 3D stresses in Cartesian coordinate compressive strength of rock mass. At the point of yield-
𝜎ci2
system are expressed as: ing, we have J1� =
𝜎ci
and J2 = . After replacing J1′ and
3 3
J2 into Eq. (10) and assuming 𝜎cl′ =
0 when yielding takes
� xyz � ⎡ 𝜎xx 𝜎xy 𝜎xz ⎤ place, Priest (2005) obtained following expressions for the
𝜎 = ⎢ 𝜎yx 𝜎yy 𝜎yz ⎥ (5)
⎢ ⎥ parameters A and B:
⎣ 𝜎zx 𝜎zy 𝜎zz ⎦
�√ �
The mean stress J1 are as follows: 3−A
√ √
3 J2 − 𝜎ci 3 𝜎ci
A= and B = (13)
𝜎xx + 𝜎yy + 𝜎zz 𝜎11 + 𝜎22 + 𝜎33 3J1� − 𝜎ci 3
𝜎m = J1 = = (6)
3 3 Priest (2005) defined the effective yield stress 𝜎zf′ for an
where 𝜎11 , 𝜎22 , and 𝜎33 are the three total principal stresses. element of material subjected
( to a general)state of effective
The effective mean stress is obtained by subtracting the pore tri-axial principal stresses 𝜎x′ ≠ 𝜎y′ ≠ 𝜎zf′ to make the DP
water pressure u from the total stress values as follows: parameters to circumscribe the Coulomb yield model. The
nonlinear representation of HB yield criterion was conceived
J1� = J1 − u (7)
by Priest (2005). The effective mean stress J1′ was expressed
The mean shear stress J2 is the independent of the pore using Eqs. (14) and (15) for HB and DP materials,
water pressure which is expressed by: respectively.

(𝜎xx − 𝜎yy )2 + (𝜎yy − 𝜎zz )2 + (𝜎zz − 𝜎xx )2

J2 = + 𝜎 2 xy + 𝜎 2 yz + 𝜎 2 xz 𝜎1� + 2𝜎3�
6 �
J1HB = (14)
(8) 3
Therefore, the above expression can be re-written as:
𝜎x� + 𝜎y� + 𝜎zf�
2 2
(𝜎11 − 𝜎22 ) + (𝜎22 − 𝜎33 ) + (𝜎33 − 𝜎11 ) 2 J1� DP = (15)
J2 = (9) 3
6
The yield criteria expressed by Eqs. (14) and (15) should
The effective collapse stress 𝜎cl′ as proposed by DP
( � compatible
be which is possible if the condition
(Drucker and Prager, 1952) is as follows: J1HB = J1� DP = J1� is satisfied. For the general condition of
)
√ effective normal stresses, where 𝜎x′ ≠ 𝜎y′ ≠ 𝜎zf′ , J2 for DP
𝜎cl� = AJ1� + B − J2 (10)
yield is expressed as follows:
where A and B are DP parameters dependent on material ( )2 ( )2 ( )2
properties. The yield occurs if 𝜎cl′ ≤ 0. The parameters A and 𝜎x� − 𝜎y� + 𝜎y� − 𝜎zf� + 𝜎zf� − 𝜎x�
(16)
B are estimated using instantaneous cohesion, c′ and instan- J2 =
6
taneous angle of the internal friction 𝜙′. These parameters

13
Modeling Earth Systems and Environment

Equations (11) and (12) are known as DP parameter Melkoumian et al. (2009). The analytical solution was vali-
equation.
( Applying
)2 the DP yield criterion at yield, we get: dated with the numerical solution suggested by Priest
J2 = AJ1� + B . The terms RHB and RDP are used to express (2005). The error 𝜀c defined in Eq. (19) can be expressed as
the distance from the central mean stress point to the HB and a function of 𝜎3′ and 𝜎zf′ . After making appropriate substitu-
DP yield surfaces in the pi-plane. These terms are expressed tions of the expressions of 𝜎1′, J1HB
′
, J1′ DP, RHB, and RDP from
in the following way: Eqs. (1), (14), (15), (17), and (18), respectively, a modified
expression of the error term 𝜀c can be obtained as follows:
[ � )a 𝜎 � + 𝜎 � + 𝜎 � ]2
m
(
𝜎 ci b 𝜎 3 x y zf
𝜀c = 𝜎3� + +s −
3 𝜎ci 3
(20)
)a )2 2
[√ √ ]
mb 𝜎3�
( )2 ( )2 (
2 1
(
� � �
+ 𝜎 +s − 2𝜎x − 𝜎y − 𝜎zf + 2𝜎y − 𝜎zf − 𝜎x + 2𝜎zf − 𝜎x − 𝜎y
� � � � � � .
3 ci 𝜎ci 3

√ ( )
(17) The expression 𝜀c 𝜎3′ , 𝜎zf′ , as shown in Eq. (20), can be
( � )2 ( )2
RHB = 𝜎1 − J1� + 2 𝜎3� − J1�
investigated for its stationary points by setting and 𝜕𝜎 ′c
𝜕𝜀c 𝜕𝜀
𝜕𝜎3′ zf

individually to zero and solving the resulting equations.

