You are on page 1of 215

Dissertation for the degree

Doctor Scientiarum
Arild Palmström

RMi – a rock mass


characterization system
for rock engineering
purposes

DEPARTMENT OF GEOLOGY
FACULTY OF MATHEMATICS
AND NATURAL SCIENCES
UNIVERSITY OF OSLO • 8/1995
i

PREFACE

This work is a contribution to the use of geological parameters in rock engineering and design. It
introduces a new system for collecting and using these parameters. Block size is used as a main
input parameter in a Rock Mass index (RMi), which characterizes the strength of rock masses.
Several methods to measure block size and other jointing characteristics have been outlined, in
addition to applications of the RMi in rock engineering.

The work is structured into the following main topics:

1. Description of important rock mass features and methods to characterize and quantify them.
This is presented in Chapters 1 - 3 and Appendices 1 - 5.
As much as possible of the existing methods for investigation and description have been used. A few
improvements of methods to quantify rock masses have been developed. Correlations between the
most common methods for joint measurements have been worked out, making acquisition of
geological information easier.

2. Selected parameters are combined in a general rock mass index (RMi) which characterizes the
compressive strength of continuous rock masses. This is found in Chapters 4 - 5 and
Appendix 6.
Although the profession has developed many qualitative and numerical methods for classification of
rock masses to assist the rock engineer, few methods have been directed towards the material that rock
masses constitute.

3. The application of RMi in various fields of rock engineering and rock mechanics is described
in Chapter 6 - 8 and Appendix 7.
RMi can be applied to assess rock support and tunnel boring (TBM) progress rate in various types of
ground. Other applications of RMi are input to existing classification systems, as well as in rock
mechanical calculations like the Hoek-Brown failure criterion and the ground response curves.

4. A contribution to communication between geologists, engineering geologists, rock


mechanicians, and rock engineers by introducing defined rock mass descriptions is presented
in part of Chapter 8.
"This would permit correlation of geological conditions between different locations and eventually
lead to more reliable methods of rock engineering. The common goal should be to provide practical
and realistic input methods relating to rock mechanical and rock design works, which could convey
the same meaning to those involved in rock construction and utilization, i.e., the contractor, the design
engineer and the engineering geologist or geotechnical engineer." (Wickham et al., 1972).

5. Some guidelines to quantify qualitative descriptions of rock masses have been worked out in
Appendix 3.
It is shown how qualitative rock mass descriptions can be 'translated' into numbers which can be used
in the RMi.

The RMi system, which has been used for different purposes during a couple of years, has shown
promising results. It can, however, be further refined and developed to also cover other fields
connected to rock engineering and construction. This opens for possible improvements in rock mass
characterization, which give benefits in rock engineering used in planning, design and follow up of
constructions in rock.
Acknowledgement

This study, which has lasted for about 4 years, has been full of challenges and problems that had to
be solved. It would not have been possible to accomplish without all the support, interest and help
from several persons and institutions.

The work has been funded by the Royal Norwegian Council for Scientific and Industrial Research
(NTNF) - from 1.1.93 the Research Council of Norway (NFR) - and I greatly appreciate their
cooperation in providing the funding.

The practical part would not have been possible without the help and support from the Norwegian
Geotechnical Institutet. NGI's library staff, Wenche Enersen, Oddny Feragen, and Liv Ström have
helped me in providing most of the papers listed in the references. Special thanks to Lloyd
Tunbridge who has gone through all chapters and made valuable comments, and to Farrokh Nadim
who has helped to develop some of the methods in Chapter 5. Thanks are also due Tore Lasse By,
Fredrik Löset, Eystein Grimstad and Rajinder Bhasin for sharing some of their experiences.

I also wish to express my sincere gratitude to all individuals who helped and contributed to the
completion of this thesis. Among them should be especially mentioned:
- Bengt Leijon and Norbert Krauland for discussions in Luleå, Sweden, and for providing me
valuable material on large scale tests in the mines of Laisvall and Långsele.
- Thomaso Lardelli, Chur, Switzerland for valuable and interesting discussion and for
providing material on tunnelling standards in Switzerland.
- Erik Dahl Johansen and Rune Rossi of Statkraft A/S for providing rock samples and
information on the experience gained during tunnel construction at Svartisen Power Plant.

Special thanks to my thesis advisor professor Arild Andresen at the Institute of geology, University
of Oslo for providing information, material, and encouragement during the work, and to Professor
Kaare Höeg for his continuous interest and his great help and valuable comments.

And last but not least, I am grateful to my family who gave me confidence and motivation with
their patience and understanding. Their support made it possible to carry out work at home without
being interrupted in evening after evening and their enthusiasm was most important during periods
of doubt and low inspiration.
)URP3DOPVWU|P$50L±DURFNPDVVFKDUDFWHUL]DWLRQV\VWHPIRUURFNHQJLQHHULQJSXUSRVHV
3K'WKHVLV2VOR8QLYHUVLW\1RUZD\S

6800$5<

,QYLHZRIWKHVFDUFLW\RIUHOLDEOHLQIRUPDWLRQRQWKHVWUHQJWKRIURFNPDVVHVDQGRIWKHYHU\KLJK
FRVWRIREWDLQLQJVXFKLQIRUPDWLRQLWLVXQOLNHO\WKDWDFRPSUHKHQVLYHTXDQWLWDWLYHDQDO\VLVRIURFN
PDVVVWUHQJWKZLOOHYHUEHSRVVLEOH6LQFHWKLVLVRQHRIWKHNH\TXHVWLRQVLQURFNHQJLQHHULQJLWLV
FOHDUWKDWVRPHDWWHPSWVKRXOGEHPDGHZKDWHYHULQIRUPDWLRQLVDYDLODEOHWRSURYLGHVRPHIRUPRI
JHQHUDOJXLGDQFHRQUHDVRQDEOHWUHQGVLQURFNPDVVVWUHQJWK
Evert Hoek and Edwin T. Brown (1980)

A rock mass is an inhomogeneous material built up of smaller and larger blocks/pieces composed of
rock material. A great variety exists, both in the composition of the rock material and in the struc-
ture and occurrence of its discontinuities. The rock mass is, in fact, a material exhibiting a wider
range in structure, composition and mechanical properties than most other construction materials.
Reliable tests of strength properties of such a complex material are impossible, or so difficult to
carry out with today’s technique, that rock engineering is based mainly on input data determined
from observations and simplified measurements of the rock mass.

As the quality of the input data determines to a great extent the success of the design, there is a
general demand for better methods in rock mass descriptions, and practical guidelines to carry out
numerical characterizations.

Faults and weakness zones often exhibit special types of rock masses in the crust. They may consist
of weathered or altered rocks found as decomposed rock and/or as clay filling, often together with
crushed rock. Such zones may cause excavation problems, which are of no concern in the volumes
between the zones. Therefore, weakness zones have been given special attention in this work.

7KHSUREOHPZLWKXQFHUWDLQWLHVLQURFNHQJLQHHULQJ

The great spatial variability and large volumes involved in rock mass utilizations result in that only
a limited number of measurements can be made in a characterization. %HIRUH construction, the
subsurface, therefore, has to be described by a limited number of imprecisely known parameters.
Considerable uncertainties may be introduced from the interpretation and extrapolation made to
describe the geological setting. Also, the fact that horizontal weakness zones and other features,
which do not outcrop, may be overlooked, is added to these errors.

Thus, although extensive field investigation and good quality descriptions will enable the
engineering geologist to predict the behaviour of a tunnel more accurately, it cannot remove the risk
of encountering unexpected features. A good quality characterization of the rock mass will,
however, in all cases except for incorrect interpretations, improve the quality of the geological input
data to be applied in evaluations, assessments or calculations and hence lead to better designs.

After the rock mass has been "RSHQHG" during excavation, the actual rock masses may be studied. In
these cases the quality of the input data used in evaluations, calculations and modelling mainly
depends on the way they are measured and characterized.
ii

The main purpose of this work has been to improve the quality in rock engineering by providing
better input data for characterizing the rock mass based on selected, well defined geological
parameters. Methods for field descriptions of outcrops, as well as logging of drill cores and
geophysical measurements, have been included.

INTACT ROCK
52&.0$66,1'(; A general, numerical
characterization
JOINTING
50L of a rock mass

PROJECT Input of features


RELATED of importance for
FEATURES the actual design

ROCK DESIGN AND


ENGINEERING IN: RMi parameters applied
- rock support in practical design
- TBM excavation and engineering
- rock blasting *)
- rock fragmentation *)

applied as Other
input in other used in applications
engineering communication of RMi
methods
QRWLQFOXGHGLQWKLVZRUN

Fig. 1 General layout of the RMi system for rock engineering and design

$PHWKRGRORJ\IRUVWUHQJWKFKDUDFWHUL]DWLRQRIURFNPDVVHV

As a rock mass basically is composed of intact rock penetrated by joints, its behaviour depends not
only on the properties of the intact material and the discontinuities separately, but also on the way
they are combined. Thus, there are several complications which arise when the parameters of a rock
mass shall be described:
1. The intact rock is inhomogeneous and variable.
2. The discontinuities are diverse in nature (i.e. their variations involve more than ten
characteristics).
3. The geometrical arrangement of discontinuities is infinitely variable.

Fig. 2 The main parameters in a rock mass.


iii

Because of these structural variations and the often large volumes involved, the properties of a rock
mass have to be determined from observations backed by laboratory tests of small specimens. Thus,
the evaluations and assessments applied in engineering are largely empirical, typically applied in the
classification schemes.

The rock mass index (RMi) developed in this work is a general characterization of rock masses in
which their main parameters are included. In principle it is based on the reduction in the strength of
the intact rock due to the presence of joints. This is expressed as
RMi = σc × JP

where σc is the uniaxial compressive strength of intact rock measured on 50 mm samples;


JP is the jointing parameter, i.e. the reduction factor from jointing. It consists of :
- the degree of jointing (given as block size); and
- the joint characteristics, representing the joint wall roughness and alteration, as well as the
size of the joint.

jointing features

Fig. 3 Combination of the selected parameters in the Rock Mass index.

It is practically impossible to carry out triaxial or shear tests on rock masses at a scale similar to that
of underground excavations. As the rock mass index, RMi, is meant to express the compressive
strength of a rock mass, a calibration has been carried out. Data from 8 large-scale tests and 1 back
analysis of a prototype situation have been applied. The data contained test results of the uniaxial
compressive strength and the inherent parameters of the rock mass. The values for Vb and jC
have been plotted in Fig. 4, and the lines representing constant jC have been drawn. In this way
Vb and jC have been combined to express the jointing parameter, JP. From the lines the jointing
parameter can be expressed as
JP = 0.2 × jC × Vb D

where jC = the joint condition factor, Vb = the block volume, and D = 0.37 jC - 0.2

Significant VFDOHHIIHFWV are generally involved when a ’sample’ is enlarged from laboratory size to
field size. From the calibration described above, RMi is tied to large samples where the scale effect
has been included in JP. For massive rock masses, however, where the jointing parameter JP ≈ 1,
the scale effect for the uniaxial compressive strength must be accounted for, as it is related to 50
mm sample size.
iv

This scale effect is expressed as


σcf = σc50 (0.05/Db) 0.2

where σc50 is the uniaxial compressive strength for 50 mm sample size, and
Db is the equivalent block diameter.

The approximate block diameter may be found from Db = 3 Vb , or, where a pronounced joint set
occurs, simply by applying the spacing (S1) of this set (Db = S1). The scale effect of compressive
strength has been included in Fig. 4 to assess the jointing parameter.
5 H G  ID FW R U IR U σF
0.1 0.2 0.3 0.5 0.7 1
scale effect of compressive 100

massive rock
strength ( σ c ) for massive rock
10

P

%ORFN YROXPH 9E
1

0.1

10


P 1
G
0.1

10

M&          P
1 F

0.1
2 3 5 7 2 3 5 7 2 3 5 7 2 3 5 7 2 3 5 7
0.00001 0.0001 0.001 0.01 0.1 1
- R L Q W L Q J S D U D P H W H U - 3

Fig. 4 Diagram for calculating the jointing parameter from block volume (Vb) and joint condition factor (jC).
Example shown: from input values Vb = 0.3 m3 and jC = 2 the value of JP = 0.2. In massive rock the
approximate scale effect is shown by the dotted line and its reduction factor.

RMi may be considered as a quality index of the rock mass as a construction material. This is
similar to the classification used for concrete (which is based on its uniaxial compressive strength).
Both hard rocks and soft rocks can be included in the RMi characterization of rock masses for rock
engineering.

7KHSDUDPHWHUVXVHGLQWKH50L
The GHJUHHRIMRLQWLQJ (i.e. the quantity of joints) is regarded as a main feature in the strength and
behaviour of a rock mass. By characterizing this parameter by the block volume (Vb) it is possible
to describe most variations in jointing, and at the same time include its three-dimensional character.

Direct description of the block volume in the field is especially useful where small blocks or irregu-
lar jointing occur; further it may improve core logging of crushed rock where the size of the
particles is small enough to be observed. Mathematical correlations between several methods for
characterizing the quantity of joints (see Fig. 5) have been worked out.

Ideally, all the FKDUDFWHULVWLFV included in the joint condition factor, i.e.:
joint roughness factor (jR)
jC = × (joint size factor (jL))
joint alteration factor (jA)
v

should be measured accurately; jR and jA from shear strength or friction tests, and jL from the
measured length and termination of the joint. Such measurements of the joints would generally be
either extremely time-consuming, or in most cases, practically impossible to carry out. Generally
only a limited amount of the joints can be characterized in a rock mass, and simplified methods
have to be applied. Therefore, the parameters in the joint condition factor have been given ratings,
which can be determined from defined descriptions. The proposed RMi system has some features
similar to those of the Q-system (Barton et al., 1974). Thus, jR and jA are almost the same as Jr
and Ja in the Q-system, The joint size and continuity factor (jL) is introduced in RMi as it often has
a significant impact on rock mass behaviour.

7+ ( 48$1 7,7< 2) -2,1 76 

0($ 685( ',1$ 1$ 5($


Y
-

7+ ( 18 0%(5 2) -2,1 76 
0($ 685( ',1$ '5 ,// 7
1
&25 (25 $ 6&$ 1 /,1(
8
2
&

7
6( ,60,&5( )5$ &7,21
1
,
9( /2&,7,(6 ,15 2&. %/2&.92/80(
2
- 9E

&
,
5
5 2&.  48$ /,7< 7
(
'(6,*1$7,21 54'
0
8
/
2
,1 '5,// &25(6 9
:(,* +7('

-2,1 7 ' (1 6,7<

0($ 685( 0(176 ,1 $1$5($

%/2&.
0($68 5(' -2,1 7 63$& ,1*6
6+ $3(

)$&725

0($ 68 5(' /(1*7+ $1 ' 7+,&.1(6 6 2) %/2 &.


β = transition formula

*) = introduced in this work

Fig. 5 Developed correlations between various methods to measure the degree of jointing and the block size.

J J
Q
L Q
L
F F
D D
S S
V
 V

W
P V (ta
X H bu
L O lar
O
G D bl
H oc
ks
P P
V
)

2
,
7
$
5

(prismatic blocks)

ODUJHVWVSDFLQJ
(equidimensional blocks) 5$7,2
VPDOOHVWVSDFLQJ

Fig. 6 The block shape factor determined from the joint spacings provided by 3 joint sets intersecting at right
angles. Example: For joint spacings 0.2 m, 0.8 m, and 3 m the spacing ratios are 4 and 15, which gives β =
135 corresponding to a ’long&flat’ block type.
vi

The RMi can also be found by simplified measurements resulting, of course, in less accurate
estimates. These simplifications may be:
- The compressive strength can be determined from simple field tests or from rock descriptions as
is described in Appendix 3.
- If the condition of the joints has not been measured, it may be assumed that jC = 1.75
("common" joint condition) for which the jointing parameter can be expressed as JP = 0.26 Vb
1/3
, or even more simply where the 3 joint sets have similar spacings (Sj): JP = 0.26 Sj.

The strong influence that the number of joint sets has on the behaviour of rock masses is expressed
in the block shape factor (β), which has been introduced to characterize the main types of blocks. It
is applied in the correlation between the volumetric joint count (Jv) and block volume as
Vb = β × Jv -3

β may be found from Fig. 6 or more roughly from


β = 20 - 7 Smax /Smin

where Smax and Smin are the longest and shortest dimensions of the block.

The methods described outline how the various parameters applied in RMi can be determined either
from commonly used measurements and/or from measurements developed in this work. In variable
rock masses, for instance where the block size varies and/or the joint condition factor is different for
the various joint sets, some guidelines have been given on how to combine these variables to find a
representative joint condition factor (JP).

7KHXVHRIWKH50LLQURFNHQJLQHHULQJ

As a relative strength index for rock masses, the RMi expresses numerically the quality of the rock
mass. RMi may be applied in various methods used in practical rock engineering. In addition, some
of the parameters in RMi can be used individually in classification systems on stability, where they
may improve the input and/or because they may be easier and or more accurately characterized.

TABLE 1 MAIN TYPES OF WORKS CONNECTED TO ROCKS AND ROCK MASSES


WITH INDICATION WHERE THE RMi CAN BE APPLIED.
7<3( $&78$/352&(662586(
- drilling (small holes)
- boring (TBM boring, shaft reaming) *)
- blasting*)
Treatment of rocks - fragmentation*)
- crushing
- grinding
- cutting*)
- rock aggregate for concrete etc.
Application of
- rock fill
rocks
- building stone
Utilization of rock - in underground excavations (tunnels, caverns, shafts) *)
masses - in surface cuts/slopes/portals *)
- excavation works
Construction works
- rock support *)
in rock masses
- water sealing
*)
Areas where the system is of particular interest.
vii

When applied in rock engineering, as well as in construction and utilization purposes, the RMi
value or its parameters are adjusted for local features of importance to the engineering purpose as
indicated in Fig. 1. Table 1 shows also other areas for applying RMi.

As the RMi expresses the inherent properties of rock masses, it is possible to compare RMi values
from various locations directly. In this way RMi may contribute to improved communication
between people involved in rock construction. Fig. 7 shows the main applications of RMi and/or its
parameters in rock mechanics and rock engineering.

&RPPXQLFDWLRQ

$33/,&$7,21 ,1 $33/,&$7,21 ,1 ,1387 ,1


52&. (1*,1((5,1* 6<67(06 )25 52&. 0(&+$ 1,&6
52&. 6833257
(9$/8$7,21
) U DJ P HQ WD WL R Q +RHN%URZQ
IDLOXUH
D Q G E O D V W L Q J
FULWHULRQ

505 V\VWHP
6WDELOLW\ DQG 1XPHULFDO
URFN VXSSRUW
PRGHOOLQJ
FDOFXODWLRQV
4  V\VWHP
'HIRUPDWLRQ
7%0 SURJUHVV PRGXOXVRI
URFNPDVVHV
HYDOXDWLRQV 1$70

*URXQG
UHVSRQVH
FXUYHV

Fig. 7 Various applications of RMi and its parameters.

50LDSSOLHGDVLQSXWWRWKH+RHN%URZQIDLOXUHFULWHULRQ
As RMi expresses the relative compressive strength of rock masses (σcm), it represents a special
case in the Hoek-Brown failure criterion of rock masses
σcm = σc ×V ½

where V is a constant representing the rock mass properties.

The determination of JP = V½ from block size and joint characteristics introduces an easier and more
direct and accurate method to find the value of V than the method presented by Hoek and Brown.
Also the P factor in Hoek-Brown failure criterion for rock masses can easily be determined from
JP and the type of rock. This improvement will indirectly be of benefit in rock mechanics as these
factors (V and P) can be applied to assess the shear strength for continuous rock masses as well as
input to ground response curves.

50LDSSOLHGLQFODVVLILFDWLRQV\VWHPV
RMi offers interesting possibilities to quantify the behaviouristic descriptions applied in the NATM.
From the shear strength and - when RMi has been further developed - the deformation modulus of
rock masses, it is possible to determine the ’Kennlinie-Bemessungsverfahren’. This is a type of
Fenner-Pacher curves (ground response curves) from which displacements as well as rock support
can be estimated in the NATM.
viii

As the RMR system of Bieniawski (1973) is based on the sum of several parameters, while RMi and
partly also the parameters involved in it are exponentially characterized, it is difficult to directly
apply RMi in RMR. The NGI Q system has, however, a similar structure and partly the same
parameters as RMi. It is the classification system, which is most similar to RMi, and in which the
parameters applied in RMi can best be utilized. If block volume (Vb) is applied instead of RQD/Jn,
the author feels that the Q system would be significantly improved. Also the use of the joint
condition factor, jC, instead of Jr/Ja may improve Q because jC includes more of the joint
characteristics.

50LDSSOLHGLQDVVHVVPHQWRIURFNVXSSRUW

Whereas the stability of a tunnel opening in a FRQWLQXRXV material can be related to the intrinsic
strength and deformation properties of the bulk material, stability in a GLVFRQWLQXRXV material
depends primarily on the character of the joints and the block size. To assess stability in
underground openings the ground has, therefore, been divided into continuous and discontinuous
ground, determined by the continuity factor, CF = tunnel diameter/block diameter

Discontinuous ground occurs where CF = approx. 5 - 100, else the ground is continuous. In this
manner, particle, fragment, block size, or joint spacing becomes indicative of tunnel behaviour.

rock mass characterization

5 2& .67 5 (1 *7+


5 2 &.0$ 6 6,1 ' ( ; 5 2 & .675 ( 66( 6

- 2,1 7&21 ' ,7 , 21


JOINT ING
5 0L 6+ $3( 2)
% / 2& .92 /8 0(
PARAMETER
7 +( 78 11 (/

BLOCK DIAM ETER


& 21 7 , 1 8 ,7< 2) continuous
7 + ( 5 2& . 0 $ 66
7 8 11 ( /', $ 0( 7 (5
discontinuous

support chart for jointed rocks

7
2 5
, 2
3
7 3
$ 8
6
5 
 <
 9
( $ 7
( 5
= + 2
, 3
6 3
8
6

2
1
WIDTH OF
THE ZONE
* 5 2 8 1 '& 2 1 ' ,7 , 2 1)$ & 72 5

675 (6 6

/ (9 (/

RMi OF ADJACENT
ROCK MASSES
input parameter additional input parameter
for weakness zones

Fig. 8 The application of the RMi to assess rock support in continuous and discontinuous ground. Weakness zones
are treated in the same way as discontinuous ground by adjustment to the ’size ratio’ and the ’ground
condition factor’.
ix

7KHFRPSHWHQF\RIFRQWLQXRXVJURXQG
The competency of continuous ground expresses whether the ground surrounding an underground
opening is overstressed or not. It is expressed as the ratio
Cg = RMi/σθ

where σθ is the tangential stresses set up in rock masses surrounding the opening. It depends on the overall
stress level, the stress anisotropy and the shape of the opening. A method to estimate the rock
stresses and σθ has been outlined.

In FRPSHWHQW ground the compressive strength of the rock mass is higher than the rock stresses.
Instability in this type of ground is mainly caused by joints or other structural weaknesses.
,QFRPSHWHQW ground appears where the rock stresses are higher than the compressive strength of the
rock mass. Squeezing develops in deformable rock masses, while rock burst, spalling etc. occurs in
massive, brittle rocks. Here, the application of RMi to assess the competency factor opens for
improved characterization and rock support of these important groups of ground.

'LVFRQWLQXRXV MRLQWHG URFNPDVVHV


Instability of this type of ground is mainly related to loosening and downfalls of one or more blocks
in the tunnel periphery mainly determined by the properties of the joints. As these are governed by
the joint condition, the block size, and the strength of the wall rock, the rock mass index, RMi, is
suitable as a main input to the ground condition factor
Gc = SL × RMi × C

where SL is the stress level factor with ratings between 0.2 and 1.5.
C is the gravity adjustment factor to compensate for the greater stability of a vertical wall
compared to a horizontal roof. It varies between 5 and 1 respectively.

The equivalent block diameter (Db) related to the size of the opening has been selected as the
second parameter in the assessment of rock support in discontinuous ground. In the tunnel periphery
Db depends on the orientation of the joints relative to the tunnel; thus, the ’size ratio’ is expressed as
Sr = (Dt/Db) Co/Nj

where Dt is the diameter of the tunnel, i.e, its span or width (Wt) or height (Ht),
Db is the equivalent block diameter, the shortest side of the block, which often is the spacing of
the main joint set.
Co is a factor for the orientation of the main joint set varying between 1 and 3.
Nj is a factor for the number of joint sets

Simplified expression have been derived for ’common’ hard rock mass conditions, from which the
factors Gc and Sr can easily be determined.

)DXOWVDQGZHDNQHVV]RQHV
Compared to the ’normal’ rock masses, weakness zones and faults form a special type of ground as
their behaviour and requirement for support often are quite different. The ground condition factor
for weakness zones is similar in structure to Gc:
Gcz = SL × RMim × C

where RMim is the modified rock mass index which includes the thickness of the zone and influence
from the adjacent rock masses.
x

Similarly, a size ratio for zones has been selected for weakness zones with width smaller than the
tunnel diameter. It is expressed as

thickness of zone
Sr z = × (orientation of zone)
block diameter (or joint spacing)

50LDSSOLHGLQSUHGLFWLRQRIWKHERULQJUDWHRI7%0V

Full face tunnel boring is highly influenced by the strength of the rock masses. This favours the use
of RMi in assessments of TBM penetration rates. Thus, the ground condition for TBMs is
characterized by the jointing parameter and the compressive strength of intact rock.

The method using RMi parameters for TBM penetration assessment has been based on the NTH
prediction model for TBM excavation.1 The drilling rate index (DRI) which represents the
properties of intact rock, has been replaced by an expression including the uniaxial compressive
strength (σc). The following expression has been developed for PDVVLYH rock masses:
keq = 0.022 E 0.72 × σc -0.43

where E is a factor representing the deformation properties of the rock - varies between 500 and 1000.

Fig. 9 Application of the parameters in RMi in the prediction of TBM boring progress.

In MRLQWHG rock masses the orientation of the main joint set relative to the tunnel axis and the rock
characteristics (E and σc ) have been applied in a TBM jointing factor expressed as:
1
Refer to the method published by the Norwegian Institute of Technology (1994): Fullface boring of tunnels (in
Norwegian). In PR 1-94, Trondheim Norway, 159 pp. An earlier version of the method has been published in English by
Movinkel T. and Johannessen O. (1986): Geologic parameters for hard rock tunnel boring. In Tunnels & Tunnelling,
April 1986, pp. 45-48.
xi

0.06 co E
keq 
JP × c0.3

where co is a correction factor for the angle between the tunnel and the main joint set. Its ratings vary
between 1 and 1.75.

The TBM net advance rate ( I ) is found from the factor keq and the equivalent thrust per cutter
(Meq) using diagrams developed by NTH. It can also be determined from the following mathe-
matical expressions, which have been deduced from the mentioned diagrams:
I = io × RPM × (60/1000) (in m/h)

where io = F × keqG with F = 0.0015 Meq1.5 and G = 25 keq- 0.4 × Meq- 0.7 (for keq ≥ 3.5).

&ORVLQJUHPDUN
Caused by the variability and complex structure of rock masses, one has to accept that the data used
in calculations, evaluations and assessments have limited accuracy. Part of this stems from the way
descriptions of the rock masses are made and combined. The work described herein may lead to
improved quality of this important feature in rock engineering.
i

Dr. thesis Arild Palmstrøm:

RMi - a rock mass characterization system for rock engineering purposes


University of Oslo, Norway; 1995

TABLE OF C O N T E N T S

Preface
Acknowledgement
Summary

1 Introduction ................................................................................................................................... 1-1


1.1 Outline of this work ................................................................................................................. 1-2
2 Rock masses as construction materials ...................................................................................... 2-1
2.1 Rocks and their main features.................................................................................................. 2-3
2.1.1 Fresh rocks........................................................................................................................2.3
2.1.2 The influence from some minerals .................................................................................. 2-4
2.1.3 The effect of alteration and weathering........................................................................... 2-5
2.1.4 Geological names and mechanical properties of rocks ....................................................2.6
2.2 Discontinuities in rock ............................................................................................................. 2-6
2.2.1 Faults ............................................................................................................................... 2-7
2.2.2 Joints and their main features.......................................................................................... 2-8
2.2.3 The main jointing characteristics .................................................................................... 2-9
2.2.4 The rock mass................................................................................................................ 2-10
2.3 Rock mass characterization for design and construction purposes ....................................... 2-11
3 Collection of geo-data - limitations and uncertainties ............................................................. 3-1
3.1 Geo-data found before, during and after excavation ............................................................... 3-2
3.2 Some methods used in geo-data collection .............................................................................. 3-4
3.2.1 Geological observations and mapping ............................................................................ 3-5
3.2.2 Joint surveys .................................................................................................................... 3-7
3.2.3 Core drilling..................................................................................................................... 3-8
3.2.4 Geophysical measurements ........................................................................................... 3-10
3.2.5 Exploratory adits and shafts.......................................................................................... 3-10
3.2.6 Laboratory and field tests .............................................................................................. 3-10
3.3 Uncertainties and errors in geo-data collection...................................................................... 3-11
3.3.1 Uncertainties caused by spatial variability of rock masses........................................... 3-11
3.3.2 Measurement errors ....................................................................................................... 3-12
3.3.3 Model uncertainties ....................................................................................................... 3-14
3.4 Summary ................................................................................................................................ 3-14
4 The combination of geo-data into a rock mass index................................................................ 4-1
4.1 The structure of a rock mass index .......................................................................................... 4-2
4.1.1 The input parameters selected ......................................................................................... 4-2
4.1.2 The Rock Mass index (RMi).......................................................................................... 4-3
4.1.3 The combination of the input parameters........................................................................ 4-4
4.2 Calibration of RMi from known rock mass strength data ....................................................... 4-5
4.3 Numerical values of the input parameters to RMi................................................................. 4-10
4.3.1 The compressive strength of intact rock (s c )................................................................ 4-10
4.3.2 The block volume (Vb).................................................................................................. 4-11
4.3.3 The joint condition factor (jC).......................................................................................4.12
4.4 Possible areas of application of the RMi............................................................................... 4-18
4.5 Discussion .............................................................................................................................. 4-19
4.5.1 Limitations of the RMi .................................................................................................. 4-19
4.5.2 Other similar rock mass characterization methods ....................................................... 4-20
ii

5 Rock masses characterized by the RMi...................................................................................... 5-1


5.1 On continuous and discontinuous rock masses ....................................................................... 5-1
5.2 Zoning of the rock masses into structural regions ................................................................... 5-3
5.3 Principles in characterizing the variations in rock masses ...................................................... 5-4
5.3.1 Variations in the rock material ........................................................................................ 5-4
5.3.2 Variations in the jointing ................................................................................................. 5-5
5.3.3 Singularities and weakness zones.................................................................................. 5-14
5.3.4 Summary of the possibilities and methods to determine the block
volume or the jointing parameter where the jointing characteristics vary .................... 5-16
6 The use of RMi in design of rock support for underground openings .................................. 6-1
6.1 Stability analyses and rock support design.............................................................................. 6-2
6.2 Instability and failure modes in underground excavations ...................................................... 6-3
6.2.1 Special modes of instability and behaviour related to weakness zones .......................... 6-5
6.2.2 Main types of failure development.................................................................................. 6-6
6.3 The main features influencing underground stability .............................................................. 6-7
6.3.1 The inherent properties of the rock mass ........................................................................ 6-8
6.3.2 The external ground features ........................................................................................... 6-9
6.3.3 The excavation features................................................................................................. 6-10
6.3.4 The time-dependent features ......................................................................................... 6-12
6.3.5 Summary of Section 6.3 ................................................................................................ 6-13
6.4 RMi applied to assess rock support....................................................................................... 6-14
6.4.1 Stability and rock support in continuous materials....................................................... 6-15
6.4.2 Stability and rock support in discontinuous (jointed) materials ................................... 6-29
6.4.3 Stability and rock support of faults and weakness zones ............................................. 6-36
6.4.4 Comments to the RMi method for assessing rock support ........................................... 6-42
7 RMi parameters applied in prediction of tunnel boring penetration ..................................... 7-1
7.1 Factors influencing the TBM performance.............................................................................. 7-2
7.2 Prediction models..................................................................................................................... 7-2
7.2.1 The NTH prognosis model .............................................................................................. 7-3
7.3 The use of RMi parameters to characterize rock masses for TBM......................................... 7-4
7.3.1 The rock material properties............................................................................................ 7-4
7.3.2 The jointing features........................................................................................................ 7-6
7.3.3 Assessment of the net advance of boring ........................................................................ 7-9
7.3.4 Example ......................................................................................................................... 7-11
7.3.5 Discussion of the RMi method for TBM penetration assessment ................................ 7-12
8 Possible other applications of the RMi in rock mechanics and rock engineering ................. 8-1
8.1 Applying RMi to determine the constants in the Hoek-Brown failure criterion ..................... 8-2
8.1.1 The original Hoek-Brown failure criterion...................................................................... 8-2
8.1.2 The modified Hoek-Brown failure criterion.................................................................... 8-4
8.2 RMi used to evaluate shear strength of rock masses............................................................... 8-5
8.3 RMi used in the input to ground response curves ................................................................... 8-7
8.4 RMi used for numerical ground characterization in the NATM ............................................. 8-9
8.4.1 The use of RMi in NATM classification....................................................................... 8-10
8.4.2 RMi used for input to Fenner-Pacher ground response diagrams................................. 8-12
8.5 The use of RMi parameters in classification systems ........................................................... 8-15
8.5.1 Input to the RMR (Geomechanics) system ................................................................... 8-16
8.5.2 Input to the Q-system .................................................................................................... 8-17
8.5.3 Input to other classification systems ............................................................................. 8-18
8.6 A contribution to improved communication.......................................................................... 8-18
8.6.1 Identification chart for geologic materials..................................................................... 8-18
8.7 Possible use of RMi in numerical models.............................................................................. 8-22
9 Discussion and conclusions 9-1
9.1 On the layout of the RMi system ............................................................................................. 9-1
9.1.1 Comparisons with the principles in other rock engineering systems .............................. 9-3
9.2 On the structure of the RMi ..................................................................................................... 9-4
iii

9.3 On the input parameters to RMi .............................................................................................. 9-5


9.3.1 The uniaxial compressive strength of the rock................................................................ 9-6
9.3.2 Jointing ............................................................................................................................ 9-6
9.4 On the variations and uncertainties in rock masses................................................................. 9-7
9.5 Comparison between RMi and other methods used in rock engineering ................................ 9-8
9.5.1 The rock quality designation (RQD) ............................................................................... 9-8
9.5.2 Rock support design systems .......................................................................................... 9-9
9.6 The need for a 'language' in rock mechanics and rock engineering ....................................... 9-12
9.7 Benefits and limitations application of the RMi system ....................................................... 9-13
9.8 Some concluding remarks ..................................................................................................... 9-14
9.8.1 Future developments...................................................................................................... 9-15
10 References .................................................................................................................................... 10-1

APPENDICES

Appendix 1: On joints and jointing


1 Joint characteristics..................................................................................................................A1-3
2 Jointing characteristics.............................................................................................................A1-5
2.1 Joint sets..........................................................................................................................A1-5
2.2 Joint spacing....................................................................................................................A1-6
2.3 Jointing pattern and block types .....................................................................................A1-7
2.3.1 Block types and sizes ................................................................................................A1-8
3 Attitude of joints ....................................................................................................................A1-10
4 Development of jointing in various rock ...............................................................................A1-10
4.1 Jointing in igneous rocks...............................................................................................A1-11
4.2 Jointing in sedimentary rocks........................................................................................A1-11
4.3 Jointing in metamorphic rocks ......................................................................................A1-12
5 Statistical distribution of joints .............................................................................................A1-12
6 Summary ................................................................................................................................A1-13
Appendix 2: On faults and weakness zones
1 Zones of weak materials ..........................................................................................................A2-2
2 Faults and fracture zones .........................................................................................................A2-3
2.1 Occurrence of faults and fractures ..................................................................................A2-6
2.2 Composition and structure of faults................................................................................A2-8
2.3 Gouge (filling materials) in faults...................................................................................A2-8
2.4 Tension fracture zones ....................................................................................................A2-9
2.5 Shear fault and fracture zones.......................................................................................A2-10
2.5.1 Coarse-fragmented crushed zones ..........................................................................A2-12
2.5.2 Small-fragmented crushed zones ............................................................................A2-12
2.5.3 Sand-rich crushed zones..........................................................................................A2-12
2.5.4 Clay-rich crushed zones ..........................................................................................A2-12
2.5.5 Foliation shear zones...............................................................................................A2-13
2.6 Altered faults.................................................................................................................A2-14
2.6.1 Altered clay-rich zones............................................................................................A2-14
2.6.2 Altered leached (crushed) zones..............................................................................A2-15
3 Recrystallized and cemented/welded zones ...........................................................................A2-15
4 Description of weakness zones..............................................................................................A2-15
5 Summary ................................................................................................................................A2-17
Appendix 3: Methods to quantify the parameters applied in the RMi
1 Methods to determine the uniaxial compressive strength of rocks .........................................A3-2
1.1 The uniaxial compressive strength (s c)...........................................................................A3-3
1.2 Effect of saturation upon strength...................................................................................A3-5
1.3 Compressive strength determined from the point-load strength ....................................A3-7
1.3.1 The point load strength index (Is).............................................................................A3-8
1.3.2 The correlation between Is and s c .........................................................................A3-8
iv

1.4 Compressive strength estimated from Schmidt hammer rebound number...................A3-10


1.5 Compressive strength assessed from simple field test .................................................A3-10
1.6 Compressive strength estimated from rock description................................................A3-11
1.6.1 Main geological characteristics...............................................................................A3-11
1.6.2 Strength assessment from rock name......................................................................A3-12
1.6.3 Reduction in strength from anisotropy ...................................................................A3-13
1.6.4 Reduction in strength from weathering and alteration............................................A3-17
1.7 Summary .......................................................................................................................A3-19
2 Methods to determine the joint condition factor (jC) ............................................................A3-20
2.1 Estimating the joint roughness factor (jR)....................................................................A3-21
2.1.1 Field measurement of large scale roughness...........................................................A3-21
2.1.2 The joint waviness factor (jw)................................................................................A3-23
2.1.3 The joint smoothness factor (js).............................................................................A3-25
2.1.4 The joint roughness factor found from jw and js.................................................A3-25
2.2 Estimating the joint alteration factor (jA).....................................................................A3-26
2.2.1 Clean joints..............................................................................................................A3-27
2.2.2. Coated joints...........................................................................................................A3-27
2.2.3 Filled joints..............................................................................................................A3-27
2.2.4 Characterization and rating of the joint alteration factor (jA)...............................A3-29
2.3 Estimating the ratio jR/jA from friction angle recordings............................................A3-30
2.4 The joint size and continuity factor (jL) .......................................................................A3-33
2.5 Summary .......................................................................................................................A3-34
3 Methods to determine the block size .....................................................................................A3-36
3.1 Types of block volume and joint density measurements ..............................................A3-36
3.2 Block volume measurements.........................................................................................A3-38
3.2.1 Block volume found from joint spacings ................................................................A3-38
3.2.2 Block volume measured directly in situ or in drill cores.........................................A3-39
3.2.3 Methods to find the equivalent block volume where joints do not delimit blocks A3-39
3.3 Block diameter registrations .........................................................................................A3-41
3.4 Rock quality designation (RQD) ..................................................................................A3-42
3.4.1 Correlation between RQD and the volumetric joint count (Jv) ..............................A3-42
3.5 The volumetric joint count (Jv).....................................................................................A3-43
3.5.1 Block volume (Vb) estimated from the volumetric joint count (Jv) ............................A3-44
3.6 Joint frequency measurements .....................................................................................A3-46
3.6.1 2-D frequency in an area or surface ........................................................................A3-46
3.6.2 1-D frequency measurements along a scanline or drill core ...................................A3-48
3.7 Joint spacing registrations.............................................................................................A3-49
3.8 Weighted joint density measurements (wJd) ...............................................................A3-50
3.8.1 Correlation between wJd and Jv ...........................................................................A3-51
3.9 Use of refraction seismic measurements to assess block volume.................................A3-52
3.9.1 Influence from the intact rock and the in situ conditions........................................A3-52
3.9.2 Methods to assess the degree of jointing from in situ seismic velocities ...............A3-53
3.9.3 Possible errors and limitations applying seismic
velocities for jointing assessments.........................................................................A3-54
3.10 Summary of the correlations to determine the block size ...........................................A3-55
4 Methods to characterize the type and shape of rock blocks ..................................................A3-57
5 "Translation" of qualitative descriptions into numerical values ...........................................A3-60
5.1 Rock material characteristics ........................................................................................A3-61
5.2 Joint characteristics .......................................................................................................A3-61
5.3 Block size or quantity of joints .....................................................................................A3-62
5.4 Faults and weakness zones............................................................................................A3-63
5.5 Examples of numerical values found from qualitative descriptions.............................A3-63
5.5.1 Example 1 (from Gjövik (underground) Stadium) ................................................A3-63
5.5.2 Example 2................................................................................................................A3-64
5.5.2 Example 3................................................................................................................A3-64
v

5.5.2 Example 4................................................................................................................A3-65


5.5.2 Example 5................................................................................................................A3-65
5.5.2 Example 6, description of a weakness zone...........................................................A3-64
5.6 Summary .......................................................................................................................A3-66
Appendix 4: An investigation of the quality of various jointing measurements
1 Layout of the investigations performed...................................................................................A4-1
2 Block shape measurement .......................................................................................................A4-2
3 2-D and 1-D joint frequency registrations...............................................................................A4-3
3.1 2-D frequency measurements..........................................................................................A4-3
3.2 1-D frequency measurements..........................................................................................A4-5
4 Weighted joint density measurements .....................................................................................A4-6
4.1 Surface observations .......................................................................................................A4-7
4.2 Bore hole logging ............................................................................................................A4-7
5 Calculations of block volume from simplified jointing measurements...................................A4-8
6 Rock quality designation (RQD) .............................................................................................A4-9
6.1 Connection between the RQD and the volumetric joint count (Jv) ..............................A4-10
6.2 Connection between block volume and "block size"
expressed as RQD/Jn in the Q system.........................................................................A4-11
7 Summary ................................................................................................................................A4-12
Appendix 5: Using refraction seismic velocities to characterize jointing
1 Features influencing the magnitude of longitudinal sonic velocities.......................................A5-2
1.1 Factors influencing the sonic velocities in intact rock ....................................................A5-2
1.2 The influence from in situ factors on measured sonic velocities ....................................A5-5
2 Earlier methods used to characterize rock masses from seismic velocities ............................A5-6
2.1 Connections between jointing and longitudinal velocities..............................................A5-7
2.2 Rock quality estimated from the seismic velocity ratio ..................................................A5-7
2.3 Correlations between seismic velocities and rock mass characteristics .........................A5-8
3 Methods for assessing the degree of jointing from in situ seismic velocities .........................A5-9
3.1 Alt. 1: Correlations between jointing and sonic velocity are not known ......................A5-10
3.2 Alt. 2: Two or more correlations exist between jointing and velocities .......................A5-11
3.3 Worked examples..........................................................................................................A5-13
4 Summary ................................................................................................................................A5-15
Appendix 6: Description of the tests and data used in the calibration of the RMi
1 Sample 1. Results from triaxial laboratory tests on Panguna andesite ...................................A6-1
2 Sample 2. Large, compressive laboratory test on granitic rock from Stripa...........................A6-3
3 Sample 3. In situ tests on mine pillars of sandstone in the Laisvall mine...............................A6-5
4 Sample 4. Strength data found from back analysis of a slide in the Långsele mine ...............A6-8
5 Sample 5 - 7. Results from large-scale laboratory triaxial tests ...........................................A6-10
5.1 Sample 5. Caledonian clay-schist from Germany.........................................................A6-11
5.2 Sample 6. Mesozoic sandstone from Germany ............................................................A6-12
5.3 Sample 7. Palaeozoic siltstone from Germany .............................................................A6-13
Appendix 7: Collected data on ground conditions and rock support in constructed underground
openings
1 Description of the locations .....................................................................................................A7-1
1.1 Gjövik Olympic mountain hall, Norway.........................................................................A7-1
1.2 Granfoss road tunnels, Oslo............................................................................................A7-2
1.3 Haukrei hydropower plant, Telemark, Norway ..............................................................A7-3
1.4 Vinstra hydropower plant, Norway.................................................................................A7-5
1.5 Horga hydropower plant, Buskerud, Norway .................................................................A7-6
1.6 Tromsö road tunnel, Norway ..........................................................................................A7-7
1.7 Nappstraumen road tunnel, Lofoten, Norway.................................................................A7-7
1.8 Stetind road tunnel, Nordland, Norway ..........................................................................A7-8
1.9 Njunis tunnel, Bardu, Norway.........................................................................................A7-8
1.10 Sumbiar road tunnel, The Faroe Islands .........................................................................A7-9
1.11 Thingbæk chalk mines, Ålborg, Denmark ....................................................................A7-10
vi

2 Calculation of ground characteristics applied in the rock support tables .............................A7-11

Appendix 8: Collected data on ground condition and TBM boring performance


1 Boring experience and ground condition at Svartisen Power plant ........................................A8-1
1.1 Measured rock properties................................................................................................A8-1
1.2 Rock mass description in tunnel locations......................................................................A8-3
2 Boring experience and ground condition at Meråker hydropower plant.................................A8-5
2.1 Observations at chainage 750 .........................................................................................A8-6
2.2 Observations at the brook intake, chainage 10020.........................................................A8-6

Appendix 9: A method to estimate the tangential stresses around underground openings


1 Estimating the magnitude of the in situ ground stresses .........................................................A9-1
2 The tangential stresses developed around an underground opening .......................................A9-4
3 A practical method to estimate the magnitude of the tangential stresses................................A9-6

Appendix 10: Symbols used


1 General...................................................................................................................................A10-1
2 Rock properties......................................................................................................................A10-1
3 Jointing and block characteristics..........................................................................................A10-1
4 Stresses and related parameters.............................................................................................A10-2
5 Refraction seismic properties and features............................................................................A10-3
6 Rock mass properties and features ........................................................................................A10-3
6.1 Classification systems and parameters .........................................................................A10-4
6.2 Parameters and features in the Rock Mass index (RMi) ..............................................A10-4
7 Parameters in the RMi rock support method.........................................................................A10-4
8 Parameters and features in the RMi method for TBM penetration assessment method.......A10-5
)URP3DOPVWU|P$50L±DURFNPDVVFKDUDFWHUL]DWLRQV\VWHPIRUURFNHQJLQHHULQJSXUSRVHV
3K'WKHVLV2VOR8QLYHUVLW\1RUZD\S

&KDSWHU

,1752'8&7,21

7KHFRUQHUVWRQHRIDQ\SUDFWLFDOURFNPHFKDQLFVDQDO\VLVRUURFNHQJLQHHULQJLVWKHJHRORJL
FDOGDWDEDVHXSRQZKLFKWKHGHILQLWLRQRIURFNW\SHVVWUXFWXUDOGLVFRQWLQXLWLHVDQGPDWHULDO
SURSHUWLHVLVEDVHG(YHQWKHPRVWVRSKLVWLFDWHGDQDO\VLVFDQEHFRPHDPHDQLQJOHVVH[HUFLVHLI
WKHJHRORJLFDOLQIRUPDWLRQXSRQZKLFKLWLVEDVHGLVLQDGHTXDWHRULQDFFXUDWH
Evert Hoek, 1986.

