Chapter 11 of Rock Mass Properties

11
Rock mass properties

11.1 Introduction
Reliable estimates of the strength and deformation characteristics of rock masses are
required for almost any form of analysis used for the design of slopes, foundations
and underground excavations. Hoek and Brown (1980a, 1980b) proposed a method
for obtaining estimates of the strength of jointed rock masses, based upon an
assessment of the interlocking of rock blocks and the condition of the surfaces
between these blocks. This method was modified over the years in order to meet the
needs of users who were applying it to problems that were not considered when the
original criterion was developed (Hoek 1983, Hoek and Brown 1988). The
application of the method to very poor quality rock masses required further changes
(Hoek, Wood and Shah 1992) and, eventually, the development of a new
classification called the Geological Strength Index (Hoek, Kaiser and Bawden 1995,
Hoek 1995, Hoek and Brown 1997). A review of the development of the criterion and
of the equations proposed at various stages in this development is given in Hoek and
Brown (1997).
This chapter presents the Hoek-Brown criterion in a form that has been found
practical in the field and that appears to provide the most reliable set of results for use
as input for methods of analysis in current use in rock engineering.
11.2 Generalised Hoek-Brown criterion
The Generalised Hoek-Brown failure criterion for jointed rock masses is defined by:
1'
3'
3'
+ ci mb
+ s
ci
(11.1)
where 1' and 3' are the maximum and minimum effective stresses at failure,
mb is the value of the Hoek-Brown constant m for the rock mass,
s and a are constants which depend upon the rock mass characteristics, and
ci is the uniaxial compressive strength of the intact rock pieces.
The Mohr envelope, relating normal and shear stresses, can be determined by the
method proposed by Hoek and Brown (1980a). In this approach, equation 11.1 is used
to generate a series of triaxial test values, simulating full scale field tests, and a
Chapter 11: Rock mass properties
162
statistical curve fitting process is used to derive an equivalent Mohr envelope defined
by the equation:
' tm
= A ci n
ci
(11.2)
where A and B are material constants

'n is the normal effective stress, and
tm is the tensile strength of the rock mass.
This tensile strength, which reflects the interlocking of the rock particles when they
are not free to dilate, is given by:
tm =
ci
2
m m 2 + 4s
b
(11.3)
In order to use the Hoek-Brown criterion for estimating the strength and
deformability of jointed rock masses, three properties of the rock mass have to be
estimated. These are
1. the uniaxial compressive strength ci of the intact rock pieces,
2. the value of the Hoek-Brown constant mi for these intact rock pieces, and
3. the value of the Geological Strength Index GSI for the rock mass.
11.3
Intact rock properties
For the intact rock pieces that make up the rock mass equation 11.1 simplifies to:
1'
3'
3'
+ ci mi
+ 1
ci
0.5
(11.4)
The relationship between the principal stresses at failure for a given rock is defined
by two constants, the uniaxial compressive strength ci and a constant mi .
Wherever possible the values of these constants should be determined by statistical
analysis of the results of a set of triaxial tests on carefully prepared core samples.
Note that the range of minor principal stress ( 3' ) values over which these tests are
carried out is critical in determining reliable values for the two constants. In deriving
the original values of ci and mi , Hoek and Brown (1980a) used a range of 0 < 3'
< 0.5 ci and, in order to be consistent, it is essential that the same range be used in
any laboratory triaxial tests on intact rock specimens. At least five data points should
be included in the analysis.
One type of triaxial cell that can be used for these tests is illustrated in Figure 11.1.
This cell, described by Hoek and Franklin (1968), does not require draining between
tests and is convenient for the rapid testing or a large number of specimens. More
sophisticated cells are available for research purposes but the results obtained from
163
the cell illustrated in Figure 11.1 are adequate for the rock strength estimates required
for estimating ci and mi . This cell has the additional advantage that it can be used
in the field when testing materials such as coals, shales and phyllites that are
extremely difficult to preserve during transportation and normal specimen preparation
for laboratory testing.
hardened and ground steel

spherical seats
clearance gap for gauge wires
mild steel cell body
rock specimen with ground ends and

with a length to diameter ratio of 2
oil inlet - maximum pressure

700 MPa
strain gauges - if required
rubber sealing sleeve
Figure 11.1: Cut-away view of a triaxial cell for testing rock specimens.
Laboratory tests should be carried out at moisture contents as close as possible to

those which occur in the field. Many rocks show a significant strength decrease with
increasing moisture content and tests on samples, which have been left to dry in a
core shed for several months, can give a misleading impression of the intact rock
strength.
Once the five or more triaxial test results have been obtained, they can be analysed
to determine the uniaxial compressive strength ci and the Hoek-Brown constant mi as
described by Hoek and Brown (1980a). In this analysis, equation 11.1 is re-written in
the form:
164
y = m ci x + s ci
(11.5)
where x = 3' and y = (1' 3' ) 2

For n specimens the uniaxial compressive strength ci , the constant mi and the
coefficient of determination r2 are calculated from:
ci2 =
mi =
r2 =
y xy ( x y n) x
n
x 2 (( x) 2 n) n
1
ci
xy ( x y n)
2
2
x (( x) n)
[ xy ( x y n]2
[ x 2 ( x) 2 n][ y 2 ( y ) 2 n]
(11.6)
(11.7)
(11.8)
A spreadsheet for the analysis of triaxial test data is given in Table 11.1. Note that
high quality triaxial test data will usually give a coefficient of determination r2 of
greater than 0.9.
When laboratory tests are not possible, Table 11.2 and Table 11.3 can be used to
obtain estimates of ci and mi .
Short-term laboratory tests on very hard brittle rocks tend to overestimate the in
situ rock mass strength. Laboratory tests and field studies on excellent quality Lac du
Bonnet granite, reported by Martin and Chandler (1994), show that the in situ
strength of this rock is only about 70% of that measured in the laboratory. This
appears to be due to damage resulting from micro-cracking of the rock which initiates
and develops critical intensities at lower stress levels in the field than in laboratory
tests carried out at higher loading rates on smaller specimens. Hence, when analysing
the results of laboratory tests on these types of rocks to estimate the values of ci
and mi , it is prudent to reduce the values of the major effective principal stress at
failure to 70% of the measured values.
Anisotropic and foliated rocks such as slates, schists and phyllites, the behaviour
of which is dominated by closely spaced planes of weakness, cleavage or schistosity,
present particular difficulties in the determination of the uniaxial compressive
strengths.
Salcedo (1983) has reported the results of a set of directional uniaxial compressive
tests on a graphitic phyllite from Venezuela. These results are summarised in Figure
11.2. It will be noted that the uniaxial compressive strength of this material varies by
a factor of about 5, depending upon the direction of loading. Evidence of the
behaviour of this graphitic phyllite in the field suggests that the rock mass properties
are dependent upon the strength parallel to schistosity rather than that normal to it.
165
Table 11.1: Spreadsheet for the calculation of ci and mi from triaxial test data
Triaxial test data
x
sig3
sig1
0
38.3
5
72.4
7.5
80.5
15
115.6
20
134.3
47.5
sumx
441.1
xy
xsq
ysq
1466.89
0.0
4542.76 22713.8
5329.00 39967.5
10120.36 151805.4
13064.49 261289.8
0.0
25.0
56.3
225.0
400.0
2151766
20636668
28398241
102421687
170680899
34523.50 475776.5
sumy
sumxy
706.3
sumxsq
324289261
sumysq
Calculation results
Number of tests
n=
Uniaxial strength
sigci =
Hoek-Brown constant
mi =
Hoek-Brown constant
s=
Coefficient of determination r2 =
5
37.4
15.50
1.00
0.997
Cell formulae
y = (sig1-sig3)^2
sigci = SQRT(sumy/n - (sumxy-sumx*sumy/n)/(sumxsq-(sumx^2)/n)*sumx/n)
mi = (1/sigci)*((sumxy-sumx*sumy/n)/(sumxsq-(sumx^2)/n))
r2 = ((sumxy-(sumx*sumy/n))^2)/((sumxsq-(sumx^2)/n)*(sumysq-(sumy^2)/n))
100
Compressive strength - MPa
90
80
70
60
50
40
30
20
10
0
0
10
20
30
40
50
60
70
80
90
Angle of schistosity to loading direction
Figure 11.2: Influence of loading direction on the strength of graphitic

phyllite tested by Salcedo (1983).
166
Table 11.2: Field estimates of uniaxial compressive strength.