√
)2 ( )2 ( )2
(18)
(
RDP 𝜎x� − J1� + 𝜎y� − J1� + 𝜎zf� − J1�
(However, ) to ascertain the true nature of the2 stationary points
=
𝜕 𝜀 𝜕2 𝜀 𝜕2 𝜀
𝜎3′ , 𝜎zf′ , the second derivative terms 𝜕𝜎 ′ 2c ,𝜕𝜎 ′ 2c and 𝜕𝜎 ′ 𝜕𝜎c ′
For HB and DP yield surfaces to predict same values, the 3 zf 3 zf

condition RHB = RDP = R must be observed. Equa- should be checked. The following
( )conditions will govern the
tions (13)–(18) have seven independent unknowns. To cal- nature of the function 𝜀c 𝜎3′ , 𝜎zf′ at the stationary points
culate these unknown, Priest (2005) proposed a spreadsheet-
( )
𝜎3′ , 𝜎zf′ . These conditions are:
based iterative strategy. The input data 𝜎ci , mb , s, a, 𝜎x′ , and
𝜎y′ should be specified first. It is required to make initial
( )
a) The solution of 𝜎3′ , 𝜎zf′ is a saddle point if following
guess for 𝜎3′ and 𝜎zf′ for which an error term defined in
condition holds:
Eq. (19) is minimized. The error term defined by Priest
(2005) as follows: ( )2
𝜕 2 𝜀c 𝜕 2 𝜀c 𝜕 2 𝜀c
− < 0. (21)
( �
𝜀c = J1HB
)2 ( )2
− J1� DP + RHB − RDP . (19) 𝜕𝜎 � 23 𝜕𝜎 � zf2 𝜕𝜎3� 𝜕𝜎zf�
( )
To minimize the value of error term 𝜀c , excel solver tech- b) The obtained solution of 𝜎3′ , 𝜎zf′ can be categorized as
nique was adopted by Priest (2005) to determine the value
either maximum or minimum )based on following condi-
of 𝜎3′ and 𝜎zf′ . (
tions. The solution 𝜎3 , 𝜎zf is a maximum point if
′ ′

Minimization of the error term 𝜕 2 𝜀c

< 0 and
𝜕 2 𝜀c
< 0 ; and it has a minimum point if
𝜕𝜎 ′ 32 𝜕𝜎 ′ zf2
𝜕2 𝜀 𝜕2 𝜀
Priest (2005) used spreadsheet-based approach to determine c
> 0 and 𝜕𝜎 ′ 2c > 0.
𝜕𝜎 ′ 32
the value of minimum effective stress 𝜎3′ and effective yield zf

stress 𝜎zf′ and other independent unknown using 3D stress The derivative terms of 𝜀c are
( calculated from Eq. (20).
with the combination of HB criterion and DP yield criterion.
)
To obtain the stationary points 𝜎3 , 𝜎zf , 𝜕𝜎 ′ and 𝜕𝜎 ′c are set
′ ′ 𝜕𝜀c 𝜕𝜀
He adopted a trial-and-error-based numerical approach to 3 zf

minimize the error term 𝜀c defined in Eq. (19). Later, an equal to zero. Thus, Eqs. (22) and (23) are formed by setting
explicit analytical solution of Eq. (19) was also presented by the derivatives equal to zero as follows:

13
 Modeling Earth Systems and Environment

[ � )a 𝜎 � + 𝜎 � + 𝜎 � ]
m
(
𝜕𝜀c 𝜎ci b 𝜎 3 x y zf
= 2 𝜎3� + +s −
𝜕𝜎3� 3 𝜎ci 3
[ )a−1 ] )a−1
amb mb 𝜎3�
√
mb 𝜎3�
( (
2 (22)
× 1+ +s +2 amb +s
3 𝜎ci 3 𝜎ci
[√ )a √ ]
mb 𝜎3�
( )2 ( )2 ( )2
2 1
(
� � �
𝜎 +s − 2𝜎x − 𝜎y − 𝜎zf + 2𝜎y − 𝜎zf − 𝜎x + 2𝜎zf − 𝜎x − 𝜎y
� � � � � � =0
3 ci 𝜎ci 3

[ � )a 𝜎 � + 𝜎 � + 𝜎 � ]
2 � 𝜎ci mb 𝜎3
(
𝜕𝜀c x y zf
=− 𝜎 + +s −
𝜕𝜎zf� 3 3 3 𝜎ci 3
[√ )a √ ]
mb 𝜎3�
( )2 ( )2 ( )2
2 1
(
� � �
−2 𝜎 +s − 2𝜎x − 𝜎y − 𝜎zf + 2𝜎y − 𝜎zf − 𝜎x + 2𝜎zf − 𝜎x − 𝜎y
� � � � � �
3 ci 𝜎ci 3
( )
2𝜎zf� − 𝜎x� − 𝜎y�
×√
( ) ( 2 ) ( 2) 2
2𝜎x� − 𝜎y� − 𝜎zf� + 2𝜎y� − 𝜎zf� − 𝜎x� + 2𝜎zf� − 𝜎x� − 𝜎y�
(23)
=0
( )
To find out the stationary points 𝜎3′ , 𝜎zf′ , it is necessary Therefore, considering the(error term ) 𝜀c to be a continu-
to solve Eqs. (22) and (23) together. The expressions of sec- ous function of its arguments 𝜎3′ , 𝜎zf′ , we obtain:
ond derivatives of the error term 𝜀c defined in Eq. (20)
should be checked to find out the conditions of maxima or
minima. The expressions are presented below:

]2
amb ma−1 𝜎ci maf
[ [ ]
𝜕 2 𝜀c f
𝜎x� + 𝜎y� + 𝜎zf�
�
=2 1 + + 2 𝜎3 + −
𝜕𝜎 �2 3 3 3 3
a(a − 1)m2b ma−2 2 a(a − 1)mb mf
2 a−2
√
f
× +2
3𝜎ci 3 3𝜎ci
[√ √ ]
)2 ( )2 ( )2
2 1
(
× 𝜎 ma − 2𝜎x� − 𝜎y� − 𝜎zf� + 2𝜎y� − 𝜎zf� − 𝜎x� + 2𝜎zf� − 𝜎x� − 𝜎y�
3 ci f 3
4
+ a2 m2b m2a−2
f
3
(24)
� �2
2 2 2𝜎zf� − 𝜎x� − 𝜎y�
𝜕 𝜀c 2
= +� �2 � �2 � �2
𝜕𝜎 �2 zf 9
2𝜎x� − 𝜎y� − 𝜎zf� + 2𝜎y� − 𝜎zf� − 𝜎x� + 2𝜎zf� − 𝜎x� − 𝜎y�
�� � �
�2 � �2 � �2
2 1
�
a � � �
−2 𝜎 m − 2𝜎x − 𝜎y − 𝜎zf + 2𝜎y − 𝜎zf − 𝜎x + 2𝜎zf − 𝜎x − 𝜎y
� � � � � �
3 ci f 3
⎡ � � ⎤
⎢ 2 2𝜎 � − 𝜎 � − 𝜎 � 2 + 2 2𝜎 � − 𝜎 � − 𝜎 � 2 − 2𝜎 � − 𝜎 � − 𝜎 � 2 ⎥
� � � �
⎢ x y zf y zf x zf x y ⎥
× ⎢� �3 ⎥
⎢ � � 2 � � 2 � � 2 2 ⎥
2𝜎x� − 𝜎y� − 𝜎zf� + 2𝜎y� − 𝜎zf� − 𝜎x� + 2𝜎zf� − 𝜎x� − 𝜎y�
(25)
⎢ ⎥
⎣ ⎦

13
Modeling Earth Systems and Environment

Table 1  Steps for solving Eqs. (22) and (23)

Step 1: Take input of the parameters mi , 𝜎x′ , 𝜎y′ , GSI, s, mb , and,a

Step 2: Calculate 𝜎1′ from Eq. (1)
Step 3: Calculate J1HB
′
from Eq. (14)
Step 4: Calculate J1DP
′
from Eq. (15)
√( )2 ( )2
RHB = 𝜎1� − J1HB
�
+ 2 𝜎3� − J1HB�
Step 5: Calculate
√ )2 ( )2
( �
)2 ( �
RDP = 𝜎x� − J1DP + 𝜎y� − J1DP + 𝜎zf� − J1DP
�
Step 6: Calculate
)2 ( )2
Step 7: Calculate epsilon = J1HB
( �
− J1� DP + RHB − RDP
Step 8: Use Maple command DF1 ∶= diff (epsilon, sigma3) to calculate 𝜕𝜎c′ . Thus, Eq. (22) will be formed
𝜕𝜀
3
𝜕𝜀c
Step 9: Use Maple command DF2 ∶= diff (epsilon, sigmazf ) to calculate 𝜕𝜎zf′
. Thus, Eq. (23) will be formed
𝜕 2 𝜀c
Step 9: Use Maple command DF11 ∶= diff (epsilon, sigma3, sigma3) to calculate 𝜕𝜎 ′ 3 as per Eq. (24)
2

𝜕2 𝜀 c

Step 10: Use Maple command D22 ∶= diff (epsilon, sigmazf , sigmazf ) to calculate 𝜕𝜎 zf as per Eq. (25)
′2

𝜕 2 𝜀c
Step 11: Use Maple command DF12 ∶= diff (epsilon, sigma3, sigmaz) to calculate 𝜕𝜎3 𝜕𝜎zf as per Eq. (26)
′ ′

Step 12: Use Maple command S ∶= solve({DF1, DF2}, {sigma3, sigmazf }) to solve Eqs. (22) and (23). Thus, the solution of unknown param-
eters (𝜎3� , 𝜎zf� ) can be obtained. The three roots are represented by S[1], S[2] and S[3], respectively
𝜕 2 𝜀c
Step 13: Use Maple command B11 ∶= subs(S[1], diff (epsilon, sigma3, sigma3)) to find out the value of 𝜕𝜎 ′ 32 for the obtained solution of (𝜎3� , 𝜎zf� )
𝜕 2 𝜀c
Step 14: Use Maple command B22 ∶= subs(S[1], diff (epsilon, sigmazf , sigmazf )) to find out the value of 𝜕𝜎 ′ zf2 for the obtained solution of
(𝜎3� , 𝜎zf� )
𝜕 2 𝜀c
Step 15: Use Maple command B12 ∶= subs(S[1], diff (epsilon, sigma3, sigmazf )) to find out the value of 𝜕𝜎3 𝜕𝜎zf for the obtained solution of
′ ′