The statement above clearly indicates the importance and need for good quality rock mass
characterization for use in rock engineering, design and construction. Proctor (1971) has shown that
the estimated cost of an underground excavation depends more upon geological data than does any
other type of civil engineering work. Any work that can contribute to better knowledge and
documentation of rock masses should therefore be of common interest. The importance and need for
improved geo-data has also later been stressed by Einstein et al. (1979) and Einstein and Baecher
(1983). Bieniawski (1984) stresses that
LWLVH[WUHPHO\LPSRUWDQWWKDWWKHTXDOLW\RIWKHLQSXWGDWDPDWFKHVWKHGHVLJQUHTXLUHPHQWV
2EYLRXVO\LWPXVWEHUHDOL]HGWKDWLILQFRUUHFWLQSXWSDUDPHWHUVDUHHPSOR\HGLQFRUUHFWGHVLJQ
LQIRUPDWLRQZLOOUHVXOW6SHDNLQJLQFRPSXWHUMDUJRQWKHIROORZLQJH[SUHVVLRQZRXOGEH
DSSURSULDWH
JDUEDJHLQJDUEDJHRXW


In the report "Definition of the most promising lines of research " presented by ISRM (1971) the
highest priority research subjects were defined as:
1. Determination of strength and deformability of fissured and massive rock masses as a function of
time.
2. Correlation between the mechanical properties of rocks and geological and petrographic data.

In the same report it was also stated that:$WSUHVHQWWLPHPRVWJHRORJLFDQGSHWURJUDSKLFGHVFULS


WLRQVRIVSHFLPHQVRUERGLHVRIURFNDUHTXDOLWDWLYHZKHUHDVURFNPHFKDQLFVGHWHUPLQDWLRQRI
PHFKDQLFDOSURSHUWLHVRIURFNDUHTXDQWLWDWLYH
%HFDXVHPDQ\HQJLQHHULQJGHFLVLRQVDUHEDVHGRQDFRPELQDWLRQRIJHRORJLFDQGURFNPHFKDQLFV
GDWDLWLVLPSRUWDQWWKDWDPRUHV\VWHPDWLFPHDQVRIFRPELQLQJDQGFRUUHODWLQJWKLVLQIRUPDWLRQ
VKRXOGEHGHYHORSHG
7KHUHLVDQHHGIRUEHWWHUGRFXPHQWDWLRQDQGFRUUHODWLRQRIJHRORJLFDODQGSHWURJUDSKLFGDWDDQG
FRUUHVSRQGLQJPHFKDQLFDOSURSHUW\GDWDREWDLQHGIURPERWKODERUDWRU\VSHFLPHQVDQGRUURFN
PDVVHVWRJHWKHUZLWKRSHUDWLQJH[SHULHQFHLQWKHVDPHURFNPDVVRUWKHVXEVHTXHQWSHUIRUPDQFH
RIVWUXFWXUHLQWKHURFNPDVVFUHDWHGE\H[FDYDWLRQ

Thus, already early in the development of rock mechanics the importance of establishing general
methods and systems for improved characterization of rock masses together with a common
language for those engaged in engineering and construction in rock was clearly expressed. This
request has to some extent, been met by some of the classification systems which are presented
later, but few of them are of a general character as they are mainly directed towards a specific
engineering function or design.
From the obvious need and interest for better geo-data mentioned above one should expect that
people involved in rock engineering and construction should have made significant efforts in
1-2

working out methods to arrive at better quality geo-data. However, in the early days in rock
mechanics and engineering geology, people paid surprisingly little interest in the geological aspects.
Karl Terzaghi, in his later years, when he had changed much of his attention from soil to rock
mechanics, wrote in his diary in 1961 about rock mechanics. ,DPPRUHDQGPRUHDPD]HGDERXW
WKHEOLQGRSWLPLVPZLWKZKLFKWKH\RXQJHUJHQHUDWLRQLQYDGHVWKLVILHOGZLWKRXWSD\LQJDWWHQWLRQWR
WKHLQHYLWDEOHXQFHUWDLQWLHVLQWKHGDWDRQZKLFKWKHLUWKHRUHWLFDOUHDVRQLQJLVEDVHGDQGZLWKRXW
PDNLQJVHULRXVDWWHPSWVWRHYDOXDWHWKHUHVXOWLQJHUURUV

Later, Poisel (1990) points out that the same trends have continued also during the last 10 years: $
ORRNDWURFNPHFKDQLFVSDSHUVVKRZVWKDWDWSUHVHQWLWLVRQO\LPSRUWDQWWRSHUIHFWPDWKHPDWLFDO
SURFHGXUHVDQGWRLQYHVWLJDWHURFNOLNHPDWHULDOV5RFNPHFKDQLFVKRZHYHUVWLOOQHHGV
LQYHVWLJDWLRQVLQWKHILHOGQRWRQO\WHVWLQJWRJHWLWVSDUDPHWHUVJHRPHFKDQLFDOPRGHOVDQG
LQWXLWLRQ
5HJUHWIXOO\ZHGRQRWDVNRXUVHOYHVRIWHQHQRXJKLIWKHUHVXOWVZHUHDFKE\PDWKHPDWLFDO
FDOFXODWLRQVUHDOO\DSSO\WRQDWXUH5RFNPHFKDQLFDOPRGHOVVKRXOGEHGHWHUPLQHGE\ZKDWQDWXUH
VKRZVQRWE\ZKDWFDQEHGRQHZLWKLQPHFKDQLFDOPDWKHPDWLFDOOLPLWV7KXVLWVKRXOGILUVWEH
DWWHPSWHGWRILQGRXWZKDWLVJRLQJRQLQQDWXUHZKLFKSURFHGXUHVRUZKLFKPHFKDQLVPVWDNHSODFH
LQWKHVWUXFWXUHDQGRQO\WKHQLWVKRXOGEHFRQVLGHUHGKRZWRPRGHOWKHPE\PHFKDQLFDO
PDWKHPDWLFDOPHDQV

From the foregoing it is clear that there has been continuous need also to establish a common
language to characterize rock mass parameters of importance for engineering for construction and
treatment of rock.

2XWOLQHRIWKLVZRUN

The opinions presented above largely explain the purpose of this work. A main goal has been to
improve the quality of geological input data, which consequently will lead to better design and rock
engineering to be applied in rock construction. No attempt has been made to analyze the science of
either geology or rock mechanics, but rather, only to use and relate geological knowledge to the
characteristics of the material called URFNPDVV. Fig. 1-1 shows the main structure of the work and
its main applications.

The first step outlined in this figure is the collection of data representative of the rock mass
composition, based on observations and measurements.

The second step includes characterization of these data into a general rock mass index (RMi). This
index and the definitions of its input parameters can be used in communication as well as for input
in existing and possible future engineering systems.

In a third step, the RMi can be adjusted for parameters or features of importance for the actual
utility or construction to assess the quality of the ground. By combining this with input from the
actual excavation or construction requirements, the system can be used in a fourth step for design
and engineering.
1-3

data on rock mass


composition
GEOLOGICAL
OBSERVATIONS 6WHS

AND MAPPING $FTXLVLWLRQRI


JHRGDWD
FIELD INVESTIGATIONS
AND TESTING

SELECTED
6WHS
PARAMETERS combination of the
ARE GIVEN ROCK MASS INDEX $JHQHUDO

selected parameters FKDUDFWHUL]DWLRQ


NUMERICAL
VALUES
RMi RIWKHURFNPDVV

6WHS
PROJECT
,QSXWRIIHDWXUHV
RELATED RILPSRUWDQFHIRU
FEATURES WKHDFWXDOGHVLJQ

ROCK DESIGN AND 6WHS


ENGINEERING IN: 50LDSSOLHGLQ
- rock support SUDFWLFDOURFN
- TBM excavation HQJLQHHULQJ
- rock blasting *)
- rock fragmentation *)

applied as
input in rock used in 2WKHU
mechanics and communication DSSOLFDWLRQV
other types of RIWKH50L
engineering
QRWLQFOXGHGLQWKLVZRUN

Fig. 1-1 Main principles of the system presented.

The work has been structured into the following main fields:
1. Collection and characterization of geological and material data (geo-data) described in Chapters
2 - 3 and in Appendices 1 - 5.
2. Combination of geo-data in RMi outlined in Chapters 4 - 5 and in Appendix 6.
3. Application of RMi in practical rock engineering. Chapters 6 - 8 and Appendix 7 describe the
direct use of RMi in rock support and in full face tunnel boring (TBM) capacity assessments.

The main part of this contribution has been to develop a general system for characterizing rock
masses in which individual parameters representative of a rock mass are combined. The selection
and combination of these parameters are based on a comprehensive study of available literature and
communication with experienced people. The system is calibrated against documented large scale
and in situ test results from real rock masses. The following features have been important during the
development of the work:
- The system should be simple and meaningful in terms, i.e. only few input data have been
selected to arrive at understandable expressions.
- Where possible, existing methods for finding the characteristics of the geo-data required should
be utilized; i.e. simple and practical methods for collecting the input values have been
described.
- The system should have a general form; i.e. constitute a platform which can be applied in rock
engineering; it should be possible to apply it as input in existing engineering systems or
methods.

With reference to the above statements by Terzaghi and by Poisel and also to statements by Müller
(1982) it is a prerequisite that the limitations and uncertainties in geo-data are always considered
when applied in calculations, design and engineering.
1-4

In addition to its contribution to improvements in practical rock engineering, it is hoped that this
work may lead to more systematic use of rock mass descriptions. Together with well defined
expressions for important geological features this may improve
- the quality of rock mass descriptions,
- the use of geo-data from field investigations, and
- the language between geologists and rock engineers involved in rock engineering and
construction.

The basis and goal of this work has, in fact, already long ago been formulated by John (1969):
5RFNPHFKDQLFVKDVWRSURYLGHPHWKRGVRIDQDO\VLVZKLFKDUHUHDOLVWLFFRPSURPLVHVEHWZHHQWKH
EHVWUHSUHVHQWDWLRQRIWKHDFWXDOJURXQGFRQGLWLRQVDQGSUDJPDWLFHQJLQHHULQJ7KHURFNPHFKDQLFV
SUDFWLWLRQHUKDVWRIDFHWKHIDFWWKDWPDQ\JHRORJLFDOHQJLQHHULQJDQGURFNPHFKDQLFVSUREOHPV
PD\EHWRRFRPSOH[WRDOORZULJRURXVDQDO\VLVEXWDWWKHVDPHWLPHDUHGHHPHGVDWLVIDFWRU\IRUWKH
FRQVWUXFWLRQRIPDMRUVWUXFWXUHV4XDOLWDWLYHHYDOXDWLRQVTXDQWLWDWLYHGHVFULSWLRQVRIJHRORJLF
IHDWXUHVDVVXFKDQGFRPSDULVRQVRIVSHFLILFWHVWUHVXOWVDUHDOZD\VRILQWHUHVWEXWPD\QRWVXIILFH
DVEDVLVIRUHQJLQHHULQJGHFLVLRQVDQGGHVLJQV
1HLWKHUDSXUHO\JHRORJLFDOQRUDFRPSOHWHO\WHFKQLFDOFODVVLILFDWLRQRIURFNPDVVHVZLOODQVZHULWV
LQWHQGHGSXUSRVHEXWVXLWDEOHSDUDPHWHUVRIURFNPDVVHVVKRXOGEHGHILQHGDQGTXDQWLILHGLQURFN
HQJLQHHULQJWHUPV

Before going further, some terms which are commonly used in this work, may need to be clarified:

'HVFULEH is to tell or write about, give a detailed account of; to picture in words. In a description,
although all the complicated technical terms are recorded, essentially the individual puts down
whatever she feels is important. Thus, in describing a rock material or rock mass it is only by
chance that a system or order is followed. Seldom can the record be compared to that of
another investigator.
&KDUDFWHUL]H is to report the particular qualities, features, or traits of. Rock mass FKDUDFWHUL]DWLRQ is
the designation of rock mass quality based on numbers and descriptive terms of certain features
in the rock mass. Such characterization can include one or more parameters.
&ODVVLI\ is to arrange or group in classes according to some system or principle. Rock mass
FODVVLILFDWLRQ is the process of combining certain features of a rock mass into classes or groups.
It follows a system and order with information being recorded in a prescribed manner. By this it
is possible to combine different features using mathematical expressions. Classification enables
useful comparisons to be made between the work of two or more investigators.

In this work the term FKDUDFWHUL]H has been selected for the process of indicating the structure,
composition and strength of rock masses. In practice there is, however, often not much difference
between the process of classification and characterization a rock mass. A main difference may be that
characterization also contains descriptive terms.
)URP3DOPVWU|P$50L±DURFNPDVVFKDUDFWHUL]DWLRQV\VWHPIRUURFNHQJLQHHULQJSXUSRVHV
3K'WKHVLV2VOR8QLYHUVLW\1RUZD\S

&KDSWHU

52&.0$66(6$6&216758&7,210$7(5,$/6

5RFNPDVVHVDUHVRYDULDEOHLQQDWXUHWKDWWKHFKDQFHIRUHYHUILQGLQJDFRPPRQVHWRI
SDUDPHWHUVDQGD FRPPRQ VHW RI FRQVWLWXWLYH HTXDWLRQV YDOLG IRU DOO URFN PDVVHV LV TXLWH
UHPRWH
Tor L. Brekke and Terry R. Howard, 1972

A rock mass is a material quite different from other structural materials used in civil engineering. It
is heterogeneous and quite often discontinuous, but is one of the materials in the earth’s crust, which
is most used in man’s construction. Ideally, a rock mass is composed of a system of rock blocks and
fragments separated by discontinuities forming a material in which all elements behave in mutual
dependence as a unit (Matula and Holzer, 1978). The material is characterized by shape and
dimensions of rock blocks and fragments, by their mutual arrangement within the rock mass, as well
as by joint characteristics such as joint wall conditions and possible filling (see Fig. 2-1).

Fig. 2-1 The main features constituting a rock mass

The complicated structure of the rock mass with its defects and inhomogeneities and the wide range
of its applications cause challenges and problems in rock engineering and construction which often
involve considerations that are of relatively little or no concern in most other branches of
engineering. One of these challenges is, according to Einstein and Baecher (1982), the uncertainties
about geological conditions and geotechnical parameters. This is perhaps one of the most distinctive
features of engineering geology compared to other engineering fields, therefore ’engineering
judgement’, adaptable design approaches, and other procedures for dealing with uncertainty or
hedging against it have been taken into use.

Important in all rock mechanics, rock engineering and design are the quality of the geo-data that
form the basis for the calculations and estimates made. This quality depends on two main features.
1. The understanding and interpretation of the geological setting of the area of interest.
2. The way the (known) rock mass at the site is described or measured.

The first feature is important mainly in the pre-construction phase and is a result of the geological
understanding based on field investigations and the experienced interpretation of available results.
To a great extent this is often wholly dependent on the skill of the geologist(s) who decide how the
2-2

investigations should be done and how the geo-data should be combined. Thus, this process can in
many instances be said to be more an "art" than a science. The details concerning the geological part
are not dealt with further here, but the influence of the geology is discussed in Chapter 3.

The second feature is mainly connected to the present work. Brown (1986) is of the opinion that
LQDGHTXDFLHVLQVLWHFKDUDFWHUL]DWLRQRIJHRGDWDSUREDEO\SUHVHQWWKHPDMRULPSHGLPHQWWRWKH
GHVLJQFRQVWUXFWLRQDQGRSHUDWLRQRIH[FDYDWLRQVLQURFN,PSURYHPHQWVLQVLWHFKDUDFWHUL]DWLRQ
PHWKRGRORJ\DQGWHFKQLTXHVDQGLQWKHLQWHUSUHWDWLRQRIWKHGDWDDUHRISULPDU\UHVHDUFKUHTXLUH
PHQWVQRWRQO\IRUODUJHURFNFDYHUQVEXWIRUDOOIRUPVRIURFNHQJLQHHULQJ

TABLE 2-1 BASIC ELEMENTS AND RELEVANT CONSIDERED AREAS (based on Natau, 1990)

%$6,&(/(0(17 6,=(5$1*( 6758&785(6 &216,'(5('$5($

Crystal lattice Angstrom size Micro structures Electron microscope


(10-7 mm)

Mineral grain µm - cm Grain structures in Microscope, hand piece, test sample of rock
rock
Hand piece, stone ornaments, building
Rock material cm - 10 m Massive rock stone, test of rock samples.

Foundations, small underground structures,


Jointed rock cm - 10 m Joint pattern, rock test samples of rock masses, test pits/adits
(composed of ’bricks’) mass
Slopes, tunnels, large underground struc-
tures, mines
Geological-tectonical 10 m - km Rock mass (geological maps and sections)
units volumes between
large faults Oil reservoirs,
(general geological maps and sections)
Geological-tectonical Several km Regional plates
large size units

,19(67,*$7,216
in laboratory in situ
STRENGTH

SIZE OF SAMPLE
Fig. 2-2 The scale factor of rock masses and the variation in strength of the material depending on the size of the
’sample’ involved. (After Janelid, 1965)
2-3

Other special features in a rock mass and its utilization in contrast to other construction materials
are:
- the VL]Hor volume of the material involved, see Fig. 2-2 and Table 2-1,
- the VWUXFWXUH and composition of the material,
- the many FRQVWUXFWLRQDQGXWLOL]DWLRQ purposes of it, see Table 2-2, and
- the difficulties in measuring the TXDOLW\RIWKHPDWHULDO (see also Appendix 4).

TABLE 2-2 MAIN TYPES OF WORKS CONNECTED TO ROCKS AND ROCK MASSES
7<3( $&78$/352&(662586(
- drilling (small holes)
- boring (TBM boring, shaft reaming) *)
- blasting*)
Treatment of rocks - fragmentation*)
- crushing
- grinding
- cutting*)

- rock aggregate for concrete etc.


Application of rocks - rock fill
- building stone

Utilization of - in underground excavations (tunnels, caverns, shafts) *)


rock masses - in surface cuts/slopes/portals *)

- excavation works
Construction works
- rock support *)
in rock masses
- water sealing
*)
Areas where the system is of particular interest.

These factors imply that other methods of data acquisition are used, and that other procedures in the
use of these data for construction purposes have been developed. Thus, the material properties of
rock masses are not measured but estimated from descriptions and indirect tests. The stress is not
applied by the engineering but is already present; the construction, however, leads to stress changes.

In the remainder of this chapter the main features of the rock mass and their effect on its behaviour
related to rock construction are briefly outlined.

 52&.6$1'7+(,50$,1)($785(6

Geologists use a classification, which reflects the origin, formation and history of a rock rather than
its potential mechanical performance. The rock names are defined and used not as a result of the
strength properties, but according to the abundance, texture and types of the minerals involved, in
addition to mode of formation, degree of metamorphism, etc. According to Franklin (1970) there are
over 2000 names available for the igneous rocks that comprise about 25% of the earth’s crust, in
contrast to the greater abundance of mudrocks (35%) for which only a handful of terms exist; yet
the mudrocks show a much wider variation in mechanical behaviour.
2-4

)UHVKURFNV

Each particular rock type is characterized by its minerals, texture fabric, bonding strength and macro
and micro structure, see Fig. 2-3.

,JQHRXVURFNV tend to be massive rocks of generally high strength. Their minerals are of a dense
interfingering nature resulting in only slight, if any, directional differences in mechanical properties
of the rock. These rocks constitute few problems in rock construction when fresh.

6HGLPHQWDU\URFNV constitute the greatest variation in strength and behaviour. The minerals of these
rocks are usually softer and their assemblage is generally weaker than the igneous rocks. In these
rocks the minerals are not interlocking but are cemented together with inter-granular matrix
material. Sedimentary rocks usually contain bedding and lamination or other sedimentation
structures and, therefore, may exhibit significant anisotropy in physical properties depending upon
the degree of their development. Of this group, argillaceous and arenaceous rocks are usually the
most strongly anisotropic. Some of the rocks are not stable in the long term, as for example
mudrocks, which are susceptible to slaking and swelling. This group of rocks therefore creates many
problems and challenges in rock construction.

0HWDPRUSKLFURFNV show a great variety in structure and composition and properties. The
metamorphism have often resulted in hard minerals and high intact rock strength; however, the
preferred orientation of platy (sheet) minerals due to shearing movements results in considerable
directional differences in mechanical properties. Particularly the micaceous and chloritic schists are
generally the most outstanding with respect to anisotropy.

7KHLQIOXHQFHIURPVRPHPLQHUDOV

Certain HODVWLFDQGDQLVRWURSLFPLQHUDOV like mica, chlorite, amphiboles, and pyroxenes may highly
influence the mechanical properties of the rocks in which they occur (Selmer-Olsen, 1964). Parallel
orientation of these minerals is often found in sedimentary and regional metamorphic rocks in
which weakness planes may occur along layers of these flaky minerals. Where mica and chlorite
occur in continuous layers their effect on rock behaviour is strongly increased. Thus, mica schists
and often phyllites have strong anisotropic mechanical properties of great importance in rock
construction. Also other sheet minerals like serpentine, talc, and graphite reduce the strength of
rocks due to easy sliding along the cleavage surfaces, see Fig. 2-3.

Quartz is another important mineral in rock construction. This mineral is grade 7 in the Mohs scale
of hardness. Sharp, obtuse-angled edges of the quartz grains have an unfavourable shape regarding
drill bit and cutter wear in percussion drilling and TBM boring respectively, while the effect from
rounded quartz grains is significantly less.

Change of moisture content in VZHOOLQJPLQHUDOVof the smectite (montmorillonite) group can cause
significant problems related to high swelling pressures (Piteau, 1970). These minerals, occurring
either as infilling or alteration products in seams or faults, have in addition to expansion, a low
shear strength, which may contribute to rock falls and, in some cases, slides in underground
openings and cuttings. Also some rocks may show swelling properties. These rocks can be
montmorillonitic shales, altered or weathered basalts, in addition to other igneous, metamorphic
rocks, or sedimentary rocks containing anhydrite.
2-5

Some rocks may slake (hydrate or "swell", oxidize), disintegrate or otherwise weather in response to
the change in humidity and temperature consequent on excavation. As mentioned above, an
abundant group of rocks, the mudrocks, are particularly susceptible to even moderate weathering
(Olivier, 1976). Refer to Fig. 2-3.

influence from some minerals some special processes acting


FLAKY SWELLING
MINERALS MINERALS ALTERATION
-mica -smectite or HYDRATIZATION
-chlorite -montmorillonite WEATHERING of mudrocks etc.
-talc -anhydrite
common rock features
MINERAL
COMPOSITION

MINERAL SIZE

TEXTURE
fresh rocks
ALTERED
HOMOGENEOUS SWELLING or SLAKING
SCHISTOSE
and ROCK ROCKS WEATHERED ROCKS
LAYERED ROCKS ROCKS

U R F N VZ L WK
UR F N VZ LWK
L V R WUR S LF
V W URQ J O \ U R F N V Z L W K U H G X F H G V W U H Q J W K D Q G G X U D E L O L W \
R UV OL JK W O\
D Q L V R WU R S LF
D Q L V R WU R S LF
S U R SH UWL H V
S UR S H UWLH V

Fig. 2-3 The main variables influencing rock properties and behaviour

7KHHIIHFWRIDOWHUDWLRQDQGZHDWKHULQJ

The processes of alteration and weathering with deterioration of the rock material have reducing
effect on the strength and deformation properties of rocks, and may completely change the
mechanical properties and behaviour of rocks (refer to Fig. 2-3). For most rocks, except for the
weaker types, these processes are likely to have great influence on engineering behaviour of rock
masses. Hence, the description and characterization of rock masses should pay particular attention
to such features.

Rocks are frequently ZHDWKHUHG near the surface, and are sometimes DOWHUHG by hydrothermal
processes. Both processes generally first affect the walls of the discontinuities1. The main results of
rock weathering and alteration are:
1. Mechanical GLVLQWHJUDWLRQ or breakdown, by which the rock loses its coherence, but has little
effect upon the change in the composition of the rock material. The results of this process are:
- The opening up of joints.
- The formation of new joints by rock fracture, the opening up of grain boundaries.
- The fracture or cleavage of individual mineral grains.

1
In this work, the following terms have been applied for the various types of discontinuities:
Joints - Minor and medium sized discontinuities, including fissures, cracks, fractures,
breaks, etc.; also some minor seams are included in this group.
Seams - Filled discontinuities, including shears; they are also named ’singularities’.
Weakness zones - Including faults, crushed zones and zones of weak rocks surrounded by stronger
rocks.
The characteristics of these features are further described in Appendices 1 and 2.
2-6

2. Chemical GHFRPSRVLWLRQ, which involves rock decay accompanied by marked changes in


chemical and mineralogical composition results in:
- Discoloration of the rock.
- Decomposition of complex silicate minerals (feldspar, amphibole, pyroxene, etc) eventually
producing clay minerals; some minerals, notably quartz, resist this action and may ’survive’
unchanged.
- Leaching or solution of calcite, anhydrite and salt minerals.

The GLVLQWHJUDWLRQ leads mainly to a greater number of joints in rock masses located in the upper
zone of weathering, while GHFRPSRVLWLRQ influences the joint condition as well as the rock material.

*HRORJLFDOQDPHVDQGPHFKDQLFDOSURSHUWLHVRIURFNV

Rocks that differ in mineral composition, porosity, cementation, consolidation, texture and
structural anisotropy can be expected to have different strength and deformation properties.
Geological nomenclature of rocks emphasizes mainly solid constituents, whereas from the
engineer’s point of view, pores, defects and anisotropy are of greater mechanical significance
(Franklin, 1970). For each type of rocks the mechanical properties vary within the same rock name.
Petrological data can, however, make an important contribution towards the prediction of
mechanical performance, provided that one looks beyond the rock names to the observations on
which they are based. It is, therefore, important to retain the names for the different rock types, for
these in themselves give relative indications of their inherent properties (Piteau, 1970).

 ',6&217,18,7,(6,152&.

Any structural or geological feature that changes or alters the homogeneity of a rock mass can be
considered as a discontinuity. Discontinuities constitute a tremendous range, from structures which
are sometimes thousands of meters in extent down to - per definition - mm size, see Fig. 2-4.
5 2 & . ' ( ) ( & 7 6 - 2 ,1 7 6 : ( $ . 1 ( 6 6 = 2 1 ( 6

IDXOWV

MRLQWV
SDUWLQJV
FUDFNV
ILVVXUHV
EHGGLQJ SODQHV
VHDPVVKHDUV

0.01 0.1 1 10 100 1000 10 000


/(1*7+ P
Fig. 2-4 The main types of discontinuities according to size. The size range (length) used for joints in this work is
indicated.

The different types, such as faults, dykes, bedding planes, tension cracks, etc. have completely
different engineering significance (Piteau, 1970). The roughness, nature of their contacts, degree
and nature of weathering, type and amount of gouge and susceptibility to ground water flow will
vary greatly from one type of discontinuity to another since their cause, age and history of develop-
ment are fundamentally different. The effect on rock masses due to these localised discontinuities
2-7

varies considerably over any given region depending on structure, composition and type of
discontinuity.
The great influence of discontinuities upon rock mass behaviour calls for special attention to these
features when characterizing rock masses for practical applications. Joints and faults have numerous
variations in the earth’s crust, this is probably the main reason that it has been so difficult to carry
out common observation and description methods (Terzaghi, 1946).

)DXOWV

Faults are breaks along which there has been displacement of the sides relative to one another
parallel to the break. Minor faults range in thickness from decimetre to meter; major faults from
several meters to, occasionally, hundreds of meters. It is important to realise that most fault zones
are the result of numerous ruptures throughout geological time, and that they quite often are
associated with other parallel discontinuities that decrease in frequency and size in the direction
away from the central zone.

Faults and fault zones often form characteristic patterns in the earth’s crust consisting of several
independent sets or systems, see Fig. 2-5. The main directions, which mainly were determined by
the state of stress, have often the same orientations as the joint sets within the same structural area.

0 1km
Weakness zone Lake

Fig. 2-5 Pattern of weakness zones and faults in the earth’s surface. (After Selmer-Olsen, 1988)

Hydrothermal activity and other processes may have caused alteration of minerals into clays, often
with swelling properties. Many faults and weakness zones thus contain materials quite different
2-8

from the ’host’ rock. The problems related to weakness zones may, therefore, depend on several
factors which may all interplay in the final behaviour.

Weakness zones and faults show numerous variations in their structures and compositions, see Fig.
2-6. In cases where the zones or faults are composed mainly of joints and seams they may be
characterized by the same descriptions as for jointing. In other cases it may be necessary to
characterize them by special descriptions and measurements or tests, as further described in
Appendix 2. The fact that faults and weakness zones of significant size can have a major impact
upon the stability as well as on the excavation process of an underground opening necessitates that
special attention, follow-up and investigations often are necessary to predict and avoid such events.

'

$%&

Fig. 2-6 Sketches of some types of weakness zones. A - C are from ISRM (1978) and D - E from Selmer-Olsen
(1950).

-RLQWVDQGWKHLUPDLQIHDWXUHV

Joints are the most commonly developed of all structures in the earth’s crust, since they are found in
all competent rocks exposed at the surface. Yet, despite the fact that they are so common and have
been studied widely, they are perhaps the most difficult of all structures to analyse. The analytical
difficulty is caused by the number of fundamental characteristics of these structures. There is,
however, abundant field evidence that demonstrates that joints may develop at practically all ages in
the history of rocks (Price, 1981).

A joint can be open or closed. Closed joints may be nearly invisible. Yet they constitute surfaces
along which there is no resistance against separation. In quarries the spacing of joints determines the
largest size of blocks of sound rock which can be obtained. Therefore, joints and joint systems have
attracted the attention of builders ever since cut stones have been used.

A joint is composed of several characteristics. In addition to length and continuity of the joint the
main are:
- roughness and strength of the joint wall surface,
- waviness or planarity of joint wall,
- alteration or coating of the joint wall, and
- possible filling. Refer to Fig. 2-7.
2-9

All these parameters influence on the shear strength of the joint (Brekke and Howard, 1972; Price,
1981; Hoek and Brown, 1980; Barton et al., 1974; Barton and Choubey, 1977; Bieniawski, 1984;
Turk and Dearman 1985; and several other authors). They also determine the amount of water that
can flow through the joint.

jo int
ity of the
o ntinu
an d c
len gth

in t
of jo
o n d it ion ce:
c s ur fa
wall oothnessoating
- s m os sible c alte ration
- p o ss ible k
- p w all roc
of

joint thickness and


possible filling material

waviness or
undulation
of joint wall

Fig. 2-7 Sketch showing the main features of a joint.

The distance between the two matching joint walls controls the extent to which these can interlock.
In the absence of interlocking, the properties of the filling of the joint determine the shear strength
of the joint. As separation decreases, the asperities of the rock wall gradually become more
interlocked, and the rock wall properties are the main contributor to the shear strength.

7KHPDLQMRLQWLQJFKDUDFWHULVWLFV

By jointing is meant the pattern and frequency or density of joints. Field studies of several workers
have shown that the joints preferentially are found in certain directions. One to three prominent sets
and one or more minor sets may occur; in addition several individual or random joints are often
present.

The joints delineate blocks. Their dimensions and shapes are determined by the joint spacings, by
the number of joint sets and by random joints. ISRM (1978), Barton (1990) and several other
authors state that the block size is as an extremely important parameter in rock mass behaviour. A
number of scale effects in rock engineering can be explained by this feature including compressive
strength, deformation modulus, shear strength, etc.

Different methods are used for measuring the jointing density. The most common are:
- Joint spacing, either in surfaces or in drill cores or scan lines.
- Density of joints, either in surfaces, or in bore holes or scan lines.
- Block size, in surfaces, and
- Rock quality designation (RQD), in drill cores.
They are further outlined in Appendix 3, where also correlation equations between them have been
developed.
2 - 10

7KHURFNPDVV

Discontinuities ranging in lengths from less than a decimetre to several kilometres divide the
bedrocks into units, volumes or blocks of different scales (Fig. 2-8):
1. The regional pattern or first order fault blocks are bounded by the larger weakness zones or
faults (see Fig. 2-5).
2. The second order blocks formed by singularities, i.e. small weakness zones or seams.
3. The third order blocks formed by normal joints.
4. The small joints in the appearance of bedding or schistosity partings form the smallest pattern
or fragments, which are of interest for engineering purposes.
5. The microcracks are responsible for making up small fragments or grains in the rock. These
discontinuities are, however, mostly considered a rock property and are therefore generally
included in the strength characterization of the rock material.

Based on this it has been found useful for engineering geological and design purposes to divide the
ground into:
- "The detailed jointing" formed mainly by the third and fourth order blocks or units, and
- "The coarse pattern of weakness zones" formed by the first order blocks or units by faults and
weakness zones.

$
2
3
1
2

4
3
 B

A 
2

2 50
- 15
0m
3
500
- 150
0m 2
5 4
%

5-
50
m

Fig. 2-8 Simplified model of various dimensions units or blocks formed by discontinuities of different size (after
Pusch and Morfeldt, 1993).
2 - 11

This corresponds with the division suggested by Selmer-Olsen (1964). The rock blocks in the
detailed jointing pattern including the rock fragments or pieces caused by the small joints/fissures is
a main feature in the rock mass characterization developed herein.

 52&.0$66&+$5$&7(5,=$7,21)25'(6,*1$1'&216758&7,21
385326(6

An important issue in rock mass description and characterization is to select parameters of greatest
significance for the actual type of design or construction. There is no single parameter or index,
which can fully designate the properties of jointed rock mass. Various parameters have different
significance and only if combined can they describe a rock mass satisfactorily (Bieniawski, 1984).
Testing of rock masses in situ has brought out very clearly the enormous variations that exist in the
mechanical behaviour of a rock mass from place to place. According to Lama and Vutukuri (1978)
the engineering properties of a rock mass depend far more on the system of geological
discontinuities within the rock mass than of the strength of the rock itself. Further, the strength of a
rock mass is often governed by the interlocking bonds of the unit "elements" forming the rock mass.

Terzaghi (1946) also concludes that, from an engineering point of view, a knowledge of the type
and frequency of the rock discontinuities may be much more important than of the types of rock
which will be encountered. Similarly, Piteau (1970) has stressed the importance of distinguishing
between the behaviour of the rock and the rock mass, especially for hard rocks. Thus, characterizing
a discontinuity system in a way that describes the variability of its geometric parameters constitutes
an essential step in dealing with stability problems in discontinuous rock masses (Tsoutrelis et al.,
1990).

This does not mean that the properties of the intact rock material should be disregarded in the
characterization. After all, if discontinuities are widely spaced, or if the intact rock is weak, the
properties of the intact rock may strongly influence the gross behaviour of the rock mass. The rock
material is also important if the joints are discontinuous. In addition, the rock description will
inform the reader about the geology and the type of material at the site. Although rock properties in
many cases are overruled by discontinuities, it should be brought to mind that the properties of the
rocks highly determine the formation and development of discontinuities.

Therefore, an adequate and reliable estimation of the nature of the rock is often a primary
requirement. For some engineering or rock mechanics purposes the mechanical characterization of
rock material alone can be used, namely for drillability, crushability, aggregates for concrete, asphalt
etc. Also, in assessment for the use of fullface boring machines (TBM), rock properties like
compressive strength, hardness, anisotropy are among the more important parameters.