Uniaxial
Comp.
Grade* Term
Strength
(MPa)
R6
Extremely > 250
Strong
Point
Load
Index
(MPa)
>10
Field estimate of
strength
Examples
Specimen can only be Fresh basalt, chert,

chipped with a
diabase, gneiss, granite,
geological hammer
quartzite
R5
Very
strong
100 - 250 4 - 10
Specimen requires
many blows of a
geological hammer to
fracture it
Amphibolite, sandstone,
basalt, gabbro, gneiss,
granodiorite, limestone,
marble, rhyolite, tuff
R4
Strong
50 - 100 2 - 4
Specimen requires
Limestone, marble,
more than one blow of phyllite, sandstone,
a geological hammer schist, shale
to fracture it
R3
Medium
strong
25 - 50
1-2
Cannot be scraped or Claystone, coal,

peeled with a pocket
concrete, schist, shale,
knife, specimen can be siltstone
fractured with a single
blow from a
geological hammer
R2
Weak
5 - 25
**
Can be peeled with a

pocket knife with
difficulty, shallow
indentation made by
firm blow with point
of a geological
hammer
Chalk, rocksalt, potash
R1
Very
weak
1-5
**
Crumbles under firm

blows with point of a
geological hammer,
can be peeled by a
pocket knife
Highly weathered or
altered rock
R0
Extremely 0.25 - 1
weak
**
Indented by thumbnail Stiff fault gouge
* Grade according to Brown (1981).

** Point load tests on rocks with a uniaxial compressive strength below 25 MPa are likely to
yield highly ambiguous results.
167
Table 11.3: Values of the constant mi for intact rock, by rock group. Note that values in
parenthesis are estimates.
Rock
type
Class
Group
Coarse
Conglomerate
(22)
SEDIMENTARY
Clastic
METAMORPHIC
Sandstone
Siltstone
19
9
Greywacke
(18)
Organic
NonClastic
Carbonate
Fine
Very fine
Claystone
4
Chalk
7
Coal
(8-21)
Breccia
(20)
Chemical
Sparitic
Limestone
(10)
Micritic
Limestone
8
Gypstone
16
Anhydrite
13
Non Foliated
Marble
9
Hornfels
(19)
Quartzite
24
Slightly foliated
Migmatite
(30)
Amphibolite
25 - 31
Mylonites
(6)
Foliated*
Gneiss
33
Schists
4-8
Phyllites
(10)
Slate
9
Granite
33
Rhyolite
(16)
Obsidian
(19)
Granodiorite
(30)
Dacite
(17)
Diorite
(28)
Andesite
19
Light
IGNEOUS
Texture
Medium
Dark
Gabbro
27
Dolerite
(19)
Basalt
(17)
Breccia
(18)
Tuff
(15)
Norite
22
Extrusive
pyroclastic type
Agglomerate
(20)
* These values are for intact rock specimens tested normal to bedding or foliation. The value
of mi will be significantly different if failure occurs along a weakness plane.
In deciding upon the value of ci for foliated rocks, a decision has to be made on
whether to use the highest or the lowest uniaxial compressive strength obtained from
168
results such as those given in Figure 11.1. Mineral composition, grain size, grade of
metamorphism and tectonic history all play a role in determining the characteristics of
the rock mass. The author cannot offer any precise guidance on the choice of ci but
suggest that the maximum value should be used for hard, well interlocked rock
masses such as good quality slates. The lowest uniaxial compressive strength should
be used for tectonically disturbed, poor quality rock masses such as the graphitic
phyllite tested by Salcedo (1983).
Unlike other rocks, coal is organic in origin and therefore has unique constituents
and properties. Unless these properties are recognised and allowed for in
characterising the coal, the results of any tests will exhibit a large amount of scatter.
Medhurst, Brown and Trueman (1995) have shown that, by taking into account the
brightness which reflects the composition and the cleating of the coal, it is possible
to differentiate between the mechanical characteristics of different coals.
11.4 Influence of sample size
The influence of sample size upon rock strength has been widely discussed in
geotechnical literature and it is generally assumed that there is a significant reduction
in strength with increasing sample size. Based upon an analysis of published data,
Hoek and Brown (1980a) have suggested that the uniaxial compressive strength cd of
a rock specimen with a diameter of d mm is related to the uniaxial compressive
strength c50 of a 50 mm diameter sample by the following relationship:
50
cd = c50
d
0.18
(11.9)
This relationship, together with the data upon which it was based, is illustrated in
Figure 11.3.
The author suggests that the reduction in strength is due to the greater opportunity
for failure through and around grains, the building blocks of the intact rock, as more
and more of these grains are included in the test sample. Eventually, when a
sufficiently large number of grains are included in the sample, the strength reaches a
constant value.
Medhurst and Brown (1996) have reported the results of laboratory triaxial tests on
samples of 61, 101, 146 and 300 mm diameter samples of a highly cleated midbrightness coal from the Moura mine in Australia. The results of these tests are
summarised in Table 11.4 and Figure 11.4.
The results obtained by Medhurst and Brown show a significant decrease in
strength with increasing sample size. This is attributed to the effects of cleat spacing.
For this coal, the persistent cleats are spaced at 0.3 to 1.0 m while non-persistent
cleats within vitrain bands and individual lithotypes define blocks of 1 cm or less.
This cleating results in a critical sample size of about 1 m above which the strength
remains constant.
It is reasonable to extend this argument further and to suggest that, when dealing
with large scale rock masses, the strength will reach a constant value when the size of
individual rock pieces is sufficiently small in relation to the overall size of the
Influence of sample size
169
Uniaxial compressive strength of specimen of diameter d

Uniaxial compressive strength of 50 mm diameter specimen
structure being considered. This suggestion is embodied in Figure 11.5 which shows
the transition from an isotropic intact rock specimen, through a highly anisotropic
rock mass in which failure is controlled by one or two discontinuities, to an isotropic
heavily jointed rock mass.
1.5
Marble
Limestone
Granite
Basalt
Basalt-andesite lava
Gabbro
Marble
Norite
Granite
Quartz diorite
1.4
1.3
1.2
1.1
1.0
0.9
0.8
0.7
0
50
100
150
200
250
300
Specimen diameter d mm
Figure 11.3 Influence of specimen size on the strength of

intact rock. After Hoek and Brown (1980a).
Table11.4 Peak strength of Moura DU coal in terms of the

parameters contained in equation (11.1) based upon a value
of ci = 32.7 MPa.
Dia.(mm)
mb
61
101
146
300
mass
19.4
13.3
10.0
5.7
2.6
1.0
0.555
0.236
0.184
0.052
0.5
0.5
0.5
0.6
0.65
170
100
Sample diameter (mm)
61
Axial strength 1 ' MPa
90
80
101
70
146
60
300
50
40
mass
30
20
10
0
0
10
Confining pressure 3' MPa
Figure 11.4 Peak strength for Australian Moura coal.

After Medhurst and Brown (1996).
The Hoek-Brown failure criterion, which assumes isotropic rock and rock mass
behaviour, should only be applied to those rock masses in which there are a sufficient
number of closely spaced discontinuities, with similar surface characteristics, that
isotropic behaviour involving failure on discontinuities can be assumed. When the
structure being analysed is large and the block size small in comparison, the rock
mass can be treated as a Hoek-Brown material.
Where the block size is of the same order as that of the structure being analysed or
when one of the discontinuity sets is significantly weaker than the others, the HoekBrown criterion should not be used. In these cases, the stability of the structure should
be analysed by considering failure mechanisms involving the sliding or rotation of
blocks and wedges defined by intersecting structural features.
Geological strength Index
171
Intact rock specimens

- use equation 11.4
One joint set - do not use

Hoek-Brown criterion
Two joint sets - do not

use Hoek-Brown criterion
Many joints - use equation

11.1 with caution
Heavily jointed rock mass

- use equation 11.1
Figure 11.5: Idealised diagram showing the transition from intact to a heavily
jointed rock mass with increasing sample size.
11.5 Geological strength Index