(𝜎3� , 𝜎zf� )
( )2
𝜕 2 𝜀c 𝜕 2 𝜀c 𝜕 2 𝜀c
2 2 − � �
Step 16: Use Maple command C ∶= B11.B22 − B12.B12 to find out the value of 3 zf for the obtained solution of (𝜎3� , 𝜎zf� ).
�
𝜕𝜎 𝜕𝜎 � 𝜕𝜎3 𝜕𝜎zf

Thus, the condition of Eq. (21) can be checked

Step 17: Use Maple command epsilon ∶= subs(S[1], epsilon) to calculate 𝜀c for the obtained solution of (𝜎3� , 𝜎zf� )
Step 18: Repeat steps 13–17 to find out the related parameters for the other roots of (𝜎3� , 𝜎zf� ) which are represented by S[2] and S[3]

Table 2  The nature of the roots for Problem 1

Sl Remark (𝜎3′,𝜎zf′ )
( ) )2
𝜕 2 𝜀c 𝜕 2 𝜀c 𝜕 2 𝜀c ( 𝜀c
𝜎3′ , 𝜎zf′ 𝜕 2 𝜀c 𝜕 2 𝜀c 𝜕 2 𝜀c
No Root 𝜕𝜎 ′ 32 𝜕𝜎 ′ zf2 𝜕𝜎3′ 𝜕𝜎zf′
𝜕𝜎 � 23 𝜕𝜎 � 2zf
− 𝜕𝜎3� 𝜕𝜎zf� as per
Priest (Priest, 2005)

1 0.465, − 10.492 182.73 1.166 8.656 138.271 2.69 × 10−16 Min. point –
2 1.029, 60.00 141.37 0.783 − 12.09 − 35.589 574.59 Saddle –
point
3 29.36, 204.46 18.18 1.52 − 5.020 2.554 1.0 × 10−16 Min. point 29.36, 204.46

( )
𝜕2 𝜀 𝜕 2 𝜀c amb a−1 ]
√ 2𝜎zf� − 𝜎x� − 𝜎y�
2 2
[
c
= =− 1+ m −2 am ma−1 × √
𝜕𝜎zf� 𝜕𝜎3� 𝜕𝜎3� 𝜕𝜎zf� 3 3 f 3 b f ( )2 ( )2 ( )2
2𝜎x� − 𝜎y� − 𝜎zf� + 2𝜎y� − 𝜎zf� − 𝜎x� + 2𝜎zf� − 𝜎x� − 𝜎y�
(26)

13
 Modeling Earth Systems and Environment

uniaxial compressive strength 𝜎ci of the intact rock material,

( m 𝜎� )
where mf = +b 3
s , After solving Eqs. (22) and (23) for
𝜎ci GSI , mb , s , and a are the HB parameters for the specified
rock type, 𝜎x′ , and 𝜎y′ are the effective principal stresses. The
( )
stationary points 𝜎3′ , 𝜎zf′ , the conditions related to designa-
unknown value of the rock mass associated with the problem
tion of saddle point, maximum and minimum point for the are 𝜎3′ and 𝜎zf′ . Priest (2005) mentioned the known value of
error term 𝜀c may be investigated. For this purpose, the equa-
tion solver of MAPLE-18 software was used. the related parameters of rock mass such as 𝜎ci = 75 MPa,
GSI = 90, mb = 13.29, s = 0.329, a = 0.5, 𝜎x� = 15 MPa and
𝜎y� = 45 MPa. Table 1 contains step-by-step procedures with
Results and discussion
MAPLE-18 commands used for solving Eqs. (22) and (23).
Equation solver command solve ({Eq. 1, Eq. 2}, {sigma3, Afterwards, the conditions pertaining to saddle point, maxi-
sigmazf}) of MAPLE-18 software is used for solving Eqs. mum or minimum are checked.
(22) and (23). Priest (2005) presented two example problems
employing 3D yield criteria proposed by him. These two It is observed that three real roots of (𝜎3′,𝜎zf′ ) are obtained.
problems are solved by the authors to validate the proposed The real roots are (0.465, -10.492), (1.029, 60.00) and
solution strategies discussed in the present study. Also, (29.3587, 204.4675), respectively. Table 2 shows the nature
another problem is solved using the proposed strategy which of all the three real roots after checking whether they repre-
was earlier solved by Melkoumian et al. (2009). sent saddle point, maximum or minimum values of the error
term 𝜀c defined in Eq. (20).
Problem 1 This problem is adopted from the work of Priest It is observed that the first root and third roots, i.e.,
(2005). In this article, authors have used MAPLE-18 soft- (0.465, -10.492) and (29.3587, 204.4675) represent local
ware to develop a code based on proposed method. The minimum values of the error function 𝜺c. The second root
known value of rock mass associated with the problem are (1.029, 60.00) is a saddle point. Priest (2005) mentioned that
if negative values of either 𝜎3′ or 𝜎zf′ are obtained after solving
Eqs. (22) and (23), they should be discarded as negative
values of 𝜎3′ and 𝜎zf′ indicate tensile stresses. Therefore, the
first root (0.4655, -10.4925) is ignored as 𝜎zf′ value is found
out to be negative. Only the third root (29.3587, 204.4675)
is retained as the acceptable solution of the problem as both
𝜎3′ and 𝜎zf′ are obtained as positive quantities. Priest (2005)
also obtained the same solution for the studied problem. For
all the three roots, the values of the error term 𝜀c are also
presented in Table 2.
After obtaining 𝜎3′ and 𝜎zf′ by solving Eqs. (22) and (23),
other related parameters such as J1′ and J2 , DP parameters A
𝜎ci2
and B are calculated using J1� = , J2 and Eq. (13), and
𝜎ci
3
= 3
the maximum effective principal stress ( 𝜎1′ ) is calculated
using Eq. (1), respectively. Other parameters such as
′
J1HB ′
, J1DP and RHB , RDP can be further calculated using Eqs.
(14), (15), (17), and (18), respectively. The results of the
Fig. 2  The plot of error function 𝜺c for Problem 1