Kirkaldie (1988) mentions a total of 28 parameters present in rock masses which may influence the
strength, deformability, permeability or stability behaviour of rock masses: 10 rock material
properties, 10 properties of discontinuities and 8 hydrogeological properties. Because it is often
difficult or impossible in a general characterization to include the many variables in such a complex
natural material, it is necessary to develop suitable systems or models in which the complicated
reality of the rock mass can be simplified by selecting only a certain number of representative
parameters. For this purpose several classification and design systems have been developed, of
which some are shown in Table 2-3 for information. Further, Table 2-4 indicates the main rock
mass and ground features and which of these that have been applied and combined in the various
systems.
2 - 12

From Table 2-4 it is seen that the following parameters are most frequently applied in design and
classification systems:
- the rock material (rock type, geological name, weathering and alteration, strength);
- the degree of jointing (joint spacing, block size, RQD); and
- in situ stresses.

Also such features as:


- orientation of main discontinuities or joint set;
- joint conditions;
- block shape or jointing pattern;
- faults and weakness zones; and
- excavation features (dimension, orientation, etc.)
have been considered as important parameters in rock masses.

TABLE 2-3 SOME OF THE MAIN DESIGN AND CLASSIFICATION SYSTEMS IN USE
1DPHRIFODVVLILFDWLRQ )RUPDQG7\SH*) 0DLQDSSOLFDWLRQV 5HIHUHQFH
The Terzaghi rock load Descriptive and For design of steel support in Terzaghi, 1946
classification system behaviouristic form tunnels
Functional type
Lauffer’s stand-up time Descriptive form For input in tunnelling design Lauffer, 1958
classification General type
The new Austrian Descriptive and For excavation and design in Rabcewicz, Müller
tunnelling method behaviouristic form incompetent (overstressed) and Pacher,
(NATM) Tunnelling concept ground 1958 - 64
Rock classification for Descriptive form For input in rock mechanics Patching and Coates,
rock mechanical purposes General type 1968
The unified classification Descriptive form Based on particles and blocks for Deere et al., 1969
of soils and rocks General type communication
The rock quality Numerical form Based on core logging; used in Deere et al., 1967
designation (RQD) General type other classification systems
The size-strength Numerical form Based on rock strength and block Franklin, 1975
classification Functional type diameter; used mainly in mining
The rock structure rating Numerical form For design of (steel) support in Wickham et al., 1972
(RSR) classification Functional type tunnels
The rock mass rating Numerical form For use in tunnel, mine and Bieniawski, 1973
(RMR) classification Functional type foundation design
The NGI Q classification Numerical form For design of support in Barton et al., 1974
system Functional type underground excavations
The typological Descriptive form For use in communication Matula and Holzer,
classification General type 1978
The unified rock Descriptive form For use in communication Williamson, 1980
classification system General type
Basic geotechnical Descriptive form For general use International Society
classification (BGD) General type for Rock Mechanics
(ISRM), 1981
'HILQLWLRQRIWKHIROORZLQJH[SUHVVLRQV
*)

'HVFULSWLYHIRUP the input to the system is mainly based on descriptions


1XPHULFDOIRUP the input parameters are given numerical ratings according to their character
%HKDYLRXULVWLFIRUP the input is based on the behaviour of the rock mass in a tunnel
*HQHUDOW\SH the system is worked out to serve as a general characterization
)XQFWLRQDOW\SH the system is structured for a special application (for example for rock support)
2 - 13

As for most other construction materials, there is also in rock engineering and construction a need
for a strength specification of the material, i.e. the rock mass. The strength of other construction
materials can be determined from the process of refining or ensured during production of the
material. In rock construction, however, the material already exists, the task is to evaluate the
strength properties it possesses (and not to produce them).

The considerations outlined above have been important in the development of the present system for
rock mass characterization.

TABLE 2-4 APPLICATION OF ROCK MASS AND GROUND PARAMETERS IN VARIOUS DESIGN AND
CLASSIFICATION SYSTEMS
&/$66,),&$7,216<67(012Å    
ROCK
- Origin, name, or type x x x x
- Weathering o  " + x
- Anisotropy

ROCK PROPERTIES
- Unit weight + x
- Porosity +
- Rock hardness x
- Strength    x x x + x x
- Deformability  o +
- Swelling o 

JOINT CONDITIONS
- Joint size/length o o
- Joint separation x o
- Joint wall smoothness x x 
- Joint waviness x x
- Joint filling x x

DEGREE OF JOINTING
- Block size x x
- Joint spacing/frequency o   o x x x x
- RQD x x
- Number of joint sets x

JOINTING GEOMETRY OR STRUCTURE


- Joint orientation with respect to excavation x +
- Jointing pattern x
- Continuity  o o
- Structure (fold, fault)  +

EXTERNAL FEATURES
- Water conditions x x x
- Rock stress conditions o  x + x
- Blasting damage +
- Excavation dimensions  x x x

&/$66,),&$7,216<67(012Å    


/HJHQG
x well defined input o very roughly defined or included
 included, but not defined " partly included (in other parameters)
+ used as additional information (in RMR as adjusted value)
&ODVVLILFDWLRQV\VWHPQR
1 Terzaghi (1946) 5 Deere et al. (1969) 8 Wickham et al. (1972) 11 Matual and Holzer (1978)
2 Lauffer (1958) 6 RQD (1966) 9 RMR (1973) 12 Williamson (1980)
3 NATM (1957-64) 7 Franklin (1970, 1975) 10 Q-system (1974) 13 BGD (1981)
4 Coates and Patching (1968)
)URP3DOPVWU|P$50L±DURFNPDVVFKDUDFWHUL]DWLRQV\VWHPIRUURFNHQJLQHHULQJSXUSRVHV
3K'WKHVLV2VOR8QLYHUVLW\1RUZD\S

&KDSWHU

&2//(&7,212)*(2'$7$/,0,7$7,216$1'
81&(57$,17,(6

7KHJHRWHFKQLFDOHQJLQHHUVKRXOGDSSO\WKHRU\DQGH[SHULPHQWDWLRQEXWWHPSHUWKHPE\
SXWWLQJWKHPLQWRWKHFRQWH[WRIWKHXQFHUWDLQW\RIQDWXUH-XGJHPHQWHQWHUVWKURXJK
HQJLQHHULQJJHRORJ\
Karl Terzaghi, 1961

The purpose of this chapter is to outline briefly how rock mass properties are determined and to
show the special problems related to the uncertainties connected to acquisition and application of
geological data.

In contrast to most other materials used for construction purposes, judgement of the quality of the
rock mass is based on observations rather than test results. The large volume of the material
involved, and the ’given properties’ of the material, in addition to the lack of access to "see" the
actual material involved, cause great challenges in the execution of investigations, interpretation of
the results as well as characterization of the complex material called rock mass. Thus, there are no
clearly established guidelines when attempting to define the extent or scope of methods to be
applied in collection of relevant geo-data for a project (Merritt and Baecher, 1981). The
investigations and procedures may vary according to the nature of the project, the complexity of the
geology, the background of the engineering company, and the experience of the individual
geologists or rock mechanics engineers involved. A general approach to collection of data on
ground conditions in different stages of a project is shown in Fig. 3-4 where also the methods
applicable to finding parameters applied in the RMi system are indicated.

The fact that geological formations are spatially variable, and that only a limited number of
measurements or observations can be made, has important consequences. The principal one is that
the subsurface must be described and characterized by a limited number of parameters, and that the
values of these parameters are imprecisely known. It is important to accept the fact that a
geotechnical parameter is expressed by a range and also that the actual range may be greater than
that observed. Thus, in most cases it is recommended not to make too large an effort to obtain
accurate values of the various parameters. Often it is better to obtain a wider statistical material
(Einstein and Baecher, 1982).

Although the types of structures that may be encountered in different parts of a proposed tunnel can
be assessed, Terzaghi (1946) points out that it is not possible to make in advance of construction a
quantitative evaluation of the difficulties. Hence the first estimate of the material and equipment
required for constructing a tunnel inevitably involves a certain amount of guesswork.
3-2

 *(2'$7$)281'%()25('85,1*$1'$)7(5(;&$9$7,21

There are two main stages in the collection of geological data:


1. From observation and investigation on surface from which interpretations and extrapolations
have to be made. These data are mainly used for planning of a project and are collected EHIRUH
construction (see Fig. 3-1).
2. From characterization of ’known’ ground conditions; either in surfaces of excavations or in
outcrops belonging to the excavation to be made. This more straight forward characterization is
further dealt with later in this chapter.

Rock mass characterization EHIRUH construction requires an understanding of the rock mass expected
to be encountered, and the influence of the geologic variables inherent in any rock mass. Later,
GXULQJDQGDIWHU construction, the characterization is carried out in the surfaces of the excavation on
known conditions. The same parameters of the rock mass may be used in the characterization in all
phases of the project.

Almost all types of geo-data collection require some degree of extrapolation, with projection from
the known to the unknown (Piteau, 1973). How well this extrapolation is performed has obvious
practical implications, since the method(s) used often can influence the amount of necessary
subsurface exploration. Some principles of extrapolation are shown in Figs. 3-1 and 3-2.

Fig. 3-1 For deep tunnels the main source of geological information before construction is from surface
observations in outcrops. Core drillings combined with the surface information can improve the accuracy
of the geological interpretation and yield additional information of the rock mass condition. But as no
information is available of the conditions where the tunnel is planned, unexpected rock mass conditions
may be encountered.
3-3

Fig. 3-2 For shallow caverns and tunnels the rock mass conditions can be found from core drillings penetrating the
planned location. Surface mapping and geophysical measurements (refraction seismic) may further
improve the quality of geo-data.

the exact composition of


rock masses outside the
tunnel surface cannot be
observed or measured

joints that do not


intersect the tunnel
cannot be oserved

observations are limited


to roof, walls and working
face surfaces of the tunnel

Fig. 3-3 In tunnels and in man made cuttings the ’real’ rock masses can be observed in the excavated surfaces. The
rock mass behaviour is often governed by the conditions in the volumes surrounding the tunnel. It is not
possible to observe or measure the properties of the rock mass exactly.

The actual rock mass conditions are not known until they are encountered in the tunnel. But still it is
not possible to determine the exact nature of the rock mass involved as parts of these are located
outside the tunnel surface, as indicated in Fig. 3-3.
3-4

The following features can influence determination of the quality of the rock mass in an excavation:
- The possible development of new discontinuities from the excavation process.
- The limited observation of the actual rock mass conditions where the surface is covered by
mud, shotcrete, lining etc.
- The development of cracks from overloading of the ground surrounding the tunnel.

 620(0(7+2'686(',1*(2'$7$&2//(&7,21

The methods for the collection of geological data have not changed very much over the past 20
years and there is still no acceptable substitute for the field mapping and core logging carried out by
an experienced engineering geologist (Hoek, 1986). On certain projects it may be justified to set out
a complete rock investigation program and generate a full suite of tests. In general, however, only a
limited amount of testing is likely, and the main investigation is restricted to field observations. The
reason is often the high cost of sub-surface exploration by core drilling or by the excavation of trial
shafts and adits. The site of a proposed underground excavation is therefore seldom investigated as
fully as a design engineer would wish.

The field observations are mainly carried out in the following types of locations:
- outcrops or cuttings;
- pilot tunnels, adits or shafts made before construction; and
- excavated tunnels/shafts/caverns.

And in addition: in bore holes made from these locations; or on drill cores. Also tests are made or
samples for tests taken from these locations. More specifically, the various kinds of observations
and tests used in description, characterization, tests and measurements of rock mass features are:
• In outcrop and surface observations, in the form of:
- geological mapping;
- engineering geological mapping; and
- joint surveying.
• In bore holes, performed as:
- core drilling; or
- percussion boring.
• Geophysical methods, the main types being:
- seismic reflection measurements;
- seismic refraction measurements;
- cross hole tomography measurements;
- resistivity measurements; as well as
- radar, electromagnetic and
- gravity measurements.
• Laboratory tests of geological and mechanical properties of rock samples and limited volumes
of rock masses, some of the tests are:
- mineral composition and texture;
- compressive strength;
- tensile strength;
- shear strength;
- elastic constants; as well as
- density, porosity, anisotropy, and
- durability.
• Various field investigations and tests, such as:
3-5

- shear strength measurements of joints:


- deformation measurements in excavations;
- modulus measurements of rock masses;
- in situ stress measurements;
- hydraulic fracturing; and
- permeability tests.

For various reasons the simple observations made on the surface provide the most reliable data of
rock mass parameters (Bieniawski, 1984). If the observation of outcrops yields insufficient
knowledge for forecasting the ground conditions for a planned site, or if unfavourable rock
conditions are anticipated, various other investigations may furnish more specific information on
the rock conditions. The following sections briefly outline the principles in applying these for geo-
data collection and how rock mass features influence the results and procedures. Methods for
numerical characterizations from various types of measurements are described in Appendix 3.

*HRORJLFDOREVHUYDWLRQVDQGPDSSLQJ

Conventional geological mapping is conducted before construction to determine rock types,


delineate major geological structures, such as faults, dykes, lithological contacts, and any other
features that present major weakness zones in the mass. Large features like faults can be
individually mapped in outcrops, trenches, or exploratory openings. Their course can be
significantly better projected through intersections with borings. Regional or local geology may be
of use in projecting faults or large shears to unexplored parts of the rock mass, but poor correlation
between predicted and encountered geology in tunnelling casts doubt on the accuracy with which
structural features at the surface can be projected into a rock mass (Wahlstrom, 1969; Dowding and
Miller, 1975).

The knowledge of the underground from surface mapping is usually limited to information found in
sporadic point observations, core drilling and occasionally in extensive outcrops or cuttings. The
engineering geologist’s task is from this information, combined with geo-logical information, to
interpolate and to estimate the properties of the rock masses and the constituent rock types at the
site. The quality of this depends upon the simplicity of the geology, the experience of the geologist
and rock engineer, and in addition the possibility of observing representative rock masses in
outcrops. Table 3-1 shows the influence, which the geological setting may have on the confidence
of the rock mass characteristics observed.

TABLE 3-1 CLASSIFICATION OF OUTCROP CONFIDENCE (from Kirkaldie, 1988)


7(50 '(6&5,37,21

High level Massive homogeneous rock units with large vertical and lateral extent. History
of low tectonic stress levels.

Intermediate Rock characteristics are generally predictable but with expected lateral and
level vertical variability. Systematic tectonic stress features.

Low level Extremely variable rock conditions due to depositional processes, structural
complexity, mass movement or buried topography. Frequent lateral and
vertical changes can be expected. Frequent and variable tectonic stress
features.
3-6

plan investigations

limited drilling program


surface mapping and ground water geophysical exploration
investigations

laboratory tests on rock


samples and index field
tests on rock cores

PRELIMINARY EVALUATIONS:
- prepare geological maps and sections:
- show favourable and unfavourable regions;
- complete input data sheets for each structural region

engineering classification of
rock masses in each region

knowledge and use classification of rock masses to


experience compare excavation stability and support
available to requirements with documented evidence
engineers from sites with similar conditions

FEASIBILITY ASSESSMENT:
- critical examination of potential tunnelling problems
- preliminary design of tunnel cross sections
- tentative alternative construction and support methods

plan investigations

detailed exploratory adits


exploratory drilling
geological mapping with test enlargements

in situ rock
geophysical testing laboratory testing
mechanics tests

measurements of groundwater tests


in situ stresses

PROCESSING OF DATA
- prepare final geological maps and sections
- analyze results of laboratory and in situ tests
- classify rock masses in regions

FROOHFWLRQRIJHRGDWD

Fig. 3-4 The main principles in acquisition of data on geological and ground conditions before construction of a
project (modified from Bieniawski, 1984). Methods useful in collection of input data to RMi are shown
3-7

Geological and engineering geological mapping provide three types of information:


1. Rock type and distribution with estimate of strength. At this stage of the planning Herget
(1982) recommends that rock strength can be judged from simple hardness tests in the field
with geological pick or rebound hammer.
2. Location of major weakness zones or larger faults, which are described individually.
3. Description of jointing including data on spacings and frictional properties. Detailed joint
measurements can be carried out in specially designed joint surveys.

During and after excavation the observations are made on ’known’ surfaces. From this it follows that
the descriptions can be much more detailed for use in various analyses and methods of design.

3.2.2 Joint surveys

Joint surveys are mainly made during the planning stage of a rock construction to provide more
detailed information on the jointing. They can also be conducted for special engineering purposes,
such as for slope stability analyses in opencast mines, rock slopes, cuttings or valley sides.

Some form of VWDWLVWLFDO approach can be beneficial in such surveys because of the inherent
stochastic nature of joints and because complete information concerning their geometry can never
be obtained. Hudson and Priest (1979) find that the ability to express block lengths, areas and
volumes by statistical distribution functions will be of great assistance in the characterization of
rock mass geometry.

The objective of statistical sampling is to infer characteristics of a large population without


measuring all its members. The joints measured or VDPSOHG are only a portion of those exposed, and
these in turn are only a small part of all the joints in the rock mass. Various survey methods may be
used to sample the jointing, but in all instances the sample will have a bias dependent upon nature
of the exposed face and the method of sampling (Einstein and Baecher, 1983)

In common joint surveys, Einstein and Baecher (1983) use three geometric properties which can be
described statistically. These, which might be recorded in a number of equivalent measures,
involve:
- Density of joints (joint spacings, numbers per rock volume or per outcrop area).
- Size (trace lengths, joint surface areas or radii).
- Planar orientation (strike or dip direction and dip).

Sampling is conducted on limited exposed rock surfaces in the form of outcrops, trenches, and
tunnels in addition to bore hole walls and drill cores. If the surface is large, the number of joints
exposed may also be large and some form of selection may have to be applied to reduce the sample
size. Techniques which rely on the geologist’s judgement for recognising the joint sets of
importance can greatly reduce the volume, but there is always the risk of missing or discounting sets
which are nevertheless of considerable importance (Robertson, 1970). This risk is greatly reduced
when using some standard sampling plan, for instance:
- the unit area method (Krumbein and Graybill, 1965),
- the detail line survey, or
- the cluster sampling plan.

Piteau (1970, 1973) and Robertson (1970) have made important contributions to joint surveys from
practical experiences in South African mines. Joint surveys are also described by Baecher and
Lanney (1978) and Einstein and Baecher (1982, 1983).
3-8

No joint survey can furnish complete information about all joints present in a body of rock, but a
properly conducted survey can furnish data which have a high probability of approximating the
orientation, spacing and condition of these joints (Terzaghi, 1965).

3.2.3 Core drilling

The recovery of core by diamond drilling is used to obtain geo-information from volumes of rock
masses that cannot be observed. The drilling technique or the problems of obtaining high core
recovery of good quality is not outlined here, but rather the value of the information gained from
this type of field investigation is described. It is one of the most important methods of sub-surface
exploration. The information from drill cores can greatly improve the results from outcrop mapping,
and can also preferably be used to improve knowledge of the underground when combined with
geophysical measurements (Hoek and Brown, 1980; Hoek, 1981). Drilling from the surface or
probing ahead of an advancing heading is the most effective means of collecting information of the
rock mass condition.

The purpose of a core drilling investigation can be to:


- Confirm the geological interpretation.
- Obtain information on the rock types and their boundaries in the rock mass.
- Obtain more information of the rock mass structure.
- Study ground water conditions.
- Provide material (samples) for rock mechanics testing and petrographic analyses.

In "hard rocks" dominated by discontinuities core drillings are often carried out to study certain
larger faults or weakness zones which are assumed to determine the stability and ground water
conditions of the opening. The bore holes will, however, also give additional information where
they penetrate the adjacent rock masses.
Considering the very high cost of good quality core recovery, Hoek and Brown (1980) comment that
it is invariably worth spending a little more to provide for good routine core examination and
carefully prepared reports with high quality photographs of the cores before they are placed in
storage. A method for improved core description is presented in Appendix 3.

Kikuchi et al. (1985) have described the use of a bore hole television camera to observe joint
location and orientation, joint aperture, and presence of joint filling, i.e. information from bore holes
can be gained without obtaining cores.

/LPLWDWLRQVDQGGHILFLHQFLHVLQFRUHORJJLQJ

Core drilling results are not necessarily typical of the overall rock mass (Terzaghi, 1965). The
jointing density measured along a bore hole through a rock mass is often quoted as a single figure,
yet the value can depend significantly on the orientation of the line through the rock mass (Hudson
and Priest, 1983). The amount of variation that could be expected along bore holes or scan lines in
different directions is a function of the jointing geometry, particularly the degree to which the joints
tend to be orientated in certain preferred directions. In general, rock masses with a few joint sets
will exhibit much greater jointing variation than those with many sets. Also Deere et al. (1969)
described that it is usually difficult to obtain an accurate estimate of the joint spacing or the fracture
spacing on the basis of exploratory borings.
3-9

Lugeon or water pressure tests in bore holes can yield information of ground water conditions. The
measurements can, however, be markedly influenced by single joints, and the results can therefore
be misleading. There are also often uncertainties connected with the execution of the test, for
instance leakage through the packer. This is especially true for double packers.

Fig. 3-5 Only a very small part of the joint plane can be observed in a drill core.

It has long been desirable to use bore hole data alone as the basis for rock mass classification
without the need for additional tests in adits or pilot tunnels. As a result of the availability of more
advanced coring techniques, such as directional drilling and oriented core sampling as well as both
bore hole and core logging procedures, bore hole data are of increasingly better standard. Knowing
the deficiencies associated with bore holes mentioned above, such data - except where several
differently orientated bore holes are drilled - can seldom provide enough relevant data; a
combination of data from bore hole and observation of exposed rock surfaces form the best input for
rock mass classification, rock design and rock engineering.

Merritt and Baecher (1981) point out another limitation with core drilling: the problem of reliably
identifying faulting or shear zones in drill cores. 3RVWFRQVWUXFWLRQDQDO\VHVRIERULQJORJVDW
SRZHUSODQWVLWHVLQGLFDWHWKDWWKHOLNHOLKRRGRIUHFRJQL]LQJHYHQPDQ\LQFKZLGHIDXOWVLQERULQJ
ORJVLVH[WUHPHO\VPDOO

Section 3.8 in Appendix 3 and Section 4 in Appendix 4 outline how information obtained from drill
cores can be improved by better quality core logging, but still the quality of the data found is limited
due to spatial variability of rock masses and the fact that a bore hole only represents one dimension,
see Fig. 3-5.

3.2.4 Geophysical measurements

The high cost of sub-surface exploration by core drilling or by excavation of exploratory adits or
shafts, results in that the use of these investigations is generally limited. Geophysical methods can
often be used to supplement the information from such explorations.
3 - 10

Of the different geophysical methods for rock mass investigation the seismic methods and crosshole
tomography seem most promising in delivering useful information of rock mass features. Seismic
methods will not give satisfactory results in all geological environments (Hoek and Brown, 1980).
When geological conditions are suitable, seismic methods can give valuable information on the
structural orientation and configuration of rock layers and on the location of major geological
discontinuities such as weakness zones and faults.

In addition to structural settings the data provided by refraction seismic measurements may also be
used to estimate the jointing density. This is further described in Appendices 3 and 5. McFeat-Smith
et al. (1986) mention an example where a seismic refraction survey provided the most cost-effective
solution for locating zones of adverse tunnelling ground in Hong Kong’s igneous rock and densely
populated terrain.

3.2.5 Exploratory adits and shafts

The unreliability of projecting geological information obtained from surface mapping and core
drillings can be such that excavation of an adit or a shaft to provide access to the rock mass for more
detailed information at the site may be required. Such investigations are used for detailed
information of the rock mass conditions in the actual area. The characterization is carried out from
observations of the excavated surfaces. Special characteristics or properties may be measured by
large-scale field tests.

When surface exposure is limited, or when it is considered that those outcrops - which are available
- have been severely altered by weathering, the excavation of a trench or a shaft is sometimes
advisable (Hoek and Brown, 1980).

3.2.6 Laboratory and field tests

These tests are generally carried out for more specific measurements of individual properties of a
rock or a rock mass. Some of the test are made for investigation of inherent properties, other are for
stress or deformation measurements. A special part of such tests is connected to control of
deformations in excavations or slopes.

Many of the large field tests are very expensive. They can be performed in exploratory adits or
shafts before project construction or in tunnels or caverns during construction. As a consequence,
the value of results found from such tests should be carefully weighted against the costs (Franklin,
1970).

Laboratory testing methods for rock material are generally well established and testing techniques
have been recommended by the International Society for Rock Mechanics (ISRM) and the
American Society for Testing and Materials (ASTM).

There are sometimes doubts about the reliability of many of the tests and how their results can be
applied. Nieto (1983) mentions, for example, that DVXUSULVLQJQXPEHURILQVLWXWHVWLQJSUR
JUDPPHVLQLQWUXVLYHLJQHRXVURFNDQGKLJKJUDGHPHWDPRUSKLFVJLYHYDOXHVRIVWDWLFPRGXOXV
UDQJLQJEHWZHHQDQG*3DXVLQJEHDULQJSODWHVDQGEHWZHHQDQG*3DXVLQJIODW
MDFNV.
3 - 11

 81&(57$,17,(6$1'(55256,1*(2'$7$&2//(&7,21

,QWKLQNLQJDERXWVRXUFHVRIXQFHUWDLQW\LQHQJLQHHULQJJHRORJ\RQHLVOHIWZLWKWKHIDFW
WKDWXQFHUWDLQW\LVLQHYLWDEOH2QHDWWHPSWVWRUHGXFHLWDVPXFKDVSRVVLEOHEXWLWPXVW
XOWLPDWHO\EHIDFHG,WLVDZHOOUHFRJQL]HGSDUWRIOLIHIRUWKHHQJLQHHU7KHTXHVWLRQLVQRW
ZKHWKHUWRGHDOZLWKXQFHUWDLQW\EXWKRZ"
Herbert H. Einstein and Gregory B. Baecher (1982)

In connection with this section the following expressions need an explanation:


8QFHUWDLQW\ or lack of absolute sureness, in geology means that observations, measurements,
calculations and evaluations made are not reliable. The consequences are that the use of
geological data often may involve some kind of guesswork.
(UURU is defined as the difference between computed or estimated result and the actual value.
A ELDV is the difference between the estimated value and the true value of a statistic obtained by
random sampling. For example, R. Terzaghi (1965) pointed out that joints sub-parallel to an
outcrop have less chance of being sampled than joints perpendicular to an outcrop. This is a
bias in sampling for orientation.

Einstein and Baecher (1982) have defined three main sources for uncertainties and errors in
engineering geology and rock mechanics:
1. Innate, spatial variability of geological formations, where wrongly made interpretations of
geological setting may be a significant consequence. This has been outlined in the beginning of
this chapter.
2. Errors introduced in measuring and estimating engineering properties, often related to sampling
and measurements.
3. Inaccuracies caused by modelling physical behaviour, including incorrect type of calculations
or models.

In any engineering study, one can never know what has been left out of an analysis. Thus, in
addition to the three major uncertainties above, there is also uncertainty due to RPLVVLRQV. The real
world has variations and properties that can never entirely be included in a characterization or an
analysis. According to Einstein and Baecher (1982) most of the major failures of constructed
facilities have been attributed to omissions.

Some of the features in rock masses, which determine the quality of geo-data, are mentioned earlier
in this chapter. Consequently, they contribute to uncertainty. In this section, additional basic factors
causing uncertainties and errors are outlined.

3.3.1 Uncertainties caused by spatial variability of rock masses

The geological subsurface is spatially variable in that it is composed of different materials which are
stratified, truncated, and in other ways separated into more or less discrete zones. It is also spatially
variable in that within an apparently homogeneous body, material properties vary from point to
point. While with sufficiently many observations this variability can be precisely characterized, the
number of observations is usually limited. Thus, uncertainty remains concerning material properties
or classification at points not observed.

The variability can be such that a construction in rock, within short distances, may encounter the
most diverse conditions. It is nearly impossible to uncover all the important variations by present-
3 - 12

day exploration techniques. A large number of case histories attest to the frequency with which
unexpected conditions occur, often with disastrous results. A larger part of these unexpected events
have been caused by wrong geological interpretation, i.e. how the main geological structures are
distributed below the surface (Merritt and Baecher, 1981). From this it is clear that the quality of the
geological interpretation is vital for characterization of the rock masses at the construction site
before construction. Where the geological interpretation is wrong or incorrect, it follows that the
assumed rock mass characterization is equally wrong - however good the description of the
observed rock masses is.

Error from the geological variability can never be avoided; it can, however, be reduced by the use of
well experienced geologists with extensive knowledge of the geology in the actual region, and also
by directing specific, appropriate investigations towards possible key geologic structures that may
occur.

3.3.2 Measurement errors

The division of errors described by Krumbein and Greybill (1965) is used to explain inconsistencies
in repeated measurements of the same quantity. They have outlined four kinds of measurement error
commonly recognized or associated with the observer, the instrument, the operational definition,
and the measurement process itself, namely:
*URVVHUURUV, attributable mainly to blunders on the part of the observer, are usually large in
magnitude and irregular in occurrence. They may result from momentary inattention, and when
subsequently noted, the observation may be discarded by the observer.
6\VWHPDWLFHUURUV arise when measurements tend to be consistently either too large or too small.
They may be produced by a miscalibrated instrument, but they also occur from such external
conditions as atmospheric moisture (in air-drying a sediment, for instance)
(UURUVRIPHWKRG occur when there is a discrepancy between the conceptual definition of the quality
to be measured and the operational definition used to make the measurement. The quantifi-
cation process in geology gives rise to many situations in which errors of method may occur.

When measurement processes are free of the errors listed above, there still remain fluctuations in
the numerical values obtained by repeated measurement of the same object. These are XQSUHGLFWDEOH
GHYLDWLRQV or UDQGRPHUURUV, which in the long run may be compensating, in that positive and
negative deviations tend to balance each other, so that the average value of the numbers tend to
approach the ’true value’.

Farmer and Kemeny (1992) show that, apart from a few simple physical property tests, virtually
none of the methods used in rock testing give reliable data. The main reason for inadequacy in test
results - which is accepted in most engineering design in rock - can be explained by the complex
and variable composition and structure of rocks and rock masses.

Another significant measurement error is associated with the angular measurement of dip and strike.
This error varies with the inclination of the joint, increasing as the joint tends to be horizontal. For
flat-lying structures of the order of 5 - 10o, where the horizontal line of projection is extremely
limited, such as for joint in a tunnel wall, Robertson (1970) has experienced that the measured strike
may vary as much as ± 20o. For attitude measurements of planar features, Friedman (1964)
estimates accuracy of ± 1o for dips greater than 70o and ± 3o for inclinations of 30 - 70o. The latter
estimates may apply to mapping of large surface outcrops, but not to observations of limited
dimensions such as in tunnels.
3 - 13

Ewan et al. (1983) reports from an interesting investigation carried out in the Kielder aqueduct
tunnels, UK, to see the reproducibility of joint spacing and orientation measurements:

Three 10 m long scanlines were set up in each of the three rock types: sandstone, mudstone and
limestone. On each scanline 6 experienced observers recorded the position and the orientation
of each joint (less than 15 m long), see Fig. 3-6.

OBSERVATIONS ALONG WALL


of joints
number

Observer 0 1 2 3 4 5 6 7 8 9 10 m

MDR 18

JT 21

GHA 19

GW 17

DAB 20

VJE 17

Fig. 3-6 Position of joints recorded by different observers on one of the scanlines (modified from Ewan et al.)

By comparing the results of the measurements carried out by the 6 persons it was found that:
- The variation in the number of joints recorded by different observers along any one
scanline varied considerably. The ratio between the highest and lowest number of joints
recorded was as high as 3.8 , but with a mean of about 2. (The maximum number of joints
along a scanline was 37.)
- The average maximum error in measurement of joint orientation was ±10o for dip direction
and ±5o for dip angle.
The fact that different observers did not identify joints at the same position underlines the
difficulty of interpretation of joints and jointing.

Piteau (1973) mentions that since many joints are highly undulating and the scale of the tunnel or
observation area often is much smaller than that of the joint, measurements of both strike and dip
may be extremely erroneous, depending where the joint is measured. Piteau has further observed
that man-made cracks from blasting in open pit mines probably involve less than one per cent of the
joints recorded. Löset (1992) has, however, experienced a significant impact from blasting on the
amount of rock support in a partly blasted, partly full face bored tunnel. The influence from blasting
may lead to higher values of the degree of jointing, joint length and joint continuity being recorded.

A more serious error may come in outcrops from joints developed by the effect of weathering.
Extrapolating data from weathered outcrops should, as mentioned earlier in Section 3.1, be done
carefully.

In addition to the errors mentioned above significant errors may be introduced by the
characterizations caused by poor definitions and/or personal interpretations.

A complete description of joints is difficult because of their three-dimensional nature and their
limited exposure in outcrops, borings or tunnels. According to Dershowitz and Einstein (1988), the
ideal characterization of jointing would involve the specific description of each joint in the rock
mass, exactly defining its position and geometric and mechanical properties. This is not possible for
a number of reasons, among others:
3 - 14

1. The visible parts of joints are limited, for instance to joint traces only, and thus prevent
complete observation.
2. Joints at a distance from the exposed rock surfaces cannot be directly observed.
3. Direct (visual or contact measurements) and indirect (geophysical) observations have limited
accuracy.
For these reasons joints in the rock mass are usually described as an assemblage rather than
individually. The assemblage has a stochastic character in that joint characteristics vary in space.

Joints show great variation in properties and some of the most significant errors due to selection of
joints to be characterized are according to Robertson (1970):
- Small joints are often disregarded.
- Very large fracture surfaces may be measured more than once.
- Joints almost parallel to the foliation or bedding may be overlooked.
Baecher and Lanney (1978) have confirmed similar trends from their studies.

3.3.3 Model uncertainties

Models used in assessments of rock mass behaviour are mainly based on theory and/or empirical
relations. As they are simplifications of reality, modelling errors are introduced. Modelling errors
are caused by errors in the theory assumed to apply to physical processes, boundaries and initial
conditions which must be chosen, errors introduced by numerical or mathematical approximations,
and important factors left out of the model. Sometimes, of course, modelling errors in predicting
engineering performance and modelling errors in estimating material properties, partially
compensate (Einstein and Baecher, 1982).

 6800$5<

As most of the geo-data collection is based on observations - either on outcrops, on tunnel surfaces
or on drill cores - it is important to know whether the condition of the rock mass observed is
representative, see Table 3-2. Some of the main parameters that determine the mechanical properties
of a rock mass have been listed in Table 3-3 along with various possibilities to collect them.

As a result of possible errors, many of the measurements and tests used in rock mechanics, while
useful in identifying rock behaviour characteristics and in empirical comparisons of rock behaviour,
have limited use in design (Farmer and Kemeny, 1992).

The great spatial variability and great volumes involved result in that only a limited number of
measurements can be made. The subsurface must, therefore, be described by a limited number of
imprecisely known parameters. Interpretations and extrapolations that are made to work out the
geological setting may introduce considerable uncertainties. Also, the fact that horizontal and other
features, which do not outcrop, may be overlooked, adds to these errors.
3 - 15

TABLE 3-2 POSSIBLE FEATURES THAT REDUCE INFORMATION AND QUALITY IN SURFACE
OBSERVATIONS
)($785(67+$70$<5('8&(0($685( &216(48(1&(6)257+(0($685(0(176
0(1725,17(535(7$7,2148$/,7<

2EVHUYDWLRQVLQRXWFURSV
- loose material, vegetation, water, snow, or ice - No or limited area of exposed rocks for
which cover the rock surface; observations
- weathered rocks occur in the outcrop (but not - The rock conditions observed are different from the
deeper underground). conditions in deeper located rock masses

2EVHUYDWLRQVLQH[FDYDWHGFXWWLQJVWUHQFKHV
DGLWVHWF
- weathered rocks occur in the surface (but not - the rock conditions observed are different from the
deeper) conditions in deeper located rock masses

2EVHUYDWLRQVLQGHHSORFDWHGXQGHUJURXQG
RSHQLQJV
- the surface has been covered by mud, shotcrete or - the cover hides the rock conditions for
other remedial (before geological mapping) observation

TABLE 3-3 INFORMATION ON CHARACTERISTIC ROCK MASS PARAMETERS OBTAINED FROM


VARIOUS POSSIBILITIES OF DATA COLLECTION.

'$7$&2//(&7(')520
52&.0$66 '5,// 81'(5 5()5$&7,21
3$5$0(7(5 &25(6 *5281' 287&5236 6(,60,&
$',76
23(1,1*6 352),/(6
5RFNV
- distribution of rocks x x x x/(x) -
- sample for strength tests x x x x/(x) -

-RLQWVDQGMRLQWLQJ
- joint spacing (x) x x x (x)
- joint length - (x) (x)/x x/(x) -
- orientation - x x x -
- waviness - x x x -
- smoothness (x) x x x/(x) -
- filling or coating (x)/x x x (x)/- -

)DXOWVDQGZHDNQHVV]RQHV
- persistence - - - (x) -
- orientation - (x) x (x) -
- thickness of zone - (x)/x x (x)/- (x)
- gouge material (x) x x - -

here is: x parameter or task can be measured


(x) parameter or task may partly or sometimes be measured
- not possible to measure the parameter or task

From the foregoing it has been found that the following features may cause uncertainties, and errors
and hence reduced quality of rock mass characterizations:
- The spatial occurrence, variations and large volume of the material (i.e. rock masses) involved
in a rock construction.
- The geological interpretation, on which the characterizations are based.
- How the investigations are performed.
- Uncertainties connected to the joints measured, as they may only be a portion of the joints
exposed which are considered to be representative of the joints within the entire rock mass.
3 - 16

- Outcrops or surfaces, where they occur, may not be representative due to weathering.
- In excavated surfaces and in drill cores it may be difficult to distinguish between natural and
artificially induced discontinuities.
- Limitations in drill core logging: soft gouge is lost during core recovery and information
relating to the waviness and the continuity of joints is minimal.
- The way the description is performed or the quality of the characterization made of the various
parameters in rock masses. As most of the input parameters in rock engineering and rock
mechanics are found from observations, additional errors may be introduced from poorly
defined descriptions.

All these aspects have important consequences in the application of geo-data in rock mechanics,
rock engineering, and construction design. The main conclusions of this chapter are therefore:
1. Although extensive field investigation and good quality descriptions will enable the engineering
geologist to predict the behaviour of a tunnel more accurately, it cannot remove the risk of
encountering unexpected features.
2. A good quality characterization of the rock mass will, however, in all cases, except for wrong
or incorrect interpretations, improve the quality of the geological input data to be applied in
evaluation, assessment or calculations and hence lead to better designs.
3. The methods, effort and costs of collecting geo-data should be balanced against the probable
uncertainties and errors.

The work behind this thesis is mainly directed to improve the characterization in the second item
listed above.
)URP3DOPVWU|P$50L±DURFNPDVVFKDUDFWHUL]DWLRQV\VWHPIRUURFNHQJLQHHULQJSXUSRVHV
3K'WKHVLV2VOR8QLYHUVLW\1RUZD\S

&KDSWHU

7+(&20%,1$7,212)*(2'$7$,172$52&.0$66,1'(;

7U\DOZD\VWRFRPELQHWKHRU\DQGSUDFWLFHDQGWRFRQIURQWLGHDVDQGH[SHULHQFH
Leopold Müller (1982)

Construction materials such as concrete, most metals, wood etc. used in civil and mining
construction are characterized or classified according to their strength properties. This basic quality
information of the material is used in the engineering and design for various construction purposes.
In rock engineering, no such specific strength characterization of the rock mass is applied. Most
rock engineering is carried out using various descriptions, classifications and unquantified
experience. Although the various utilizations of rocks and rock masses have different purposes and
are subjected to various problems, the suitability and quality of the rock mass depends largely on its
strength properties.

Hoek and Brown (1980), Bieniawski (1984), Nieto (1983) and several other authors have indicated
the need for a VWUHQJWKFKDUDFWHUL]DWLRQ of rock masses. Williamson and Kuhn (1988) are of the
opinion that QRFODVVLILFDWLRQV\VWHPFDQEHGHYLVHGWKDWGHDOVZLWKDOOWKHFKDUDFWHULVWLFVRIDOO
SRVVLEOHURFNPDWHULDOVRUURFNPDVVHV:KDWZHDUHDLPLQJIRUWKHUHIRUHLVDV\VWHPWKDWZRXOG
JHQHUDOO\JURXSWKHURFNVLQVXFKDZD\WKDWWKRVHSDUDPHWHUVZKLFKDUHRIPRVWXQLYHUVHFRQFHUQ
DUHFOHDUO\GHDOWZLWKDQGWKHQXPEHURIV\PEROVDUHNHSWWRDPLQLPXP

The Rock Mass index, RMi, has been developed as a general strength characterization of the
structural material that a rock mass represents. It therefore includes only the inherent features or
parameters in the rock mass. This prerequisite is in accordance with the ideas by Patching and
Coates (1968) who presented a general characterization based on the intrinsic parameters of a rock
ZKLFKDUHWKHVDPHLUUHVSHFWLYHRISODFHRUFLUFXPVWDQFHV)RUWKLVUHDVRQLWZDVFRQVLGHUHG
QHFHVVDU\WRRPLWIDFWRUVUHODWHGWRHQYLURQPHQWIURPWKHFODVVLILFDWLRQDOWKRXJKVWUHVV
DSSOLFDWLRQVSRUHZDWHUDQGRWKHULQIOXHQFHVKDYHDSURQRXQFHGHIIHFWRQWKHEHKDYLRXURIDURFN
LQDQ\JLYHQVLWXDWLRQ-XVWDVDVWUXFWXUDOHQJLQHHUZKRLVGHVLJQLQJDVWHHOVWUXFWXUHZLOOHVWDEOLVK
WKHVWUHVVGLVWULEXWLRQVRIWKHVWUXFWXUHVHSDUDWHO\IURPWKHVSHFLILFDWLRQVRIWKHVWHHOVRLQDQ\
VSHFLILFSUREOHPLQURFNPHFKDQLFVWKHHQYLURQPHQWDOIDFWRUVZLOOEHFRQVLGHUHGDQGHVWDEOLVKHG
IRUWKDWSUREOHPLQDGGLWLRQWRWKHGHWHUPLQDWLRQRIWKHQDWXUHRUFODVVLILFDWLRQRIWKHURFN The
classification of Patching and Coates (1968) was, however, descriptive and therefore not useful in
calculations.