The strength of a jointed rock mass depends on the properties of the intact rock pieces
and also upon the freedom of these pieces to slide and rotate under different stress
conditions. This freedom is controlled by the geometrical shape of the intact rock
pieces as well as the condition of the surfaces separating the pieces. Angular rock
pieces with clean, rough discontinuity surfaces will result in a much stronger rock
mass than one which contains rounded particles surrounded by weathered and altered
material.
The Geological Strength Index (GSI), introduced by Hoek (1995) and Hoek,
Kaiser and Bawden (1995) provides a system for estimating the reduction in rock
mass strength for different geological conditions. This system is presented in Table
11.5 and Table 11.6. Experience has shown that Table 11.5 is sufficient for field
172
observations since the letter code that identifies each rock mass category can be
entered into a field log. Later, these codes can be used to estimate the GSI value from
Table 11.6.
Once the Geological Strength Index has been estimated, the parameters that
describe the rock mass strength characteristics, are calculated as follows:
GSI 100
mb = mi exp
28
(11.10)
For GSI > 25, i.e. rock masses of good to reasonable quality, the original HoekBrown criterion is applicable with
GSI 100
s = exp
(11.11)
and
a = 0.5
(11.12)
For GSI < 25, i.e. rock masses of very poor quality, the modified Hoek-Brown
criterion applies with
s=0
(11.13)
and
a = 0.65
GSI
200
(11.14)
The choice of GSI = 25 for the switch between the original and modified criteria is
purely arbitrary. It could be argued that a switch at GSI = 30 would not introduce a
discontinuity in the value of a, but extensive trials have shown that the exact location
of this switch has negligible practical significance.
For better quality rock masses (GSI > 25), the value of GSI can be estimated
directly from the 1976 version of Bieniawskis Rock Mass Rating, with the
Groundwater rating set to 10 (dry) and the Adjustment for Joint Orientation set to 0
(very favourable) (Bieniawski 1976). For very poor quality rock masses the value of
RMR is very difficult to estimate and the balance between the ratings no longer gives
a reliable basis for estimating rock mass strength. Consequently, Bieniawskis RMR
classification should not be used for estimating the GSI values for poor quality rock
masses.
If the 1989 version of Bieniawskis RMR classification (Bieniawski 1989) is used,
then GSI = RMR89 - 5 where RMR89 has the Groundwater rating set to 15 and the
Adjustment for Joint Orientation set to zero.
One of the practical problems which arises when assessing the value of GSI in the
field is related to blast damage. As illustrated in Figure 11.6, there is a considerable
difference in the appearance of a rock face which has been excavated by controlled
blasting and a face which has been damaged by bulk blasting. Wherever possible, the
undamaged face should be used to estimate the value of GSI since the overall aim is
to determine the properties of the undisturbed rock mass.
173
Table 11.5: Characterisation of rock masses on the basis of interlocking and joint alteration1
In earlier versions of this table the terms BLOCKY/SEAMY and CRUSHED were used, following
the terminology used by Terzaghi (1946). However, these terms proved to be misleading and they have
been replaced, in this table by BLOCKY/DISTURBED, which more accurately reflects the increased
mobility of a rock mass which has undergone some folding and/or faulting, and DISINTEGRATED
which encompasses a wider range of particle shapes.
174
Table 11.6: Estimate of Geological Strength Index GSI based on geological descriptions.
175
Figure 11.6: Comparison between the results achieved using controlled blasting (on the left)
and normal bulk blasting for a surface excavation in gneiss.
Where all the visible faces have been damaged by blasting, some attempt should be
made to compensate for the lower values of GSI obtained from such faces. In recently
blasted faces, new discontinuity surfaces will have been created by the blast and these
will give a GSI value that may be as much as 10 points lower than that for the
undisturbed rock mass. In other words, severe blast damage can be allowed for by
moving up one row in Table 11.5 and Table 11.6.
Where blast damaged faces have been exposed for a number of years, it may also
be necessary to step as much as one column to the left in order to allow for surface
weathering which will have occurred during this exposure. Hence, for example, a
badly blast damaged weathered rock surface which has the appearance of a
BLOCKY/DISTURBED and FAIR (BD/F in Table 11.5) rock mass may actually be
VERY BLOCKY and GOOD (VB/G) in its unweathered and undisturbed in situ
state.
An additional practical question is whether borehole cores can be used to estimate
the GSI value behind the visible faces? For reasonable quality rock masses (GSI >
25) the best approach is to evaluate the core in terms of Bieniawskis RMR
classification and then, as described above, to estimate the GSI value from RMR. For
poor quality rock masses (GSI < 25), relatively few intact core pieces longer than 100
mm are recovered and it becomes difficult to determine a reliable value for RMR. In
these circumstances, the physical appearance of the material recovered in the core
should be used as a basis for estimating GSI.
176
11.6 Mohr-Coulomb parameters
Most geotechnical software is written in terms of the Mohr-Coulomb failure criterion

in which the rock mass strength is defined by the cohesive strength c and the angle of
friction . The linear relationship between the major and minor principal stresses,
1' and 3' , for the Mohr-Coulomb criterion is
1' = cm + k 3'
(11.15)
where cm is the uniaxial compressive strength of the rock mass and k is the slope of
the line relating 1' and 3' . The values of and c can be calculated from
sin ' =
c =
'
k 1
k +1
cm (1 sin ' )
2 cos '
(11.16)
(11.17)
There is no direct correlation between equation 11.15 and the non-linear Hoek-Brown
criterion defined by equation 11.1. Consequently, determination of the values of c
and for a rock mass that has been evaluated as a Hoek-Brown material is a difficult
problem.
The author believes that the most rigorous approach available, for the original
Hoek-Brown criterion, is that developed by Dr J.W. Bray and reported by Hoek
(1983). For any point on a surface of concern in an analysis such as a slope stability
calculation, the effective normal stress is calculated using an appropriate stress
analysis technique. The shear strength developed at that value of effective normal
stress is then calculated from the equations given in Hoek and Brown (1997). The
difficulty in applying this approach in practice is that most of the geotechnical
software currently available provides for constant rather than effective normal stress
dependent values of c and .
Having evaluated a large number of possible approaches to this problem, it has
been concluded that the most practical solution is to treat the problem as an analysis
of a set of full-scale triaxial strength tests. The results of such tests are simulated by
using the Hoek-Brown equation 11.1 to generate a series of triaxial test values.
Equation 11.15 is then fitted to these test results by linear regression analysis and the
values of c and are determined from equations 11.17 and 11.16. The steps
required to determine the parameters A, B, c and are given below. A spreadsheet
for carrying out this analysis, with a listing of all the cell formulae, is given in Figure
11.7.
The relationship between the normal and shear stresses can be expressed in terms
of the corresponding principal effective stresses as suggested by Balmer (1952):
'n
3'
1' 3'
1' 3' + 1
(11.18)
Mohr-Coulomb parameters
177
= (1' 3' ) 1' 3'
(11.19)
For the GSI > 25, when a = 0.5:

1'
3'
= 1+
mb ci
(11.20)
2(1' 3' )
For GSI < 25, when s = 0:

a 1
'
a 3
= 1 + amb
ci
3'
1'
(11.21)
The tensile strength of the rock mass is calculated from:

ci
2
tm =
m m 2 + 4s
b
(11.22)
The equivalent Mohr envelope, defined by equation 11.2, may be written in the form:
Y = log A + BX
(11.23)
where

Y = log
ci
'n tm
, X = log
ci
(11.24)
Using the value of tm calculated from equation 11.22 and a range of values of
and 'n calculated from equations 11.19 and 11.18 the values of A and B are
determined by linear regression where :
B=
XY ( X Y ) T
X 2 ( X ) 2 T
A = 10^ ( Y T B( X T ) )
(11.25)
(11.26)
and T is the total number of data pairs included in the regression analysis.
The most critical step in this process is the selection of the range of 3' values. As
far as the author is aware, there are no theoretically correct methods for choosing this
range and a trial and error method, based upon practical compromise, has been used
for selecting the range included in the spreadsheet presented in Figure 11.7.
For a Mohr envelope defined by equation 11.2, the friction angle i' for a specified
normal stress 'ni is given by:
178
i'
'
tm
= arctan AB ni
ci
B 1
(11.27)
The corresponding cohesive strength ci' is given by:
ci' = 'ni tan i'
(11.28)
and the corresponding uniaxial compressive strength of the rock mass is :
cmi =
2ci' cos i'