Table 3  The nature of the roots for Problem 2

Sl Remark (𝜎3′,𝜎zf′ )
( ) )2
𝜕 2 𝜀c 𝜕 2 𝜀c 𝜕 2 𝜀c ( 𝜀c
𝜎3′ , 𝜎zf′ 𝜕 2 𝜀c 𝜕 2 𝜀c 𝜕 2 𝜀c
No Root 𝜕𝜎 ′ 32 𝜕𝜎 ′ z2 𝜕𝜎3′ 𝜕𝜎zf′
𝜕𝜎 � 23 𝜕𝜎 � 2zf
− 𝜕𝜎3� 𝜕𝜎zf� as per
Priest (Priest, 2005)

1 0.196, 57.019 192.33 0.732 − 14.661 − 74.032 930.207 Saddle point –

2 0.918, − 14.655 154.33 1.410 9.157 133.886 −16
1.09 × 10 Min. point –
3 34.74, 225.198 16.071 1.544 − 4.711 2.625 1.09 × 10−14 Min. point 34.74, 225.198

13
Modeling Earth Systems and Environment

After obtaining 𝜎3′ and 𝜎zf′ by solving Eqs. (22) and (23),
other related parameters such as J1′ and J2 , DP parameters A
𝜎ci2
and B are calculated using J1� = and J2 = and Eq. (13)
𝜎ci
3 3
and the maximum effective principal stress (𝜎1′) is calculated
using Eq. (1), respectively. Other parameters such as
′
J1HB ′
, J1DP and RHB , RDP can be further calculated using Eqs.
(14), (15), (17), and (18) in tandem. The results of the
related parameter for the accepted real roots (34.74, 225.198)
are: J1′ = 25.00 MPa, J2 = 12,158.42 MPa, A = 0.912, B =
20.49 MPa, 𝝈 ′1 = 225.73 MPa, J′1HB = J′1DP = 98.40 MPa,
RHB = RDP = 155.93 MPa which also match perfectly with
those found out by Priest (2005).
Furthermore, the surface represented by 𝜀c is plotted in
Fig. 3  The plot of error function 𝜺c for Problem 2 Fig. 3 for the third real root obtained for this problem. This
root is also designated as the acceptable solution of the 3D
related parameter for the accepted real roots (29.36, 204.46) yield surface proposed by Priest (2005) as both roots are
are: J1′ = 25.00 MPa, J2 = 10,371.30 MPa, A = 0.926, B = positive. Plotting of the surface is carried out within a range
20.12 MPa, 𝝈 ′1 = 205.75 MPa, J′1HB = J′1DP = 88.155 MPa, of 𝜎3′ and 𝜎zf′ values (𝜎3′ = 0 – 300; 𝜎zf′ = 0 – 300) keeping in
RHB = RDP = 144.022 MPa. Similar values were reported by mind that the solution should be captured in the plot.
Priest (2005).
Furthermore, the surface represented by 𝜀c is plotted in Problem 3 Melkoumian et al. (2009) also derived an analyti-
Fig. 2 for the third real root obtained for this problem. This cal solution of the 3D yield criterion expressed in Eq. (19).
root is also designated as the acceptable solution of the 3D They compared the proposed analytical solution with the solu-
yield surface proposed by Priest (2005) as both roots are tion obtained using the numerical solution strategy proposed
positive. Plotting of the surface is carried out within a range by Priest (2005) for a range of input stress values of 𝜎x′ and𝜎y′ .
of 𝜎3′ and 𝜎zf′ values (𝜎3′ = 0 – 300; 𝜎zf′ = 0 – 300) keeping in The analytical solution ( 𝜎3′ and𝜎zf′ ) of the error term 𝜀c
mind that the solution should be captured in the plot. matched closely with the numerical solution with differences
less than 5 kPa. The numerical values of related terms are
Problem 2 This problem is adopted from the work of Priest 𝜎ci = 75 MPa,GSI = 90 , mb = 13.29, s = 0.329, a = 0.5 ,
(2005). In this article, authors have used MAPLE-18 soft-
𝜎x� = 25 MPa and 𝜎y� = 45 MPa. In the present work, the solu-
ware to develop a code based on proposed method. The input
parameters are same as those in problem 1. The numerical tion of 𝜎3′ and 𝜎zf′ obtained from the exact solution (as shown
values of all input parameters are same except a new value in Table 4) proposed by the authors are also compared with
of 𝜎x′ is considered now. The unknown values of the strength the analytical solution given by Melkoumian et al. (2009). It
criterion associated with the problem are 𝜎3′ and 𝜎zf′ . The is observed that the present solution matches very well with
numerical values of related terms are: 𝜎ci = 75 MPa, the solution reported in the literature.
GSI = 90,mb = 13.29, s = 0.329, a = 0.50, 𝜎x� = 25 MPa, and
𝜎y� = 45 MPa. The solution of rock mass strength criterion From Table 4, it is evident that three real roots are again
proposed by Priest (2005) is presented in Table 3, from obtained by solving Eqs. (22) and (23) for each input pairs
which it is observed that three real roots are again obtained of 𝜎x′ and 𝜎y′ . The 1st and the 3rd roots correspond to mini-
𝜕2 𝜀
for the new data. The first root (0.196, 57.019) is a saddle mum value of 𝜀c as the conditions laid down in 𝜕𝜎 ′ 2c < 0 and
3
point. The second and third roots, i.e., (0.918, -14.655) and 𝜕 2 𝜀c
< 0 are satisfied for both 1st and 3rd root. All 2nd roots
𝜕𝜎 ′ zf2
(34.74, 225.198) are obtained as two local minimum values ( )2
𝜕2 𝜀 𝜕2 𝜀 𝜕2 𝜀
of the error term 𝜺c . The second root is discarded as the are identified as saddle point. As 𝜕𝜎 � 2c 𝜕𝜎 � 2c − 𝜕𝜎 � 𝜕𝜎c � <0
value of 𝜎zf′ is found out to be negative. The third root (34.74,
3 zf 3 zf