Based on the author’s own experience and published papers in this context the following
considerations have been important during the development of RMi and its input data:
- Few input data should be included to arrive at a simple expression.
- Existing methods should be applied for geo-data acquisition where possible.
- Simple and practical methods for finding the input values should be preferred.
- Guidelines should be developed for adequate descriptions so that they can be "translated" to
numerical values.
4-2

- Correlations should be developed so that input data from various types of measurements can be
used.
7+(6758&785(2)7+(52&.0$66,1'(;

From the outline in Chapters 2 and 3 and in Appendix 1 it is clear that a rock mass is a material
much more complex in composition, structure, variability than most other structural materials. The
presence of various defects (discontinuities) in a rock mass, which tend to reduce the inherent
strength of the rock, constitutes the main feature in its behaviour. This fact is the main principle of
the Rock Mass index (RMi), as explained in this chapter.

7KHLQSXWSDUDPHWHUVVHOHFWHG

Although external forces acting on a rock mass, such as induced rock stresses or water pressures are
justifiably included in some classification systems for designing tunnel support (see Table 2-3), the
strength properties of the rock mass is not a direct function of these features. Deere et al. (1969)
proposed to use parameters related to the character of discontinuities as the PDLQ feature for both a
general and a diagnostic classification system. This use of jointing as the major input does not,
however, exclude the importance of the rock material on the behaviour of the rock masses. For
example, if joints are widely spaced or if the rock material is weak, the properties of the intact rock
may strongly influence the gross behaviour of the rock mass. In addition, the rock material is also
important if the joints are not continuous. This view of rock mass behaviour and strength has been
further outlined by Wood (1991) as:
- Better quality rock masses are determined by the geometry of the rock mass structure, specially
block size and block shape.
- Fair to poor quality rock masses are determined by the inter-block shear strength and
deformational characteristics.
- Very poor quality rock masses mainly depends on the low strength of the intact material.

For jointed rock masses, Hoek et al. (1992) are of the opinion that the strength characteristics are
controlled by the block shape and size as well as their surface characteristics determined by the
intersecting discontinuities. They recommend that these parameters are selected to represent the
average condition of the rock mass. Similar ideas have been set forth by Tsoutrelis et al. (1990),
Matula and Holzer (1978), Coates and Patching (1968) and Milne et al. (1992). These
considerations have been used in the selection of the following input parameters in a general
strength characterization of a rock mass:
- the size of the blocks delineated by joints, - measured as block volume;
- the strength of the block material, - measured as uniaxial compressive strength;
- the shear strength of the block faces, - measured as friction angle, and
- the size and termination of the joints, - measured as length and continuity.
This is shown schematically in Fig. 4-1. It is considered, however, that taken together, they provide
a fairly complete indication of the strength of a given rock mass. Spesific features such as faults,
dykes and shear zones, should be considered separately (Bieniawski, 1984, 1989).

An additional, also important rock mass parameter, WKHEORFNVKDSH, is not directly included in RMi.
The main reason for this is the objective to maintain a simple structure of RMi. Block shape, being
a geometric delineation of the three-dimensional pattern of jointing, is, however, indirectly included
in the block volume, as the block volume varies with its shape. This is further described in
Appendices 3 and 4.
4-3

Fig. 4-1 An aggregate of blocks delineated by joints indicating the parameters selected for a general rock mass
characterization

Numerical values alone are seldom sufficient for characterizing the properties of such a complex
material as a rock mass. Therefore, numerical values should be accompanied by supplementary
descriptions as presented in Appendix 3 where the requirements for extracting numerical values
from descriptions are outlined.

7KH5RFN0DVVLQGH[ 50L

The main principle in the development of RMi has been focussing on the effects of the defects in a
rock mass in reducing the strength of the intact rock. The RMi is thus defined as
RMi = σc × JP eq. (4-1)

Here σc = the uniaxial compressive strength of the intact rock material, and
JP = the jointing parameter, see Fig. 4-2. It is a reduction coefficient representing the
block size and the condition of its faces represented by their friction properties and
the size of the joints, see Fig. 4-1. The value of JP varies from almost 0 for crushed
rocks to 1 for intact rock. Its value is found by combining the block size, and the joint
conditions as described in the next section in this chapter.

JOINT
ROUGHNESS
JOINT
JOINT CONDITION
ALTERATION FACTOR
jC
JOINT SIZE AND -2,17,1*
TERMINATION 3$5$0(7(5
-3
DENSITY BLOCK VOLUME
OF JOINTS Vb
52&. 0$ 66 ,1'(;
81,$;,$/ 50L
ROCK &2035(66,9(
MATERIAL 675(1*7+
σF

Fig. 4-2 The principle of the RMi characterizing the material properties of a rock mass.

As may be noticed, RMi is not dimensionless, but has the units of σc . The individual input
components of RMi, the rock strength, and the jointing features are measured and combined to
deduce this rock mass strength. Results from large scale and field measurements of rock mass
4-4

strengths have been used to develop an expression for RMi as close as possible to reality. This is
further described below in Section 4.2.
7KHFRPELQDWLRQRIWKHLQSXWSDUDPHWHUV

The importance of the two main contributors to RMi, the compressive strength of intact rock (σc)
and the jointing parameter (JP) is further described in this section.

The parameter for the rock material (σc), the uniaxial compressive strength of intact rock, is used
directly. Its value can be determined from laboratory tests. Estimates of (σc) can also be obtained as
described in Appendix 3.

The jointing parameter (JP) is a combination of the block size, measured as its volume (Vb), and the
joint condition factor (jC), see Fig. 4-2:
The block volume, Vb, is a measure of the degree of jointing or the density (amount) of joints.
As it is a 3-dimensional measure, it indirectly also is an expression of the overall geometry of
the rock mass. It can be determined from field measurements of the block dimensions as further
described in Appendix 3.
The joint condition factor, jC, represents the inter-block frictional properties. Barton et al.
(1974) have in their Q-system chosen the roughness and alteration factors (Jr and Ja) to
represent the importance of dilatancy and shear strength of joints. The ratio of the two
parameters (Jr/Ja) represents, with the ratings they are given in the Q-system, a fair
approximation to the actual VKHDUVWUHQJWKSURSHUWLHV of the joint within normal variations of
these factors (Barton et al, 1974; Barton and Bandis, 1990). It appears, therefore, logical to
make use of the same values and combination of these parameters for the joint condition factor
in the RMi. 1
A joint size factor (jL) has been chosen as a size correction factor for joints. The reason for this
is the fact that larger joints have a markedly stronger impact on the behaviour of a rock mass
than smaller. In addition, the continuity or termination of the joint has been included. This part
of the factor is divided between joints that terminate in massive rock, i.e. discontinuous joints,
and other joints. The effect of a discontinuous joint is much less as the failure plane must partly
pass through intact rock.

The factors included in the joint condition factor are combined in the following way:
jC = jL × jR/jA eq. (4-2)

Here jR = the joint roughness factor of the joint wall surface and its planarity (similar to Jr in
the Q-system),
jA = the joint alteration factor, representing the character of the joint wall (the presence
of coating or weathering and possible filling characteristics). It is similar to Ja in
the Q system, and
jL = the joint size and continuity factor.

The numeric values of these components of jC can be found from various field observations
and measurements as described in Appendix 3.

1
The symbols Jr and Ja have been changed into jR and jA because some minor modifications have
been made in their definitions.
4-5

 &$/,%5$7,212)50L)520.12:152&.0$66675(1*7+'$7$

7KH SXUSRVH RI VFLHQFH LV WR VLPSOLI\ QRW WR FRPSOLFDWH 7KH IXQFWLRQ RI DQ HQJLQHHULQJ
JHRORJLVWJHRWHFKQLFDORUURFNHQJLQHHULVWRH[DPLQHDQGREVHUYHWKHFRPSOH[YDULDEOHVRIDQ
DUHDRUSURMHFWVLWHDQGIURPWKLVHIIRUWDUULYHDWDVHWRIVLPSOHVLJQLILFDQWJHQHUDOL]DWLRQV
Douglas A. Williamson and C. Rodney Kuhn (1988)

It is practically impossible to carry out triaxial or shear tests on rock masses at a scale similar to that
of surface or underground excavations (Hoek and Brown, 1988). The numerous attempts made to
overcome this problem by modelling generally suffer from the limitations and simplifications,
which have to be made in order to permit construction of the models. Consequently, the possibility
of predicting the strength of jointed rock masses on the basis of direct in situ tests or of model
studies is very limited.

This problem resulted in that Hoek and Brown (1980), during development of the Hoek-Brown
failure criterion for rock masses, had very few strength data available (see Section 8.1). Their
criteria for jointed rock masses are, therefore, based almost wholly on the laboratory tests carried
out on Panguna andesite described by Jaeger (1969). In addition to these data on the Panguna
andesite, for working out the RMi it has been possible to make use of some few more results of
triaxial laboratory tests on large scale samples of rock masses, including:
- clay schist, sandstone, and siltstone from various locations in Germany, and
- granite from the Stripa test mine, Sweden.
Also results from in situ tests of quartzitic sandstone in the Laisvall mine in Sweden, and a back
analysis from a large slide in quartzite and schist in the Långsele mine in Sweden have been used in
the calibration.

As the rock mass index is meant to express the compressive strength of a rock mass (σcm ) it can be
expressed as RMi = σcm = σc × JP. The uniaxial compressive strength of intact rock (σc ) is defined
and can be determined within a reasonable accuracy. The jointing parameter (JP), however, is a
combined parameter made up of the following features:
- the block volume (Vb) which can be found from measurements, and
- the joint condition factor (jC) which is the result of three independent joint parameters
(roughness, alteration and size).

The results from the tests and back analysis have been used to determine how Vb and jC can be
combined to express JP (and RMi accordingly when σc is known). This calibration has been
performed in the following way:
1. From the known results of the tests or back analysis, namely of
- the uniaxial compressive strength of the rock mass (σcm ) and
- the uniaxial compressive strength of the intact rock (σc )
the value of the jointing parameter is by definition
JP = RMi/σc = σcm /σc eq. (4-3)
2. Numerical characterizations have been made of the joints and the block characteristics in the
actual rock mass 'sample' tested to find
- the block volume (Vb), and
- the joint condition factor (jC) found from eq. (4-2).
3. The data from the tests described in Appendix 6 and shown in Table 4 -1 have been plotted
on the diagram in Fig. 4-3. Log. scales have been used both for the jointing parameter (JP)
along the x-axis and for the block volume (Vb) along the y-axis. As joint spacing (i.e. block
size) generally has an exponential distribution (see Appendix 1, Section 5), the lines
representing jC are expected to be straight.
4-6

4. From the values of block volume (Vb) and jointing parameter (JP) the position of the
corresponding joint condition factor (jC) is found for each of the data sets. As a best fit to these
data the lines representing jC have been drawn, as shown in Fig. 4-3.

TABLE 4-1 THE RESULTS FROM LARGE SCALE TESTS ON ROCK MASSES FURTHER DESCRIBED IN
APPENDIX 6.
6DPSOH/RFDWLRQ 5RFNW\SH σF M& 9E -3
QR MPa
1 Panguna andesite 265 4-6 2 - 6 cm3 0.014
2 Stripa granitic rock 200 1.5 - 2.5 5 - 15 dm3 0.04
3 Laisvall mine sandstone 210 0.75 -1 0.1 - 0.3 m3 0.095
4 Långsele mine grey schist, greenstone 110 - 1600 2 - 0.3 8 - 20 dm³ 0.01
5a Thüringer wald clay-schist 55 1.5 - 2 5 - 10 dm3 0.055 *)
5b " " " 100 2 - 2.5 5 - 10 dm3 0.08 **)
6 Hessen sandstone/claystone 10.5/4.8 5 - 10 (?) 1 - 5 dm3 0.17
7 Hagen siltstone 65 3.5 - 4.5 5 - 10 dm3 0.10
*) **)
Tests parallel to schistosity Tests normal to schistosity

 



5

2 P


% O R F N Y R O X P H 9 E
2
SAM PLE 3: jC = 0 .7 5 - 1.5


SAM PLE 7: jC = 3 .4 - 4.5 5

2
SAM PLE 4: jC = 0.2 - 0.3


5
SAM PLE 5: jC = 2 - 2.5
2
SAM PLE 6: jC = 5 - 10? 
P
G
SAM PLE 2: jC = 1.5 - 2.5 5



  2
 
SAM PLE 1: jC = 4 - 6 M& 
 
 


 5


  2
 

 


& 5
M


2 3 5 7 2 3 5 7 2 3 5 7 2 3 5 7
                  

- R L Q W L Q J S D U D P H W H U - 3

Fig. 4-3 The connection between block volume, joint condition and jointing parameter determined from plotting of
the data sets described in Appendix 6.
4-7

0.1 0.2 0.3 0.5 0.7 1

1 0.8 

0.8 1 P approx. scale effect of 5

ma ss iv e ro c k
0.6
compressive strength ( σ c )
(joints/m )

% / 2 & . 9 2 / 8 0 ( 9 E

3

1 
1.5 0.5 J
2
Q
1.5 L
0.4 F
2 D
1.5
2 S
 V 5
 0.3 
W 2 W
Q 3 Q
L
X
R
3 R
M   2
F 0.2 
 4 H
W 3
4 J
Q
L 5 D
R U 5
M 4 5 H
 6
F
L 6
Y
'
4
5 $
U 2
5
W 0.1
H 6 8
P 8 0.08
X 10
O 8
R 10 5
9 0.06
10
15 0.05 2
15
20 0.04
15
20
0.03  5
20
30
30 2
0.02
30
50
5
50 60
50 60
2
80 0.01
60
80
100
80 100  5
100
2

2 3 5 7 2 3 5 7 2 3 5 7 2 3 5 7 2 3 5 7 2 3 5 7

Fig. 4-4 Diagram for finding the value of the jointing parameter (JP) from the joint condition factor (jC) and
various measurements of jointing density (Vb, Jv, RQD).

Examples shown in Fig. 4-4:


 For Vb = 0.00005 m3 (50 cm3 ) and jC = 0.2 the JP = 0.0006;
 For two joint sets with jC = 1.5 and the volumetric joint count Jv = 3.3 the JP = 0.3;
 For one joint set with spacing S = 0.25 m and jC = 8 the JP = 0.5 (determined by the scale
effect of compressive strength in massive rock)
 For RQD = 50 and jC = 1 the JP = 0.03.

The jointing parameter can also be determined by the following expression which has been derived
from the lines representing jC in Fig. 4-3:
4-8

JP = 0.2 jC × VbD eq. (4-4)

where Vb = the block volume, given in m3, and


D = 0.37 jC - 0.2, which has the following values:
jC = 0.1 0.25 0.5 0.75 1 1.5 2 2.5 3 4 6 9 12 16 20
D= 0.586 0.488 0.425 0.392 0.37 0.341 0.322 0.308 0.297 0.28 0.259 0.238 0.225 0.213 0.203

Fig. 4-4 shows the same diagram where also other measurements than block volume can be applied
directly. These are located in the upper left part of the figure. Here, the volumetric joint count (Jv)
for various block shapes can be used instead of the block volume. Also, RQD can be used directly
with the limitations of the accuracy in this measure as given in Appendix 4.

RMi is a material characterization of the structural material called "rock mass" as it involves only its
inherent features. As it has a general form, RMi is not a quality characterization, but merely, within
its limitations, a rock mass index strength, as further outlined in Section 4.5. The classification
presented in Table 4-2 is suggested for RMi.

TABLE 4-2 CLASSIFICATION OF THE RMi


&+$5$&7(5,=$7,21
50L9$/8(
Term for RMi Term related to rock
(MPa)
mass strength
Extremely low Extremely weak < 0.001
Very low Very weak 0.001 - 0.01
Low Weak 0.01 - 0.1

Moderately high Moderately strong 0.1 - 1


High Strong 1 - 10
Very high Very strong 10 - 100
Extremely high Extremely strong > 100

For the most common joint conditions where jC = 1 - 2, the jointing parameter will vary between
JP = 0.2 Vb 0.37 and JP = 0.28 Vb 0.32. For jC = 1.75 the jointing parameter can simply be
expressed as
JP = 0.25 3 Vb eq. (4-5)
and for jC = 1:
JP = 0.2 Vb 0.37 eq. (4-6)

The graphical solution of eq. (4-6) is presented in Fig. 4-6, from which estimates of RMi can be
quickly made from the block volume and the uniaxial compressive strength of the rock

As shown in Fig. 4-5 significant VFDOHHIIHFWV are generally involved when a ’sample’ is enlarged
from laboratory size to field size. After the calibration described above, RMi is tied to large samples
where the scale effect has been included in JP. The joint size factor (jL) is also a scale variable. For
massive rock masses, however, where the jointing parameter JP ≈ 1 the scale effect for the uniaxial
compressive strength (σc) must be accounted for, as σc is related to a 50 mm sample size. Barton
(1990) suggests from data presented by Hoek and Brown (1980) and Wagner (1987) that the actual
compressive strength for large ’field samples’ with diameter (d is measured in mm) may be
determined from
σcf = σc50 (50/d) 0.2 = σc50 (0.05/Db) 0.2 = σc50 × fσ eq. (4-7)
4-9

where σc50 = the uniaxial compressive strength for 50 mm sample size, and
fσ = (0.05/Db) 0.2 is the scale factor for compressive strength
400
UNIAXIAL COMPRESSIVE STRENGTH (MPa)

300 Hoek and Brown curve


σc = σc50 (50/d)
0.18

200

Experimental data
σ = σ50 (50/d)
0.22
100

0
1000 2000 3000
SIZE (mm)

Fig. 4-5 Empirical equations for the scale effect of uniaxial compressive strength (from Barton, 1990, based on
data from Hoek and Brown, 1980 and Wagner, 1987). Barton suggests to apply a value of 0.2 for the
exponent.

100 ) 2 5
& 2 0 0 2 1
 -2 , 17 & 2 1 ' , 7 , 2 1 M &    

2
RM
10 i=
10
3
m

5 0
60
2 40
1 20
5 15
%ORFN YROXPH 9E

10
2
6
100 4
5
2
2
3
dm

10 1
0.
5 6
0.
4
2
0.
1 2
5 0.
1
0.
2 06
0.
04
100
0.
5 02
2 0.
01
cm3

0.
10 00
0.
5 00 6
4
0.
2 00
2
1
1 2 3 5 7 10 20 30 50 70 200 300 500 MPa
100
8QLD[LDO FRPSUHVVLYH VWUHQJWK RI LQWDFW URFN σ F

Fig. 4-6 RMi for ’common’ joint condition (jC = 1 - 2) based on the uniaxial compressive strength of intact rock
and the block volume. Example: for Vb = 0.2 m3 and σc = 100 MPa the RMi = 25 MPa.
4 - 10

Eq. (4-7) is valid for sample diameter up to some metres, and may therefore be applied for massive
rock masses as indicated in Fig. 4-4. The block diameter (Db) may be found from
o 3 27 3
Db = Vb = Vb (eq. (6-8))

3
as presented in Appendix 3 and in Chapter 6, or more approximately as Db = Vb
or simply by applying the spacing for the main joint set.

 180(5,&$/9$/8(62)7+(,13873$5$0(7(567250L

The various parameters used in RMi are shown in Fig. 4-2. Several simplifications had to be made
in its structure to maintain an overview of the many properties of a rock mass. The volumes
involved in a rock excavation and the size of the input parameters are generally so large that their
numerical values mainly have to be determined from field observations. An exception is the
compressive strength of intact rock. Well defined and practical usable descriptions are important for
a good result. In this section it is shown how the ratings of these parameters have been determined.
A more detailed description on how to find the values of these parameters is given in Appendix 3.

7KHFRPSUHVVLYHVWUHQJWKRILQWDFWURFN σ F

Several authors have stressed the importance of compressive strength of rock material as a
classification parameter (Deere et al., 1969; Coates, 1964; Bieniawski, 1973, 1984, 1989; Piteau,
1970).

INFLUE NCE
F ROM
MOISTURE
laboratory

UN IAX IA L
test

COM P RES SIV E


TE ST
&2035(66,9( 675(1*7+ 2) 52&.
field or laboratory test

POINT LOAD TEST

SC HMIDT
HA MMER
TEST

SIMP LE FIEL D
(HA MMER) TE ST

RO CK NAM E
AN D STRUCT URE
observations and
strength tables

ANISO TROPY

WE ATH ERING
or ALTERATION

= transition/correlation

Fig. 4-7 Various methods to assess the uniaxial compressive strength.


4 - 11

The uniaxial compressive strength of rock can be determined in the laboratory according to the
specifications given by the ISRM. Other ways of assessing this strength is indicated in Fig. 4-7. Wet
specimens are used where the location of interest is below the ground water table. It should be noted
whether the strength value used represents wet or dry conditions. Where no indication is given, dry
specimens have normally been tested.

Where anisotropic rocks occur, WKHORZHVWFRPSUHVVLYHVWUHQJWKVKRXOGEHDSSOLHG which generally


will be a test direction at 25 - 45o to the schistocity or layering as outlined in Appendix 3, Section 1.

The value of the uniaxial compressive strength (σc ) in MPa from a 50 mmdiameter sample, is
applied directly in RMi. For massive rocks (see Fig. 4-4) the scale effect of σc shown in eq. (4-7)
should be applied.

7KHEORFNYROXPH 9E

The discontinuities cut the rock masses in various directions and delineate a bulk unit, which is
simply referred to as the block. The block size is, therefore, intimately related to the degree of
jointing. Each one of such blocks is more or less completely separated from others by various types
of discontinuities. If all blocks in a rock mass volume could be measured or "sieved", a block size
distribution similar to that for granular soils is found (Fig. 4-8).

100

80
75
6PDOOHU

60

3
Vbmin = 0.01 m
40 Vb25 = 0.07 m3
3
Vb50 = 1.15 m
25 3
20 Vb75 = 0.3 m
Vbmax = 2 m3

0
2 5 2 5 2 5 2 5
0.001 0.01 0.1 1 10 m3
Vb25 Vb75
%ORFN YROXPH 9E

Fig. 4-8 Example of block size distribution

A great variation range will mostly be found between the sizes of the smaller and the larger blocks
in a location; the characterization of the block size should, therefore, indicate their size range.
Simplifications have often to be made during this measurement, as it is not possible to measure all
blocks and their dimensions. Block size is, however, often the most important parameter in the
RMi, and emphasis should be placed on this measurement. Possible ways for estimating the block
volume are shown in Fig. 4-9.

The block volume (Vb) is used directly in the calculation of the jointing parameter (JP). As shown
in Appendix 3, it can also be found from other measurements of the jointing density such as the
volumetric joint count (Jv) and the rock quality designation (RQD). An improved technique for
block size registration in surfaces and drill cores - the weighted joint measurement - is developed
4 - 12

and described here together with a method for finding the block volume from refraction seismic
measurements.

DIRECT VOLUME
MEASUREMENT

BLOCK
MEASURED SHAPE
in surfaces

JOINT SPACINGS FACTOR

2-D FREQUENCY

%/2&. 92/80(
MEASUREMENT

REFRACTION
SEISMIC 

9E
7
MEASUREMENT 1
8
2
&
in surfaces

drill cores


WEIGHTED
or from

7
JOINT DENSITY 1
,
MEASUREMENT 2
- Y
 -
&
,
5
7
1-D FREQUENCY
from drill cores

(
0
MEASUREMENT 8
/
ASSUMED

FACTOR
BLOCK
SHAPE

2
9
ROCK QUALITY
DESIGNATION (RQD)

= transition / correlation

Fig. 4-9 Various methods to assess the block volume (Vb). See Appendix 3, Section 3.

7KHMRLQWFRQGLWLRQIDFWRU M&

The joint condition factor is meant to represent the friction properties of the block faces (i.e. joints)
and the relative scale effect imposed by the joints.

The works of Patton (1966) have emphasized the importance of the surface characteristics of joints
in determining their shear strength. Of particular importance was Patton’s recognition that the shear
resistance resulting from asperities on the joint surfaces had to be overcome during deformation
either by sliding over or by shearing through.

The strength of the rock in which the discontinuities occur, has a direct bearing on the strength
characteristics of the discontinuities, particularly where the walls are in direct rock to rock contact
as in the case of unfilled joints ( ISRM, 1978). The nature of DVSHULWLHV, particularly those of
roughness and hardness, are likely to be dependent on the mineralogical and lithological make-up of
the rock. Mineral FRDWLQJV will affect the shear strength of discontinuities to a marked degree if the
walls are planar and smooth (Piteau, 1970).

The distance between the two matching joint walls controls the extent to which these can interlock.
In the absence of interlocking, the shear strength of the joint is that of the filling material. As
separation decreases, the asperities of the rock wall gradually become more interlocked, and both
the filling and the rock material contribute to the shear strength. According to Barton et al. (1974)
the function tan-1(Jr/Ja) in the Q system is a fair approximation to the friction angle of the joint. As
shown in Appendix 3, the ratio jR/jA is similar to the ratio Jr/Ja.

The author has experienced during several years of geological engineering practice that the length of
joints often has a significant influence on the behaviour of rock masses. Both Lardelli (1992) and
4 - 13

Kleberger (1992) have also stressed the importance of this observation, in particular the difference
between partings and normal joints.

The properties or effects of a joint depend therefore on the following basic factors (Bieniawski,
1984; Piteau, 1970):
1. The FRQGLWLRQRIWKHMRLQW, i.e.
− The strength (hardness) of the wall rock material in FOHDQ joint surfaces, or the friction angle
of the minerals in the coating.
− Weathering of the wall rock of the planes of weakness.
− The small scale asperities and large scale planarity of the joint surface (unevenness and
waviness).
− The distribution, thickness and nature of the gouge materials in ILOOHG joints.
− Size (persistence) and termination of the joint.
2. The H[WHUQDOIHDWXUHV, such as
− The shear movement that has occurred.
− The presence or absence of water on the joint.

Each of the following three main parameters representing the joint condition is given a numerical
value from well defined and simple field registrations based mostly on existing methods:
A. 7KHURXJKQHVVIDFWRU (jR) representing the unevenness of the joint surface which consists of:
− the smoothness (js) of the joint surface, and
− the waviness (jw) or planarity of the joint wall.
B. 7KHDOWHUDWLRQIDFWRU (jA) expressing the characteristics of the joint (Barton 1974):
− the strength of the wall rock, or
− the thickness and strength of a possible filling.
C. 7KHVL]HIDFWRU (jL) representing the influence of the size and termination of the joint.

The joint condition factor is found from the following expression:


jC = jL × jR/jA = jL(js × jw)/jA eq. (4-8)

Often, rough and inexpensive investigations are carried out where only an approximate estimate of
the rock mass characteristics is sufficient. In such cases, there is often limited information on the
parameters in jC. The parameters included in this factor have each been given unit values for
common occurrences. Most commonly the value of jC = 1.5 - 2; using jC = 1 may generally be
somewhat conservative, i.e. ’on the safe side’.

This entails that rough characterization of RMi can be made even if some of the parameters in the
joint condition are absent. The RMi value will, of course, be less accurate in such cases. The benefit
of this is that where the jointing condition is not known - for example in the case of refraction
seismic measurements - the RMi can be estimated (of course, with limited accuracy) from input of
only the rock strength and block size alone.

4.3.3.1 The joint roughness factor (jR)

The roughness factor is, as mentioned, similar to Jr in the Q-system. Roughness here includes both
the small scale asperities (smoothness) on the joint surface and the large scale planarity of the joint
plane (waviness). It has been found appropriate to divide the roughness into these two different
features, as it is often easier to characterize them separately in the joint survey.
4 - 14

Surface VPRRWKQHVVor unevenness is the nature of the asperities in the joint surface, which can be
felt by touch. This is an important parameter contributing to the condition of joints. Asperities that
occur on joint surfaces interlock, if the surfaces are clean and closed, and inhibit shear movement
along joint surfaces. Asperities usually have a wave length and amplitude measured in millimetres
and are readily apparent on a core-sized exposure of a discontinuity. The applicable descriptive
terms are defined in Table 4-3.

TABLE 4-3 CHARACTERIZATION OF THE SMOOTHNESS FACTOR (js). THE DESCRIPTION IS PARTLY
BASED ON BIENIAWSKI (1984) AND BARTON ET AL. (1974).
7(50 '(6&5,37,21 IDFWRUMV
Very rough Near vertical steps and ridges occur with interlocking effect on the 3
joint surface.
Rough Some ridge and side-angle steps are evident; asperities are clearly 2
visible; discontinuity surface feels very abrasive (like sandpaper
grade approx.< 30)
Slightly rough Asperities on the discontinuity surfaces are distinguishable and can 1.5
be felt (like sandpaper grade approx. 30 - 300).
Smooth Surface appear smooth and feels so to the touch (smoother than sand-1
paper grade approx. 300).
Polished Visual evidence of polishing exists, or very smooth surface as is of- 0.75
ten seen in coatings of chlorite and specially talc.
Slickensided Polished and often striated surface that results from friction along 0.6 - 1.5
a fault surface or other movement surface.

:DYLQHVV of the joint wall appears as undulations from planarity. It is defined by


max. amplitude (a max) from planarity
U =
length of joint (Lj)

The maximum amplitude or offset (amax) can be found using a straight edge which is placed on the
joint surface. The length of the edge should be of the same size as the joint, provided that this is
practically possible. As the length of the joint seldom can be observed or measured, simplifications
in the determination of (U) have to be done. A procedure described by Piteau (1970) can be applied
with a standard 0.9 m long edge. Barton (1982) has used a length of 200 mm for joint roughness
coefficient (JRC) measurements. For the smallest joints even shorter lengths can be applied. The
simplified waviness or undulation is found as
measured max. amplitude (a)
u =
measured length along joint (L)

The ratings of the waviness factor are shown in Table 4-4. Often it is found sufficient to determine
the waviness by visual observation as described in Appendix 3, because undulation measurements
are time-consuming.

TABLE 4-4 CHARACTERIZATION OF WAVINESS FACTOR (jw).


7(50 XQGXODWLRQ X ZDYLQHVVIDFWRU(jw)
Interlocking (large scale) 3
Stepped 2.5
Large undulation u>3% 2
Small undulation u = 0.3 - 3 % 1.5
Planar u < 0.3 % 1
4 - 15

The joint roughness factor is found from jR = js × jw, or it can also be determined from Table 4-5.
As the ratings of these parameters are based on the Q system, the joint roughness factor (Jr) in the Q
system is, as mentioned, similar to jR.

TABLE 4-5 JOINT ROUGHNESS FACTOR (jR) FOUND FROM SMOOTHNESS AND WAVINESS. THE
VALUES ARE SIMILAR TO Jr IN THE Q SYSTEM.
ZDYLQHVV*)
VPRRWKQHVV *) planar slightly strongly stepped interlocking
undulating undulating (large scale)

very rough 3 4 6 7.5 9


rough 2 3 4 5 6
slightly rough 1.5 2 3 4 4.5
smooth 1 1.5 2 2.5 3
polished 0.75 1 1.5 2 2.5
slickensided**) 0.6 - 1.5 1-2 1.5 - 3 2-4 2.5 - 5
For irregular joints a rating of jR = 5 is suggested
*)
For filled joints in Table 4-6 jR = 1
**)
For slickensided joints the Jr value depends on the presence and outlook of the striations the highest value
is used for marked striations.

Joint roughness includes the condition of the joint wall surface both for filled and unfilled (clean)
joints. For joints with filling thick enough to avoid contact of the two joint walls, any shear
movement will be restricted to the filling, and the joint roughness will then have minor or no
importance (See Appendix 3, Section 2). In the cases of filled joints it is often difficult or
impossible to measure the smoothness and often also the waviness. Therefore the roughness factor
is defined as jR = 1 as in the Q system (where Jr = 1).

4.3.3.2 The joint alteration factor (jA)

This factor is for a major part based on Ja in the Q-system. It represents both the strength of the joint
wall and the effect of filling and coating materials. The strength of the surface of a joint is a very
important component of shear strength and deformability where the surfaces are in direct rock to
rock contact as in the case of unfilled (clean and coated) joints (Bieniawski, 1984, 1989). The
strength of the joint surface is determined by the following:
- the condition of the surface in clean joints,
- the type of coating on the surface in closed joints,
- the type, form and thickness of filling in joints with separation.

When ZHDWKHULQJ or DOWHUDWLRQ has taken place, it can be more pronounced along the joint wall than
in the block. This results in a wall strength that is often some fraction of what would be measured
on the fresher rock found in the interior of the rock blocks. The state of weathering or alteration of
the joint surface where it is different from that of the rock material, is therefore essential in the
characterization of the joint condition.

TABLE 4-6 CHARACTERIZATION AND RATING OF JOINT ALTERATION FACTOR ( jA)


(partly based on Ja in the Q-system)
4 - 16

$&217$&7%(7:((17+(7:252&.:$//685)$&(6

 7(50  '(6&5,37,21 M$
--------------------------------- ---------------------------------------------------------------------------------- --------------
&OHDQMRLQWV
-Healed or "welded" joints . Softening, impermeable filling (quartz, epidote etc.) ......................... 0.75
-Fresh rock walls ............... No coating or filling on joint surface, except of staining ................... 1
-Alteration of joint wall:
1 grade more altered ........ The joint surface exhibits one class higher alteration than the rock 2
2 grades more altered ...... The joint surface shows two classes higher alteration than the rock 4
&RDWLQJRUWKLQILOOLQJ
-Sand, silt, calcite etc. ....... Coating of friction materials without clay ......................................... 3
-Clay, chlorite, talc etc. ...... Coating of softening and cohesive minerals ...................................... 4

%),//('-2,176:,7+3$57/<2512&217$&7%(7:((17+(52&.:$//685)$&(6

3DUWO\ZDOO 1RZDOO
FRQWDFW FRQWDFW
7<3(2)),//,1*  '(6&5,37,21 thin fillings thick filling
0$7(5,$/ (< 5 mm*)) or gouge
M$ M$
--------------------------------- -------------------------------------------------------------- -------------- --------------
-Sand, silt, calcite etc. ...... Filling of friction materials without clay ................ 4 8
-Compacted clay materials "Hard" filling of softening and cohesive materials .. 6 10

-Soft clay materials ........... Medium to low over-consolidation of filling ........... 8 12


-Swelling clay materials .... Filling material exhibits clear swelling properties... 8 - 12 12 - 20
*)
Based on division in the RMR system (Bieniawski, 1973)

TABLE 4-7 ENGINEERING CHARACTERIZATION OF WEATHERING/ALTERATION. (from Lama &


Vutukuri, 1978)
7(50)25
*5$'( :($7+(5,1* '(6&5,37,21
25$/7(5$7,21

, )UHVK No visible signs of weathering. Rock fresh, crystals bright. Few discontinuities
may show slight staining.

,, 6OLJKWO\ Penetrative weathering developed on open discontinuity surfaces but only slight
weathering of rock material. Discontinuities are discoloured and dis-coloration
can extend into rock up to a few mm from discontinuity surface.

,,, 0RGHUDWHO\ Slight discoloration extends through the greater part of the rock mass. The rock
material is not friable (except in the case of poorly cemented sedimentary rocks).
Discontinuities are stained and/or contain a filling comprising altered materials.

,9 +LJKO\ Weathering extends throughout rock mass and the rock material is partly friable.
Rock has no lustre. All material except quartz is discoloured.Rock can be
excavated with geologist’s pick.

9 &RPSOHWHO\ Rock is totally discoloured and decomposed and in a friable condition with only
fragments of the rock texture and structure preserved. The external appearance is
that of a soil.

9, 5HVLGXDOVRLO Soil material with complete disintegration of texture, structure and mineralogy of
the parent rock.

The alteration factor (jA) is, as seen in Table 4-6, somewhat different from (Ja) in the Q system.
Some changes have also been made in an attempt to make field observations easier and quicker. The
4 - 17

values of Ja can be used - provided the alteration of the joint wall is the same as that of the intact
rock material.

The various classes of rock weathering/alteration that can be determined from field observations,
are shown in Table 4-7.

4.3.3.3 The joint length and continuity factor (jL)

Several writers have experienced during many years of geological engineering that the size and
continuity of the joints often have great influence on the properties of rock masses. Both Lardelli
(1992) and Kleberger (1992) have stressed this, in particular the difference in importance between
partings and normal joints upon rock mass behaviour.

The MRLQWOHQJWKcan be crudely quantified by observing the discontinuity trace lengths on surface
exposures. It is an important rock mass parameter, but is one of the most difficult to quantify in
anything but crude terms. Frequently, rock exposures are small compared to the length of persistent
discontinuities, and the real persistence can only be guessed. However, the difficulties and
uncertainties involved in the field measurements will be considerable for most rock exposures
encountered. The size or the length of the joint is often a function of the thickness or separation of
the joint, and can sometimes be evaluated from this feature. This is further described in Appendix 3
(Section 2.4 and Fig. A3-18).

As the exact length of a joint seldom can be found, the most important task is to estimate the size
range of the joint. Often it is no problem to observe the difference between partings and medium or
larger sized joints during field observations. -RLQWFRQWLQXLW\ is divided into two main groups:
- continuous joints that terminate against other joints
- discontinuous joints that terminate in massive rock.

TABLE 4-8 THE JOINT SIZE AND CONTINUITY FACTOR (jL).


-2,17 5$7,1*2)M/)25
/(1*7+ 7(50$1'7<3( continuous discontinuous
,17(59$/ joints  joints
<1m very short bedding/foliation partings 3 6

0.1 - 1.0 m short/small joint 2 4


1 - 10 m medium " 1 2
10 - 30 m long/large " 0.75 1.5

> 30 m very long/large filled joint or seam*) 0.5 1


*)
Often a singularity, and should in these cases be treated separately.

The joint size factor for continuous joints can also be expressed as
jL = 1.5 × L - 0.3 eq. (4-9)

where L = the length of the joint in metre. The ratings of jL are shown in Table 4-8.

 3266,%/($5($62)$33/,&$7,212)7+(50L

The main purpose during the development of the RMi has been to work out a system to characterize
rock masses, which is applicable in rock engineering. As RMi is linked to the material, it represents
only the inherent properties of a rock mass. Thus, it does not express external loads or forces acting
on the material, such as:
4 - 18

• the in situ rock stresses;


• the presence of ground water;
• the orientation of:
- loads or stresses,
- structural elements (joints, anisotropy, etc.),
- permeability or ground water flow; and
• the impacts from human activity.
For application of RMi in practical rock mechanics and engineering for civil or mining, one or more
of these features usually have to be included where they have influence or impact on the ground
conditions.

The main activities where rocks and rock masses involved are shown in Table 4-9, which also
indicates the fields in which the RMi may be of interest.

As RMi is a strength index it is suitable for application in rock engineering, design or other
evaluations connected with utilization of rocks. This is fully shown in Chapter 6 on stability
assessments for rock support analysis and in Chapter 7 on capacity evaluation of tunnel boring
machines (TBM). For these applications the RMi is adjusted for local features of importance to
determine the behaviour of the rock mass.

The RMi value can not be used directly in existing classification systems as many of them are
systems of their own. Some of its input parameters are sometimes similar to those used in these
classifications and may then be applied more or less directly, see Chapter 8.

TABLE 4-9 A BRIEF VIEW OF MAIN INTERNAL ROCK MASS PARAMETERS AND THEIR IMPORTANCE
IN ROCK AND ROCK MASS UTILITIES.
5(/$7,9(,03257$1&(2)
87,/,7< $&7,9,7< rock strength jointing singularities *)
⋅ drilling (small holes) x -/(x) -
+ boring (TBM boring, reaming) x x (x)
Treatment of rocks = blasting x (x) (x)
and rock masses = fragmentation (x) x (x)
⋅ crushing x - -
⋅ grinding x - -
= cutting x x (x)

Application of ⋅ rock aggregate for concrete x (x) -


rock material ⋅ rock fill x x -
⋅ natural stone/building stone x x (x)

Utilization of + underground excavations (x)/x x x


rock masses + surface cuts and slopes (x) x x
= foundations for dams etc. (x) x x

/HJHQG: + Suitable for characterization x great influence


= May partly be suitable for characterization (x) limited influence
⋅ Generally not suitable - little or no influence
*)
These are seams, shears, weakness zones   

Hoek (1983, 1986) and Hoek and Brown (1988) mention that further work is required to improve
the Hoek-Brown failure criterion, since the use of classification systems developed for the design of
tunnel support has been found to have some limitations when used for estimating rock mass strength
parameters. They suggest that it may be necessary to develop a system specifically for this purpose.
As described in Section 5.2 in this chapter and further dealt with in Chapter 8, the jointing
4 - 19

parameter (JP) in the RMi is similar to one of the main parameters in the Hoek-Brown failure
criterion for rock masses. The RMi may, therefore, contribute to such a future system.