1 sin i'
(11.29)
Note that the cohesive strength ci' given by equation 11.29 is an upper bound value
and that it is prudent to reduce this to about 75% of the calculated value for practical
applications.
The values of c and obtained from this analysis are very sensitive to the range
of values of the minor principal stress 3' used to generate the simulated full-scale
triaxial test results. On the basis of trial and error, it has been found that the most
consistent results are obtained when 8 equally spaced values of 3' are used in the
range 0 < 3 < 0.25ci.
An example of the results, which are obtained from this analysis, is given in Figure
11.8. Plots of the values of the ratio c ' ci and the friction angle , for different
combinations of GSI and mi are given in Figure 11.9.
The spreadsheet includes a calculation for a tangent to the Mohr envelope defined
by equation 11.2. A normal stress has to be specified in order to calculate this tangent
and, in Figure 11.8, this stress has been chosen so that the friction angle is the same
for both the tangent and the line defined by c = 3.3 MPa and = 30.1, determined
by the linear regression analysis described earlier. The cohesion intercept for the
tangent is c = 4.1 MPa which is approximately 25% higher than that obtained by
linear regression analysis of the simulated triaxial test data.
Fitting a tangent to the curved Mohr envelope gives an upper bound value for the
cohesive intercept c. It is recommended that this value be reduced by about 25% in
order to avoid over-estimation of the rock mass strength.
There is a particular class of problem for which extreme caution should be
exercised when applying the approach outlined above. In some rock slope stability
problems, the effective normal stress on some parts of the failure surface can be quite
low, certainly less than 1 MPa. It will be noted that in the example given in Figure
11.8, for values of 'n of less than about 5 MPa, the straight line, constant c and
method overestimates the available shear strength of the rock mass by increasingly
significant amounts as 'n approaches zero. Under such circumstances, it would be
prudent to use values of c and based on a tangent to the shear strength curve in the
range of 'n values applying in practice.
179
Figure 11.7 Spreadsheet for calculation of Hoek-Brown and equivalent Mohr-Coulomb

parameters
Hoek-Brown and equivalent Mohr Coulomb failure criteria
Input:
Output:
sigci =
85
mb =
sigtm =
k=
sigcm =
1.40
-0.13
3.01
11.36
signt =
15.97
Tangent:
MPa
mi =
10
GSI =
45
a=
B=
coh =
0.5
0.70
3.27
MPa
MPa
s = 0.0022
A=
0.50
phi = 30.12 degrees
E = 6913.7 MPa
MPa
phit=
coht =
4.12
MPa
MPa
30.12
degrees
Calculation:
sig3
1E-10
3.04
6.07
9.1
12.14
15.18
18.21
21.25
sig1
4.00
22.48
33.27
42.30
50.40
57.91
64.98
71.74
ds1ds3
15.89
4.07
3.19
2.80
2.56
2.40
2.27
2.18
sign
0.24
6.87
12.56
17.85
22.90
27.76
32.50
37.13
tau
0.94
7.74
11.59
14.62
17.20
19.48
21.54
23.44
x
-2.36
-1.08
-0.83
-0.67
-0.57
-0.48
-0.42
-0.36
y
-1.95
-1.04
-0.87
-0.76
-0.69
-0.64
-0.60
-0.56
xy
4.61
1.13
0.71
0.52
0.39
0.31
0.25
0.20
xsq
5.57
1.17
0.68
0.45
0.32
0.23
0.17
0.13
sig3sig1
0.00
68.23
202.01
385.23
612.01
878.92 1183.65 1524.51
sig3sq
0.00
9.22
36.86
82.94
147.45
230.39
331.76
451.56
taucalc
0.96
7.48
11.33
14.45
17.18
19.64
21.91
24.04
sig1sig3fit 11.36
20.51
29.66
38.81
47.96
57.11
66.26
75.42
signtaufit
3.41
7.26
10.56
13.63
16.55
19.38
22.12
24.81
tangent 4.25309 8.10321 11.4032 14.4729 17.3991 20.2235 22.9702 25.655
Sums
85.00
347.08
35.35
157.80
116.55
-6.77
-7.11
8.12
8.74
4855
1290
Cell formulae:
mb = mi*EXP((GSI-100)/28)
s = IF(GSI>25,EXP((GSI-100)/9),0)
a = IF(GSI>25,0.5,0.65-GSI/200)
sigtm = 0.5*sigci*(mb-SQRT(mb^2+4*s))
A = acalc = 10^(sumy/8 - bcalc*sumx/8)
B = bcalc = (sumxy - (sumx*sumy)/8)/(sumxsq - (sumx^2)/8)
k = (sumsig3sig1 - (sumsig3*sumsig1)/8)/(sumsig3sq-(sumsig3^2)/8)
phi = ASIN((k-1)/(k+1))*180/PI()
coh = (sigcm*(1-SIN(phi*PI()/180)))/(2*COS(phi*PI()/180))
sigcm = sumsig1/8 - k*sumsig3/8
E = IF(sigci>100,1000*10^((GSI-10)/40),SQRT(sigci/100)*1000*10^((GSI-10)/40))
phit = (ATAN(acalc*bcalc*((signt-sigtm)/sigci)^(bcalc-1)))*180/PI()
coht = acalc*sigci*((signt-sigtm)/sigci)^bcalc-signt*TAN(phit*PI()/180)
sig3 = Start at 1E-10 (to avoid zero errors) and increment in 7 steps of sigci/28 to 0.25*sigci
sig1 = sig3+sigci*(((mb*sig3)/sigci)+s)^a
ds1ds3 = IF(GSI>25,(1+(mb*sigci)/(2*(sig1-sig3))),1+(a*mb^a)*(sig3/sigci)^(a-1))
sign = sig3+(sig1-sig3)/(1+ds1ds3)
tau = (sign-sig3)*SQRT(ds1ds3)
x = LOG((sign-sigtm)/sigci)
y = LOG(tau/sigci)
xy = x*y
x sq = x^2
sig3sig1= sig3*sig1
sig3sq = sig3^2
taucalc = acalc*sigci*((sign-sigtm)/sigci)^bcalc
s3sifit = sigcm+k*sig3
sntaufit = coh+sign*TAN(phi*PI()/180)
tangent = coht+sign*TAN(phit*PI()/180)
180
80
70
60
Major principal stress ' MPa

1
50
1' = cm +
40
30
1'
= 3'
1 + sin '
1 sin
'
'
+ ci mb 3 + s
ci
3'
20
10
cm
0
0
10
20
30
Minor principal stress 3' MPa
Shear strength MPa
30
tangent
20
' tm
= A ci n
ci
10
= c ' + 'n tan '

0
0
10
20
Normal stress
30
n' MPa
40
Figure 11.8: Plot of results from simulated full scale triaxial tests on a rock mass defined by
a uniaxial compressive strength ci = 85 MPa, a Hoek -Brown constant mi = 10 and a
Geological Strength Index GSI = 45.
181
0.10
0.08
0.06
0.05
0.04
0.03
mi
35
30
25
20
16
13
10
0.02
Cohesive strength / Uniaxial strength of intact rock
0.20
0.01
5
10
20
30
40
50
60
70
80
0.008
90
Geological Strength Index GSI
a. Plot of ratio of cohesive strength c to uniaxial

compressive strength ci .
55
mi
50
35
30
25
45
16
Friction angle - degrees
20
13
40
10
35
30
25
20
15
10
10
20
30
40
50
60
70
80
90
b. Plot of friction angle

Figure 11.9: Plots of cohesive strength and friction angles for different GSI and mi values.
182
11.7 Deformation modulus
Serafim and Pereira (1983) proposed a relationship between the in situ modulus of
deformation and Bieniawskis RMR classification. This relationship is based upon
back analysis of dam foundation deformations and it has been found to work well for
better quality rocks. However, for many of the poor quality rocks it appears to predict
deformation modulus values which are too high. Based upon practical observations
and back analysis of excavation behaviour in poor quality rock masses, the following
modification to Serafim and Pereiras equation is proposed for ci < 100 :
GSI 10
40
ci
Em =
10
100
(11.30)
Note that GSI has been substituted for RMR in this equation and that the modulus Em
is reduced progressively as the value of ci falls below 100. This reduction is based
upon the reasoning that the deformation of better quality rock masses is controlled by
the discontinuities while, for poorer quality rock masses, the deformation of the intact
rock pieces contributes to the overall deformation process.
Based upon measured deformations, equation 11.30 appears to work reasonably
well in those cases where it has been applied. However, as more field evidence is
gathered it may be necessary to modify this relationship.
11.8 Post-failure behaviour
When using numerical models to study the progressive failure of rock masses,
estimates of the post-peak or post-failure characteristics of the rock mass are required.
In some of these models, the Hoek-Brown failure criterion is treated as a yield
criterion and the analysis is carried out using plasticity theory (e.g. Pan and Hudson
1988). No definite rules for dealing with this problem can be given but, based upon
experience in numerical analysis of a variety of practical problems, the post-failure
characteristics illustrated in Figure 11.10 are suggested as a starting point.
11.8.1
Very good quality hard rock masses
For very good quality hard rock masses, such as massive granites or quartzites, the
analysis of spalling around highly stressed openings (Hoek, Kaiser and Bawden 1995)
suggests that the rock mass behaves in an elastic brittle manner as shown in Figure
11.10(a). When the strength of the rock mass is exceeded, a sudden strength drop
occurs. This is associated with significant dilation of the broken rock pieces. If this
broken rock is confined, for example by rock support, then it can be assumed to
behave as a rock fill with a friction angle of approximately = 38 and zero cohesive
strength.
Typical properties for this very good quality hard rock mass may be as shown in
Table 11.7. Note that, in some numerical analyses, it may be necessary to assign a
very small cohesive strength in order to avoid numerical instability.
Post-failure behaviour
183
70
60
Stress
50
Elastic-brittle
40
30
20
10
0
0.000
0.001
0.002
0.003
Strain
(a) Very good quality hard rock mass