225.198) is chosen as the acceptable solution for this prob- for the solved values of 𝜎3′ and 𝜎zf′ . All the 1st roots are dis-
lem. Priest (2005) also reported the same solution. carded as both 𝜎3′ and 𝜎zf′ are found to be negative. The 3rd

13
13
Table 4  The nature of the roots for Problem 3
Known Analytical Solution (Melkoumian Solutions of Proposed method
Input et al. 2009)
Parameters 1st root of Eq. (20) 2nd root of Eq. (20) 3rd root of Eq. (20)
𝜎x′ 𝜎y′ 𝜎3′ 𝜎zf′ 𝜎3′ 𝜎zf′ 𝜀c 𝜎3′ 𝜎zf′ 𝜀c 𝜎3′ 𝜎zf′ 𝜀c
(MPa) (MPa) (MPa) (MPa) (MPa) (MPa) (Min. Point) (MPa) (MPa) (Saddle (MPa) (MPa) (Min. Point)
point)

2.57 18.16 10.085 118.655 − 1.444 − 4.777 4.0 × 10−18 − 0.939 31.784 86.112 10.086 118.647 8.0 × 10−16
2.56 18.21 10.104 118.752 − 1.443 − 4.768 0.00 − 0.934 31.856 86.264 10.104 118.744 2.6 × 10−15
−16
2.55 18.37 10.173 119.128 − 1.437 − 4.756 1.2 × 10 − 0.920 32.077 87.000 10.173 119.121 1.6 × 10−15
2.53 18.64 10.289 119.751 − 1.427 − 4.732 1.0 × 10−16 − 0.897 32.449 88.210 10.289 119.744 1.0 × 10−15
−18
2.54 19.00 10.463 120.689 − 1.413 − 4.724 1.0 × 10 − 0.867 32.925 90.254 10.463 120.682 2.0 × 10−16
−16
2.57 19.45 10.690 121.906 − 1.395 − 4.726 4.4 × 10 − 0.830 33.507 93.037 10.690 121.900 2.5 × 10−15
−18
2.65 19.97 10.977 123.438 − 1.373 − 4.757 1.0 × 10 − 0.790 34.155 96.838 10.977 123.432 1.3 × 10−15
2.78 20.54 11.314 125.226 − 1.348 − 4.820 1.0 × 10−18 − 0.748 34.841 101.601 11.314 125.221 5.0 × 10−16
−16
2.94 21.15 11.687 127.181 − 1.320 − 4.901 1.4 × 10 − 0.704 35.562 107.072 11.687 127.176 5.0 × 10−16
−16
3.15 21.78 12.095 129.307 − 1.290 − 5.013 1.3 × 10 − 0.661 36.283 113.445 12.095 129.303 9.0 × 10−16
3.38 22.41 12.515 131.466 − 1.259 − 5.137 1.3 × 10−16 − 0.618 36.992 120.256 12.515 131.462 1.0 × 10−15
−18
3.63 23.02 12.936 113.610 − 1.227 − 5.274 4.0 × 10 − 0.578 37.664 127.407 12.936 133.606 8.0 × 10−16
−16
3.88 23.59 13.338 135.637 − 1.197 − 5.411 1.2 × 10 − 0.541 38.283 134.499 13.339 135.634 0.00
4.12 24.11 13.712 137.501 − 1.169 − 5.543 1.0 × 10−18 − 0.508 38.842 141.302 13.712 137.499 1.2 × 10−14
−16 1.0 × 10−14
4.33 24.55 14.032 139.083 − 1.145 − 5.658 1.2 × 10 − 0.480 39.311 147.287 14.032 139.081
−16
4.51 24.92 14.303 140.414 − 1.124 − 5.755 1.01 × 10 − 0.456 39.703 152.462 14.303 140.412 1.0 × 10−16
−18
4.65 25.18 14.500 141.380 − 1.108 − 5.832 9.0 × 10 − 0.441 39.973 156.362 14.500 141.379 1.0 × 10−16
4.73 25.35 14.623 141.977 − 1.099 − 5.875 9.0 × 10−18 − 0.430 40.154 158.753 14.623 141.976 4.0 × 10−16
−18
4.76 25.40 14.664 142.176 − 1.096 − 5.891 1.0 × 10 − 0.427 40.204 159.563 14.663 142.170 1.2 × 10−14
Modeling Earth Systems and Environment
Modeling Earth Systems and Environment

root is accepted as the desired solution as both 𝜎3′ and 𝜎zf′ are Brady BHG, Brown ET (2004) Rock mechanics for underground min-
ing, 3rd edn. pp 628
found to be positive. All the 3rd roots are seen to closely Chang C, Haimson B (2000) True triaxial strength and deformability
match with the analytical solution reported by Melkoumian of the German continental deep drilling program (KTB) deep hole
et al. (2009). amphibolite. J Geophys Res Solid Earth 105:18999–19013
Colmenares L, Zoback M (2002) A statistical evaluation of intact rock
failure criteria constrained by polyaxial test data for five different
rocks. Int J Rock Mech Min Sci 39:695–729. https://​doi.​org/​10.​
Summary and conclusion 1016/​S1365-​1609(02)​00048-5
Deng DP (2020) Limit equilibrium solution for the rock slope stability
Priest (2005) proposed a 3D yield criterion for rocks incor- under the coupling effect of the shear dilatancy and strain soften-
ing. Int J Rock Mech MinSci 34:104421
porating GHB material properties. Application of this yield Drucker DC, Prager W (1952) Soil mechanics and plastic analysis or
criterion necessitates the solution of two minor principal limit design. Q Appl Math 10:157–165
stresses 𝜎3′ and normal stress value 𝜎zf′ . For this purpose, he Haimson B (2006) True triaxial stresses and the brittle fracture of rock.
Pure Appl Geophys 163:1101–1130
defined an error term 𝜀c and recommended to minimize the
Hoek E (1983) 23rd Rankine-Lecture-strength of jointed rock masses.
same. Priest (2005) presented an iterative trial-and-error Geotechnique 33:185–223