The rock mass strength characteristics found from RMi can also be further applied in NATM
classification and rock support design as well as in ground response curves, as demonstrated in
Chapter 8. Finally, it should be mentioned that the system for characterizing block geometry
(volume, shape factor, angles) may be of use in numerical models.

 ',6&866,21

The following discussion is limited to the structure and development of the RMi as dealt with in this
chapter. A discussion of the RMi system, its use, and a comparison with other engineering systems
and methods is presented in Chapter 9.

/LPLWDWLRQVRIWKH50L

The RMi is meant to express the relative variation in the strength between different rock masses. As
determination of the strength of an in situ rock mass by laboratory type testing for many reasons is
not practical, the RMi makes use of input from geological observations and test results on
individual rock pieces or rock surfaces which have been removed from the actual rock mass.

RMi is restricted to expressing only the compressive strength. Hence, it has been possible to arrive
at a simple expression, contrary to, for example, the general failure criterion for jointed rock masses
developed by Hoek and Brown (1980) and Hoek et al. (1992). Because simplicity has been preferred
in the structure and in selection of parameters in RM, it is clear that such an index may result in
inaccuracy and limitations, the most important of which are connected to:
A. 7KHUDQJHDQGW\SHVRIURFNPDVVHVFRYHUHGE\WKH50L.
Both the intact rock material as well as the joints exhibit great directional variations in
composition and structure, which results in an enormous range of properties of rock masses. It
is, therefore, not possible to characterize all these combinations in one, single number.
However, it should be added that the RMi probably may characterize a wider range of materials
than most other classification systems. Characterization of rock masses by the RMi is presented
in Chapter 5.
B. 7KHDFFXUDF\LQWKHH[SUHVVLRQRIWKH50L
The value of the jointing parameter (JP) is calibrated from a few large scale compression tests.
Both the evaluation of the various factors (jR, Ja and Vb) in JP and the size of the samples
tested, which in some of the cases had less than 5x5x5 blocks, have resulted in that there
certainly are errors connected to the expression developed for the JP. In addition, the test results
used were partly made on dry, partly on wet samples (Stripa on wet). The influence of moisture
may have reduced the accuracy of the data used.
Also, the uniaxial tests are encumbered with errors as pointed out by Farmer and Kemeny
(1992) and in Appendix 3, Section 1. The value of RMi found can, therefore, be very
approximate. In some cases, however, the errors in the various parameters may partly cancel
out.

C. 7KHHIIHFWRIFRPELQLQJSDUDPHWHUVWKDWYDU\LQUDQJH.
The input parameters to the RMi express generally a certain range of variation related to
changes in the actual representative volume of the rock mass. The combination of such ranges
4 - 20

in RMi may cause additional errors. Chapter 5 briefly outlines some methods to reduce this
effect.

The result of the foregoing is that RMi in many cases will give an inaccurate value for the strength
of such a complex assemblage of different materials and defects as in a rock mass. For this reason,
the RMi is regarded as a UHODWLYH expression of the rock mass strength. It should preferably be used
in communication and characterization.

Being valid for minimum 125 blocks the direct use of the RMi involves larger volumes of rock
masses than is actually required. RMi can, therefore, seldom be directly applied in engineering and
design. Some modification or supplementary adjustments have generally to be made as shown in
Chapters 6 on rock support and Chapter 7 on TBM.

2WKHUVLPLODUURFNPDVVFKDUDFWHUL]DWLRQPHWKRGV

The RMi has been developed during a process that has involved a critical examination of rock mass
characteristics and available literature. The main philosophy has been to take account of the effect
of discontinuities in reducing the strength of intact rock.

Earlier, a similar approach to a strength characterization of rock masses has been proposed by
Hansagi (1965, 1965b), who introduced a similar reduction factor to the jointing parameter (JP) to
arrive at an expression for the FRPSUHVVLYH strength of the rock mass, given as
σmc = σc × Cg eq. (4-10)
where σc = compressive strength,
Cg = the reduction factor which Hansagi named ’gefüge-factor' (joint factor) being
"UHSUHVHQWDWLYHIRUWKHMRLQWHGHIIHFWRIDURFNPDVV".

The Cg factor consists of two inputs: a factor for the "structure of jointing" (core length), and a
scale factor. Hansagi (1965b) mentions that the value of Cg is 0.7 for massive rock and 0.47 for
jointed rock (from small joints) for two test locations in Kiruna, Sweden. Hansagi did not, however,
- as far as the author knows - publish more on his method.

From Fig. 4-10 it is seen that the expression for the RMi is also similar in structure to the expression
of unconfined FRPSUHVVLYHVWUHQJWKRIURFNPDVVHV (σcm), which is a part of the Hoek-Brown failure
criterion for rock masses, and is expressed as
σcm = σc × V ½ eq. (4-11)

Here σc = the uniaxial compressive strength of the intact rock material, and
s = an empirical constant. The value of V ranges from 0 for jointed rock masses to 1 for
intact rock. The value of s is found from the RMR or the Q classification system as
described by Hoek (1983), Hoek and Brown (1980, 1988), and Wood (1991).

Thus, the jointing parameter( JP) is similar to V ½ in the Hoek-Brown failure criterion. The process
of finding JP is, however, more direct and clear as it only involves features that have a direct
impact on this parameter. This is further described in Chapter 8.
4 - 21

4.5

4.0

3.5
σ1’

σ3’
3.0
Major principal stress σ1’

Triaxial compression
2.5 2 1/2
σ1’ = σ3’ + (P σc σ3’ + V σc )

2.0

σ1’
1.5
Uniaxial compression
β 2 1/2
σcm = (V σc )

1.0
Uniaxial tension
2 1/2
σtm = ½ σc (m - (m + 4s) )
0.5 σct

0.5 1.5
Minor principal stress σ3’

Fig. 4-10 The uniaxial compressive strength is one special mode of the Hoek-Brown failure criterion for rock
masses (from Hoek, 1983).
)URP3DOPVWU|P$50L±DURFNPDVVFKDUDFWHUL]DWLRQV\VWHPIRUURFNHQJLQHHULQJSXUSRVHV
3K'WKHVLV2VOR8QLYHUVLW\1RUZD\S

&KDSWHU

52&.0$66(6&+$5$&7(5,=('%<7+(50L

+RZFRQYLQFHGDUHZHWKDWWKHURFNSURSHUWLHVZHIHHGLQWRWKHGHVLJQPRGHOVDFWXDOO\
GHVFULEHWKHZD\WKHURFNPDVVUHDFWVWRDSHUWXUEDWLRQOLNHDWXQQHORUDFDYHUQ"
Ulf E. Lindblom (1986)

RMi is an expression, which covers intact rock as well as the aggregates of rock blocks formed by
joint planes. As RMi is tied to the resulting effect of interacting blocks, it is not meant to
characterize single blocks or joints. Therefore, it is, from its definition and structure, only applicable
in continuous volumes of rock masses. Ideally, such a volume should be homogeneous.

Where RMi is used as a general characterization of rock masses, the size of the ’sample’ or the
volume involved is not related or limited by the excavation constructed. Thus, the rock mass is
considered continuous if the ’sample’ size is not limited by geologic boundaries or other in situ
features. For some types of rock masses with widely spaced joints such a ’sample’ will, however, be
of a considerable size.

The actual project and the specific calculation or engineering work connected with it, determines the
size of the rock mass involved. In such cases the rock mass may be considered as continuous or
discontinuous as described in the next section.

 21&217,18286$1'',6&217,1828652&.0$66(6

When the ’sample’ of a rock mass being considered is such that only a few joints are contained in
this volume, its behaviour is likely to be highly anisotropic, and it is considered as GLVFRQWLQXRXV. If
the sample size is many times the size of the individual fragments, the effect of the each particle
(and hence the joints) is statistically levelled out, and the sample may be considered FRQWLQXRXV
(Deere et al., 1969). See Fig. 5-1.

This is the case when none of the discontinuities or joint sets is significantly weaker than any of the
others within the volume of rock under consideration. If a discontinuity is very weak when
compared to the others, as could be the case when dealing with a fault passing through a jointed
rock mass, the rock mass may be characterized as continuous, and the fault must be characterized
separately as a singularity if it is relatively small, or else as a weakness zone, as described in
Appendix 2.

The volume required for a sample of a rock mass to be considered continuous is a matter of
judgement. It depends on the range of block or particle sizes making up the ’sample’ volume. This
matter has been discussed by several authors:

• John (1962) suggests that a sample of about 10 times the average (linear) size of the single
units may be considered a uniform continuum. It is clear that this will depend to a great extent
on the uniformity of the unit sizes in the material or the uniformity of the spacings of the
discontinuities. For a unit of 1 m3 the size of the sample would be 103 = 1000 m3.
5-2

Intact rock

Single discontinuity

Two discontinuities

Several discontinuities
UNDERGROUND EXCAVATION ROCK SLOPE

Jointed rock mass

Fig. 5-1 Various volumes of rock masses involved in a ’sample’. Continuous rock masses are: ’intact rock’, ’jointed
rock mass’, and possibly ’several discontinuities’. (From Hoek, 1983) The tunnel shown involves relatively
few discontinuities, i.e. a discontinuous rock mass.

• Deere et al. (1969) have tied the ’sample’ size to the size of a tunnel from its stability
behaviour. Whereas the stability of a tunnel opening in a continuous material can be related to
the intrinsic strength and deformation properties of the bulk material, stability in a
discontinuous material depends primarily on the character and spacing of the discontinuities.
In this connection they have found that the size of the ’sample’ related to a tunnel should be
considered discontinuous ZKHQWKHUDWLRRIIUDFWXUHVSDFLQJWRDWXQQHOGLDPHWHULVEHWZHHQ
WKHDSSUR[LPDWHOLPLWVRIDQG)RUDUDQJHRXWVLGHWKHVHOLPLWVWKHURFNPD\EH
FRQVLGHUHGFRQWLQXRXVWKRXJKSRVVLEO\DQLVRWURSLF (See Fig. 5-2).

• Another approximate indication is based on the experience from large sample testing at
Karlsruhe, Germany (see Appendix 6), where a volume containing at least 5 × 5 × 5 = 125
blocks is considered continuous (Mutschler, 1993).

σ2

Fig. 5-2 The size for a ’sample’ to be characterized as continuous for the actual tunnel. In this case the stability is
governed by the blocks around the opening, i.e. the ground around the tunnel is discontinuous (modified
from Barton, 1990).
5-3

From this it is clear that it is important to determine whether a material should be considered
continuous or discontinuous in a particular case. Accordingly, the type of behaviour of the material
may be predicted, from which suitable theories and methods of design may be employed.

In this connection it may also be mentioned that the current approach to modelling engineering
projects in a jointed rock mass is to treat the rock as a discontinuum (controlled by individual joints)
in the near field of an opening, and as a continuum in the far field (when the volumes are
significantly larger).

=21,1*2)52&.0$66(6,1726758&785$/5(*,216

To facilitate the characterization of the variation of rock masses within a region or along a borehole,
it is often necessary and convenient to distinguish a number of VWUXFWXUDOUHJLRQV (Piteau, 1973),
wherein the rock and joints have similar composition. Each part selected can then be considered and
treated individually for its particular characteristics. ISRM (1980) has applied the term ]RQLQJRI
URFNPDVVHV for such division into structural regions in their proposed method for basic geotechnical
description (BGD) of rock masses. Designation of a structural region implies that the GHWDLOHG
MRLQWLQJ (Selmer-Olsen, 1964) within the region selected is similar, assuming that the individual
joint sets have the same characteristics, as the joint sets most likely - at least on a local basis - have
been developed under similar conditions of stress (Piteau, 1973).

A zone may include differing volumes of rock masses, such as interbedded layers of sedimentary or
volcanic formations exhibiting the same geotechnical characteristics (ISRM, 1980). In the case of
rock masses which vary continuously from place to place, for example due to weathering, ISRM
advises delineating arbitrary zone boundaries in such a way that the properties of each zone may be
considered relatively uniform. Generally, it is found that structural regions of similar jointing will
juxtapose at major geological structures. The boundaries of a zone will therefore often be defined by
faults, dykes and rock boundaries.

As the RMi expresses the inherent characteristics, it is well suited to be applied in the zoning. The
input data from the survey is then, manually or by the use of a computer, utilized to find the rock
mass quality values for each of the structural regions.
NOTE
The classification into rock classes is based
sum of the investigation
Location of section shown in FIG. 1
a ce
su rf
und
Gro CLASS5 (b)

CLASS 5 (b)
DI
VE
RS
IO
ROCHES FAU

CLASS 3
N
TU
NN

CLASS 5 (a)
EL
FA
LT

UL

CLASS 4
T

0 10 50 100 m

10 Scale

Fig. 5-3 Example of zoning into rock classes in a profile (from Chappell, 1990)
5-4

Chappell (1990) has, through zoning of rock masses into structural regions (Fig. 5-3), arrived at
models where strength characteristics and deformational response of the intact rock material are
combined with associated discontinuity parameters.

 35,1&,3/(6,1&+$5$&7(5,=,1*7+(9$5,$7,216,152&.0$66(6

In spite of a zoning of the rock volumes into structural regions of similar characteristics, the various
parameters of the rock mass within a zone may still show variations, refer to Fig. 5-4. As described
in Chapter 4 the parameters selected has been divided into three main groups:
- the intact rock features,
- the features contributing to the size and shape of rock blocks, and
- the features connected to the joints and their condition.

variation in block size


and rock strength
variation in joint length
and termination (jL)
variation in shear strength (jR, jA)
in and between set 1, 2 and 3
1

Fig. 5-4 Sketch showing possible variations in rock mass parameters.

The numerical characterization of these parameters is described in Appendix 3 where different


methods of finding their values are described. As many or all of the parameters may show variations
within the actual zone or rock mass volume, their conditions may be compared and their values
determined from evaluation and judgement based on understanding of the geological setting. A
short description of the rock mass may help to a clearer knowledge of the site conditions.

9DULDWLRQVLQWKHURFNPDWHULDO

The distribution of rocks in a location is generally determined from field observations based on
geological classification. Within the same type of rock strength (σc) and anisotropy may vary,
which may be caused by variation(s) in the following features:
- composition/structure;
- texture; and/or
- weathering/alteration.
5-5

In addition folding and alternation between different rocks may contribute to the variation in σc
values in the location. The range of some of these features can be found from the tests or
assessments described in Appendix 3 on numerical determination of input parameters to the RMi.
Some of the variations in the rock properties may sometimes be viewed in connection with the
jointing features as described in Appendix 1.

Such variation in the compressive strength should, as recommended by the ISRM (1978), be given
as a range. ,QURFNVZKHUHWKHVWUHQJWKYDULHVZLWKGLUHFWLRQRIWKHWHVWLQJWKHORZHVWYDOXHVKRXOG
EHDSSOLHGIRUσFDVLQSXWWR50L

The requirements for the quality of the geo-data to be applied and the actual type of engineering will
indicate whether the different rocks in a location may be characterized as separate volumes or not.

9DULDWLRQVLQWKHMRLQWLQJ

Jointing is generally much more complex in its variation and influence in rock mass behaviour than
the intact rock. In most types of observations and measurements the joints are characterized and not
the blocks they delineate. In Appendix 3 methods to asses the block size from such measurements
are shown.

long, smooth & planar joints with clay filling;


partly joint wall contact

GLI IHUHQW MRLQW DOWHUDWLR Q


IDFWRU DQG MRLQW OHQJWK
EHWZHHQ WKH MRLQW VHWV

short, smooth & planar


and fresh joints

long smooth & slightly


undulating joints

GLIIHUHQW MRLQW VL]H


WHUPLQDWLRQ DQG URXJK QHVV
IRU WKH MRLQW VHWV

discontinuous,
rough & planar joints

discontinuous, short foliation


partings; fresh, smooth & undulating

GLI IHUHQ W MRLQ W VL]H DQG


WHUPLQDWLRQ IRU WKH
WZR MR LQW VHWV

long, continuous rough & planar,


slightly altered joints

Fig. 5-5 Some examples of variations in jointing.


5-6

The detailed jointing may constitute the various patterns by which joint spacings determine the
block sizes and their variation. Jointing can be divided into the following:
1. Common types of jointing mainly consisting of tectonic joints:
− Mainly joint sets and few random joints (often with one of the joint sets along the
bedding or the foliation).
− Few joint sets plus many random joints.
− Mostly random joints (i.e. irregular jointing).
2. Special types of jointing which can make up a smaller or larger part of the jointing:
− Foliation jointing in anisotropic (schistose) rocks.
− Columnar jointing in basalts.
− Cleavage jointing in some granites.
− Sheet jointing (exfoliation jointing) caused by stresses near the surface.
− Desiccation jointing in sedimentary mudstones.
− Jointing in the zone of weathering.
− Jointing in tectonic zones (crushed zones).

Some examples of variations in rock masses are given in Fig. 5-5.

The variations in one or more of these factors result in that, in reality, regular geometric shapes are
the exception rather than the rule. Jointing in sedimentary rocks usually produces the most regular
block shapes.

There are so many variations in jointing that it has not been possible to work out one single method
to characterize all these in a common jointing parameter. Therefore, different methods are shown,
and it is up to the user to select the method that is best for the actual case.

5.3.2.1 The block size (Vb)

The block size and its variation depend on the density of the jointing influenced also by the number
of joint sets and the spacings in these sets. In addition random joints may contribute, especially
where irregular jointing occurs.

The variation in block size may be graphically shown in a sieve curve similar to that shown in Fig.
A3-21 or in Fig. 4-8. This variation can be measured and reported in different ways, depending on
the number of joint registrations and the accuracy required of the measurements. One possible
method is, - based on the efforts required and the availability of measurements, - to use Vb25 and
Vb75 as the range (see Fig. A3-21), similar to what is used in soil mechanics.

The block size, which is another measure for the quantity of joints, can be found from several types
of measurements by using the relations described in Section 3 in Appendix 3.

$%ORFNYROXPHIRXQGIURPMRLQWVSDFLQJRUMRLQWGHQVLW\PHDVXUHPHQWV

Spacing may be given as a range (Smin and Smax ) for each joint set. The minimum block volume is
found from the minimum values for each set Vbmin = S1min × S2min × S3min provided the joint sets
intersect at right angles. The maximum block volume is found accordingly.
5-7

Example 5-1: Block volume determined from joint spacings.


The following joint spacings have been observed in a location:
joint set 1, spacing S1 = 0.3 - 0.5 m
joint set 2, spacing S2 = 0.5 - 1 m
joint set 3, spacing S3 = 1 - 3 m
Provided the joints intersect at right angles the block volume is
Vbmin = 0.3 × 0.5 × 1 = 0.15 m3 , Vbmax = 0.5 × 1 × 3 = 1.5 m3

Example 5-2: Block volume found from 1) the quantity of joints, and 2) from the joint spacings.
Measured on the outcrop of approximately 25 m2 shown in Fig. 5-6, the following number
of joints have been found:
7 joints with length > 5 m (= na1);
6 joints with length approx. 3 m (= na2); and
45 small joints with length approx. 1.5 m (= na3).

Fig. 5-6 Jointed Ordovician mudstone. The rulers shown are 1 m long (from Hudson and Priest 1979).

1) Block volume found from the quantity of joints. Most joints are shorter than the dimension of
the observation area and their quantity should therefore be adjusted using eq. (A3-32a)
na* = na × Lj / 25 (see Appendix 3, Section 3). This gives
na1* = 1, na2* = 3.6 and na3* = 13.5
The density of joints is then Na = (na1* + na2* + na3*) / 25 = 4.8 joints/m. As it is not
known how the surface is oriented with respect to the main joint set an average value of ka =
1.5 is applied to find the volumetric joint count (see eq. (A3-32) and Fig. A3-25 ):
Jv = Na × ka ≈ 7.2 joints/m3.

Assuming that the blocks are mainly compact (equidimensional) with a shape factor of β = 30
the average block volume is found as (eq. (A3-27):
Vb = β × Jv-3 = 0.08 m3
5-8

2) Block volume found from the following spacings roughly measured in Fig. 5-6:
set 1: S1 = 1.3 - 2.2 m (average = 1.75 m)
set 2: S2 = 0.8 - 1.8 m (average = 1.3 m)
set 3: S3 = 1.2 - 2 m (average = 1.4 m)
In addition 45 random joints can be seen within the observation area of 25 m2. These joints are
mainly short (approximately 1.5 m long), therefore their number has been adjusted according to
eq. (A3-32a) na* = 45 × 1.5 / 25 = 13.5. The density of random joints is then
Na = 13.5 / 25 = 2.7 joints/m
An average value of ka = 1.5 has been chosen to find the contribution of random joints to Jv:
Jv random = Na × ka = 4 joints/m3

The resulting average volumetric joint count is found as


Jv = (1/1.75) + (1/1.3) + (1/1.4) + Jvrandom = 6.05 joints/m3
Using β = 30 the block volume is
Vb = β × Jv-3 = 30 × 6.05 - 3 = 0.13 m3

The second method (the combined spacing and joint density method) gives block volume 60%
larger than the first method. This may be explained by the very approximate joint spacing in
method no. 2.
(A rough estimate from the photo indicates an average block volume of
Vb = 0.2 m3 for compact (equidimensional) blocks).

%3UREDELOLW\FDOFXODWLRQVWRGHWHUPLQHWKHYDULDWLRQLQEORFNYROXPH

As the joint spacings generally are independent random variables, probability calculations may be
applied to determine the range of the block volume from the variations in spacings for each joint set.
Suppose the three joint sets intersect at right angles, then the block volume is
Vb = S1 × S2 × S3
where S1, S2, S3 are the joint spacings for the three sets.

Within each set the spacing varies within certain limits. In the derivations below it is assumed that
the minimum value of the spacing corresponds to
mean value - α standard deviations
and the maximum value to
mean value + α standard deviations.

The expression above for the block volume can be written as


ln Vb = ln S1 + ln S2 + ln S3 eq. (5-1)

Assume that the joint spacings have a log-normal distribution. This is often the case for jointing as
shown in Section 5 in Appendix 1. Then eq. (5-1) can be expressed by its
mean ln value
µln Vb = µln S1 + µln S2 + µln S3 eq. (5-2)
and the standard deviation as
σlnVb = {(σlnS1)2 + (σlnS2)2 + (σlnS3)2}½ eq. (5-3)
where µln S1 ≈ (ln S1min + ln S1max )/2 (and similar for µln jR and µln jA) eq. (5-4)
σlnS1 ≈ (ln S1max - ln S1min )/2α eq. (5-5)
5-9

Applying α standard deviations from the mean ln-value (µln Vb) and a log-normal distribution, the
block volume will be
Vblow ≈ e µlnVb σ lnVb
( -α )
eq. (5-6)
and
Vb high ≈ e µlnVb σ lnVb
( +α )
eq. (5-7)

For practical purposes α = 1 standard deviation may be applied.

Example 5-3: The variation range of block volume found from joint spacings.
The following spacings have been measured:
- joint set 1, spacing S1 = 1 - 2 m ( ln S1 = 0 - 0.693)
- joint set 2, spacing S2 = 2 - 3 m ( ln S2 = 0.693 - 1.099)
- joint set 3, spacing S3 = 3 - 4 m ( ln S3 = 1.099 - 1.386)
For α = 1 standard deviation the mean ln values of the spacings and the volume are found as
µln S1 = ½ (0.693) = 0.347
µln S2 = ½ (0.693 + 1.099) = 0.896
µln S3 = ½ (1.099 + 1.386) = 1.243
µln Vb = 0.347 + 0.896 + 1.243 = 2.486
The standard deviation is:
σlnS1 = (0.693)/2 = 0.347
σlnS2 = (1.099 - 0.693)/2 = 0.203
σlnS3 = (1.386 - 1.099)/2 = 0.144
σlnVb = (0.347 2 + 0.2032 + 0.1442) ½ = 0.427
This gives
Vblow ≈ e (2.486 - 0.427) = 7.84 m3
Vbhigh ≈ e (2.486 + 0.427) = 18.40 m3
and
Vbmean ≈ e µ ⋅ lnVb = 12.0 m3

(If the lowest, mean, and highest values for all spacing had been chosen
Vbmin = 6 m3, Vbmean = 13.13 m3, Vbmax = 24 m3 )

As mentioned in Appendix 1, there may be cases where only one joint set or two joint sets occur,
hence no blocks are delineated which means that the block volume is infinite. In other cases most of
the joints terminate in solid rock so that blocks are not clearly delimited. An example of this is
schists in which foliation joints and partings are the only joints present. In such situations the length
of the joints may be applied for calculating the effect of the jointing. A useful method is shown in
Example 5-7.

5.3.2.2 The joint condition factor (jC)

The joint condition factor (jC) is composed of 4 variables: the smoothness, waviness, size and
alteration of each joint in the actual volume of rock mass. Thus jC may show significant variations,
and it may be difficult to estimate its range of variation.

Ideally the value of jC for each of the joints or joint sets should be used in RMi. As it was found
impossible to include the jC value for all joints and at the same time maintain the simple structure
of RMi, this factor is only represented as one number (or range). This means that where there are
5 - 10

different conditions for the various joint sets, some simplifications have to be made to combine
them as shown in the following:

$M&GHWHUPLQHGIRURQHMRLQWRUMRLQWVHWZKHUHWKHSDUDPHWHUVLQYROYHGLQLWYDU\

Alt.1.
The variation range of jC is found from combination of the parameter values so that the
minimum value and the maximum value is found for the actual joint or joint set
jCmin = jLmin × jRmin /jAmax eq. (5-8)
jCmax = jLmax × jRmax /jAmin eq. (5-9)

Example 5-4: jC determined from variations in joint characteristics.


The following values have been found for a joint set:
- the joint wall surface is smooth to slightly rough, jw = 1 - 1.5
- the waviness is slightly to strongly undulating, js = 1.5 - 2
- silty coating on the joint wall, part wall contact, jA = 3
- the continuous joints vary between 0.5 - 5 m in length, jL = 1 - 1.5
From this the roughness factor is found as
jR = jw × js = 1.5 - 3
and the minimum and maximum joint condition will be (from the lowest and highest value
of jR and jL)
jCmin = 0.5, jCmax = 1.5

Alt. 2.
The variation range of jC is found from SUREDELOLW\FDOFXODWLRQV similar to that described for
block volume, provided the three parameters in the joint condition are independent random
variables. The joint condition factor is jC = jL × jR/jA

Each parameter varies within certain limits. The expression above can be written as
ln jC = ln jL + ln jR - ln jA eq. (5-10)

Assuming that the joint condition parameters have a log-normal distribution,


eq. (5-10) has the following mean ln-value:
µln jC = µln jL + µln jR - µln jA eq. (5-11)
where µln jL ≈ (ln jLmin + ln jLmax )/2 eq. (5-12)
(and similar for µln jR and µln jA)

Applying ±1 standard deviations from this mean ln-value, µln jC , the standard deviation is
σlnjC = {(σln jL)2 + (σlnjR)2 + (σlnjA)2}½ eq. (5-13)

where σln jL ≈ (ln jLmax - ln jLmin )/2 eq. (5-14)


(and similar for σlnjR and σlnjA)

For a log-normal distribution of µln jC the joint condition will be


jClow ≈ e( µlnjC - σ lnjC ) eq. (5-15)

and
jChigh ≈ e( µlnjC + σ lnjC ) eq. (5-16)
5 - 11

%7KHUHVXOWLQJM&RU-3IRUWKHURFNPDVVZKHQM&YDULHVIRUHDFKMRLQWRUMRLQWVHW

Alt. 1. Where the joint sets have approximately the same spacings:
Use the average value for jC for all sets.

Alt. 2. Where the spacings are different for the joint sets:
a. Simply apply the (assumed) average jC for all the joint sets.

b. Use the joint set with the most unfavourable value for jC.
This method is applied in the Q system. Also ISRM (1980) suggests thatZKHQMRLQWVHWV
VKRZGLIIHUHQWVKHDUVWUHQJWKWKHVHWZKLFKVKRZVWKHVPDOOHVWPHDQDQJOHRIIULFWLRQ
VKRXOGEHDGRSWHGXQOHVVVSHFLILFFLUFXPVWDQFHVZDUUDQWRWKHUZLVH$UHFRUGRIWKH
DQJOHVRIIULFWLRQFRUUHVSRQGLQJWRRWKHUIUDFWXUHVHWVPD\SURYHRILQWHUHVW

c. Sometimes one joint set is significantly more important than the others. In such cases the
data for this set may be applied directly.

d. Carry out an assessment of jC or JP as shown for the following two main cases:

&DVH
For every joint set with its jC and spacing it is assumed that the effect of jC varies with the size
(area) of the joint plane. As the area of joint planes in a volume depends on the spacing (see Fig. 5-
7), it is assumed that jC depends on the second power of the spacing.

If the spacing and the joint condition factor for the joint sets are S1, S2, S3 and jC1, jC2, jC3
respectively, the resulting jC may be expressed as
jC = Σ{(1/Si)2 × jCi} / Σ(1/Si)2 eq. (5-17)

the influence from characteristics of


the main surfaces of the main joints
dominates the joint condition factor

Fig. 5-7 The joint set with smallest spacing has the largest area in the block surface and hence the greatest impact
on the jC.

or where joint frequencies are measured; for 2-D measurements


jC = Σ{(Nai)2 jCi}/ Σ(Nai)2 eq.(5-18)

And for 1-D measurements


jC = Σ{(Nli)2 jCi}/ Σ(Nli)2 eq. (5-19)

Here, Na will give more accurate values than Nl because it is adjusted for the length of the joints.
5 - 12

Example 5-5 Finding the average jC representative for all joint sets.
The observed data on joint spacings and joint conditions are given in Table 5-1.

TABLE 5-1 OBSERVATION DATA ON JOINT SPACING AND JOINT CONDITION


Joint Average Joint Average joint
set Spacing spacing condition condition factor
no. Si factor jC 1/Si2 (1/Si2) jC
1 0.5 - 1.5 1 1-2 1.5 1 1.5
2 1-2 1.5 0.25 - 0.5 0.35 0.44 0.15
3 2-3 2.5 2-3 2.5 0.16 0.4

Σ1/Si2 = 1.6 Σ(1/Si2) jC = 2.05

In this case eq. (5-17) may be used. From the values in Table 5-1 the resulting joint condition
factor for all joint sets is then
jC = Σ(1/Si)2 jCi / Σ(1/Si)2 = 2.05/1.6 = 1.3

Example 5-6: jC found from various joints and joint conditions in an outcrop.
Though one joint set may be seen (vertical), the jointing pattern in Fig. 5-8 may be
characterized as irregular.

Fig. 5-8 Jointed outcrop of Carboniferous sandstone (from Hudson and Priest 1979). The rulers shown are 1 m
long

It is assumed that all joints are fresh, slightly rough and planar. This means that jR = 1.5 and
jA = 1. Most joints are continuous, i.e. terminate against each other. Because they have
different size, jC will vary as given in the table on next page for a measurement area of 1 m2.

As many of the joints are smaller than the length of the dimension of the observation area (1
m2), their quantity has been adjusted in Table 5-2 using eq. (5-18).
5 - 13

TABLE 5-2 ’OBSERVATIONS’ MADE ON FIG. 5-9. THE JOINT CONDITION FACTORS HAVE BEEN
ASSUMED.
’Observed Average Adjusted number Assumed joint
Approx. of joints
’ joints joint length condition
joint length
(na) ( Lj) Na* = na⋅Lj/ A factor(jC) (Na*)2 jC(Na*)2
1 > 1.5 m >1.5 m 1 1.5 1 1.5
5 0.5 - 1 m 1m 5 2.5 25 62.5
20 0.2 - 0.5 m 0.3 m 6 4 36 144
40 < 0.2 m 0.15 m 6 6 36 216
ΣNa* = 18 Σ(Na*)2 = 98 ΣjC(Na*)2 = 424

The total amount of adjusted joints is Na* = 18 joints/m. Using ka = 1.5 in


eq. (A3-32b) and Jv = Na × ka = 27 joints/m3, the estimated block volume is
Vb ≈ β× Jv-3 = 50 × 27-3 = 2.5 dm3
(The blocks seem to mainly to be flat; therefore β = 50 is assumed.)

The resulting joint condition factor may be found from eq. (A3-32a) using the adjusted joint
densities :
jC = Σ(Nai*)2 jCi / Σ(Nai*)2 = 4.3
The jointing parameter is JP = 0.2 √jC VbD = 0.08 (D = 0.37 jC - 0.2 = 0.276)

&DVH
Consider that the rock mass is composed of (flat) blocks formed by only of one of the joint sets
having its jointing parameter JP1. JP1 can be regarded as the strength of the material in the
blocks formed by joint set 2 for which the JP2 can be found. The same principle can be applied
for the remaining joint sets as is described below:
⇒ Find the jointing parameter JP1 related to joint set 1 (with spacing S1 and average
length L1) from its jC and volume Vb1 = S1 × L1 2.
⇒ The same procedure is carried out also for the other joint sets and block volumes.
⇒ The resulting jointing parameter JP is the product of the jointing parameters found for
each set: JP = JP1 × JP2 × JP3

Example 5-7 JP found for various joint spacings and joint conditions.

In Fig. 5-9 the jointing consists mainly of joints and partings along the foliation of a mica
schist. There are mainly two types of these joints:

Fig. 5-9 Large foliation joints and small foliation partings


5 - 14

⇒ Foliation partings (set 1a). These are small (L1a = 0.1 - 1.5 m long) and discontinuous
joints, with spacing S1a = 0.1 - 0.3 m (average 0.2 m) with:
− joint smoothness factor, js = 1.5 (slightly rough
− joint waviness factor, jw = 2 (strongly undulating)
− joint alteration factor, jA = 1 (fresh joint walls)
− joint length and continuity factor, jL = 4.
The joint condition factor for this set is jC1a = js × jw × jL/jA = 12

⇒ Foliation joints (set 1b); these are pervasive joints (L1b > 5 m) with spacing S1b = 2 - 3 m
(average 2.5 m) having:
− joint smoothness factor,js = 1 (smooth)
− joint waviness factor, jw = 2 (strongly undulating)
− joint alteration factor, jA = 2 (slightly altered rock in the joint wall; one grade more
than the rock)
− joint length and continuity factor, jL = 0.7.
The joint condition factor for this set is jC1b = js × jw × jL/jA = 0.7.

It is difficult to apply the method outlined in case 1 as the joints do not delineate defined
blocks. A possible way to characterize this type of ground is to consider that it is composed of
two sorts of blocks formed by the two types of joints. The jointing parameter is found as the
product of the jointing parameter for each of the two types of blocks as is shown in the
following:
− For joint set 1a - the foliation partings - the average block volume is determined by the
spacing and length of the joints
Vb1a = S1a × L1a2 = 0.13 m3
With jC1a = 12 the jointing parameter for this set is JP1a = 0.44.
− Similarly, for joint set 1b the block volume1 Vb1a = S1b × 42 = 40 m3
and the jointing parameter JP1b = 0.72 based on jC1b = 0.7.
The resulting jointing parameter for the rock mass is JP = JP1a × JP1b = 0.32.
(If the method shown in case 1 had been applied, the jointing parameter would be JP = 0.43,
because the effect of the foliation joints (set 1b) will not be fully included.)

Also for EHGGLQJMRLQWV with variation in spacings and joint characteristics the same method
as shown for foliation joints may be applied. Where both bedding joints and cross joints occur
this method may be useful.

6LQJXODULWLHVDQGZHDNQHVV]RQHV

6LQJXODULWLHV, i.e. seams or filled joints and small weakness zones, should be mapped and
considered separately where they occur as single features, see Fig. 5-9. If they occur in a kind of
pattern spaced less than about 5 m, they may sometimes be included in the detailed jointing
measurement.

The type and thickness of the filling is generally a main characteristic of singularities.

1
Here the length L = 4 m has been applied as outlined in Section 3.2.3 in Appendix 3
5 - 15

TABLE 5-3 ASSUMED APPROXIMATE RANGE OF JP AND/OR RMi VALUES FOR THE MAIN TYPES OF
WEAKNESS ZONES. THE VALUES DO NOT INCLUDE THE EFFECT OF SWELLING.
Jointing parameter Rock mass index
7<3(2):($.1(66=21( -3 50L

=RQHVRIZHDNPDWHULDOV
• /D\HUVRIVRIWRUZHDNPLQHUDOV, such as:
- clay materials 1)................................................. ** 0.01 - 0.05
- mica, talc, or chlorite layers and lenses 2)........... ** 0.05 - 5
- coal seams......................................................... 0.04 - 0.1 0.6 - 3
• =RQHVRIZHDNURFNVRUEUHFFLDWHGURFNV such as:
- some dolerite dykes 3)........................................ 0.005 - 0.05 *
- some pegmatites, often heavily jointed............... 0.005 - 0.05 *
- some brecciated zones and layers which
have not been "healed"...................................... 0.005 - 0.05 *
• :HDWKHUHGVXUIDFHRUQHDUVXUIDFHGHSRVLWV.......... 0.005 - 0.05 0.05 - 3

)DXOWVDQGIDXOW]RQHV
• 7HQVLRQIDXOW]RQHV
- feather joints and filled zones, such as:
> clay-filled zones 1)......................................... ** 0.01 - 0.05
> calcite-filled zones 2)..................................... ** 0.5 - 5
• 6KHDUIDXOW]RQHV
- coarse-fragmented, crushed zones...................... 0.01 - 0.1 *
- small-fragmented, crushed zones....................... 0.001 - 0.02 *
- sand-rich crushed zones..................................... 0.0005-0.005 0.0005 - 0.005
- clay-rich, crushed zones, such as:
> simple, clay-rich zones................................. 0.001 - 0.015 *
> complex, clay-rich zones.............................. 0.0005 - 0.01 *
> unilateral, clay-rich zones............................ 0.002 - 0.02 *
- foliation shears 4)
• $OWHUHGIDXOWV
- altered, clay-rich zones...................................... 0.005 - 0.05 0.006 - 3.5
- altered, leached (crushed) zones........................ 0.002 - 0.02 0.003 - 2
- altered veins/dykes............................................ 0.01 - 0.1 0.0003 - 0.3

5HFU\VWDOOL]HGDQGFHPHQWHGZHOGHG]RQHV It is difficult to assume general numerical values


for these types of zones

* Varies with the type of rock


** Massive rock is assumed(a scale factor of 0.5 has been applied for the compressive strength of rock)
1)
The clay is assumed very soft - firm
2)
No strength data found. The values given are assumed
3)
Assumed that the joints are without clay
4)
When occurring alone the foliation shear is probably a singularity; else probably a simple or complex clay-
rich zone
5 - 16

Fig. 5-9 Example of the influence from a singularity on stability (from Cecil, 1970)
Large and moderateZHDNQHVV]RQHV should, as previously mentioned, be characterized as one type
of rock mass having its own RMi value. In Appendix 3 the various features of weakness zones and
faults are further described. Not only the central part is of importance in the behaviour of the zone,
but also the transitional part and the composition of the surrounding rock masses should be
identified and given numerical values based on observation of block volume, joint condition and
rock material.

In many weakness zones most of the discontinuities are filled. Thus, the properties of the filling
material may dominate the behaviour of the zone.
Approximate RMi or JP values for weakness zones are shown in Table 5-3.

 6XPPDU\RIWKHSRVVLELOLWLHVDQGPHWKRGVWRGHWHUPLQHWKHEORFNYROXPHRU
WKHMRLQWLQJSDUDPHWHUZKHUHWKHMRLQWLQJFKDUDFWHULVWLFVYDU\

A summary of the possibilities for characterizing different joint condition parameters in various
types of rock masses is shown in Table 5-4.

TABLE 5-4 VARIOUS TYPES OF JOINTING AND JOINT CONDITION INDICATING THEIR SUITABILITY
TO BE CHARACTERIZED IN THE RMi
-2,17&21',7,21)$&725 (jC)
7<3(2)-2,17,1* DIFFERENT jC
SAME jC FOR
- consisting of between the between the between the
ALL JOINTS
sets only joints in the sets single joints
5HJXODUMRLQWLQJ
- mainly of joint sets x x (x) -
- columnar jointing x (x) ? -
0L[HGMRLQWLQJ
- joint sets + random joints x x  
,UUHJXODUMRLQWLQJ
- mainly random or irregular joints x - - 
)ROLDWLRQMRLQWLQJ
- long joints + short partings x x (x) ?
%HGGLQJMRLQWLQJ
- long joints + short cross joints x x (x) -
x Well suited for RMi characterization; i.e. jC can be used directly from field registrations
(x) Can be characterized satisfactorily; i.e. jC is assessed according to the method described
 Can be roughly characterized; i.e. jC may be estimated provided simplifications are made
? This type of jointing occurs seldom
- This type of jointing does not occur

The value of jC, which is connected to the different types of joints or joint sets, forms a vital part of
geo-data acquisition. It is, therefore, important that the observations are carried out by experienced
persons with knowledge of the geological conditions, and that the selection of parameters is tied to
well defined classes.
A verbal description of the joint condition is of great help here as additional information. This is
further explained in Appendix 3 in connection with ’translation’ of descriptions into numerical
values.