15
Stress
10
Strain softening
5
0
0.000
0.001
0.002
0.003
Strain
(b) Average quality rock mass

2.0
Elastic-plastic
Stress
1.5
1.0
0.5
0.0
0.000
0.001
0.002
0.003
Strain
(c) Very poor quality soft rock mass

Figure 11.10: Suggested post failure characteristics for different quality rock masses.
184
Table 11.7: Typical properties for a very good quality hard rock mass
Intact rock strength
Hoek-Brown constant
Geological Strength Index
Friction angle
Cohesive strength
Rock mass compressive strength
Rock mass tensile strength
Deformation modulus
Poissons ratio
Dilation angle
Post-peak characteristics
Friction angle
Cohesive strength
Deformation modulus
11.8.2
ci
mi
GSI
c
cm
tm
Em
150 MPa
25
75
46
13 MPa
64.8 MPa
-0.9 MPa
42000 MPa
0.2
/4 = 11.5
f
cf
Efm
38
0
10000 MPa
Average quality rock mass
In the case of an average quality rock mass it is reasonable to assume that the postfailure characteristics can be estimated by reducing the GSI value from the in situ
value to a lower value which characterises the broken rock mass.
The reduction of the rock mass strength from the in situ to the broken state
corresponds to the strain softening behaviour illustrated in Figure 11.10(b). In this
figure it has been assumed that post failure deformation occurs at a constant stress
level, defined by the compressive strength of the broken rock mass. The validity of
this assumption is unknown.
Typical properties for this average quality rock mass may be as follows:
Table 10.8: Typical properties for an average rock mass.
Hoek-Brown constant
Friction angle
Cohesive strength
Deformation modulus
Poissons ratio
Dilation angle
Broken rock mass strength
Deformation modulus
11.8.3
ci
mi
GSI
c
cm
tm
Em
80 MPa
12
50
33
3.5 MPa
13 MPa
-0.15
9000 MPa
0.25
/8 = 4
fcm
Efm
8 MPa
5000 MPa
Very poor quality rock mass
Analysis of the progressive failure of very poor quality rock masses surrounding
tunnels suggests that the post-failure characteristics of the rock are adequately
represented by assuming that it behaves perfectly plastically. This means that it
Reliability of rock mass strength estimates
185
continues to deform at a constant stress level and that no volume change is associated
with this ongoing failure. This type of behaviour is illustrated in Figure 10.10(c).
Typical properties for this very poor quality rock mass may be as follows:
Table 11.9: Typical properties for a very poor quality rock mass
Hoek-Brown constant
Friction angle
Cohesive strength
Deformation modulus
Poissons ratio
Dilation angle
Broken rock mass strength
Deformation modulus
ci
mi
GSI
c
cm
tm
Em
20 MPa
8
30
24
0.55 MPa
1.7 MPa
-0.01 MPa
1400 MPa
0.3
zero
fcm
Efm
1.7 MPa
1400 MPa
11.9 Reliability of rock mass strength estimates

The techniques described in the preceding sections of this chapter can be used to
estimate the strength and deformation characteristics of isotropic jointed rock masses.
When applying this procedure to rock engineering design problems, most users
consider only the average or mean properties. In fact, all of these properties exhibit
a distribution about the mean, even under the most ideal conditions, and these
distributions can have a significant impact upon the design calculations.
In the text that follows, a slope stability calculation and a tunnel support design
calculation are carried out in order to evaluate influence of these distributions. In each
case the strength and deformation characteristics of the rock mass are estimated by
means of the Hoek-Brown procedure, assuming that the three input parameters are
defined by normal distributions.
11.9.1
Input parameters
Figure 11.11 has been used to estimate the value of the value of GSI from field
observations of blockiness and discontinuity surface conditions. Included in this
figure is a crosshatched circle representing the 90% confidence limits of a GSI value
of 25 5 (equivalent to a standard deviation of approximately 2.5). This represents
the range of values that an experienced geologist would assign to a rock mass
described as BLOCKY/DISTURBED or DISINTEGRATED and POOR. Typically,
rocks such as flysch, schist and some phyllites may fall within this range of rock mass
descriptions.
186
Figure 11.11: Estimate of Geological Strength Index GSI based on geological descriptions.
187
In the authors experience, some geologists go to extraordinary lengths to try to

determine an exact value of GSI (or RMR). Geology does not lend itself to such
precision and it is simply not realistic to assign a single value. A range of values, such
as that illustrated in Figure 11.11 is more appropriate. In fact, in some complex
geological environments, the range indicated by the crosshatched circle may be too
optimistic.
The two laboratory properties required for the application of the Hoek-Brown
criterion are the uniaxial compressive strength of the intact rock ( ci ) and the intact
rock material constant mi. Ideally these two parameters should be determined by
triaxial tests on carefully prepared specimens as described by Hoek and Brown
(1997).
It is assumed that all three input parameters can be represented by normal
distributions as illustrated in Figure 11.12. The standard deviations assigned to these
three distributions are based upon the authors experience of geotechnical programs
for major civil and mining projects where adequate funds are available for high
quality investigations. For preliminary field investigations or low budget projects, it
is prudent to assume larger standard deviations for the input parameters.
11.9.2
Output parameters
The values of the friction angle , the cohesive strength c ' , the uniaxial compressive
strength of the rock mass cm and the deformation modulus E m of the rock mass
were calculated by the procedure described in previous sections of this chapter. The
Excel add-on program @RISK2 was used for a Monte Carlo analysis in which 1000
calculations were carried out for randomly selected values of the input parameters.
The results of these calculations were analysed using the program BESTFIT1 and it
was found that all four output parameters could be adequately described by the
normal distributions illustrated in Figure 11.12.
In several trials it was found that the output parameters , c ' and cm were always
well represented by normal distributions. On the other hand, for GSI values of more
than 40, the deformation modulus E m was better represented by a lognormal
distribution.
11.9.3
Slope stability calculation
In order to assess the impact of the variation in output parameters, illustrated in

Figure 11.12, a calculation of the factor of safety for a homogeneous slope was
carried out using Bishops circular failure analysis in the program SLIDE3. The
geometry of the slope and the phreatic surface, the rock mass properties and the
critical failure surface for the average properties are shown in Figure 11.13.
2
3
From Palisade Corporation, 31 Decker Road, Newfield, New York 14867, USA.
Available from Available from Rocscience Inc., 31 Balsam Avenue, Toronto, Ontario, Canada M4E
3B5, Fax 1 416 698 0908, Phone 1 416 698 8217, Email: software@rocscience.com, Internet
http://www.rocscience.com.
188
INPUT
OUTPUT
0.35
0.16
ci
Mean 10 MPa
Stdev 2.5 MPa
0.08
0.30
Probability
Probability
0.12
0.04
0.25
Mean 22.85
Stdev 1.31
0.20
0.15
0.10
0.05
0.00
0.00
0
10
15
20
18
Intact rock strength - MPa
22
24
26
28
Friction ange - degrees

7.0
0.40
6.0
Mean 8
Stdev 1
0.20
Probability
mi
0.30
Probability
20
0.10
5.0
4.0
Mean 0.23 MPa

Stdev 0.068 MPa
3.0
2.0
1.0
0.00
0.0
4
10
12
0.0
Hoek-Brown constant mi
0.1
0.2
0.3
0.5
Cohesive strength - MPa
0.16
2.0
cm
GSI
1.5
Mean 25
Stdev 2.5
0.08
Probability
0.12
Probability
0.4
0.04
Mean 0.68 MPa

Stdev 0.21 MPa
1.0
0.5
0.00
0.0
15
20
25
30
35
0.0
0.5
1.0
1.5
Rock mass strength - MPa