strategy to obtain the parameters of interest. In the present Hoek E, Brown ET (1980) Underground excavations in rock. Inst. Min.
work, a rigorous solution of the parameters of interest (i.e., Metall, London, p 156
Hoek E, Brown ET (1997) Practical estimates of rock mass strength.
𝜎3′ and 𝜎zf′ ) is proposed. Complete analytical expressions of
Int J Rock Mech Min Sci. https://d​ oi.o​ rg/1​ 0.1​ 016/S
​ 1365-1​ 609(97)​
related conditions are presented in the current study. 80069-X
After minimizing the error term, three real roots are Hoek E, Kaiser PK, Bawden WF (1995) Support of underground exca-
obtained. For the studied problems, one of the three real vations in hard rock. AA Balkema, Rotterdam, pp 84–97
Hoek E, Carranza-Torres C, Corkum B (2002a) Hoek-Brown criterion
roots represents a saddle point. Other two roots represent - 2002a edition, in: NARMS-TAC Conference, Toronto. https://​
two local minimum values of the error term and thus, they doi.​org/​10.​1016/​0148-​9062(74)​91782-3
represent two possible solutions of the problem at hand. Hoek E, Carranza-Torres C, Corkum B (2002b) Hoek-Brown crite-
However, the root with only positive values of both 𝜎3′ and rion—2002b edition., in: Proceedings of the 5th North American
Rock Mechanics Symposium (NARMS-TAC 2002b).
𝜎zf′ are finally chosen as the acceptable solution. The Hoek E (1988) The Hoek-Brown failure criterion-a 1988 update. in:
obtained solutions for all the studied problems from the Proc. 15th Canadian Rock Mech. Symp. Toronto, Dept. Civil
existing literatures match very well with already reported Engineering, University of Toronto, pp. 31–38
Hoek E (1990) Estimating Mohr-Coulomb friction and cohesion values
solutions. Any root with negative values for either of 𝜎3′ or from the Hoek-Brown failure criterion. In: International Journal of
𝜎zf′ is ignored as it indicates tensile failure of the rock mass. Rock Mechanics and Mining Sciences & Geomechanics Abstracts.
Pergamon, pp. 227–229
Jiang H, Zhao J (2015) A simple three-dimensional failure criterion
for rocks based on the Hoek–Brown criterion. Rock Mech Rock
Author contributions VK: Formal analysis, Writing—original draft;
Eng 48:1807–1819
BR: Writing—original draft; A. Burman: Conceptualization, Investi-
Jiang H, Wang X, Xie Y (2011) New strength criteria for rocks under
gation, Methodology, Writing—original draft, Writing—review and
polyaxial compression. Can Geotech J 48:1233–1245
editing; SK: Writing—original draft; A. Bardhan: Investigation, Visu-
Liu S, Su Z, Li M, Shao L (2020) Slope stability analysis using elastic
alization, Writing—review and editing.
finite element stress fields. Eng Geol 273:105673
Marinos P, Hoek E (2001) Estimating the geotechnical properties of
Declarations heterogeneous rock masses such as flysch. Bull Eng Geol Environ
60:85–92. https://​doi.​org/​10.​1007/​s1006​40000​090
Conflicts of interest There is no conflict of interest. Matsuoka H (1982) A new failure criterion for soils in three-dimen-
sional stresses.
Melkoumian N, Priest SD, Hunt SP (2009) Further development of
the three-dimensional Hoek-Brown yield criterion. Rock Mech
References Rock Eng 42:835–847
Pan XD, Hudson JA (1988) A simplified three dimensional Hoek–
Agar JG, Morgenstern NR, Scott JD (1987) Shear strength and stress– Brown yield criterion, in: ISRM International Symposium.
strain behaviour of Athabasca oil sand at elevated temperatures OnePetro.
and pressures. Can Geotech J 24:1–10 Paterson MS (1978) Experimental rock deformation. brittle field.
Al-Ajmi AM, Zimmerman RW (2005) Relation between the Mogi and Springer, Berlin. Doi: https://​doi.​org/​10.​1029/​gm036​p0297
the Coulomb failure criteria. Int J Rock Mech Min Sci 42:431–439 Priest SD (2005) Determination of shear strength and three-dimen-
Anyaegbunam AJ (2015) Nonlinear power-type failure laws for geoma- sional yield strength for the Hoek-Brown criterion. Rock Mech
terials: synthesis from triaxial data, properties, and applications. Rock Eng 38:299–327
Int J Geomech 15:4014036 Priest S (2012) Three-dimensional failure criteria based on the Hoek-
Benz T, Schwab R, Kauther RA, Vermeer PA (2008) A Hoek-Brown Brown criterion. Rock Mech Rock Eng 45:989–993
criterion with intrinsic material strength factorization. Int J Rock
Mech Min Sci 45:210–222