Table 5-5 shows a summary of the methods to determine the block volume and joint condition
factor where the joint characteristics and joint spacings vary.
5 - 17

TABLE 5-5 SOME OF THE METHODS AND EXAMPLES TO DETERMINE THE VALUE OF INPUT
PARAMETERS TO RMi ON JOINTING
9DULDWLRQVLQWKHEORFNVL]H 9E
$7KHEORFNYROXPHIRXQGIURPMRLQWVSDFLQJRUMRLQWGHQVLW\PHDVXUHPHQWV
Example 5-1 shows how the block volume can be determined from joint spacings.
Example 5-2 outlines how the block volume can be found from 1) the quantity of joints, and
2) from joint spacings.

%3UREDELOLW\FDOFXODWLRQVWRGHWHUPLQHWKHYDULDWLRQLQEORFNYROXPH
Example 5-3 shows a method to determine the variation range of block volume from joint spacings.

9DULDWLRQVLQWKHMRLQWFRQGLWLRQIDFWRU M&
$7KHM&GHWHUPLQHGIRURQHMRLQWRUMRLQWVHWZKHUHLWVSDUDPHWHUVYDU\
Alt.1. The variation range of the jC is found from combination of the parameter values so that the minimum
value and the maximum value is found for the actual joint or joint set.
Example 5-4: The jC determined from variations in joint characteristics.
Alt. 2. The variation range of jC is found using a SUREDELOLW\FDOFXODWLRQ similar to example 5-3.

%7KHUHVXOWLQJM&RU-3IRUWKHURFNPDVVZKHQM&YDULHVIRUHDFKMRLQWRUMRLQWVHW
Alt. 1. Where the joint sets have approximately the same spacings:
Use the average value of jC for all sets.
Alt. 2. Where the spacings are different for the joint sets:
a. Simply apply the (assumed) average jC for all the joint sets.
b. Use the joint set with the most unfavourable value for jC.
c. Sometimes one joint set is significantly more important than the others. In such cases the data for
this set may be applied directly.
d. Carry out an assessment of jC as described in the following two cases:
&DVH For every joint set with its jC and spacing it is assumed that the effect of jC varies
with the size (area) of the joint plane.
Example 5-5 shows how the average jC representative for all joint sets can be found.
Example 5-6: The jC found from various joints and joint conditions in an outcrop.
&DVH Consider that the rock mass is composed of (flat) blocks formed by only one of the
joint sets with its own jointing parameter (JP). The resulting JP for the rock mass is the
product of the jointing parameters found in the same way for each of the joint sets:
JP = JP1 × JP2 × JP3 × ...
Example 5-7: Determination of JP for various joint spacings and joint characteristics.
)URP3DOPVWU|P$50L±DURFNPDVVFKDUDFWHUL]DWLRQV\VWHPIRUURFNHQJLQHHULQJSXUSRVHV
3K'WKHVLV2VOR8QLYHUVLW\1RUZD\S

&KDSWHU

7+(86(2)50L,1'(6,*12)52&.6833257,181'(5*5281'
23(1,1*6

7KHEDVLFDLPRIDQ\XQGHUJURXQGH[FDYDWLRQGHVLJQVKRXOGEHWRXWLOL]HWKHURFNLWVHOIDVWKH
SULQFLSDO VWUXFWXUDO PDWHULDO FUHDWLQJ DV OLWWOH GLVWXUEDQFH DV SRVVLEOH GXULQJ WKH H[FDYDWLRQ
SURFHVVDQGDGGLQJDVOLWWOHDVSRVVLEOHLQWKHZD\RIFRQFUHWHRUVWHHOVXSSRUW,QWKHLULQWDFW
VWDWHDQGZKHQVXEMHFWHGWRFRPSUHVVLYHVWUHVVHVPRVWKDUGURFNVDUHIDUVWURQJHUWKDQFRQFUHWH
DQGPDQ\DUHRIWKHVDPHRUGHURIVWUHQJWKDVVWHHO&RQVHTXHQWO\LWGRHVQRWPDNHHFRQRPLF
VHQVHWRUHSODFHDPDWHULDOZKLFKPD\EHSHUIHFWO\DGHTXDWHZLWKRQHZKLFKPD\EHQREHWWHU
Evert Hoek and Edwin T. Brown (1980)

The purpose of this chapter is to show the use of the RMi in stability analysis and rock support
estimates for underground excavations. The first two sections summarize some of the current
knowledge on stability and failure modes in underground openings. Based on this, a system using
RMi parameters in rock support design has been developed.

To clarify, definitions of a few expressions related to the behaviour of rock masses underground are
presented:
6WDELOLW\is here used to express the behaviour of rock masses related to their OLNHOLKRRGRI
EHLQJIL[HGLQSRVLWLRQ(Webster’s dictionary).
Stability may be felt as a relative expression. In hard rock tunnelling where often a
considerable part of the tunnel can be left unsupported during the construction period as well
as during operation, any instability that requires support may be regarded as being a stability
problem. Tunnelling in poor rock conditions where continuous use of support or lining is
required, "stability problems" are often associated only with those parts of the ground where
the "standard" excavation procedure and method of support is inadequate and special
measures or solutions are required. In either case, a major objective is to assess the stability
behaviour correctly and select safe and economical methods for excavation and support.
Stability is a relative term also in other respects as it may be connected to the required level of
safety, which may vary with the use of the construction. The level of safety may also be
different in the various countries according to regulations for working conditions and safety,
as well as the experience the contractor possesses.

)DLOXUH, ’the losing of strength’, may, in contrast to instability (’the lack of being fixed in
position’) be regarded as the follower of instability. It may be simply said that failure is the
result of instability. Both failure and instability are used rather inconsistently in the literature
as they often overlap.

The term JURXQG is frequently used in this chapter. By ground is here meant ’the in situ rock
mass subjected to stress and water’.

5RFNPDVV is as mentioned earlier ’rocks penetrated by discontinuities’, i.e. the structural


material which is being excavated and in which the underground opening is located.
6-2

Another expression introduced in this chapter is the FRPSHWHQF\RIWKHJURXQG. It has a similar


meaning to ’competency of the rock’, i.e. related to the strength of the material compared to the
forces acting. While a competent rock bed is defined as "DURFNOD\HUZKLFKGXULQJIROGLQJIOH[HV
ZLWKRXWDSSUHFLDEOHIORZRULQWHUQDOVKHDU (Dictionary of geology, 1972), competent ground is a
rock mass or soil having higher strength than the stresses acting on it.

 67$%,/,7<$1$/<6(6$1'52&.6833257'(6,*1

There are no universal standard analyses for determining rock support design, because each design
is specific to the circumstances (scale, depth, presence of water, etc.) at the actual site and the
national regulations and experience. Support design for a tunnel in rock often involves problems
that are of relatively little or no concern in most other branches of solid mechanics. The material
and the underground opening forms an extremely complex structure. ",WLVVHOGRPSRVVLEOHQHLWKHU
WRDFTXLUHWKHDFFXUDWHPHFKDQLFDOGDWDRIWKHJURXQGDQGIRUFHVDFWLQJQRUWRWKHRUHWLFDOO\
GHWHUPLQHWKHH[DFWLQWHUDFWLRQRIWKHVH" (Hoek and Brown, 1980).

Therefore, the rock engineer is generally faced with the need to arrive at a number of design
decisions in which judgement and practical experience must play an important part. Prediction
and/or evaluation of support requirements for tunnels is largely based on observations, experience
and personal judgement of those involved in tunnel construction (Brekke and Howard, 1972). Often,
the estimates are backed by theoretical approaches in support design of which three main groups
have been practised in recent years, namely
- the classification systems,
- the ground-support interaction analysis (and the Fenner-Pacher curves used in NATM),
- the key block analysis.

The complex dilemma of structural analyses of tunnels is described in the guidelines for the design
of tunnels of an ITA Working Group edited by Duddeck (1988), from which the following is
extracted: 7KHUHVXOWRIDQDQDO\VLVGHSHQGVYHU\PXFKRQWKHDVVXPHGPRGHODQGWKHYDOXHVRIWKH
VLJQLILFDQWSDUDPHWHUV7KHPDLQSXUSRVHVRIWKHVWUXFWXUDODQDO\VLVDUHWRSURYLGHWKHGHVLJQ
HQJLQHHUZLWK
  $EHWWHUXQGHUVWDQGLQJRIWKHJURXQGVWUXFWXUHLQWHUDFWLRQLQGXFHGE\WKHWXQQHOOLQJSURFHVV
  .QRZOHGJHRIZKDWNLQGVRISULQFLSDOULVNVDUHLQYROYHGDQGZKHUHWKH\DUHORFDWHG
  $WRROIRULQWHUSUHWLQJWKHVLWHREVHUYDWLRQVDQGLQVLWXPHDVXUHPHQWV
7KHDYDLODEOHPDWKHPDWLFDOPHWKRGVRIDQDO\VLVDUHPXFKPRUHUHILQHGWKDQDUHWKHSURSHUWLHVWKDW
FRQVWLWXWHWKHVWUXFWXUDOPRGHO+HQFHLQPRVWFDVHVLWLVPRUHDSSURSULDWHWRLQYHVWLJDWH
DOWHUQDWLYHSRVVLEOHSURSHUWLHVRIWKHPRGHORUHYHQGLIIHUHQWPRGHOVWKDQWRDLPIRUDPRUH
UHILQHGPRGHO

The design of excavation and support systems for rock, although based on some scientific
principles, has to meet practical requirements. In order to select and combine the parameters of
importance for stability in an underground opening the main features determining the stability are
reviewed including various modes of failure.
6-3

 ,167$%,/,7<$1')$,/85(02'(6,181'(5*5281'(;&$9$7,216

Basically, the instability of rock masses surrounding an underground opening may be divided into
two main groups (Hudson, 1989):
1. One is block failure, where pre-existing blocks in the roof and side walls become free to move
because the excavation is made. These are called ’structurally controlled failures’ by Hoek and
Brown (1980) and involve a great variety of failure modes (as loosening, ravelling, block falls
etc.).
2. The other is where failures are induced from overstressing, i.e. the stresses developed in the
ground exceed the local strength of the material, which may occur in two main forms, namely:
a. Overstressing of massive or intact rock (which takes place in the mode of spalling,
popping, rock burst etc.).
b. Overstressing of particulate materials, i.e. soils and heavy jointed rocks (where squeezing
and creep may take place).

Various modes of failures are connected to these groups. Terzaghi (1946) has in his classification
worked out a behaviouristic description based on failure modes. Also the new Austrian tunnelling
method (NATM) contains a similar description of the ground behaviour which summarizes the main
types of instability in underground openings. Both descriptions are shown in Table 6-1.

Additional modes of rock mass behaviour in underground excavations described by Terzaghi (1946)
are:
Spalling 1, which refers to the falling out of individual blocks, primarily as a result of damage
during excavation.
Running ground, which occurs when a material invades the tunnel until a stable slope is formed
at the face. Stand-up time is zero or nearly zero. Examples are clean medium to coarse sands
and gravels above ground water level.
Flowing ground, which is a mixture of water and solids, which together invade the tunnel from
all sides, including the bottom. It is encountered in tunnels below ground water table in
materials with little or no coherence.

This has been envisaged in Fig. 6-1 where the various modes of failures in the 6 main groups of
ground are indicated. The groups are defined by the continuity of rock masses as described in
Section 1 in Chapter 5 and the quality of the ground, related to jointing for discontinuous rock
masses and to the stress/strength ratio for continuous materials.

Most types of rock masses fall within this scheme. In addition to the continuity and the competency
of the ground the time factor, the way the particle or blocks move, and the presence of water
determine the development and mode of a failure.

Input from experience and knowledge of the behaviour of various types of rock masses in
underground openings is important in stability analysis and rock support evaluations. Further, the
understanding of how possible failure modes are related to ground conditions is a prerequisite in the
estimates of rock support.

1
This term is often used by other authors as synonymous with popping or mild rock burst.
6-4

TABLE 6-1 VARIOUS MODES OF ROCK MASS BEHAVIOUR IN UNDERGROUND OPENINGS; TERMS
DEFINED BY TERZAGHI (1946) AND NATM (1993)
7HUPVDSSOLHGE\7HU]DJKL 7HUPXVHGLQ7KHQHZ$XVWULDQWXQQHOOLQJ
PHWKRG 1$70

Firm ground is a material which will stand unsupported 6WDEOH


in a tunnel for several days or longer. The term includes Elastic behaviour of the surrounding rock mass.
a great variety of materials: sands and sand-gravels with Small, quickly declining deformations. No relief
clay binder, stiff unfissured clays at moderate depths, features after scaling.
and massive rocks. The rock masses are long-term stable.

/RRVHQLQJ
Elastic behaviour of the rock mass, with small
deformations which quickly decline. Some few small
structural relief surfaces from gravity occur in the
roof.

5DYHOOLQJ
Ravelling ground indicates a material which gradually Far-reaching elastic behaviour of the rock mass with
breaks up into pieces, flakes, or fragments. The process small deformations that quickly decrease. Jointing
is time-dependent and materials may be classified by the causes reduced rock mass strength, as well as limited
rate of disintegration as VORZO\ or UDSLGO\ ravelling. For a stand-up time and active span*) . This results in relief
material to be ravelling it must be moderately coherent and loosening along joints and weakness planes,
and friable or discontinuous. Examples are jointed rocks, mainly in the roof and upper part of walls.
fine moist sands, sands and sand-gravel with some
binder, and stiff fissured clays. 6WURQJO\UDYHOOLQJ
Deep, non-elastic zone of rock mass around the tun-
nel. The deformations will be small and quickly
reduced when the rock support is quickly installed.
The low strength of rock mass results in possible
loosening effects to considerable depth followed by
gravity loads. Stand-up time and active span are
small with increasing danger for quick and deep
loosening from roof and working face.

Squeezing ground. 6TXHH]LQJRUVZHOOLQJ


6TXHH]LQJ rock slowly advances into the tunnel without The "plastic" zone of considerable size with
perceptible volume increase. It is merely due to a slow detrimental structural defects such as joints, seams,
flow of the material towards the tunnel, at almost shears results in plastic VTXHH]LQJ as well as URFN
constant water content. The manifestations and the EXUVWSKHQRPHQD
causes of the squeeze can be very different for different Moderate, but clear time-dependent squeezing with
clays and decomposed rocks. only slow reduction of deformations (except for rock
Prerequisite for squeeze is a high percentage of burst). The total and rate of displacements of the
microscopic and sub-microscopic particles of micaceous opening surface is moderate. The rock support can
minerals or of clay minerals with a low swelling sometimes be overloaded.
capacity.

Popping or rock burst is the sudden, violent detachment


of thin rock slabs from sides or roof, and is caused
primarily by the overstressing of hard, brittle rock.

Swelling ground advances into the tunnel chiefly by 6WURQJO\VTXHH]LQJRUVZHOOLQJ


expansion from water adsorption. The capacity to swell Development of a deep squeezing zone with severe
seems to be limited to those rocks which contain clay inwards movement and slow decrease of the large
minerals such as montmorillonite, and to rocks with deformations. Rock support can often be overloaded.
anhydrite.
*)
Active span is the span of the tunnel or the distance from the lining to the working face if this is smaller. (Lauffer, 1958)
6-5


& R Q W L Q X L W \ R I J U R X Q G 

(he a vil y jo in t ed r ock s)


U
H U
W
H
H
CON TI NU OU S

W
P H 
D P ) / 2 :, 1* * 52 81 '
L & 2/ / $ 36 (
D
G L
 
G 5 81 1, 1 * * 52 81 '
J 
Q N  
L F 6 48 ( ( = ,1 * 
Q R 5 $9 ( / / ,1 *
H O
S E

R  / 22 6 ( 1, 1*
5 $9 ( / / ,1 *


 

I N COM PETEN T GR OU ND COMPE TENT GR OU ND


(overstressed rock mass es)

D I SC ON TIN U OUS
(j oi n te d ro cks )


 5 $ 9 ( / / , 1 *


 / 2 2 6 ( 1 , 1 *


% / 2 & . ) $ / / 6
6 7$ %/ (

 poor c onditions  good c ondition s 


(m as siv e roc ks)
C ON TI N U OUS


6 3$ / / , 1*
 
3 23 3 , 1*

5 2& . %8 56 7
6 7$ %/ (

& 5( ( 3

6 48 ( ( =,1 *



  

I N COM PETEN T GR OU ND COMPE TENT GR OU ND


(overstressed rock mass es)

4XD O L W \ RI JU RXQG

1)
Swelling may increase instability; water must be present for this process
2)
In connection with little or no water
3)
Depends on the presence of water
4)
Depends on the type of rock

Fig. 6-1 Main types of instability in underground excavations.

 6SHFLDOPRGHVRILQVWDELOLW\DQGEHKDYLRXUUHODWHGWRIDXOWVDQGZHDNQHVV
]RQHV

Faults and weakness zones often require special attention in underground works, because their
structure, composition and properties may be quite different from the surrounding rock masses.
Zones of significant size can have a major impact upon the stability as well as on the excavation
process of an underground opening, for instance from possible flowing and running, as well as high
ground water inflow. These and several other possible difficulties connected with such zones
require that special investigations often are necessary to predict and avoid such events. Bieniawski
(1984, 1989) therefore recommends that they are mapped and treated as regions of their own.
6-6

Many faults and weakness zones contain materials quite different from the ’host’ rock from
hydrothermal activity and other geologic processes. The instability of weakness zones may depend
on other features than the surrounding rock which all interplay in the final failure behaviour. An
important inherent property in this connection is the character of the JRXJHRUILOOLQJPDWHULDO.
Brekke and Howard (1972) has described the main types of fillings in seams and weakness zones
and their possible behaviour, as shown in Table 6-2

TABLE 6-2 BEHAVIOUR OF FILLING MATERIAL AND GOUGE IN SEAMS AND WEAKNESS ZONES
(revised from Brekke and Howard, 1972)
'RPLQDQWPDWHULDOLQ 327(17,$/%(+$9,285&$86('%<7+(*28*(0$7(5,$/
IDXOWJRXJHILOOLQJ $WIDFH /DWHU
Swelling clay Free swell, sloughing. Swelling Swelling pressure and squeeze on support or
pressure and squeeze on shield. lining, free swell with down-fall or washing
if lining is inadequate.

Inactive clay Slaking and sloughing caused by Squeeze on supports or lining where un-
squeeze. Heavy squeeze under protected, slaking and sloughing due to
extreme conditions. environmental changes.

Chlorite, talc, graphite, Ravelling. Heavy loads may develop due to low
serpentine strength, in particular when wet.

Crushed rock fragments Ravelling or running. Flowing if Loosening loads on lining, running and
or sand-like gouge surplus of water. Stand-up time may ravelling if unconfined.
be extremely short.

Porous or flaky calcite, Favourable conditions. May dissolve, leading to instability of rock
gypsum mass.

Table 6-2 indicates that most modes of failures can take place in such zones; often two or more may
act at the same time making stability evaluations of weakness zones a very difficult task. Fault
gouge is normally impervious, with a major exception for sand-like gouge. Otherwise, high
permeability may occur in the jointed rock masses adjacent to the fault zone. High water inflows
encountered in underground openings when excavating from the weak impervious gouge in the
zone, is one of the most adverse conditions associated with faults (Brekke and Howard, 1972).

 0DLQW\SHVRIIDLOXUHGHYHORSPHQW

The main modes of failures or instability in underground openings may develop in basically three
different ways:
1. Loosening and falls of single blocks or fragments.
2. Collapse; i.e. the tunnel is filled with the fallen blocks which become wedged together and
provide support to the remaining loose, unstable blocks.
3. Limited deformations on the surface of the opening caused by the redistribution of the stresses
forming a stable arch in the surrounding rock masses.

The first and the last type - though dangerous for the tunnel workers - have generally limited
consequences, while the second is the most serious mode of failure and may cause severe
construction problems. It may start as a progressive failure, which develops into a collapse in
particulate materials (for example highly jointed or altered rock masses).
6-7

Terzaghi (1946) has described the typical development as: ([SHULHQFHVKRZVWKDWWKHJURXQGGRHV


QRWFRPPRQO\UHDFWDWRQFHWRWKHFKDQJHRIVWUHVVSURGXFHGE\WKHEODVW7KHURXQGEODVWFUHDWHV
DQXQVXSSRUWHGVHFWLRQRIURRIORFDWHGEHWZHHQWKHQHZIDFHDQGWKHODVWVXSSRUWRIWKHWXQQHO$V
VRRQDVWKHQDWXUDOVXSSRUWLQJURFNLVUHPRYHGE\EODVWLQJVRPHEORFNVGURSRXWRIWKHURRI
OHDYLQJDVPDOOJDSLQWKHKDOIGRPH,IWKHQHZO\H[SRVHGURRIVHFWLRQLVOHIWZLWKRXWVXSSRUWVRPH
PRUHEORFNVGURSRXWDIWHUDZKLOHWKXVZLGHQLQJWKHJDS)LQDOO\WKHHQWLUHPDVVRIURFN
FRQVWLWXWLQJWKHKDOIGRPHGURSVLQWRWKHWXQQHODQGDQHZKDOIGRPHLVIRUPHG7KHQHZKDOIGRPH
DOVRVWDUWVWRGLVLQWHJUDWHDQGWKHSURFHVVFRQWLQXHVXQWLOWKHWXQQHOVHFWLRQDGMRLQLQJWKHZRUNLQJ
IDFHLVILOOHGZLWKURFNGHEULV
7KHUDWHDWZKLFKWKHSURJUHVVLYHGHWHULRUDWLRQRUUDYHOOLQJRIWKHKDOIGRPHWDNHVSODFHGHSHQGV
RQWKHVKDSHDQGVL]HRIWKHEORFNVEHWZHHQMRLQWVRQWKHZLGWKRIWKHMRLQWVRQWKHMRLQWILOOLQJDQG
WKHDFWLYHVSDQ

These observations stress the need to closely evaluate the timing for installation of rock support and
the need to follow-up the development of tunnel behaviour where low stand-up time occurs. This is
further described in Section 6.4.4.

 7+(0$,1)($785(6,1)/8(1&,1*81'(5*5281'67$%,/,7<

The various types of failures described in the foregoing section may be the result of numerous
variables in the ground. Both the composition of the rock mass and the forces acting upon it
contribute to the result. Wood (1991) and several other authors find that the EHKDYLRXU of ground in
an underground excavation depends on
- the generic or internal features of the rock mass;
- the external forces acting, (the ground water and stresses); and
- the activity of man in creating the opening and its use.
Based on published papers, especially the work by Cecil (1970) and Hoek and Brown (1980), and
from own experience the following factors have been found most decisive for the stability of
underground constructions in jointed rock masses:

1. The LQKHUHQWSURSHUWLHVRIWKHPDWHULDO URFNPDVVHV surrounding the opening. They consist


mainly of:
a. Intact rock properties.
b. Properties of jointing and discontinuities.
c. Structural arrangement of joints and other discontinuities.
d. Swelling properties of rocks and minerals.
e. Durability of the material.
2. The H[WHUQDOIRUFHVDFWLQJ in the ground:
a. Magnitude and anisotropy of horizontal and vertical stresses in undisturbed rock.
b. Ground water.
3. The H[FDYDWLRQIHDWXUHV, such as:
a. Shape and size of the underground opening.
b. Method(s) and timing of rock support.
c. Method of excavation.
d. Ratio of joint spacing/span width.
4. The WLPHGHSHQGHQWIHDWXUHV, mainly consisting of:
a. The effect of stand-up time.
b. The long-term behaviour (caused by changes in 1. and 2.)
6-8

The influence of these features and their possible application in a method for stability and rock
support design are described in the following.

 7KHLQKHUHQWSURSHUWLHVRIWKHURFNPDVV

The geological conditions have generally greater influence on the stability than any other single
factor. The exact rock mass conditions at the site will, as mentioned in Chapter 3, not be known
until the excavation is made. Some of the variation in rock masses, their composition, occurrence
and characteristics have been described in Appendices 1 and 2, and their numerical characterization
in Chapter 5.

6.3.1.1 Properties of the intact rock and the discontinuities

The mechanical properties of the rock material can be characterized by the strength of the intact
rock, while the joint characteristics and the block size are representative of many of the properties of
the discontinuities. The latter are by several authors defined as the main contributors to instability
underground. The parameters for rock strength, block volume and joint characteristics are all
included in the RMi. They are further described in Chapter 4 and in Appendix 3.

6.3.1.2 Structural arrangement of geologic discontinuities

By the structural arrangement of discontinuities is meant the joint pattern and the geometry of
blocks. Also the orientation of discontinuities with respect to the periphery of the opening and the
intersection geometry of discontinuities are included. Their intersecting angle with the tunnel
determines whether they influence the stability in the walls or the roof of a tunnel.

Cecil (1970) observed that multiple joint sets are most often associated with support and that single
sets of joints were frequently of no concern to the stability of an opening. This may be reasonable as
many joint sets generally will result in smaller blocks and hence reduced stability. Cecil also noticed
that a single random joint may have a very drastic effect on an otherwise stable jointed rock mass
(Fig. 6-2).

12.5 m

Fig. 6-2 Example of the effect of single joints on stability (from Cecil, 1970).

Deere et al. (1969) indicate from experience that two orientations of joints are particularly
important:
- steeply dipping joints (45-90o) which are parallel or subparallel to the tunnel axis,
- flat-lying (0-30o) joints occurring in the tunnel roof.
6-9

Bieniawski (1984) has classified importance of the orientation of joints in relation to an


underground opening, as shown in Table 6-3. The influence from orientation is similar also for
weakness zones and singularities.

TABLE 6-3 CHARACTERIZATION OF DISCONTINUITY ORIENTATION RELATED TO AN


UNDERGROUND EXCAVATION (revised from Bieniawski, 1984).
675,.( 'LS R 'LS R 'LS R
favourable* very favourable*
Strike across tunnel axis fair
unfavourable** fair**
Strike parallel to tunnel axis fair fair very unfavourable
* for drive with dip, ** for drive against dip

6.3.1.3 Swelling properties of rocks and minerals

This is a special property of rocks and soils caused by the expansion of special minerals like
smectite (montmorillonite, vermiculite, etc.) and anhydrite upon access to water. The swelling
pressure will exert loads on the support in addition to the load from the ground stresses and gravity.
Swelling may on some occasions highly influence the stability as well as the problems during the
tunnel excavation (Selmer-Olsen 1964, 1988; Brekke and Selmer-Olsen, 1965; Selmer-Olsen and
Palmström 1989, 1990).

In addition to several types of soils containing swelling clay minerals (bentonite, etc.) swelling
materials can for example occur in:
- altered rock containing smectite;
- sedimentary rock containing anhydrite; or
- clay material in seams or filled joints either occurring singly or as parts of a fault or
weakness zone.

6.3.1.4 Durability of the material

Durability is the resistance of a rock against slaking or disintegration when exposed to weathering
processes. Some rocks may hydrate ("swell"), oxidize, or disintegrate or otherwise weather in
response to the change in humidity and temperature consequent on excavation. An abundant group
of rocks, the mudrocks, are particularly susceptible to even moderate weathering (Olivier, 1976).
This will change the mechanical properties of rock and hence influence the stability.

7KHH[WHUQDOJURXQGIHDWXUHV

The external forces acting on a rock mass surrounding an underground opening are related to its
depth and geological location. Their magnitude and orientation may be influenced by the
topography in the area, climate, and geological history.

The two main external features are mentioned below. The effect of vibrations from earth quakes or
from near-by blastings, or local drainage from other near located tunnels are other features which in
addition may influence the stability. Their occurrence and effect are highly connected to local
features of the site and the construction to be made.
6 - 10

6.3.2.1 Magnitude of horizontal and vertical stresses in undisturbed ground

The in situ stress level at the location of an underground excavation may have a great impact on its
stability where the stresses set up in the rock mass around the opening exceed its strength. Not only
high stresses may cause instability problems, also a low stress level may increase instability in
jointed rock masses because of reduced shear strength on joints from the low normal stress.

The excavation of the opening disturbs the original, virgin stresses and the stresses set up around the
opening may be quite different, depending upon the ratio between horizontal and vertical stresses
and the size and the shape of the opening. After the excavation has been mined, these stresses will
redistribute. The ratio between the strength of the material and the stresses acting is important in
this process. The stresses can be measured where a stress cell can be placed, or it can be estimated
from topography, overburden and knowledge of the general stress situation in the region.

6.3.2.2 Ground water

Excessive ground water pressure or flow can occur in almost any rock mass, but it would normally
only cause serious stability problems when this takes place in crushed or sand-like materials, or if
associated with other forms of instability. Cecil (1970) mentions that the effect of ground water on
stability is caused by reducing both the strength of rock material and the shear strength of the
discontinuities. As mentioned by Selmer-Olsen (1964), Brekke and Howard (1972) and Selmer-
Olsen and Palmström (1989), water can significantly reduce the strength of the filling or gouge in a
fault, weakness zone or seam if swelling takes place. Swelling and the following softening also
leads to reduced frictional resistance.

Although water may have little influence on stability, the presence of significant quantities of
ground water can cause disruption in the excavation process. The most serious problems with
ground water occur when it is encountered unexpectedly. Terzaghi (1946) mentions that quantities
of 61.3 m3/min have been experienced in granitic rocks at a depth of 265 m. Large inflows of water
can also be experienced in karstic limestones. According to Brekke and Howard (1972) real hazards
arise where large quantities of water in a permeable rock mass are released when an impervious
fault gouge is punctured through excavation. In this instance, large quantities of gouge and rock can
be washed into the tunnel.

Relatively few of the described failures in the literature are specifically related to joint water
pressure, it is, however, very possible that ground water can contribute to instability in weak
ground. High ground water pressures built up near the excavation have in some occasions caused
instability. The impact from ground water pressure should be evaluated in cases where it has
significant influence.

7KHH[FDYDWLRQIHDWXUHV

Excavation features are the man-made disturbances in the ground. The creation of an excavation
includes a number of factors influencing on the stability in the underground opening. The most
important of these are mentioned below.
6 - 11

6.3.3.1 The size and shape of the opening

Terzaghi (1946) found that the loads will increase linearly with size of the opening, except for
swelling ground where high swelling pressures may develop regardless of the size. For excavations
exposed to high rock stresses, Selmer-Olsen (1988) points out the importance of excavating a
simple shape of the opening without ledges and overhang to reduce the amount of loosening and
spalling. He has shown that, by shaping the tunnel with a reduced curve radius in the roof where the
largest in situ tangential stress occurs, it is possible to reduce the area of over-stressing and hence
the extent of instability (see Fig. 6-10).

Also, in jointed rocks without overstressing, the stability is improved where a simple shape of the
opening is chosen (Selmer-Olsen, 1964, 1988). The stability diagram in the Q-system clearly shows
that the amount of rock support depends significantly on the size of the opening. Further, Hoek and
Brown (1980) and Hoek (1981) have shown that the shape of the opening has a significant influence
on the magnitude of the stresses set up in the rocks surrounding the opening.

6.3.3.2 Method of excavation

It is commonly accepted that any method used to excavate a tunnel will cause some disturbance of
the surrounding rock structure, which in turn will affect the stability. The various excavation
techniques used may exert different influence on the tendency of blocks to loosen and fall out of the
tunnel walls or roof. For example, mechanical tunnel excavation would tend to disturb the blocks
much less than drill and blast excavation.

In most cases, it is very difficult to distinguish between the "before" and "after" conditions and
whether impact from the excavation may have had an effect on the amount of rock support.
Fracturing from blasting leads to reduced block size; actual loosening of rock caused by blasting is,
however, often more reflected as ’overbreak’ than by additional support requirements. The influence
of blasting can be substantially reduced by controlled SHULPHWHUEODVWLQJ. Another result is a
smoother surface of the opening.

6.3.3.3 Method(s) and timing of rock support

In overstressed ground where yielding and squeezing take place, the rate and size of deformations
depend on the timing and strength of the confinement (method, amount and stiffness of support)
placed. This is clearly shown in the ground - support interaction curves (see Chapter 8, Section 8.3).

6.3.3.4 Ratio of joint spacing and tunnel diameter

Deere et al. (1969) suggested that for a tunnel in a jointed or particulate material, a characteristic
dimension, such as the tunnel diameter, may be compared with the size of the individual fragments
or the joint spacing of the material. This expresses the continuity of the ground as described in
Chapter 5, and is regarded as an important parameter representing the effect of block loosening, see
Fig. 6-3.
6 - 12

Fig. 6-3 The difference between discontinuous (left) and continuous materials (revised from Barton, 1990b).
Increasing the number of blocks in the tunnel surface increases the likelihood of blocks to loosen and the
volume involved in a possible failure.

7KHWLPHGHSHQGHQWIHDWXUHV

When time-dependent behaviour of soil or rock around an underground opening is considered, there
are two separate influences:
short term:
- the variation of the stress field as the face advances away from the point concerned in the
tunnel (stand-up time), and
long term:
- the creep factors, i.e. creep under constant shear stress,
- the influence from the environment,
- the durability of the rock mass, and
- the effect of ground water.

6.3.4.1 Short term behaviour and the effect of stand-up time

Significant changes in tunnel stability may occur as a result of readjustment of stresses in the walls
and roof of a tunnel as the face is advanced. The deformation and stress re-distribution after
excavation requires time. The stability effect of this is acknowledged in the ’stand-up time’ which
was first systematized by Lauffer (1958). Lauffer showed that the property of the rock mass, the
active span of the excavation, and the time elapsed until unstable conditions occur, are related to
each other. These relations, which are key-points in stability assessments, have been applied in the
new Austrian tunnelling method (NATM). Also Bieniawski (1973) has selected the principles of
Lauffer in the stability diagram applied in the geomechanics (RMR) classification system.

Earlier, Terzaghi (1946) described the effect of tunnel span on what he called ’the bridge-action
period’ (stand-up time). 7KHEULGJHDFWLRQSHULRGIRUDJLYHQPDWHULDOLQFUHDVHVYHU\UDSLGO\ZLWK
GHFUHDVLQJGLVWDQFHEHWZHHQVXSSRUWV7KXVIRULQVWDQFHDYHU\ILQHPRLVWDQGGHQVHVDQGFDQ
EULGJHDVSDQRQHIRRWZLGHIRUVHYHUDOKRXUV<HWWKHVDPHVDQGZRXOGDOPRVWLQVWDQWDQHRXVO\
GURSWKURXJKDJDSEHWZHHQVXSSRUWVZLWKDZLGWKRIILYHIHHW
6 - 13

The stand-up time diagram by Lauffer is also based on the behaviour of rock mass. A main point in
this diagram is that an increase in tunnel size leads to a drastic reduction in stand-up time. The
effect of time, therefore, plays an important role when stability evaluation of placing rock support
are being made, especially at face when low stability (short stand-up time) conditions are
encountered. This is further dealt with in Section 6.4.4.

6.3.4.2 Long-term behaviour

There are several geologic factors that may influence the long-time dependent behaviour of rock
masses in an underground excavation. The influence of possible alteration of rock and gouge, or
swelling, softening and weakening along discontinuities are factors that must be specially evaluated
in each case. These effects are not necessarily obvious during construction, therefore, the long time
stability may easily be underestimated during the construction period. Brekke and Howard (1972)
have shown the effect of long-time behaviour of fillings in faults and weakness zones (see Table 6-
2).

In highly stressed rocks the effect of long-time creep may change the strength of the material as
described by Lama and Vutukuri (1978). Another long-term effect is the slaking of mudrocks as
mentioned in Chapter 2, Section 2.1.2.

The hydraulic effect of ground water may wash out joint filling materials through piping action, and
thus, in this way, influence the long-term stability of the rock masses surrounding the opening.

From this it is clear that the effect of time depends on the conditions at the specific site, and it is
difficult to include this effect in a general method of stability analysis.

6XPPDU\RI6HFWLRQ

It is not possible to include all the factors mentioned above in a practical system for assessing
stability and rock support. Therefore, only the most important features should be selected. Based on
the published material mentioned in the foregoing and the author’s own experience in this field, the
factors mentioned in Table 6-4 are considered the generally most important ones regarding stability
and rock support.

Regarding other factors, which influence the stability in underground openings, the following
comments are made:
- The effect from VZHOOLQJ of some rocks, and some gouge or filling material in seams and faults
has not been included.2 The swelling effect highly depends on local conditions and should
preferably be linked to a specific design carried out for the actual site conditions.
- The ORQJWHUPeffects must be evaluated in each case from the actual site conditions. These
effects may be creep effects, durability (slaking etc.), and access to and/or influence of water.
- In the author’s opinion it is very difficult to work out a general method to express the VWDQGXS
WLPH accurately as it is a result of many variables - among others the geometric constellations.
Such variables are generally difficult to characterize by a simple number or value.

2
The influence from weakening and loss of friction in swelling clays is, however, included in the joint alteration
factor (jA) as input to the joint condition factor (jC) in RMi.
6 - 14

TABLE 6-4 THE GROUND PARAMETERS OF MAIN INFLUENCE ON STABILITY OF UNDERGROUND


OPENINGS
7+(*5281'&21',7,216 &+$5$&7(5,=('%<
7KHLQKHUHQWSURSHUWLHVRIWKHURFN
PDVV
- The intact rock strength * The uniaxial compressive strength (included in RMi)

- The jointing properties * The joint characteristics and the block volume (represented in the
jointing parameter (JP))

- The structural arrangement of the (*) 1) Block shape and size (joint spacings )
discontinuities * 2) The intersection angle between discontinuity and tunnel surface

- The special properties of * 1) Width, orientation and gouge material in the zone
weakness zones 2) The condition of the adjacent rock masses

7KHH[WHUQDOIRUFHVDFWLQJ
- The stresses acting * The magnitude of the tangential stresses around the opening, deter-
mined by virgin rock stresses and the shape of the opening

- The ground water (*) Although ground water tends to reduce the effective stresses acting in
the rock mass, the influence of water is generally of little importance
where the tunnel tends to drain the joints. Exceptions are in weak
ground and where large inflows disturb the excavation and where
high ground water pressures can be built up close to the tunnel
7KHH[FDYDWLRQIHDWXUHV
- The shape and size of the opening * The influence from span, wall height, and shape of the tunnel

- The excavation method (*) The breaking up of the blocks surrounding the opening from blasting

- Ratio tunnel dimension/block size * Determines the amount of blocks and hence the continuity of the
ground surrounding the underground opening
* Applied in the system for stability and rock support (Section 6.4) (*) Partly applied

There are features linked to the specific case, which should be evaluated separately. They are the
safety requirements, and the vibrations from earthquakes or from nearby blasting or other
disturbances from the activity of man.

 50L$33/,('72$66(6652&.6833257

,WLVHVVHQWLDOWRNQRZZKHWKHUWKHSUREOHPLVWKDWRIPDLQWDLQLQJVWDELOLW\ZLWKWKHSUHH[LVWLQJ
MRLQWLQJSDWWHUQRUZKHWKHULWLV WKH YHU\ GLIIHUHQW SUREOHP RI D \LHOGLQJ URFN PDVV 7KH VWUHVV
VLWXDWLRQLVWKHUHIRUHRQHRIWKHPDLQSDUDPHWHUVLQVWDELOLW\DQGURFNVXSSRUWHYDOXDWLRQV
Sir A.M. Muir Wood (1979)

Methods of applying the RMi value directly or some of the parameters in the RMi in the design of
rock support are described in this section. The developments made are based on the previous
sections in this chapter, published papers, in addition to the author’s own practical tunnelling
experience.

The behaviour of the rock mass surrounding an underground opening is the combined result of
several of the parameters mentioned in the foregoing sections. The influence or importance of each
of them will vary with the opening, its location, and with the composition of the rock mass at this
location. In a selection of these parameters it has been found beneficial to combine parameters
which have similar effects on the stability into two main groups. These are the FRQWLQXLW\IDFWRU and
the ground FRQGLWLRQIDFWRU:
6 - 15

- The continuity, i.e. the ratio tunnel size/block size, of the ground surrounding the tunnel
which determines whether the volume of rock masses involved can be considered
discontinuous or not, see Fig. 6-4. This is important both as a parameter in the
characterization of the ground, but also in the determination of appropriate method of
analysis. As mentioned in Chapter 5, discontinuous rock masses may have a continuity
factor between 5 and 100, else the ground is continuous.
1000 G
H G
W
tunnel dimension
block dimension

H
Q
L K
R V
M
 X
\ U
O F
K 
U
J R
L
K

overstressed competent
100


)
& G


 H
W
5 Q
L
2 R
M
7 10
&
$ poor good
) 5

<
7
,
8
H
1
, Y
L
7 V
V
1 1 D
2 P
& overstressed competent

0.01 1 100
*5281' 48$/,7<

Fig. 6-4 The classification of ground into continuous and discontinuous rock masses.

- The quality of the ground, is composed of important properties of the rock mass and the
external features of main influence on the stability of the opening. As pointed out previously,
there are different parameters that determine stability in continuous and discontinuous
ground. Therefore, in FRQWLQXRXV ground the competency factor has been applied. It is
expressed as
strength of the rock mass
Cg  eq. (6-2)
tangential stress around the opening

In discontinuous ground, and where weakness zones are involved, a ground condition factor is
introduced as further described in Sections 6.4.2 and 6.4.3.

6WDELOLW\DQGURFNVXSSRUWLQFRQWLQXRXVPDWHULDOV

Fig. 6-4 shows that continuous rock masses involves two categories
1. Slightly jointed (massive) rock with continuity factor (tunnel size/block size), continuity factor
CF < approx. 5.
2. Highly jointed and crushed rocks, continuity factor CF > approx. 100.