0.0100
Figure 11.12: Assumed normal distributions

for input parameters and calculated
distributions for output parameters.
Probability
0.0075
0.0050
Em
Mean 373 MPa
Stdev 48 MPa
0.0025
0.0000
200 250 300 350 400 450 500 550
Deformation modulus - MPa
189
139.3,324.4
Critical failure circle

Factor of Safety = 1.32
300,160
400,160
400,127
240,120
0,100
Phreatic surface
100,100
Rock mass properties:
Friction angle = 22.9 degrees
Cohesive strength c = 0.23 MPa
Unit weight = 23 kN/m 3
400,0
0,0
Figure 11.13: Slope and phreatic surface geometry, rock mass properties and
critical failure surface for a homogeneous slope.
The distribution of the factor of safety was determined by Rosenbleuths Point

Estimate method (Rosenbleuth 1976, Harr 1987) in which the two values are chosen
at one standard deviation on either side of the mean for each variable. The factor of
safety is calculated for every possible combination of point estimates, producing 2m
solutions, where m is the number of variables considered. The mean and standard
deviation of the factor of safety are then calculated from these 2m solutions.
This calculation of the mean and standard deviation is given in Table 11.10. Based
upon the fact that the two variables included in this analysis are defined by normal
distributions and considering the form of the equations used to calculate the factor of
safety, it is reasonable to assume that the factor of safety will be adequately
represented by a normal distribution. This distribution is illustrated in Figure 11.13.
Table 11.10: Calculations for Rosenbleuths Point Estimate method using
one standard deviation.
Case
-,c+,c+
-, c+
+, c-
Friction
Angle
21.19
24.16
21.19
24.16
Safety
Factor
1.215
1.407
1.217
1.406
5.245
Cohesion
0.162
0.298
0.298
0.162
sums
(SF-SFi)2
0.00922
0.00922
0.00884
0.00912
0.0364
n
_
Mean Safety Factor = SF = 1 SFi = 1.31
n i =1
Standard deviation = S 2 =
1 n _
( SF SFi )2
n 1 i =1
= 0.11
190
4
Factor of Safety
Mean = 1.31
Stdev = 0.11
Probability
3
2
1
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1.0
0.9
Factor of Safety
Figure 11.14: Normal distribution of the factor of safety of the

slope defined in Figure 11.13.
The mean factor of safety for this slope is 1.3 that is a value frequently used in the
design of slopes for open pit mines. It is interesting that the probability of failure,
given by the portion of the distribution curve for SF < 1, is very small. This suggests
that, for a high quality geotechnical investigation such as that assumed in this study, a
safety factor of 1.3 is adequate to ensure stability under the assumed conditions.
11.9.4
Tunnel stability calculations
Consider a circular tunnel of radius ro in a stress field in which the horizontal and
vertical stresses are both po. If the stresses are high enough, a plastic zone of
damaged rock of radius rp surrounds the tunnel. A uniform support pressure pi is
provided around the perimeter of the tunnel. This situation is illustrated in Figure
11.15.
Assuming that the rock mass fails with zero plastic volume change, the critical
stress level pcr at which failure initiates is given by :
p cr =
2 p o cm
1+ k
(11.31)
1 + sin
1 sin
(11.32)
where
k=
Where the support pressure pi is less than the critical pressure pcr, the radius rp of
the plastic zone and the inward deformation of the tunnel wall uip are given by:
1
( k 1)
2( po (k 1) + cm )
=
ro (1 + k )((k 1) pi + cm )
rp
rp
(1 + )
=
2(1 )( po pcr ) (1 2)( po pi )
ro
E
ro
uip
(11.33)
(11.34)
191
Figure 11.15: Development of a plastic zone around a circular

tunnel in a hydrostatic stress field.
In order to study the influence of the variation in the input parameters, a Monte Carlo
analysis was performed using the program @RISK in an Excel spreadsheet that had
been programmed to perform the analysis defined above. It was assumed that a 5 m
diameter tunnel (ro = 2.5 m) was subjected to uniform in situ stress of po = 2.5 MPa.
The rock mass properties were defined by the normal distributions for , c, cm and E
defined in Figure 11.12.
This analysis was carried out for a tunnel with no support. A second analysis was
performed for a tunnel with a support pressure of pi = 0.3 MPa which is
approximately that which can be achieved with a closed ring of 50 mm thick shotcrete
with a uniaxial compressive strength of 14 MPa (after 1 day of curing). This would
represent the early support that would be achieved by the immediate application of
shotcrete behind the advancing face. A third analysis was performed for a support
pressure pi = 0.8 MPa. This is approximately the support which can be achieved in
this size of tunnel by a 75 mm thick shotcrete lining with a uniaxial compressive
strength of 35 MPa (cured for 28 days). The results of these analyses are summarised
graphically in Figures 11.16 and 11.17.
Figures 11.16 and 11.17 show that the size of the plastic zone and the tunnel
deformation can be represented by lognormal distributions. As would be expected,
the mean values for the size of the plastic zone and the magnitude of the sidewall
displacements are reduced significantly by the installation of support.
What is surprising is the dramatic reduction in the standard deviations with
increasing support pressure. This is because of the strong dependence of the size of
the plastic zone upon the difference between the critical pressure pcr and the support
pressure pi. A detailed discussion on this dependence is beyond the scope of this
technical note and is the subject of ongoing research by the author.
From the results of the analysis described above it is evident that the installation of
a relatively simple support system is very effective in controlling the behaviour of
this tunnel. Without support there is an approximate 50% probability of severe
instability and possible collapse of the tunnel. A plastic zone diameter of 15 m and a
tunnel closure of 50 mm in a 5 m diameter tunnel would certainly cause visible signs
of distress. The fact that a relatively thin shotcrete lining can control the size of the
192
plastic zone and the closure of the tunnel provides confirmation of the effectiveness
of support.
A word of warning is required at this point. The example described above is for a 5
m diameter tunnel at a depth of approximately 100 m below surface. For larger
tunnels at greater depths, the plastic zone and the displacements can be significantly
larger. The demands on the support system may be such that it may be very difficult
to support a large tunnel in poor ground at considerable depth below surface.
support pressure = 0.8 MPa

lognornal distribution
mean = 1.31, stdev = 0.11
4.0

lognormal distribution
mean = 1.94, stdev = 0.30
Probability
3.0
no support
mean 2.88, stdev = 0.75
2.0
1.0
0.0
1
Plastic zone radius / tunnel radius

2.5
5.0
7.5
10.0
12.5
Plastic zone radius - m
Figure 11.16: Lognormal distributions representing the range of

plastic zone radii for different supporting pressures.
mean = 0.0016, stdev = 0.0005
1000

mean = 0.004, stdev = 0.0019
Probability
750
no support
mean = 0.0102, stdev = 0.0079
500
250
0
0.00
0.01
0.02
0.03
0.04
Tunnel deformation / tunnel radius
25
50
75
100
Tunnel deformation - mm
Figure 11.17: Lognormal distributions representing the range of

tunnel deformations for different support pressures.
Practical examples of rock mass property estimates

11.9.5
193
Conclusions
The uncertainty associated with estimating the properties of in situ rock masses has a
significant impact or the design of slopes and excavations in rock. The examples that
have been explored in this section show that, even when using the best estimates
currently available, the range of calculated factors of safety or tunnel behaviour are
uncomfortably large. These ranges become alarmingly large when poor site
investigation techniques and inadequate laboratory procedures are used.
Given the inherent difficulty of assigning reliable numerical values to rock mass
characteristics, it is unlikely that accurate methods for estimating rock mass
properties will be developed in the foreseeable future. Consequently, the user of the
Hoek-Brown procedure or of any other equivalent procedure for estimating rock mass
properties should not assume that the calculations produce unique reliable numbers.
The simple techniques described in this section can be used to explore the possible
range of values and the impact of these variations on engineering design.
11.10 Practical examples of rock mass property estimates
The following examples are presented in order to illustrate the range of rock mass
properties that can be encountered in the field and to give the reader some insight of
how the estimation of rock mass properties was tackled in a number of actual
projects.
11.10.1 Massive weak rock
Karzulovic and Diaz (1994) have described the results of a program of triaxial tests
on a cemented breccia known as Braden Breccia from the El Teniente mine in Chile.
In order to design underground openings in this rock, attempts were made to classify
the rock mass in accordance with Bieniawskis RMR system. However, as illustrated
in Figure 11.18, this rock mass has very few discontinuities and so assigning realistic
numbers to terms depending upon joint spacing and condition proved to be very
difficult. Finally, it was decided to treat the rock mass as a weak but homogeneous
almost intact rock and to determine its properties by means of triaxial tests on large
diameter specimens.
A series of triaxial tests was carried out on 100 mm diameter core samples,
illustrated in Figure 11.19. The results of these tests were analysed by means of the
regression analysis presented in Section 11.3. Back analysis of the behaviour of
underground openings in this rock indicate that the in-situ GSI value is approximately
75. From the spreadsheet presented in Figure 11.7 the following parameters were
obtained:

Hoek-Brown constant
Hoek-Brown constant
ci
mi
GSI
s
51 MPa
16.3
75
0.062
Friction angle
Cohesive strength
Deformation modulus
c
Em
42
4.32 MPa
30000 MPa
194
Figure 11.18: Braden Breccia at El Teniente Mine in Chile.