13
 Modeling Earth Systems and Environment

Serrano A, Olalla C (1994) Ultimate bearing capacity of rock masses. XD P, JA H (1988) A simplified three dimensional Hoek–Brown yield
In: International Journal of Rock Mechanics and Mining Sciences criterion. Rom. M Rock Mech. power plants. Balkema, Rotterdam
& Geomechanics Abstracts. Elsevier, pp. 93–106. 95–103.
Serrano A, Olalla C (2006) Discussion of “nonlinear mohr envelopes Yuan W, Li JX, Li ZH, Wang W, Sun XY (2020) A strength reduction
based on triaxial data” by R. Baker. J Geotech Geoenviron Eng method based on the Generalized Hoek-Brown (GHB) criterion
132:128–129 for rock slope stability analysis. Comput Geotech 117:103240
Shen J, Karakus M, Xu C (2012) Direct expressions for lineariza- Zhang L, Zhu H (2007) Three-dimensional Hoek-Brown strength crite-
tion of shear strength envelopes given by the Generalized Hoek- rion for rocks. J Geotech Geoenviron Eng 133:1128–1135
Brown criterion using genetic programming. Comput Geotech
44:139–146 Publisher's Note Springer Nature remains neutral with regard to
Sofianos AI (2003) Tunnelling Mohr-Coulomb strength parameters for jurisdictional claims in published maps and institutional affiliations.
rock masses satisfying the generalized Hoek-Brown criterion. Int
J Rock Mech Min Sci 40:435–440 Springer Nature or its licensor holds exclusive rights to this article under
Sofianos AI, Halakatevakis N (2002) Equivalent tunnelling Mohr- a publishing agreement with the author(s) or other rightsholder(s);
Coulomb strength parameters for given Hoek-Brown ones. Int J author self-archiving of the accepted manuscript version of this article
Rock Mech Min Sci 39:131–137 is solely governed by the terms of such publishing agreement and
Taylor D (1937) Stability of earth slopes. J Bost Soc Civ Eng applicable law.
24:197–246

13