Instability in continuous ground can, as mentioned in Section 6.2, be both stress-controlled and
structurally influenced. The structurally released failures, which occur in the highly jointed and
crushed rock masses, are described in Section 6.4.2 for discontinuous materials. According to Hoek
and Brown (1980) they are generally overruled by the stresses where overstressing occurs.

The system for assessment of stability and rock support is presented in Fig. 6-5.
6 - 16

&2035(66,9( 675(1*7+
TYPE OF ROCK
2) 52&.
5 2 & . 0 $ 6 6 , 1 ' ( ;

- 2 , 1 7 & 2 1 ' , 7 , 2 1 )$ & 7 2 5

JOINTING PARAMETER 50L


% / 2 & . 9 2 / 8 0 (


6
6
H
O
GROUND WATER
(
L
W
TANGENTIAL
DIAMETER

1 F STRESSES
X
BLOCK

) 5 2& . 675(66(6
)
,
G
 AROUND
7 U
6 R
 6 + $ 3( 2 )
THE OPENING

H
. O 7 + ( 2 3( 1 , 1 *
W
& W
L
2 U
5 E ', $0(7(5 2)

7+ ( 23(1, 1*

CONTINUITY OF &20 3( 7( 1&<

THE GROUND 2) 7 +( *5281'

&) &J
fo r
co
roc k nti nuo
mas us
ses

input
parameter
main input
parameter

Fig. 6-5 The principle and the parameters involved in assessment of stability and rock support in continuous rock
masses.

6.4.1.1 The competency of continuous ground

The excavation of a tunnel disturbs the original rock stresses and the ground water situation in the
ground. After the opening is mined, the original stresses are redistributed in the remaining rock
mass. This results in local increases in the stresses in the immediate vicinity of the excavation.

As the stress-controlled failures generally dominate the instability in this group of ground the
competency factor (Cg) in eq. (6-2) has been selected to characterize the quality of the ground. Cg
may be found by combining the induced stresses acting in the rock masses around the opening and
the strength of the rock mass. As RMi is valid in continuous ground, and expresses the strength of
the rock mass (as outlined in Chapter 4, Section 4.5.1), it can be used in assessing the FRPSHWHQF\
IDFWRU
RMi
Cg = eq. (6-3)
σθ

where σθ is the tangential stress at different points around the underground opening. It can be
found from the vertical rock stress (pz ), the ground water pressure (uz ), and the shape
of the opening as outlined in Appendix 9.

Cg indicates whether the material around the tunnel is overstressed or not. This term has earlier
been proposed by Muir Wood (1979) as the ratio of uniaxial strength of rock to overburden stress,
6 - 17

to assess the stability of tunnels. This parameter has also been used by Nakano (1979) to recognize
the squeezing potential of soft-rock tunnelling in Japan.

The greatest influence of the stresses occurs when they exceed the strength of the material, creating
incompetent ground, as further outlined in Sections 6.4.1.2 - 6.4.1.4. Such incompetent ground
leads to failure if confinement by rock support is not established (Fig. 6-6). If the deformations take
place instantaneously (often accompanied by noise), the phenomenon is called URFNEXUVWLQJ; if the
deformations caused by overstressing occur more slowly, VTXHH]LQJ occurs, as further described in
the following.
WD
V Q H
V
WU J V
H H V
P D
VXSSRUWOLQLQJ V Q
V WL IUR P
 D W N
F
O H
Q R
U

σθ
P J
H LQ
LQ G
I Q
Q X
R
F UR
U
X
V

P
R
IU J
W LQ
Q Q
H OL
P UW
H R
Q
IL S
Q S
R X
F V

7811(/

Fig. 6-6 The principle of confinement from rock support of an overstressed element in incompetent ground.

6.4.1.2 Continuous, massive ground

overstressed
2
,
7
$
5

(
massive

=
,
6

overstressed

*5281'48$/,7<

In massive rock with few joints the rock mass index is RMi = f × σc . Thus, the competency of
ground is
Cg = RMi/σθ = fσ × σc /σθ eq. (6-4)

where fσ is a scale factor for the compressive strength, see Section 4.2 in Chapter 4.

In FRPSHWHQW massive ground, i.e. Cg > 1, where the new stress condition around the tunnel does
not exceed the ground strength at any time, the ground moves into a new position of equilibrium.
Structural reinforcement is only required to support possible loosened blocks from unfavourable
combinations of the few joints present or of spalls from extension cracking.
6 - 18

In LQFRPSHWHQW massive ground, i.e. Cg < 1, the overstressing of the rock mass will cause some
form of stress induced instability:
- InEULWWOH rocks they may cause breaking up into fragments or slabs ’expressed’ as rock burst 3
in hard, strong rocks such as quartzites and granites. This is further described in Section
6.4.1.4.
- In the more GHIRUPDEOHIOH[LEOHRUGXFWLOH rocks such as soapstone, evaporites, clayey rocks
(mudstones, clay schist, etc.) or weak schists, the failure by overstressing may act as
squeezing; a slow inward movements of the tunnel surface, as outlined in Section 6.4.1.5.

Thus, in overstressed, massive rocks the deformation properties or the stiffness of the material
mainly determine which one of the two types of stress problems that can take place.

6.4.1.3 Continuous ground in the form of particulate (highly jointed) materials


crushed

overstressed
2
,
7
$
5

(
=
,
6
overstressed
*5281'48$/,7<

This type of ground consists of highly jointed and crushed rock masses as well as soil materials.
Instability in such particulate rock masses may develop as two modes:
1) the stress dependent, and 2) the structurally induced failures.

At a relative ORZWRPRGHUDWHVWDWHRIVWUHVV instability is dominated by VWUXFWXUDOO\FRQWUROOHG


gravity-induced loosening or ravelling of blocks. This disintegration may be slow or rapid. Though
the blocks in heavily jointed rock masses in general are smaller than those in discontinuous rock
masses (see Section 6.4.2), the properties responsible for their structurally induced instability are
similar. This type of instability, which is further described in Section 6.4.2, is often experienced in
weakness zones.

InRYHUVWUHVVHGground instability in the form of squeezing takes place, as further described in


Section 4.1.5.

Here, it should be noted that rapid failures in the form of UXQQLQJJURXQG and IORZLQJJURXQG also
may take place. They occur in earth-like rock masses and in particulate soils. Due to the very serious
problems and consequences they may produce for tunnel excavation, special considerations
regarding investigations, excavation and rock support are generally required. No general use of
RMi seems relevant for these conditions.

3
Here, ’rock burst’ is related to overstressing of the rock. Stress induced failures caused by lower stresses
are known as ’spalling’, ’popping’, ’slabbing’, etc.
6 - 19

6.4.1.4 Rock burst and spalling in brittle rocks

Stress induced failures in brittle rocks are known as VSDOOLQJ 4SRSSLQJ or URFNEXUVW, but also a
variety of other names are in use, among them ’splitting’ and ’slabbing’. They often take place at
depths in excess of 1,000 m below surface, but can also be induced at shallow depth where high
horizontal stresses are acting.

Selmer-Olsen (1964) and Muir Wood (1979) mention the importance of differences in magnitudes
of the horizontal and vertical stresses. Selmer-Olsen (1964, 1988) has experienced that in the hard
rocks in Scandinavia stresses might cause spalling in tunnels located inside valley sides steeper than
20o and where the top of the valley sides are higher than 400 m above the level of the tunnel. The
main reason for this is explained by the very great anisotropy between the maximum and minimum
principal stresses (σ1/σ2 >10), as described later in Section 6.5.

σθ σθ

Incipient
Extension Crack
b) Extension
a) Shear Failure
Strain Slabbing

Fig. 6-7 Rock burst in the form of shear failure and ’extension-strain slabbing’ in massive rock (from Deere et al.,
1969).

The failure illustrated in Fig. 6-7 may consist of small rock fragments or slabs of many cubic
metres. The latter may involve the movement of the whole roof, floor or both walls. These failures
do not involve progressive failures, except for heavy rock burst. However, they often cause
significant problems and reduced safety for the tunnel crew during excavation.

$,QVWDELOLW\DVVHVVPHQWV

Hoek and Brown (1980) have made studies of the stability of square tunnels in various types of
massive quartzite in South Africa. In this region where k = ph /pz = 0.5, the maximum tangential
stress in the walls is σθ ≈ 1.4 pz (see Appendix 9) where the main stability problems occurred.
Thus, the rock burst activity can be classified as:
σc /σθ >7 Stable
σc /σθ = 3.5 Minor (sidewall) spalling
σc /σθ =2 Severe spalling
σc /σθ = 1.7 Heavy support required
σc /σθ = 1.4 Possible rock burst conditions
σc /σθ < 1.4 Severe (sidewall) rock burst problems.

4
Terzaghi (1946), Proctor (1971) and several other authors use the term ’spalling’ for "drop off of spalls
or slabs of rock from tunnel surface several hours or weeks after blasting".
6 - 20

Similarly, Russenes (1974) has shown the relations between rock burst activity, tangential stress on
in the tunnel surface and the point load strength of the rock (Fig. 6-8).
POINT LOAD STRENGTH, Is (MPa)

12

y
it
tiv
ac
t
rs
bu
ck

w
ro

8 Lo e
at
o

er
N

d
Mo y
ivit
act
rst
k bu
4
h r oc
Hig

20 40 60 80 100
TANGENTIAL STRESS, σt (MPa)

Fig. 6-8 The level of rock burst related to point load strength of the rock and the tangential stress (σt = σθ) in the
tunnel surface calculated from Kirsch’s equations (from Nilsen, 1993, based on data from Russenes,
1974).

The following classification 5 was found for horseshoe shaped tunnels:


σc /σθ >4 No rock spalling activity
σc /σθ =4-3 Low rock spalling activity
σc /σθ = 3 - 1.5 Moderate rock spalling activity
σc /σθ < 1,5 High rock spalling/rock burst activity
As seen, these results fit relatively well with the results of Hoek and Brown.

Later, Grimstad and Barton (1993) made a compilation of rock stress measurements and laboratory
strength tests and arrived at the following relation, which supports the findings of Hoek and Brown
as well as Russenes:
σc /σθ > 100 Low stress, near surface, open joints
σc /σθ = 3 - 100 Medium stress, favourable stress condition
σc /σθ = 2 - 3 High stress, very tight structure. Usually favourable to stability, maybe unfavourable
to wall stability
σc /σθ = 1.5 - 2 Moderate slabbing after > 1 hour
σc /σθ = 1 - 1.5 Slabbing and rockburst after minutes in massive rock
σc /σθ < 1 Heavy rockburst (strain-burst) and immediate dynamic deformations in massive rock

The value for σc referred to above is related to the strength of 50 mm diameter samples. In massive
rock the ’sample’ or block size is significantly larger - in the order several m3. The scale effect for
compressive strength determines the value of RMi = fσ × σc (see eq. (4-7)). For block sizes in the
range of 1 - 15 m3 fσ = 0.45 - 0.55. This means that σc = RMi/fσ ≈ 2 RMi; hence, the values of the
ratio RMi /σθ as shown in Table 6-5 are approximately half of the values for σc /σθ listed above.
The table has been worked out based on this.

The uniaxial compressive strength σc has been calculated from the point load strength (Is) using σc =
5

20 Is.
6 - 21

TABLE 6-5 CHARACTERIZATION OF FAILURE MODES IN BRITTLE, MASSIVE ROCK


&RPSHWHQF\IDFWRU  )$,/85(02'(6
Cg = fσ⋅σc /σθ = RMi /σθ   LQPDVVLYHEULWWOHURFNV

> 2.5 No rock stress induced instability


2.5 - 1 High stress, slightly loosening
1 - 0.5 Light rock burst or spalling
< 0.5 Heavy rock burst

Ideally, The strength of the rock should be measured in the same direction as the tangential stress is
acting. Strength anisotropy in the rock may, however, cause that the values of the competency factor
in Table 6-5 may not always be representative.

High stresses in massive rock cause new cracks that form slabs parallel to the periphery.
Measurements carried out by SINTEF (1990) in the 10 m wide Stetind road tunnel in Norway show
that the maximum stresses occur 5 m radially outwards from the tunnel after relief joints have
developed around the tunnel, see Figs. 6-9 and 6-10.

This is in accordance with the theories of stress redistribution that the stress peak moves inward in
the surrounding rock mass as deformation and cracking take place.

0.5 1.0 1.5 2.0 2.5 3.0


Depth of hole (m)

Observed joint

Measuring point

Fig. 6-9 Registration of relief joints in a core drill hole outward from the upper part of the wall. Most joints have
been developed within 2.5 m from the tunnel surface (from SINTEF, 1990).

20

15
Stress (MPa)

10 σ1

0
σ2
-5
0.5 1.5 2 2.5 3
Depth of hole from the tunnel surface (m)

σ1 = major principal stress


σ2 = minor principal stress

Fig. 6-10 Ground stresses measured in a drill hole from the upper part of the wall. The highest stress was measured
5 m from the tunnel surface (from SINTEF, 1990).
6 - 22

In Scandinavia, tunnels having spalling and rock burst problems are, in most cases, supported by
shotcrete (often fibre reinforced) and rock bolts, as this has in practice been found to be most
appropriate confinement. The general trends in support design are shown in Table 6-6. Earlier, wire
mesh and rock bolts in addition to scaling, were used as reinforcement in this type of ground. This is
only occasionally applied in Norway today.

TABLE 6-6 THE GENERAL AMOUNT OF ROCK SUPPORT IN OVERSTRESSED, BRITTLE ROCKS IN
NORWAY
6WUHVVSUREOHP &KDUDFWHULVWLFEHKDYLRXU 5RFNVXSSRUWPHDVXUHV

High stresses May cause loosening of a few fragments Some scaling and occasional spot bolting

Light rock burst Spalling and falls of thin rock fragments Scaling plus rock bolting

Heavy rock burst Loosening and falls, often as violent Scaling + rock bolt spaced 0.5 - 2 m, plus 50 -100
detachment of fragments and platy blocks mm thick shotcrete, often fibre reinforced

Only two cases have been studied during this work as described in Appendix 7. Their characteristics
are shown in Table 6-7. Fig. 6-11 summarizes the various modes of rock burst and appropriate rock
support.

TABLE 6-7 ROCK MASS AND GROUND CHARACTERISTICS, AND APPLIED ROCK SUPPORT IN TUNNEL
ROOF IN CONTINUOUS GROUND (from descriptions in Appendix 7). THE VALUES HAVE BEEN
PLOTTED IN FIG. 6-11
*URXQGFKDUDFWHULVWLFV $SSOLHGURFNVXSSRUW
3URMHFW Continuity factor Rock mass index Competency factor B( ) = rock bolt (spacing)
&) 50L 50Lσθ F( ) = fibrecrete (thickness)
Stetind road tunnel B(1.5 m)
4.2 41.5 0.2
-chainage 15750 F(50 - 80 mm)

Haukrei headrace t.
3.0 54.8 3.4 no support
-chainage 200

6&$/,1* 6 & $/, 1 *


6+ 27& 5(7(    PP
 

52&. %2/76 6 3 27 % 2/7, 1*
52&. %2/76
VSDFHG    P
VSDFHG    P

B(1.5)+F(50-80) no support

Fig. 6-11 Relationship between the competency factor, failure modes and rock support in continuous, massive and
brittle materials.
6 - 23

%3RVVLEOHPHDVXUHVWRUHGXFHURFNVWUHVVSUREOHPV

There is usually some rock breakage from excavation in drill and blast tunnels which contributes to
form a zone of relaxation around the skin of the opening (Goodman, 1989). Thus, the cracks from
the blasting result in that the stresses redistribute away from the opening. This may explain the
experience gained in Scandinavia that rock burst is less developed in blasted tunnels than in TBM
tunnels. Increased development of joints and cracks from additional blasting in the periphery of the
tunnel is, therefore, sometimes used in Scandinavia to reduce rock burst problems. This experience
indicates that rock with joints or fissures is less subject to rock burst than massive rock under the
same stress level, as is further described in the next section.

The importance of the shape and size of an excavation upon the magnitude of the stresses and on the
stability has been shown by several authors. Through an example Hoek and Brown (1980) show
how the amount of rock support can be greatly reduced by optimizing the shape and layout of a
cavern. Selmer-Olsen (1964, 1988) mentions that in highly anisotropic stress regimes with rock
burst, a method of reducing the extent of rock support is to reduce the radius in the roof where the
largest in situ tangential stress occur. In this way it is possible to reduce the overstressed area where
highest amount of support is required, see Fig. 6-12.

ROCK BOLT ROCK BOLT

SPALLING CRACK

$ %

Fig. 6-12 If high anisotropic stresses occur, the extent of spalling (or rock burst) may be reduced by favourably
shaping the tunnel. ’A’ shows the situation in a tunnel with symmetric shape, and ’B’ the situation in the
tunnel with an asymmetric shape with reduced radius (from Selmer-Olsen, 1988).

6.4.1.5 Squeezing ground

Squeezing can occur both in massive (weak and deformable) rock and in highly jointed rock masses
as a result of overstressing. It is characterized by yielding under the redistributed state of stress after
excavation. The squeezing can be very large; according to Bhawani Singh et al. (1992) deformations
as much as 17% of the tunnel diameter have been measured in India. The squeezing can occur not
only in the roof and walls, but also in the floor of the tunnel.

Squeezing is related to time dependent shearing, i.e. shear creep. A general opinion is that squeezing
is associated with volumetric expansion (dilation),as the radial inward displacement of the tunnel
surface develops. Einstein (1993) writes, however, that squeezing does not necessarily involve
volume increase, and that it often may be associated with swelling.
6 - 24

a b c
complete shear failure, buckling failure, tensile splitting shearing and sliding

Fig. 6-13 Main types of failure modes in squeezing ground (from Aydan et al., 1993).

Aydan et al. (1993) have pointed out three possible developments of squeezing failures in the
ground surrounding an underground opening (Fig. 6-13):
a) Complete shear failure.
This involves the complete process of shearing of the medium in comparison with the rock-
bursting, in which the initiation by shearing process is followed by splitting and sudden
detachment of the surrounding rock as shown in Fig. 6-13 (a). It is observed in continuous ductile
rock masses or in masses with widely spaced discontinuities.
b) Buckling failure.
This type of failure is generally observed in metamorphic rocks (i.e. phyllite, mica-schist) or
thinly bedded ductile sedimentary rocks (i.e. mudstone, shale, siltstone, sandstone, evaporitic
rocks)
c) Shearing and sliding failure.
This is observed in relatively thickly bedded sedimentary rocks and involves sliding along
bedding planes and shearing of intact rock.

The above division has been worked out from 21 tunnels in Japan in which squeezing occurred. The
most common rock types are mudstones, tuffs, shales, and serpentinites. Most rocks have
compressive strength σc < 20 MPa. From their observations Aydan et al. (1993) have pointed out
five main states of straining in the rock masses surrounding the tunnel (Fig. 6-16):
1. Elastic state Rock behaves almost linearly and no cracking is visible.
2. Hardening state Microcracking starts to occur and the orientation of microcracks generally
coincide with the maximum loading direction.
3. Yielding state After exceeding the peak of the stress - strain curve, micro-cracks tend to coalesce
to initiate macro-cracks.
4. Weakening state Initiated macro-cracks grow and align in the most critical orientations.
5. Flowing state Macro-cracks along the most critical orientations completely coalesce and
constitute sliding planes or bands, and fractured material flow along these planes.

Other examples of squeezing behaviour are shown in Figs. 6-14 and 6-15.

Fig. 6-17 shows the experience gained from the practical studies made by Aydan et al. (1993). They
have used the compressive strength of the rock as the parameter for the materials (which have
strengths σc < 20 MPa). No description of the rocks is presented in their paper; it is in the following
assumed that the rocks contain few joints, as the presence of joints is not mentioned.
6 - 25

Shear
crack
weathered Sound
rock rock

β
Boundary
shear 2.5
planes m

Pressure on plates

No pressure on plates
Rock bolt
4m
Original contour

Sheared Shotcrete lining


wedge 4m

Principal β Relative
stresses
displacement

Fig. 6-14 Principle of shear failure in overstressed ground based on ideas from Rabcewicz (revised from
Hagenhofer, 1991)

progressive
sidewall segment heave
fracturing
invert segment

invert overstressing
bedding
planes

Fig. 6-15 Example of overstressing mechanism in the lower sidewall and in invert of a tunnel in Cyprus (from Sharp
et al., 1993)

Applying straight lines instead of the slightly curved ones in Fig. 6-17, the division given in Table
6-8 has been found. In this evaluation the following assumptions have been made:
• k = ph /pz = 1 and pz = γ × z = 0.02 z (in MPa). (Aydan et al. measured γ = 18 - 23MN/m3
)
• Circular tunnels for which the ratio σθ /pz ≈ 2.0 in roof (Hoek and Brown, 1980). The
tangential stresses can be found from the method presented by Hoek and Brown (1980) as
outlined in Appendix 9.
• The expressions above are combined into σc /z = (2 × 0.02)σc /σθ . It is probable that scale
effect of compressive strength has been inluded in Fig. 6-17; therefore σc has been
replaced by RMi, and the values for the ratio RMi /σθ in Table 6-8 have been found. This
table is based on a limited amount of results and should, therefore, be updated when more
data from practical experience in squeezing ground - especially in highly jointed ground -
can be made available.
6 - 26

Bhawani Singh et al. (1992) developed another empirical criterion, based on the Q-system, which
constitutes another possibility for evaluating the competency of rock masses. Incompetency
resulting in squeezing may occur if the height above the excavation is
z > 350 Q1/3 eq. (6-5)

where Q is the rock mass quality in the Q-system.

This expression has several limitations as it is restricted to deformable (ductile) rock masses.
Neither the influence of tectonic or residual stresses, which in many parts of the world results in
considerable horizontal stresses leading to stability problems, is included.
8
NORMALISED STRAIN LEVELS ηp, ηs, ηf

εp
ηp = εe σp
7
εs
ηs = εe
6
εf
ηf = εe
5 5

1 2 3 4 5
4
Eq. (6)
3
4

2 3

2
1
1

0
0 2 4 6 8 10 12 14 16 18 20
UNIAXIAL STRENGTH σc (MPa)

1 2 3 4 5 1. Elastic state
2. Hardening state
3. Yielding state
4. Weakening state
5. Flowing state

Fig. 6-16 Idealized (left) stress-strain curves with corresponding development of squeezing and plots (right) of
normalized strain levels (from Aydan et al., 1993).

0
1 2 3 4 5 6 7 8

σc (MPa)

100
OVERBURDEN H (m)

NS

200
LS

For straight lines between:


FS NS (no squeeze and light squeeze) σc/H = 1/25
300
LS and FS (fair squeeze) σc/H = 1/35
FS and HS (high squeeze) σc/H = 1/50
HS

400

Fig. 6-17 A chart for estimating the possibility for squeezing (after Aydan et al., 1993)
6 - 27

According to Seeber et al. (1978) the rock support in squeezing ground may be as shown in Table 6-
9. These data have been used in Fig. 6-18 where the rock support has been related to the
competency factor.
TABLE 6-8 CLASSIFICATION OF SQUEEZING (based on Aydan et al., 1993)
6TXHH]LQJFODVV
7KHWXQQHOEHKDYLRXUDFFRUGLQJWR$\GDQHWDO 
competency range
1RVTXHH]LQJ The rock behaves elastically and the tunnel will be stable as the face effect ceases.
RMi/σθ > 1

/LJKWVTXHH]LQJ The rock exhibits a strain-hardening behaviour. As a result, the tunnel will be stable
RMi/σθ = 0.7 - 1 and the displacement cease.

)DLUVTXHH]LQJ The rock exhibits a strain-softening behaviour, and the displacement will be larger.
RMi/σθ = 0.5 - 0.7 However, it will cease away from the face effect.

+HDY\VTXHH]LQJ The rock exhibits a strain-softening behaviour at much higher rate. Subsequently,
RMi/σθ = 0.5 - 0.35*) displacement will be larger and will not tend to cease away from effect.

9HU\KHDY\VTXHH]LQJ The rock flows resulting in very large displacements; the medium will collapse if not
RMi/σθ < 0.35*) supported appropriately, and it will then be necessary to re-excavate the opening and
install heavy support.
*)
This value has been roughly estimated

TABLE 6-9 CONVERGENCE AND ROCK SUPPORT IN SQUEEZING GROUND (based on Seeber et al., 1978)
$SSUR[FRQYHUJHQFHDQGURFNVXSSRUWDFFRUGLQJWR6HHEHUHWDO 
1$70 IRUWXQQHOZLWKGLDPHWHUP
(QJOLVKWHUP Without support With support installed
ÖNORM Support
B 2203 (1983) Convergence Convergence Possible rock support
pressure
min. 2 ⋅ 5 cm = 10 cm 2 ⋅ 3 cm = 6 cm 0.2 MPa bolts1) spaced 1.5 m
Stark gebräch Squeezing or ----- ----- ----- - - - - --
oder druckhaft swelling max. 2 ⋅30 cm = 60 cm 2 ⋅ 5 cm = 10 cm 0.7 MPa bolts1) spaced 1.5 m
shotcrete 10 cm
min. 2 ⋅ 40 cm = 80 cm 2 ⋅ 10 cm = 20 cm 0.8 MPa bolts1) spaced 1 m
Heavy shotcrete 10 cm
Stark druckhaft squeezing or ----- ----- ----- ------
swelling max. > 2 m 2 ⋅ 20 cm = 40 cm 1.5 MPa bolts2) spaced 1 m
shotcrete 20 cm
1) 2)
bolt length 3 m bolt length 6 m

6+27&5(7( 6+27&5(7( )RUKLJKO\MRLQWHGURFNPDVVHV


PP PP
  Use support chart
52&.%2/76 52&.%2/76 for discontinuous
VSDFHGP VSDFHGP rock masses

Fig. 6-18 Relationship between the competency factor, failure modes and support (12 m diameter tunnel) in highly
jointed rock masses and in massive, 'ductile' rocks.
6 - 28

$7KHXVHRIDQDO\WLFDOPHWKRGVWRGHWHUPLQHURFNVXSSRUWLQVTXHH]LQJJURXQG

As it is considered theoretically that a plastic zone is formed, elastic-plastic solutions similar to the
ground response interaction analysis may be applicable in calculating the behaviour. There is,
however, a limit at which the problems of rock behaviour and support may be considered in plane
strain in two dimensions (Muir Wood, 1979). The advance of a tunnel develops a complicated
three-dimensional stress pattern in the vicinity of the face. Even for the simple case of a circular
tunnel in ground considered as isotropic and elastic with a hydrostatic stress distribution only
simplified analysis can be used. The designer has the difficult task of determining realistic values of
the strength parameters φ and c of the ground (Deere et al., 1969). By applying the RMi, the values
of P may be easier and better characterized. The actual analyses may involve the ground response
curves as applied in the NATM support system (Seeber et al., 1978), or the Hoek-Brown criterion,
refer to Chapter 8.

Also, for the rock stresses applied in the analysis there are uncertainties connected to their measured
magnitudes and directions. It may be difficult to carry out reliable rock stress measurements in deep
drill holes from the ground surface to the actual location before construction. Therefore, rough
estimates of the stress level as described in Section 6.5 have often been applied, based on the weight
of the overburden.

The stand-up time is a main feature during excavation in incompetent, continuous ground. The close
timing of the excavation and the rock support carried out as initial support plays an important part in
weak ground tunnelling as manifested in the NATM concept.

Another important feature in tunnelling is the influence on the rock load from the arching effect of
the ground surrounding a tunnel. Terzaghi (1946) introduced the term DUFKDFWLRQ for this capacity
of the rock located above the roof of a tunnel to transfer the major part of the total weight of the
overburden onto the rock located on both sides of the tunnel. By allowing the material to yield and
crush to some extent in such incompetent ground while the inward redistribution of stresses takes
place, its potential strength can be mobilized. The high ground stresses close to the tunnel dissipate
as the rock masses dilate or bulk (increases in volume). In this way only a reduced support is needed
to contain the cracked rock surrounding the tunnel. Terzaghi (1946) mentions that because of this
DUFKDFWLRQ in completely crushed but chemically intact rock and even in some sands, the rock load
on the roof support does not exceed a small fraction of the weight of the ground located above the
roof. The utilization of this effect is one of the main principles in the NATM.

There has not been time to work further on squeezing ground to develop a support chart from data
on tunnel, stresses, and rock support for the squeezing ground. From case examples including
characterization of the rock masses combined with analytical and modelling works, a similar chart
as for discontinuous (jointed) rock may prove to be appropriate for this type of ground.
6 - 29

6WDELOLW\DQGURFNVXSSRUWLQGLVFRQWLQXRXV MRLQWHG PDWHULDOV

3DUDGR[LFDOO\WKHH[FDYDWLRQRIDQXQGHUJURXQGRSHQLQJLQDKLJKO\VWUHVVHGHQYLURQ
PHQWLVOLNHO\WREHOHVVKD]DUGRXVZKHQWKHURFNLVMRLQWHGWKDQZKHQLWLVLQWDFW
Nick Barton (1990)

1000
crushed

overstressed
100
2
,
7
G
HW
$ Q
LR
5
 M
( 10
=
, SRRU JRRG
6
massive

1
overstressed

0.01 1 1000
*5281' 48$/,7<

The failures in this group of jointed rocks occur when wedges or blocks, limited by joints, fall or
slide from the roof or sidewalls. They develop as local sliding, rotating, and loosening of blocks and
may occur in excavations in jointed rocks at most depths. The properties of the intact rock are of
relatively little importance as these failures, in general, do not involve development of fracture(s)
through the rock (Hoek, 1981). The strength of the rock influences, however, often the wall strength
of the joint and may in this way contribute to the stability.

Fig. 6-19 The influence from discontinuities on block loosening and overbreak (from Stini, 1950). Upper figure:
Overbreak caused by smooth foliation partings in a quartzphyllite. Lower figure: Layering joints and long
cross joints cause instability in the roof.

The stability in jointed rock masses may be divided between instability of an individual block and
cases in which failure involve two or more blocks. Two examples of overbreak from failure caused
by joints are shown in Fig. 6-19.
6 - 30

The key block method may be used as analysis in this group as it applies knowledge of orientation
and condition of significant, joints and weakness planes in the rock mass; refer to Goodman (1989).
The principles in this analysis and the methods for data collection have been also been described by
Hoek and Brown (1980).

As the condition, orientation, frequency and location of the joints in the rock mass relative to the
tunnel are the main controlling factors, the stability can generally not be predicted by equations
derived from theoretical considerations (Deere et al., 1969). A common solution is to apply charts
or tables in which the amount and types of support are found from combination of several rock mass
and excavation parameters. This principle has been applied among others in the Q and the RMR
systems.

TYPE OF ROCK &20 35(66, 9(


6 7 5 ( 1 * 7+
5 2 & . 0 $ 6 6 , 1 ' ( ;

- 2 , 1 7 & 2 1 ' , 7, 2 1 ) $ & 7 2 5


5 0L
JOINTING PARAMETER
% / 2 & . 9 2 / 8 0 (
D I A ME TER
BLOCK

G R O U N D W AT E R ROCK
S TR E S S
5 2 & . 6 7 5 ( 6 6 ( 6
L E V EL

6 , = ( 2 )

7 + ( 2 3 ( 1 , 1 *

C ON T I N UI T Y OF
ADJUSTMENT FOR THE
TH E G R OU N D WALLS OF THE OPENING
&)
25,(1 7$7,21 2)
7+ ( 23(1 ,1 *

6, =( 2) 7+ (

:($ .1 (66 =21 (

&21',7, 21 2) 52&.


0$ 66(6 $'-$&(1 7
72 7+ ( =21 (

IMPACT FROM
THE EXCAVATION

USE OF
& 2 0 3 ( 7 ( 1 & <
6 ,= ( 5$7 , 2 TH E O P EN I N G
2 ) 7 + ( * 5 2 8 1 '
6U &J
fo r
disc
ontin
uou
s gro
und
input
parameters
main input
parameters
additional input for
weakness zones

not included

Fig. 6-20 The parameters involved in stability and rock support assessment in discontinuous ground.

As shown in Fig. 6-20 the following main parameters influencing stability have been selected to
characterize discontinuous rock masses for design of rock support:
- a factor for the ground conditions, and
- the size features, which is related to the dimensions and orientation of the opening relative to
the blocks and the joints.
6 - 31

6.4.2.1 The ground condition factor (Gc)

The ground conditions consist of the general LQKHUHQWrock mass features of main influence on
stability and the H[WHUQDO stresses acting. The main features are:
- The block volume, Vb, )
(representing the quantity of joints) )
- The condition of joints, jC ) expressed in the rock
(smoothness, waviness, size, etc.) ) mass index (RMi)
- The strength of the joint surface )
(compressive strength of rock) )
- The stress level in the ground

The combination of these parameters is indicated in Fig. 6-20.

The rock mass index (RMi) has been selected to represent the inherent rock mass properties as it
contains all their main internal factors.

$(IIHFWRIVWUHVVOHYHOLQWKHJURXQG

In addition to the inherent properties of the material the stability is influenced by the stresses acting
across the joints in the rock mass surrounding the tunnel. A relatively high stress level will
contribute to a ’tight structure’ with increased shear strength along joints and, hence, increased
stability. This has often been observed in deep tunnels. For the same reason a low stress level is
unfavourable to stability. This effect is frequently seen in portals and tunnels near the surface where
the low stress level often is ’responsible’ for loosening and falls of blocks.

The Q-system uses the impact of stresses in jointed rock in its ’stress reduction factor’ (SRF) as
shown in Table 6-10. From this it is seen that there is a factor of 5 between the most and the least
favourable SRF for jointed rock masses.

In a jointed rock mass containing variable amount of joints with different orientations it is not
possible in a simple way to calculate and incorporate the stresses acting across the joints. Therefore,
a general stress level factor (SL) similar to that in the Q-system has been chosen.

TABLE 6-10 CLASSIFICATION OF STRESS LEVEL (FOR ANY SHAPE OF OPENING) AND SRF VALUES
(from Barton et al., 1974)
------------------------------------------------------------------------------------------------------------------------------------------------
Low stress, near surface σc /σ1 > 200 SRF = 2.5
Medium stress σc /σ1 = 200 - 10 SRF = 1
High stress level, very tight structure σc /σ1 = 10 - 5 SRF = 0.5 - 2
(usually favourable to stability, may be unfavourable to wall stability)
------------------------------------------------------------------------------------------------------------------------------------------------

The stress level referred to here is the WRWDO stresses. The influence of pore pressure or joint water
pressure is generally difficult to incorporate in the stress level. Often, the joints
around the tunnel will drain the ground water in the volumes nearest to the tunnel, hence the
influence from ground water pressure on the effective stresses is limited. The total stresses have,
therefore, been selected. The ratings of SL have roughly been chosen as
SL ≈ 1/SRF. In some cases, however, where unfavourable orientation of joints combined by high
ground water pressure will tend to reduce the stability by extra loading on key blocks, the stress
6 - 32

level factor should be reduced. The reduction of SL given in Table 6-11 for these cases are roughly
assumed.

TABLE 6-11 THE RATINGS OF THE STRESS LEVEL FACTOR (SL)


0D[LPXP $SSUR[LPDWH 6WUHVVOHYHO
7HUP VWUHVV RYHUEXUGHQ IDFWRU
σ (for k ≈ 1) 6/
average
Very low stress level (in portals etc.) < 0.25 MPa < 10 m 0 - 0.25 0.1
Low stresses level 0.25 - 1 MPa 10 - 35 m 0.25 - 0.75 0.5
Moderate stress level 1 - 10 MPa 35 - 350 m 0.75 - 1.25 1.0
High stress level > 10 MPa > 350 m 1.25**) - 2.0 1.5**)
*)
In cases where ground water pressure is of importance for stability, it is suggested to:
- divide SL by 2.5 for moderate influence
- divide SL by 5 for significant influence
**)
For stability of high walls a high stress level may be unfavourable. Possible rating SL = 0.5 - 0.75

There is an obvious greater stability of a vertical wall compared to a horizontal roof. Milne et al.
(1992) have introduced a gravity adjustment factor to compensate for this where the wall is given a
factor of 5 and horizontal backs 1. Similarly, Barton et al. (1975) has applied a wall/roof factor as an
adjustment of the Q-value. This factor depends, however, on the quality of the ground. Its value is 5
for good quality (Q > 10); 2.5 for medium (Q = 0.1 - 10); and 1.0 for poor quality ground (Q < 0.1).

Based on Milne and Potvin (1992) the ground condition factor (Gc) is adjusted by a gravity
adjustment factor
C = 5 - 4 cosβ eq. (6-6)

where β = angle (dip) of the surface from horizontal. (C = 1 for horizontal surfaces,
C = 5 for vertical walls.)

Based on the considerations above the ground condition factor is thus expressed as
Gc = SL × RMi × C eq. (6-7)

%3RVVLEOHLQVWDELOLW\LQGXFHGIURPKLJKJURXQGVWUHVVHV

The experience is, as mentioned earlier in this section, that rock bursting is less developed in jointed
rock than in massive rock under the same stress level. At depths where the stresses developed
around the excavation may exceed the strength of the rock, both stress induced and structurally
controlled failures may occur simultaneously. According to Hoek (1981) one of these two forms,
tends to dominate at a particular site where they both occur.

Terzaghi (1946) describes this type of stress controlled failures in jointed rock as ,IWKHURFN
PDVVHVDURXQGWKHWXQQHOLVLQDVWDWHRILQWHQVHHODVWLFGHIRUPDWLRQWKHFRQQHFWLRQVRULQWHUORFNV
EHWZHHQEORFNVVXFKDV$DQG%LQ)LJDQGWKHLUQHLJKERXUVPD\VXGGHQO\VQDSZKHUHXSRQ
WKHEORFNLVYLROHQWO\WKURZQLQWRWKHWXQQHO,IVXFKDQLQFLGHQWRFFXUVLWLVQHFHVVDU\WRSURYLGHWKH
WXQQHOZLWKWKHVXSSRUWSUHVFULEHGIRUSRSSLQJ

Little information has, however, been found in the literature on this effect. Barton (1990) has
experienced that, if jointing is present in highly stressed rock, extensional strain and shear strain can
be accommodated more readily and are partially dissipated. The result is that stress problems under
6 - 33

high stress levels are less in jointed than in massive rock. This has also been clearly shown in
tunnels where destress blasting is carried out in the tunnel periphery with the purpose to develop
additional cracking and in this way reducing the amount of rock bursting.

Fig. 6-21 Possible instability in jointed rock masses exposed to high rock stress level (from Terzaghi, 1946).

Under high stress level in moderately to slightly jointed rock masses cracks may develop in the
blocks and cause reduced stability from the loosening of fragments. This phenomenon has been
observed in the Thingbæk chalk mine in Denmark described in Appendix 7.

6.4.2.2 The size ratio

The size ratio is meant to represent the geometrical conditions at the actual location. It includes the
dimension of the blocks and the underground opening and is expressed as:
Sr = (Dt/Db) × (Co/Nj) eq. (6-8)

where Dt is the diameter (span or wall height) of the tunnel.


Db is the block diameter represented by the smallest dimension of the block, which often
turns out to be the spacing of the main joint set. Often theHTXLYDOHQWEORFNGLDPHWHU is
applied where joints do not delimit separate blocks (for instance where less than 3
joint sets occur). In these cases Db may be found from the following expression which
involves the block volume (Vb) and the block shape factor (β) as shown in Appendix
3, Section 4: 6
Db = (βo /β) Vb_ = (27/β) 3 Vb eq. (6-8)

Nj is a factor representing the number of joint sets as an adjustment to Db in eq. (24)


where more or less than three joint sets are present. As described by Barton et al.
(1974) the degree of freedom determined by theQXPEHURIMRLQWVHWV significantly
contributes to stability. The value of Nj is found from the expression
Nj = 3/nj eq. (6-9)

where nj = the number of joint sets (nj = 1 for one set; nj = 1.5 for two sets plus random joints; nj = 2
for two sets, etc.)

6
(β0 /β) has been chosen in eq. (6-8) as a simple expression to find the smallest block diameter. It is most appropriate
for β < 150. For higher values of β a dominating joint set will normally be present for which the average joint spacing
(S1) should be applied.
6 - 34

Co is an orientation factor representing the influence from the RULHQWDWLRQ of the joints on
the block diameter encountered in the underground opening. The ratings of Co in
Table 6-12 are based on Bieniawski (1984) and Milne et al. (1992). The strike and dip
are measured relative to the tunnel axis. As the jointing is three-dimensional, the effect
of joint orientation is often a matter of judgement, often the orientation of the main
joint set is has the main influence and is applied to determine Co.

TABLE 6-12 THE ORIENTATION FACTOR FOR JOINTS AND ZONES. THE DIVISION IS BASED ON TABLE
6-3.
IN WALL IN ROOF 5DWLQJRI
o o TERM RULHQWDWLRQIDFWRU
for strike > 30 for strike <30 all strike values
&R
dip < 20o dip < 20o dip > 45o favourable 1
dip = 20 - 45o dip = 20 - 45o dip = 20 - 45o fair 1.5

dip > 45o - dip < 20o unfavourable 2


- dip > 45o - very unfavourable 3

6.4.2.3 Rock support chart for discontinuous materials

The rock support chart shown in Fig. 6-23 is developed from case examples in Table 6-13 and from
experience gained in numerous tunnels excavated in hard rock, mainly in Norway. To simplify and
limit the size of the support diagram Vb = 10-6 m3 (= 1 cm3 ) has been chosen as the minimum
block (or fragment) size. This means that where smaller particles than medium gravel occur, Vb = 1
cm3 or block diameter Db = 0.01 m is used.