This rock is a cemented breccia with practically no joints. It
was dealt with in a manner similar to weak concrete and tests
were carried out on 100 mm diameter specimens illustrated
in Figure 11.19.
Fig. 11.19. 100 mm diameter by 200 mm long specimens of

Braden Breccia from the El Teniente mine in Chile.
195
11.10.2 Massive strong rock masses

The Rio Grande Pumped Storage Project in Argentina includes a large underground
powerhouse and surge control complex and a 6 km long tailrace tunnel. The rock
mass surrounding these excavations is a massive gneiss with very few joints. A
typical core from this rock mass is illustrated in Figure 11.20. The appearance of the
rock at the surface is illustrated in Figure 11.6, which shows a cutting for the dam
spillway.
Figure 11.20: Excellent quality core with

very few discontinuities from the
massive gneiss of the Rio Grande project
in Argentina.
Figure 11.21: Top heading

of the 12 m span, 18 m
high tailrace tunnel for the
Rio
Grande
Pumped
Storage Project.
196
The rock mass can be described as BLOCKY/VERY GOOD and the GSI value, from
Table 11.6, is 75. Typical characteristics for the rock mass are as follows:
110 MPa
Hoek-Brown constant
ci
mi
GSI
75
Hoek-Brown constant
mb
7.25
Hoek-Brown constant
0.062
Constant
0.5
17.7
Friction angle
Cohesive strength
43
9.4 MPa
Rock mass compressive

strength
cm
43 MPa
Rock mass tensile

strength
Deformation modulus
tm
-0.94 MPa
Em
42000 MPa
Figure 11.21 illustrates the 8 m high 12 m span top heading for the tailrace tunnel.
The final tunnel height of 18 m was achieved by blasting two 5 m benches. The top
heading was excavated by full-face drill and blast and, because of the excellent
quality of the rock mass and the tight control on blasting quality, most of the top
heading did not require any support.
Details of this project are to be found in Moretto et al (1993). Hammett and Hoek
(1981) have described the design of the support system for the 25 m span
underground powerhouse in which a few structurally controlled wedges were
identified and stabilised during excavation.
11.10.3 Average quality rock mass
The partially excavated powerhouse cavern in the Nathpa Jhakri Hydroelectric project
in Himachel Pradesh, India is illustrated in Fig. 14. The rock is a jointed quartz mica
schist, which has been extensively evaluated by the Geological Survey of India as
described by Jalote et al [23]. An average GSI value of 65 was chosen to estimate the
rock mass properties which were used for the cavern support design. Additional
support, installed on the instructions of the Engineers, was placed in weaker rock
zones.
The assumed rock mass properties are as follows:
Hoek-Brown constant
Hoek-Brown constant
Hoek-Brown constant
Constant
ci
mi
GS
I
mb
s
a
30 MPa
15.6
65
4.5
0.02
0.5
Friction angle
Cohesive strength
Rock mass compressive
strength
Deformation modulus
c
cm
40
2.0 MPa
8.2 MPa
tm
Em
-0.14 MPa
13000 MPa
Two and three dimensional stress analyses of the nine stages used to excavate the
cavern were carried out to determine the extent of potential rock mass failure and to
provide guidance in the design of the support system. An isometric view of one of
the three dimensional models is given in Figure 11.23.
197
Figure 11.22: Partially completed 20 m

span, 42.5 m high underground
powerhouse cavern of the Nathpa
Jhakri Hydroelectric Project in
Himachel Pradesh, India. The cavern is
approximately 300 m below the
surface.
Figure 11.23: Isometric view of the 3DEC4 model of the

underground powerhouse cavern and transformer gallery of the
Nathpa Jhakri Hydroelectric Project, analysed by Dr. B. Dasgupta5.
4
Available from ITASCA Consulting Group Inc., Thresher Square East, 708 South Third Street, Suite 310,
Minneapolis, Minnesota 55415, USA. Fax 1 612 371 4717
5
Formerly at the Institute of Rock Mechanics (Kolar), Kolar Gold Fields, Karnataka, now with of Advanced
Technology and Engineering Services, Delhi. India.
198
The support for the powerhouse cavern consists of rockbolts and mesh reinforced
shotcrete. Alternating 6 and 8 m long 32 mm diameter bolts on 1 x 1 m and 1.5 x 1.5
m centres are used in the arch. Alternating 9 and 7.5 m long 32 mm diameter bolts are
used in the upper and lower sidewalls with alternating 9 and 11 m long 32 mm
rockbolts in the centre of the sidewalls, all at a grid spacing of 1.5 m. Shotcrete
consists of two 50 mm thick layers of plain shotcrete with an interbedded layer of
weldmesh. The support provided by the shotcrete was not included in the support
design analysis, which relies upon the rockbolts to provide all the support required.
In the headrace tunnel, some zones of sheared quartz mica schist have been
encountered and these have resulted in large displacements as illustrated in Figure
11.24. This is a common problem in hard rock tunnelling where the excavation
sequence and support system have been designed for average rock mass conditions.
Unless very rapid changes in the length of blast rounds and the installed support are
made when an abrupt change to poor rock conditions occurs, for example when a
fault is encountered, problems with controlling tunnel deformation can arise.
The only effective way known to the authors for anticipating this type of problem
is to keep a probe hole ahead of the advancing face at all times. Typically, a long
probe hole is percussion drilled during a maintenance shift and the penetration rate,
return water flow and chippings are constantly monitored during drilling. Where
significant problems are indicated by this percussion drilling, one or two diamonddrilled holes may be required to investigate these problems in more detail. In some
special cases, the use of a pilot tunnel may be more effective in that it permits the
ground properties to be defined more accurately than is possible with probe hole
drilling. In addition, pilot tunnels allow pre-drainage and pre-reinforcement of the
rock ahead of the development of the full excavation profile.
Figure 11.24: Large displacements

in the top heading of the headrace
tunnel of the Nathpa Jhakri
Hydroelectric project. These
displacements are the result of
deteriorating rock mass quality
when tunnelling through a fault
zone.
199
11.10.4 Poor quality rock mass at shallow depth

Kavvadas et al (1996) have described some of the geotechnical issues associated with
the construction of 18 km of tunnels and the 21 underground stations of the Athens
Metro. These excavations are all shallow with typical depths to tunnel crown of
between 15 and 20 m. The principal problem is one of surface subsidence rather than
failure of the rock mass surrounding the openings.
The rock mass is locally known as Athenian schist which is a term erroneously
used to describe a sequence of Upper Cretaceous flysch-type sediments including
thinly bedded clayey and calcareous sandstones, siltstones (greywackes), slates,
shales and limestones. During the Eocene, the Athenian schist formations were
subjected to intense folding and thrusting. Later extensive faulting caused extensional
fracturing and widespread weathering and alteration of the deposits.
The GSI values, estimated from Bieniawskis 1976 RMR classification, modified
as recommended by Hoek, Kaiser and Bawden (11995) ranges from about 15 to about
45. The higher values correspond to the intercalated layers of sandstones and
limestones, which can be described as BLOCKY/DISTURBED and POOR (Table
11.6). The completely decomposed schist can be described as DISINTEGRATED and
VERY POOR and has GSI values ranging from 15 to 20. Rock mass properties for
the completely decomposed schist, using a GSI value of 20, are as follows:
Hoek-Brown constant
Geological Strength
Index
Hoek-Brown constant
Hoek-Brown constant
ci
mi
GSI
5-10 MPa
9.6
20
Constant
Friction angle
Cohesive strength
0.55
22.4
0.09-0.18 MPa
mb
s
0.55
0
Rock mass strength

Deformation modulus
cm
Em
0.27-0.53 MPa
398-562 MPa
The Academia, Syntagma, Omonia and Olympion stations were constructed using
the New Austrian Tunnelling Method twin side drift and central pillar method as
illustrated in Figure 11.25. The more conventional top heading and bench method,
illustrated in Figure 11.26, was used for the excavation of the Ambelokipi station.
These stations are all 16.5 m wide and 12.7 m high. The appearance of the rock mass
in one of the Olympion station side drift excavations is illustrated in Figures 11.27
and 11.28.
Numerical analyses of the two excavation methods showed that the twin side drift
method resulted in slightly less rock mass failure in the crown of the excavation.
However, the final surface displacements induced by the two excavation methods
were practically identical.
Maximum vertical displacements of the surface above the centre-line of the
Omonia station amounted to 51 mm. Of this, 28 mm occurred during the excavation
of the side drifts, 14 mm during the removal of the central pillar and a further 9 mm
occurred as a time dependent settlement after completion of the excavation.
According to Kavvadas et al (1996), this time dependent settlement is due to the
dissipation of excess pore water pressures which were built up during excavation. In
the case of the Omonia station, the excavation of recesses towards the eastern end of
200
the station, after completion of the station excavation, added a further 10 to 12 mm of

vertical surface displacement at this end of the station.
Figure 11.25: Twin side drift and central

pillar excavation method. Temporary support
consists of double wire mesh reinforced 250 300 mm thick shotcrete shells with
embedded lattice girders or HEB 160 steel
sets at 0.75 - 1 m spacing.
Figure 11.26: Top heading and bench method

of excavation. Temporary support consists of
a 200 mm thick shotcrete shell with 4 and 6
m long untensioned grouted rockbolts at 1.0 1.5 m spacing
Figure 11.27: Side drift in the

Athens
Metro
Olympion
station excavation that was
excavated by the method
illustrated in Figure 11.25. The
station has a cover depth of
approximately 10 m over the
crown.
201
Figure 11.28: Appearance of the very poor quality Athenian Schist at the face of
the side heading illustrated in Figure 11.27.
11.10.5 Poor quality rock mass under high stress

The Yacamb Quibor tunnel in Venezuela is considered to be one of the most
difficult tunnels in the world. This 26 km long water supply tunnel through the Andes
is being excavated in sandstones and phyllites at depths of up to 1200 m below
surface. The graphitic phyllite is a very poor quality rock and gives rise to serious
squeezing problems which, without adequate support, result in complete closure of
the tunnel. A full-face tunnel-boring machine was completely destroyed in 1979 when
trapped by squeezing ground conditions.
At its worst, the graphitic phyllite has an unconfined compressive strength of about
15 MPa and the estimated GSI value is about 24. Typical rock mass properties are as
follows:
Hoek-Brown constant
Hoek-Brown constant
Hoek-Brown constant
ci
mi
GSI
mb
s
15 MPa
10
24
0.66
0
Constant
Friction angle
Cohesive strength
Rock mass strength
Deformation modulus
c
cm
Em
0.53
24
0.34 MPa
1 MPa
870 MPa
Various support methods have been used on this tunnel and only one will be
considered here. This was a trial section of tunnel, at a depth of about 600 m,
constructed in 1989. The support of the 5.5 m span tunnel was by means of a
complete ring of 5 m long, 32 mm diameter untensioned grouted dowels with a 200
mm thick shell of reinforced shotcrete. This support system proved to be very
effective but was later abandoned in favour of yielding steel sets (steel sets with
sliding joints) because of construction schedule considerations.
202
Examples of the results of a typical numerical stress analysis of this trial section,
carried out using the program PHASE2, are given in Figures 11.28 and 11.29. Figure
11.28 shows the extent of failure, with and without support, while Figure 11.29 shows
the displacements in the rock mass surrounding the tunnel. Note that the criteria used
to judge the effectiveness of the support design are that the zone of failure
surrounding the tunnel should lie within the envelope of the rockbolt support, the
rockbolts should not be stressed to failure and the displacements should be of
reasonable magnitude and should be uniformly distributed around the tunnel. All of
these objectives were achieved by the support system described earlier.
8 MPa
12 MPa
In situ stresses
Failure zone
with support
Deformed
profile with
no support
Failure zone with

no support
Figure 11.28: Results of a numerical analysis

of the failure of the rock mass surrounding
the Yacambu-Quibor tunnel when excavated
in graphitic phyllite at a depth of about 600
m below surface.
Figure 11.29: Displacements in the rock mass

surrounding the Yacambu-Quibor tunnel.
The maximum calculated displacement is
258 mm with no support and 106 mm with
support.
11.10.6 Slope stability considerations

When dealing with slope stability problems in rock masses, great care has to be taken
in attempting to apply the Hoek-Brown failure criterion, particularly for small steep
slopes. As illustrated in Figure 11.30, even rock masses that appear to be good
candidates for the application of the criterion can suffer shallow structurally
controlled failures under the very low stress conditions which exist in such slopes.
As a general rule, when designing slopes in rock, the initial approach should
always be to search for potential failures controlled by adverse structural conditions.
These may take the form of planar failures on outward dipping features, wedge
failures on intersecting features, toppling failures on inward dipping failures or
complex failure modes involving all of these processes. Only when the potential for
structurally controlled failures has been eliminated should consideration be given to
treating the rock mass as an isotropic material as required by the Hoek-Brown failure
criterion.
203
Figure 11.30: Structurally controlled

failure in the face of a steep bench in a
heavily jointed rock mass.
Figure 11.31 illustrates a case in which the base of a slope failure is defined by an
outward dipping fault which does not daylight at the toe of the slope. Circular failure
through the poor quality rock mass overlying the fault allows failure of the toe of the
slope. Analysis of this problem was carried out by assigning the rock mass at the toe
properties that had been determined by application of the Hoek-Brown criterion. A
search for the critical failure surface was carried out utilising the program XSTABL6
which allows complex failure surfaces to be analysed and which includes facilities for
the input of non-linear failure characteristics as defined by equation 11.2.
Tension crack
Failure through weak rock
mass
Groundwater
surface
Failure on fault dipping at 25

Failure through weak rock mass
at slope toe
Figure 11.31: Complex slope failure

controlled by an outward dipping basal
fault and circular failure through the poor
quality rock mass overlying the toe of the
slope.
100 m
Available from Interactive Software Designs, Inc., 953 N. Cleveland Street, Moscow, Idaho, USA 83843, Fax +
1 208 885 6608

Chapter 11 of Rock Mass Properties

Uploaded by

Copyright:

Available Formats

Chapter 11 of Rock Mass Properties

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 11 of Rock Mass Properties

Uploaded by

Copyright:

Available Formats

11

Rock mass properties

Chapter 11: Rock mass properties

where A and B are material constants

Intact rock properties

Intact rock properties

hardened and ground steel

rock specimen with ground ends and

oil inlet - maximum pressure

strain gauges - if required

rubber sealing sleeve

Laboratory tests should be carried out at moisture contents as close as possible to

Chapter 11: Rock mass properties

where x = 3' and y = (1' 3' ) 2

Intact rock properties

Compressive strength - MPa

Angle of schistosity to loading direction

Figure 11.2: Influence of loading direction on the strength of graphitic

Chapter 11: Rock mass properties

Table 11.2: Field estimates of uniaxial compressive strength.

Specimen can only be Fresh basalt, chert,

Cannot be scraped or Claystone, coal,

Can be peeled with a

Chalk, rocksalt, potash

Crumbles under firm

Indented by thumbnail Stiff fault gouge

* Grade according to Brown (1981).

Intact rock properties

Chapter 11: Rock mass properties

Influence of sample size

Uniaxial compressive strength of specimen of diameter d

Figure 11.3 Influence of specimen size on the strength of

Table11.4 Peak strength of Moura DU coal in terms of the

Chapter 11: Rock mass properties

Sample diameter (mm)

Axial strength 1 ' MPa

Confining pressure 3' MPa

Figure 11.4 Peak strength for Australian Moura coal.

Geological strength Index

Intact rock specimens

One joint set - do not use

Two joint sets - do not

Many joints - use equation

Heavily jointed rock mass

11.5 Geological strength Index

Chapter 11: Rock mass properties

Geological strength Index

Chapter 11: Rock mass properties

Geological strength Index

Chapter 11: Rock mass properties

Most geotechnical software is written in terms of the Mohr-Coulomb failure criterion

For the GSI > 25, when a = 0.5:

For GSI < 25, when s = 0:

The tensile strength of the rock mass is calculated from:

Chapter 11: Rock mass properties

The corresponding cohesive strength ci' is given by:

ci' = 'ni tan i'

and the corresponding uniaxial compressive strength of the rock mass is :

2ci' cos i'