Roughly, for
FRPPRQ
KDUGURFNPDVVFRQGLWLRQV, i.e. SL = 1, 3 joint sets (nj = 3) and
3
RMi = 40 Vb (for σc = 160 MPa and jC = 1.75), the following simplified expressions can be
applied:

The ground condition factor:


3
Gc = 40C Vb eq. (6-10)

(C = 1 for horizontal roofs and C = 5 for vertical walls)

The size ratio (for β= 40 and Co =1.5 (fair joint orientation))


Wt Ht
Sr = 3 or Sr = 3 eq. (6-11)
Vb Vb

The various excavation techniques used may disturb and to some degree change the rock mass
conditions. This may increase the tendency of blocks to loosen and fall out of the tunnel walls or
roof. Especially, excavation by blasting tends to develop new cracks around the opening. This will
cause that the size of the original blocks will be reduced, which will cause an increase of the size
ratio (Sr) and a reduction of the ground condition factor (Gc). Knowing or estimating the change in
block size from excavation it is, therefore, easy to calculate the adjusted values for (Sr) and (Gc)
and thus include the impact from excavation in the assessments of rock support.
6 - 35

TABLE 6-13 SUMMARY OF GROUND CHARACTERISTICS AND INSTALLED ROOF SUPPORT IN


DISCONTINUOUS ROCK MASSES FROM DESCRIPTIONS IN APPENDIX 7. THE VALUES FOR
Gc AND Sr HAVE BEEN PLOTTED IN FIG. 6-23.
*URXQGFKDUDFWHULVWLFV $SSOLHGURFNVXSSRUW
Jointing parameter Ground
3URMHFWDQGORFDWLRQ & condition Size ratio B( ) = rock bolt (spacing)
Rock mass index factor F( ) = fibrecrete (thickness)
-3 & 50L *F 6U S( ) = shotcrete (thickness)
Gjövik Olympic mountain hall 0.21 & 17.2 17.2 189 B(2 m) length 5 m
B(5 m) length 12 m
F(100 mm)
Granfoss road tunnel, chainage 400 0.21 & 8.4 6.7 23.8 B(1.5 m)
F(70 mm)

chainage 1875 0.13 & 8.1 8.1 37.8 B(1.5 m)


F(70 mm)

chainage 1320 0.11 & 4.3 4.3 52.4 B(1.5 m)


F(80 mm)

chainage 1420 0.19 & 11.6 11.6 26.7 B(1.5 m)


F(70 mm)

chainage 1700 0.19 & 11.5 11.5 25.2 B(1.5 m)


F(70 mm)
Haukrei headrace tunnel, chainage 200 0.46 & 54.8 54.8 3.0 no support
Horga headrace tunnel, chainage 470 0,13 & 13.5 13.5 15 B(3 m)

chainage 1485 0.27 & 27 27 16.8 B(3 m)


Tromsö road tunnel, 0.4 & 39.5 33 13.5 B(2.5 m)

roundabout 0.4 & 39.5 33 27 B(2 m)


F(50 mm)
Nappstraumen road tunnel 0.35 & 42.4 36.4 15 B(2.5 - 3 m)
Njunis, access tunnel, chainage 6250 0.24 & 48.5 72.7 12 spot bolting
Sumbiar road tunnel, chainage 650 0.24 & 49 49 12.5 spot bolting

chainage 1315 0.21 & 41.9 42 61 S(50 mm)

chainage 2100 0.05 & 10.7 10.7 48.4 S(50 mm)


Thingbæk chalk mine 0.84 & 0.84 0.8 6.7 B(spot)

The support chart in Fig. 6-23 covers the rock support of walls as well as roof in the underground
opening. Examples of calculating the RMi value and the input factors Gc and Sr to the support chart
are shown in Appendix 7.
6 - 36

600

400

1 .5
(B1.5, F100)
200
&R

2
100
:WRU+W

80

3
H
'E

(S50)

discontinuous (jointed) materials


60 (B1.5, F80)

(S50)
40
(B1.5, F70)
(B2, F50)
6L]HUDWLR6U

(B1.5, F70)

(B1.5, F70)
20
(B3)
(B2.5-3)
(B3) (B2.5)
(B spot)
[ [ (B spot)
10
8
[ (B spot)
6

4 1.
5
(B spot)
[
shotcrete
2

fibrecrete and rock bolts


2 rock bolts
S(50) shotcrete 50 mm thick
3

F(100) fibrecrete 100 mm thick

B(2) rock bolts spaced 2 m

1
0.01 0.02 0.04 0.06 0.1 0.2 0.4 0.6 1 2 4 6 10 20 40 60 100 200 400 600 1000

*URXQGFRQGLWLRQIDFWRU*F 6/x50Lx&
C is a factor for other surfaces of the excavation than horizontal roof: C = 5 - 4 cosβ
(β = angle of the surface from horizontal)

Fig. 6-23 Rock support chart for discontinuous (jointed) rock masses.

6WDELOLW\DQGURFNVXSSRUWRIIDXOWVDQGZHDNQHVV]RQHV

Weakness zones consist of rock masses with properties significantly poorer than those of the
surrounding rock masses. Included in the term weakness zones are faults, zones or bands of weak
rocks in strong rocks, etc. as described in Appendix 2. Weakness zones occur both geometrically
and structurally as special features in the ground. The following features in the zones are of main
importance for stability:

1. The JHRPHWU\DQGGLPHQVLRQVof the zone.


The instability and problems in weakness zones will generally increase with the ZLGWKof the
zone. However, this feature should always be assessed in relation to the attitude of the zone and
to the frequency, orientation, and character of adjacent joint sets, the existence of adjacent seams
or faults (if any), and the quality of the adjacent rock mass. Brekke and Howard (1972) note that
several severe slides in tunnels have occurred where each individual seam or fault has been of
small width, but where the interplay between them has led to failure.
6 - 37

The orientation of the zone relative to the tunnel can have a considerable influence on the
stability of the opening. As for joints, the problems in general increase as the strike becomes
more parallel to the opening and when the zone is low-dipping. This comes also from the fact
that for such orientations the zone affects the tunnel over a longer distance.

2. The UHGXFHGVWUHVVHV in the zone compared to the overall ground stresses.


An important effect in weakness zones is the fact that the stresses in and near the zone will be
other than normal. Selmer-Olsen (1988) has experienced that faults and weakness zones may
cause large local variations in the rock stresses. Although the overall stresses in an area may
indicate that a weakness zone should be overstressed and behave as incompetent (squeezing)
ground when encountered in an excavation, this will seldom be the case. The reason is the greater
deformability in the zone and transfer of stresses onto the adjacent rock masses. Failures in
weakness zones will, therefore, seldom be squeezing, but gravity induced. Very wide zones,
however, are expected to have stresses and behaviour equal to those of the surrounding ground.

Also for this case, the quality of the rock masses surrounding the weakness zone may contribute
to the stability of the zone.

3. The DUFKLQJHIIHFWfrom the surrounding rock masses


Terzaghi (1946) explained that the rock load on the roof support, even in sand and in completely
crushed rock, is only a small fraction of the weight of rock located above the tunnel because of
the DUFKDFWLRQ or silo effect. Where the width of the zone is smaller than the tunnel diameter,
additional arch action from the stronger, adjacent rock masses leads to reduced the load exerted
on the rock support compared to that of a rock mass volume with the same composition.

4. Possible occurrence of VZHOOLQJVORXJKLQJRUSHUPHDEOH materials in the zone.


These features are further discussed in Section 6.4.3.3.

The composition of weakness zones and faults can be characterized by RMi and/or by its
parameters. Many weakness zones occur as continuous materials when compared to the the tunnel
size or zone, and may be considered as such in the calculations. Based on the comments above a
similar system as has been presented for discontinuous (jointed) rock masses in Section 6.4.2, has
been found to cover most types of zones. It applies a ground condition factor and a size ratio
adjusted for features of the zone as shown in Fig. 6-24.

As mentioned in the introduction to this section, the interplay between the properties of the zone
and the properties of the adjacent rock masses plays an important role, especially for small and
medium sized zones. The inherent features of both can be characterized by their respective RMi
values. The RMiz for the zone is adjusted for the size (Tz) of the zone and for the quality of
adjacent rock masses expressed by their jointing parameter (RMia ).

6.4.3.1 The ground condition factor for zones

Löset (1990) has developed an expression to characterize weakness zones for application in the Q-
system. The expression includes the size of the zone and combines the quality of the zone with the
quality of the rock masses on both sides of the zone. The 'combined' quality is
6 - 38

log Qm = (Tz × log Qz + log Qa)/(Tz +1) eq. (6-14)

where Tz = the width of the zone in metres, Qz = the quality of the zone, and Qa = the quality of
the adjacent rock masses.

&2 03 5( 6 6 ,9(
7 < 3 ( 2 ) 5 2 & .
6 75 (1 *7 +

- 2 , 1 7 & 2 1 ' , 7 , 2 1 ) $ & 7 2 5

-2 , 17, 1*3$ 5$ 0 ( 7( 5

% / 2 & . 9 2 / 8 0 (

5
(
* 5 2 8 1 ' : $ 7 ( 5
.
5 2 & .

& 7
2 ( 6 7 5 ( 6 6

/ 0 5 2 & . 6 7 5 ( 6 6 ( 6
/( 9( /

% ,$
'
6 , = ( 2 )

7 + ( 2 3 ( 1 , 1 *

& 2 1 7 , 1 8 , 7 < 2 )
$ ' -8 6 70 ( 17 ) 25 7+(
7 + ( * 5 2 8 1 '
:$ / /6 2) 7+( 2 3 ( 1 , 1 *

&)
2 5, ( 17$7, 21 2 )
7+( 2 3 ( 1 , 1 *

6 , = ( 2 ) 7+ (

: ( $ .1( 6 6 =2 1 (

&2 1 ', 7, 21 2 ) 52 &.

0 $ 6 6 ( 6 $ '-$ &( 1 7

72 7+( = 2 1(

, 0 3$ & 7 ) 5 20
7+( ( ; & $9$7, 2 1

6,=( 5$7,2
8 6 ( 2 )

7 + ( 2 3 ( 1 , 1 *
& 2 0 3( 7( 1 & <
2 ) 7 + ( * 5 2 81 '
6U &J

input parameter

main input
parameter
additional input for
weakness zones
not included

Fig. 6-24 The parameters and their combination for assessing the stability and rock support of weakness zones.

The same principles can be applied for rock masses characterized by the RMi

Tz × log RMiz + log RMia


log RMim  eq. (6-15)
Tz + 1

or

Tz ×log RMi z +log RMi a


RMim 10 Tz+1 eq. (6-16)
6 - 39

For Tz = 0 (no weakness zone) eq. (6-16) is RMim = RMia

As an alternative to the complicated eq. (6-16) a simplified expression has been developed
RMi m = (10Tz2 × RMi z + RMi a)/(10Tz2 + 1) eq. (6-17)

The correlation between the two expressions for RMim is shown in Fig. 6-25.

10
R Mi = 5 in adjacent rock

values found from simplified expression

values found from Löset (1990)

0.1

R Mi = 0.02 for zone

0.01

0.001
0.01 0.1 1 10 100 1000
T hicknes s of weaknes s s zone (m)

Fig. 6-25 The variation of RMim with thickness of weakness zone (Tz) using the expression based on Löset (1990)
(eq. (6-11)) and the simplified expression (eq. (6-13)). Input values: RMiz = 0.02 for the zone, and RMia =
5 for the adjacent rock masses.

For larger zones the effect of arching is limited; the ground condition for such zones should
therefore be that of the zone (= RMiz). From eqs. (6-14) to (6-16) and Fig. 6-25 is found that for 20
m thick zones RMim ≈ RMia . The stress reduction in zones may probably take place also in larger
zones than this.

An expression for the ground condition factor has been chosen for weakness zones similar to that
for discontinuous (jointed) rock masses.
Gcz = SL × RMim × C eq. (6-18)

It may be discussed if the stress level factor (SL) has significant influence on stability in weakness
zones since they, as mentioned in the beginning of this section, often exhibit reduced stresses
compared to those in the adjacent rock masses. However, stresses influence on the shear strength
along joints and hence the stability, especially in clay-free crushed zones. Another argument for
including SL is the benefit of simplicity to apply similar expressions for Gc and Gcz .

6.4.3.2 The size ratio for zones

It is earlier mentioned that weakness zones show increased arching effect compared to the overall
rock mass when they have thickness less than approximately the diameter (span) of the tunnel. For
such zones the size ratio (Dt/Db)(Co/Nj) is adjusted for the zone ratio Tz/Dt to form the size ratio
for zones 7

7
This ratio is applied provided Tz/Dbzone < Dt/Dbadjacent
6 - 40

diameter of tunnel size (width) of zone Tz Co


Sr ]  × ×(orientation of zone)  × eq. (6-19)
block size (in zone) diameter of tunnel Db ]RQH Nj

here, Tz = thickness of zones smaller than the diameter (span or height) of the tunnel;
Co = factor for the orientation of the zone as shown in Table 6-12
Dt = the diameter (span or wall height) of the tunnel.
Nj = the rating for the number of joint sets in the zone.

For zones thicker than the tunnel diameter the size ratio described for discontinuous (jointed) rock
masses (eq. (6-9)) should be applied (Srz = Sr = Dt/Db × Co/Nj).

Similarly, as for jointed rock masses, a minimum block size Vb = 1 cm3 or block diameter Db =
0.01 m has been chosen. The support chart for weakness zones is shown in Fig. 6-26. For the few
data collected for this type of ground (Table 6-14) there is relatively good agreement between the
ground characteristics and the applied rock support.

TABLE 6-14 SUMMARY OF GROUND CHARACTERISTICS AND APPLIED ROOF SUPPORT IN WEAKNESS
ZONES (from descriptions in Appendix 7).
*URXQGFKDUDFWHULVWLFV
Jointing parameter Ground $SSOLHGURFNVXSSRUW
3URMHFWDQGORFDWLRQ & condition Size ratio
Rock mass index factor B( ) = rock bolt (spacing)
-3 & 50L *F 6U F( ) = fibrecrete (thickness)
Haukrei headrace tunnel, chainage 110 0.04 & 3.7 4.7 14.4 B(1.5 m)
(3.7) (21.5) F(80 mm)
Vinstra headrace tunnel 0.01 & 0.12 0.1 311 B(1 m)
(0.09) (311) F(200 mm) + ribs
Horga headrace tunnel, chainage 810 0.008 & 0.75 1.6 67.8 F(120 mm)
(0.75) (67.8)
Njunis acces tunnel, chainage 6300 0.026 & 5.3 4.2 22.8 B(1.5 m)
(2.7) (34.2) F(60 mm)
Sumbiar road tunnel, chainage 600 0.05 & 10.6 18.1 4.2 B(1.5 m)
(10.6) (42) straps + wire mesh
The numbers in brackets are corresponding values applying expressions for discontinuous (jointed) ground

6.4.3.3 Problems related to special features in weakness zones

Faults and weakness zones have been further described in Appendix 2 and in Section 6.2.1 of this
chapter. Many faults and weakness zones contain materials quite different from the surrounding
rock. Various geologic processes may have caused alteration of the materials in the zone into clays,
often with swelling properties. The special properties of swelling clays having very low friction and
loss of strength in addition to heavy loads on support structures from swelling, can strongly
influence and often overshadow other properties of the zone. Zones showing moderate and low
swelling properties may behave similarly to moderate and low squeezing ground. In such cases the
rock support may also be similar. The long-time effect of swelling, dissolving, and outwash may
easily be underestimated during the construction period, and the permanent rock support
recommendations taken to ensure long time stability may, therefore, prove inadequate.
6 - 41

600

400
(B1, F200 + ribs)

1.5
200

2
6L]HUDWLR6U  'E &R

100
]
7]

80

3
(F120)

discontinuous (jointed) materials


60
]

40

(B1.5, F60)

20

(B1.5, F80)

10
8

6
(B1.5+ straps)

4 1.
5
2

2 F(100) fibrecrete 100 mm thick


B(2) rock bolts spaced 2 m)
3

1
0.01 0.02 0.04 0.06 0.1 0.2 0.4 0.6 1 2 4 6 10 20 40 60 100 200 400 600 1000

*URXQGFRQGLWLRQIDFWRU*F 6/x50L x& P

2 2
RMim = (10Tz x RMiz + RMia ) / (10Tz +1)
C is a factor for other surfaces of the excavation than horizontal roof: C = 5 - 4 cos β
(β = angle of the surface from horizontal)

Sr = Dt x Co / Db is applied where Srz > Sr and Tz > Dt (i.e. Wt or Ht )

Fig. 6-26 Rock support chart for moderate and small weakness zones, excluding zones with high swelling properties
or high permeability. The diagram is similar to the chart in Fig. 6-23.

High water inflow in underground openings encountered when excavating from the weak central
part of impervious gouge is, as mentioned earlier, one of the most adverse conditions associated
with faults. Such ground conditions can seldom be characterized and assessed by a general system.

With the many varieties in structure and composition of faults and weakness zones (see Appendix
2) it will in many cases be relevant to carry out observations, tests and calculations adapted to each
individual zone, and treating each of them as a special case in the engineering design process.

Brekke and Howard (1972) wrote that it is not uncommon that the construction problems associated
with faults or seams have been described as "unexpected" while the fact has been that one knew of
their existence, but their behaviour was incorrectly assessed prior to or during construction.
Inadequate characterization may be a possible explanation for this.
6 - 42

&RPPHQWVWRWKH50LPHWKRGIRUDVVHVVLQJURFNVXSSRUW

The main features and parameters influencing stability of underground openings were discussed in
Section 6.3. Based on this the parameters and features of importance for stability have been
selected. The two main groups of ground applied in the methods are:
- FRQWLQXRXVJURXQG, which can be either massive rock or heavily jointed (particulate) rock
masses; and
- GLVFRQWLQXRXVJURXQG, composed of jointed rock masses.

6.4.4.1 On the input parameters applied

The behaviour of the materials in the two groups is completely different. Therefore, the two
approaches to assess the rock support are different. Common for both groups is that the com-
position and inherent properties of the structural material (i.e. rock mass) can be characterized by
the rock mass index, RMi. The influence from stresses is, however, different. For continuous
ground the tangential stresses (σθ) set up in the ground surrounding the opening are applied, while
for discontinuous ground a stress level factor has been selected. The magnitude of the stresses in the
ground can be estimated by applying the method outlined in Appendix 8.

The influence/impact from ground water has not yet been outlined. In FRQWLQXRXV ground it can be
included in the effective stresses applied to calculate the tangential stresses set up in the rock masses
surrounding the underground opening. In GLVFRQWLQXRXV ground the direct effect of water is often
minor, hence this feature has not been selected. It is, however, possible to adjust the stress level
factor where water pressure has a marked influence on stability.

Various methods to collect and determine the values of the parameters applied in the RMi have been
described in Appendix 3. Among these, the block volume (Vb) is the most important, as it is also
included in the continuity factor. Great care should be taken when this parameter is determined.

As described in Chapter 5 there is, however, often a problem in characterizing the variations in rock
mass composition. The block size varies within wide limits and the calculations must often be based
on a variation range. Where less than three joint sets occur, it is a general problem in measuring the
block volume. Methods of assessing the equivalent volume have been shown in Chapter 5 based on
the tools presented in Appendix 3. As shown in the latter the (equivalent) block volume can be
found from the volumetric joint count and a block shape factor which is defined from the ratio
between joint spacings. Common values for β are given in Table A3-27a. The block shape can also
be estimated from the longest and shortest dimension of the block using eq. (A3-42)
β = 27+7(a3/a1 - 1)

The compressive strength (σc ) of the rock can, for support assessments of discontinuous (jointed)
rock masses, often be found with sufficient accuracy from simple field tests (Schmidt hammer,
simple hammer test) or it is in many cases sufficient to estimate σc from the name of the rock using
for example Table A3-8.
6 - 43

6.4.4.2 The support charts

The support charts in Figs. 6-11, 6-23 and 6-26 cover most types of rock masses except squeezing.
They are mainly based on Scandinavian practice where shotcrete (wet method, often fibre
reinforced) plays an important part. The charts have been worked out from the author’s own
experience in addition to the 24 cases presented in Appendix 7 from visits to Norwegian and Danish
tunnels. The compressive strength of the rocks varies from 2 MPa to 200 MPa and the jointing
intensity from crushed to massive. In squeezing ground work remains to develop more adequate
support charts. Also for this group of ground the application of RMi in the stability and support
calculations seems very promising.

All support charts presented in the foregoing have been combined in Fig. 6-27, which covers most
types of rock masses for estimating the types and amount of rock support. It is based on the
condition that loosening and falls, which may involve blocks or large fragments, should be avoided.
This also includes appropriate timing of rock support and excavation as is discussed later in this
section.

In this connection it should be pointed out that, as the loosening or failures in jointed rock is mainly
geometrically related, i.e. determined by the size of the blocks and the orientation and location of
each joint, it is impossible to develop a precise support chart. Generally, support charts can only
give the average amount of rock support. They can, therefore, be considered as an expression for the
’statistical average’ of appropriate rock support. Further, a support chart can only give the amount
and methods for support based on the support regulations and experience in the region. In other
regions where other methods and applications have been developed, other support charts can be
worked out based on the current practice.

The required level stability and rock support is determined from the utility of the underground
opening. The Q-system applies the ESR (excavation support ratio) as an adjustment of the span to
include this feature. From the current practice in underground excavations it is, however, difficult to
include the different requirements for stability and rock support in a multiplication factor. For
example, the roof in underground power houses will probably never be left unsupported even for a
Q-value higher than 100; the same practice which is applied in traffic tunnels in Central Europe,
seems to be gradually more common also in Norwegian road tunnels. Also, in large underground
storage caverns in rock the roof is generally shotcreted before benching, because falls of even small
fragments may be harmful in high caverns. Caused by this, a chart is preferably worked out for each
main category of excavation.

6.4.4.3 What is new in the RMi support method compared to existing methods?

The support method developed differs from the existing support classification systems. While these
directly combine all the selected parameters to arrive at a quality or rating for the ground conditions,
the RMi method applies an index to characterize the construction material, i.e. the inherent rock
mass and its properties. This index is applied as input in the ground conditions. The splitting up in
the RMi support method into continuous and discontinuous rock masses and the introduction of the
size ratio (tunnel size/block size) are new features in this method compared to existing methods.
6 - 44

URFNVGXFWLOH
heavy fair light

PDVVLYHURFNV
KLJKO\MRLQWHG
5(,1)25& (' 5(,1)25& (' )250$66,9(52&.6QRVXSSRUW
6+27&5(7( 6+27&5 (7(
   PP    PP

  )25+,*+/<-2,17('52&.0$66(6

52&. %2/76 52&. %2/76 Use support chart for discontinuous materials
VSDFH G    P VSDFHG  P

6+27&5 (7(  P P


6 &$ / ,1 *

PDVVLYH
 12 5 2&.

EULWWOH

URFNV

52&. % 2/76 52&. %2/76 683325 7
6 32 7 % 2 /7 ,1 *
VSDF HG     P VSDFHG  P

600

G *
Q 1
D 1, WH

- R L Q W H G U R F N V D Q G Z H D N Q H V V ] R Q H V
UH

fib rce
WH ,

fo
400 / WF

re m e
UH 

RO 5
(

0.
R
WF 7

re nt
 K

5
in-
R ( * V

CK
K 5

-1
V & ,1 HG
1

.
1 , Q

BO .5
J
2  / VL
& (

LT
200 7 GH
( O

1
SP
5 D
& FL

AC
1 H
2 S
& V
:WRU+W &R

IN
U
1M

G
100

(m
2

)
80
'E

3
60

di scontinuous ground
40
6L]H UDWLR 6U

20
0.5

10
-
1.5

6
1.5

4
2

2
3

1
0.01 0.02 0.04 0.06 0.1 0.2 0.4 0.6 1 2 4 6 10 20 40 60 100 200 400 600 1000

*URXQG FRQGLWLRQ IDFWRU *F  6/ x 50L x &

Fig. 6-27 Combination of the support chart in Figs. 6-11, 6-18, 6-23 and 6-26.
6 - 45

The application of the RMi in rock support involves a systemized collection and application of input
data. RMi makes also use of a clearer definition of the types of ground covered. It probably includes
a wider range of ground than the two existing main support classification systems; the RMR and the
Q-system. A comparison between the parameters selected in the different methods is given in
Chapter 9.

Using the RMi in assessment of rock support may seem complicated at first glance. Possible
beginner problems using the support chart should be relatively quickly overcome. Descriptions and
collection of input data require, however, involvement of experienced engineering geologists, as is
the case for most rock engineering projects.

The structure of RMi and its use in rock support engineering allows for accurate calculations where
high quality data are available. It is, however, also possible in this method to apply simplified
expressions for the ground conditions and size ratio when only rough sup-port estimates are
required. Thus, for
FRPPRQ
KDUGURFNPDVVFRQGLWLRQV, i.e. SL = 1 and RMi = 40 3 Vb (for σc =
150 MPa and jC = 1.75), the simplified expressions in eqs. (6-10) to (6-13) can be applied. As they
only require input from the block volume, alternatively the spacing of the dominating joint set, the
support estimates can quickly be carried out.

6.4.4.4 On the timing of rock support installation

It is common to distinguish between the primary or initial rock support installed to ensure stability
during construction, and the final or permanent rock support usually added to ensure stability for the
lifetime of the structure. The time expired between the two stages can, however, vary considerably
as the latter often is performed after completion of the excavation works. In the following some
comments are given for installation of the initial support.

$,QORZVWDELOLW\JURXQG

The great influence of time in tunnel construction was first clearly formulated by Lauffer (1958).
Since then several papers have been published to further develop the stand-up time concept of
Lauffer. The best known are the works by Bieniawski (1974, 1984, 1989) where the stand-up time is
related to rock mass quality in the RMR system.

Also Fairhurst (1988) writes that time is a variable of potentially major significance in tunnel
excavation. Delay of the installation of a support system can result in increased instability that can
lead to collapse of the excavation. In rocks with very short stand-up time at the face it is always a
problem to design a support system because of:
- the variability and uncertainty of the structural properties (strength and deformability) of the
rock mass being encountered; and
- the uncertainties regarding stresses and loads.

There is neither any time nor accessibility to carry out necessary observations to make the
calculations that are needed to analyse the stability problem. The decisions have often to be made
quickly based on experience and available equipment.

In squeezing ground it is essential that a confining rock support be applied in proper time. This can
increase the stability of the rock behind the excavation surface very effectively so that the rock
remains in position to create an arching effect of the tunnel system (Müller, 1982). In NATM, which
6 - 46

originally was developed for squeezing ground, timing of the rock support installation is one of the
main features. Müller (1982) is of the opinion that WLPLQJ is a factor in tunnelling, that can KDUGO\
EHFRPSXWHGRUHYHQDVVHVVHGE\DURFNPHFKDQLFVVSHFLDOLVWLIKHLVQRWSURYLGHGZLWKGHHS
JHRORJLFDONQRZOHGJHRULIKHGRHVQRWLQWLPDWHO\FROODERUDWHZLWKDQHQJLQHHULQJJHRORJLVW
Experience of the people involved may be the most important contribution in such situations.

According to Müller (1982) there are two main possibilities to solve the problem of unstable
conditions shortly after blasting:
- One is to reduce the rounds - a measure which is very effective.
- The other is to divide the excavation face from fullface to heading and benching, or by
dividing even the heading section in two or three parts.

Also the method(s) of support and the skill of the contractor determine how quickly the support can
be installed after excavation.

%,QKDUGURFNUHJLPHV

From studies of several tunnels during construction in Scandinavian hard rocks Cecil (1971) has
worked out the following time-stability-support classification:
Type 1. Stable at blasting:
- no anticipated falls, no support;
- minor rock falls or overbreak at blasting, support not considered necessary for
prevention of loosening;
- support in anticipation of loosening;
- unsupported, gradual deteroriation and subsequent support.
Type 2. Falls at blasting:
- support in anticipation of progressive loosening.
- no support immediately after blasting, progressive loosening, support applied
to prevent further loosening.
- support shortly after blasting to prevent or stop progressive loosening.
Type 3. Support installation shortly after blasting:
- failure of support, thereafter, additional support.

From these types it is the experience that:


Type 1 requires mainly limited amount of immediate support.
Type 2 requires immediate support to be strengthened by the permanent support.
Type 3 occurs in ground with low stand-up time where quick execution of support and reduced
length of round often are necessary. Later, this support must be strengthened by
permanent support.

The installation of support is often decided by the mining crew and is to a great extent determined
from similar behaviouristic observations as shown here. The experience is further that initial rock
support is often installed without any documentation of the ground condition, either because of the
limited access to the face and the short time available, or because it is selected by the tunnel crew.
Where the surface in the tunnel is covered by shotcrete concrete lining or other materials before a
full description of the rock mass conditions has been made, it is not possible to determine the
appropriate permanent rock support.

The support charts are valid only where the rock support is applied at the right time.
)URP3DOPVWU|P$50L±DURFNPDVVFKDUDFWHUL]DWLRQV\VWHPIRUURFNHQJLQHHULQJSXUSRVHV
3K'WKHVLV2VOR8QLYHUVLW\1RUZD\S

&KDSWHU

50L3$5$0(7(56$33/,(',135(',&7,212)7811(/%25,1*
3(1(75$7,21

1RWKLQJKDVEHHQPRUHGLIILFXOWWKDQHYDOXDWLQJWKHURFNPDVVFKDUDFWHULVWLFVDQGDSSO\LQJ
WKHHYDOXDWLRQWRDIRUPXODSUHGLFWLQJSHQHWUDWLRQUDWH
Richard Robbins (1980)

The first tunnel boring machine (TBM) was probably made in Italy in 1846. It used percussive
drilling for drilling slots in hard rock (Rostami, 1992). In 1851 Charles Wilson invented a boring
machine with disc type cutters. Another machine was built for boring the English channel tunnel
between England and France in 1865.

Hard rock tunnel boring came into use after the World War II when in 1947 Jim Robbins redesigned
one of the coal borers when he was working as a consultant to a coal company. The disc cutters
were first introduced into boring in 1957 on a Robbins TBM. This has allowed the excavation of
harder materials, resulting in a wider use of TBM in underground construction.
The new design high performance (HP) TBMs have provided additional improvements in tunnel
boring technology in geologies consisting of alternating very hard and relatively soft rocks. In this
connection it can be mentioned that from a cutter-ring load of 40 - 50 kN in the beginning of the
70s, the strongest hard rock (HP) machines of today have a maximum load of 320 kN. In the same
period the weekly tunnelling advance has increased from a few tens of metres to several hundred
metres.

Today, tunnel boring is successfully carried out in rocks with uniaxial compressive strengths
exceeding 300 MPa, and with tunnel diameters of 10 m and larger. Technically, TBMs can now be
said to have reached a stage of development where a tunnel can practically be bored in any rock and
ground (Nilsen and Ozdemir, 1993). Still, however, performance prediction is an important part of
any TBM project. This is due to the general need for cost- and schedule-evaluations at the various
planning stages of a tunnel project, as well as to develop the information necessary for a reliable
comparison between alternative tunnel construction methods (TBM versus drill and blast).

As effective TBM boring is achieved with working thrusts above the critical thrust for the rock mass
being bored, the strength of the rock mass will have a marked influence on the boring performance.
The RMi - being a relative indicator of the compressive strength of the rock mass - should therefore
be suitable in assessment of the tunnel boring penetration in hard and moderately hard rock masses.
7-2

)DFWRUVLQIOXHQFLQJWKH7%0SHUIRUPDQFH

Some of the main factors influencing on the TBM performance are listed in Table 7-1.

TABLE 7-1 HARD ROCK MASS AND MACHINE FACTORS INFLUENCING TBM PERFORMANCE (from
Lislerud, 1988)
5RFNPDVVIDFWRUV 0DFKLQHIDFWRUV

- Rock mass jointing (ks) - Thrust per cutter (M)


° type and continuity - Cutter edge bluntness (br)
° frequency - Cutter spacing (A)
° orientation - Cutter diameter (d)
- Rock porosity - Torque capacity and RPM
- Rock drillability (DRI) - The machine’s capacity for handling large chips or blocks
- Rock hardness/abrasiveness - General solidity against blows and vibrations
- Stress in rock - Cutterhead curvature and diameter (D)
- Backup equipment

In this work only the effect of the rock mass factors has been analyzed. These can be divided into:
• Intact rock properties, which for excavation with TBM can be characterized in different ways.
Chen and Vogler (1992) mention the following methods which are mainly used today:
− Strength properties, such as uniaxial compressive strength, tensile strength, point load
strength index.
− Hardness, such as Schmidt hardness, total hardness, Mohr hardness, Shore scleroscope,
NBC cone indenter.
− Energy properties, such as fracture toughness, toughness index, critical energy release rate,
and acoustic emission properties.
− Rock internal texture, such as grain size, grain shape, porosity, cementation and orientation
of grains.
− Empirical parameters, such as drillability index, Goodrich drillability, Morris’ drillability,
specific energy test by instrumented cutting, NTH drillability test, direct cutting testing, etc.
• Jointing properties, which includes the quantity of joints and the joint characteristics. It has long
been known that the frequency and orientation of joints in a rock mass is an important factor in
TBM tunnelling (Graham, 1976). Rostami (1992) mentions, however, that due to the complexity
of jointing, little success in applying this parameter has been achieved to date.

3UHGLFWLRQPRGHOV

In general, methods for TBM performance prediction are based on one or more of the following
main principles:
1. Field mapping and/or -testing
2. Small scale laboratory testing ("index testing")
3. Large scale laboratory testing
4. Empirical methods
5. Theoretical models

Many researchers have independently worked on their own indices and tests to be able to predict the
performance and economic factors associated with boring rock tunnels. Therefore, a wide variety of
7-3

performance prediction methods and principles are used in different countries and by the various
research institutes and TBM manufacturers. Some of the methods are based mainly on one or two
rock parameters (for instance uniaxial compressive strength and a rock abrasion value), while others
are based on a combination of comprehensive laboratory, field- and machine-data.

The effect on the boring rate from jointing is a factor, which has been pursued for many years. It has
always been recognized that the presence of joints improves the boring rate. However, LQWKH
LQWHUHVWRIFRQVHUYDWLVPLQPRVWDQDO\VHVWKHLPSURYHPHQWLQERULQJUDWHGXHWRMRLQWLQJKDVEHHQ
QHJOHFWHGE\WHVWLQJXQIUDFWXUHGVSHFLPHQVRIVROLGURFNDQGEDVLQJSUHGLFWLRQVRQWKHVWUHQJWK
FKDUDFWHULVWLFVRILQWDFWURFN (Robbins, 1980). This has probably caused some of the problems in
comparing the various models, which have been mentioned in published papers.

Of the many models presented, the NTH model (Norwegian Institute of Technology, 1994) is
considered to be the closest relation to the RMi system and its parameters. The NTH performance
prediction model is a combination of the main principles nos. 1, 2, and 4 shown above. A short
description of this method is given in the following.

7KH17+SURJQRVLVPRGHO

The main advantages of the NTH model for TBM performance prediction are the generally very
comprehensive empirical data-base, where the important influence of rock jointing can be easily
taken into account (Nilsen and Ozdemir, 1993).

The model is based primarily on empirical correlations between geological/rock mechanical


parameters and actual tunnelling performance. Time and cost curves for the various tunnelling
operations have been established by collecting and analysing a large amount of data on tunnelling
performance and rock mass properties from tunnelling in Europe. The model has been continuously
revised and improved as new tunnelling data and TBM modifications become available. Today’s
model, version no. 5, is based on data from about 230 km of bored tunnels.

Geological field mapping, rock sampling and rock testing form the basis for the performance
prediction. It does not deal directly with cutting force requirements; but rather uses data on rock and
machine specifications to provide an estimate of machine performance. The model uses the
following information as input:
a. Rock parameters, including jointing, drillability index, and abrasiveness. (Abrasiveness is
used in the bit wear evaluation.)
b. Machine parameters, consisting of cutter shape and size, cutterhead RPM, cutterhead
curvature, number of discs on the cutterhead, and the applied thrust and power on the
machine.

The following tests are carried out to find the so-called ’drilling rate index ’ (DRI) (refer to Movinkel
(1986) and Lislerud (1988)):
- Brittleness test - a rock aggregate impact test.
- Siever’s J-value test, a miniature drill test expressing the hardness of the rock surface.
With these input parameters and a parameter for the jointing, the model then produces an estimate
of machine advance using empirically developed relationships.

7KHXVHRI50LSDUDPHWHUVWRFKDUDFWHUL]HURFNPDVVHVIRU7%0
7-4

According to the NTH model the penetration rate can be estimated by combining the rock material’s
drilling properties with the jointing of the rock mass and the representative machine factors. The
system for applying the RMi to evaluate the TBM boring capacity is shown in Fig. 7-1. Separate
parameters have been chosen for:
- The rock material, represented by its compressive strength, σc.
- The jointing, represented by the jointing parameter (JP)
- The tunnel /shaft boring machine properties (K), represented by the utilized thrust per disc
cutter, and the size of the cutters.

rock mass factors TBM machine factors


J O I N T I N G

-2, 1 7 &2 1', 7, 21


$ 33/ ,('7+ 5 8 67
)$ &725 M& -2 ,1 7, 1 *
3(5' ,6&
3$ 5 $ 0(7 (5
%/ 2&. 92/ 80( 9E
-3 0%

&255(&7,21)25
7%0- 2,1 7,1 *
&877(56,=( N G 
)$&725

NV &255(&7,21)25

&877(563$&,1* N D
MATERIAL

7< 3( 2) 52&. (IDFWRU


ROCK

(4 8 ,9$ / (1 77 % 0 (4 8 ,9$ / (1 77+5 8 6 7

-2 ,1 7, 1 *)$ & 725 3( 5& 8 77 (5

NHT 0HT

input parameter

Fig. 7-1 Layout of a method to predict TBM penetration using RMi parameters based on the NTH model.

7KHURFNPDWHULDOSURSHUWLHV

As effective TBM boring is achieved with working thrusts above the critical thrust for the rock
being bored, the strength of the intact rock is considered as a main parameter influencing the boring
performance. The uniaxial compressive strength test is, therefore, the most widely used (and
perhaps misused) rock property to determine the drillability of a rock masses. Fig. 7- 2 shows the
effect of σc in some of the published models.

For rocks with uniaxial compressive strength between 140 MPa and 200 MPa Graham (1976) found
that the rate of penetration can be roughly estimated as:
p = 3.94 T/σc eq. (7-1)

where p is penetration per revolution (mm)


T is thrust per cutter (kN)
σc is compressive strength of intact rock (MPa)
This relationship is approximate for a standard disc cutterhead and will vary with the design and
type of cutterhead to be used.
7-5

1.00 Penetration Per Pass


.90
Penetration Per Pass (inches)

A HOWARD HANDEWITH
.80 B COLORADO SCHOOL OF MINES
C D.B. SUGDEN
.70

.60

.50
-0.55
.40 A B C curve A: penetration = 75 σc
-1.24
curve B: penetration = 2000 σc
.30

.20

.10

0.00
10000

30000
20000

35000

50000
5000

15000

25000

40000

45000
0

Rock Uniaxial Compressive Strength (psi)

Fig. 7-2 The penetration per pass of a TBM equipped with 15½" disc cutters and 164 kN thrust/cutter, found in
some prediction models (from Robbins, 1980).

The compressive strength, together with other physical properties of the rock are used by the
Robbins Company (now a subsidiary of Atlas Copco) for estimating the penetration rates of tunnel
and raise boring machines built by the Robbins Company. This is done using computer models of
performance developed within the Robbins Company. The models are theoretically derived and
have been extensively checked against data from both laboratory disc cutting tests and field
performance measurements.

quartzite
Uniaxial compressive strength (σc )

MPa
300 -0.6 300 -0.6
DRI = E x σ DRI = E x σ
c c

200 200
medium to fine
grained granite
sandstone coarse grained granite

mica gneiss
100 100 mica schist

siltstone
phyllite shale

0 0
0 20 40 60 80 100 0 20 40 60 80 100
Drilling rate index (DRI) Drilling rate index (DRI)

Fig. 7-3 Correlation between drilling rate index DRI and the compressive strength of the rock (from Movinkel and
Johannessen, 1986).

Though Hustrulid (1971), Graham (1976), and other authors have found some correlation between
the rock compressive strength and the cutting performance, Rostami (1992) is of the opinion that, in
general, the compressive strength is not a good indicator of boreability. This might be due to the fact
mentioned earlier that many of the prediction models do not include the effect of jointing in their
models.
7-6

As mentioned earlier, the NTH model uses the drilling rate index (DRI) to represent the properties
of intact rock. The correlation between the DRI and the compressive strength of the rock material is
shown in Fig. 7-3. The three dotted curves here can be expressed as
DRI = E × σc - 0.6 eq. (7-2)

where E is a factor representing various groups of rocks. It has the following values:
E = 1000 for most non-schistose, hard rocks (compressive strength σc > 40 MPa)
E = 750 for metamorphic schists (σc = 30 - 150 MPa)
E = 500 for argillaceous rocks (σc = 10 - 100 MPa)

7.3.2 The jointing features

Jointing is often the most important parameter regarding the drillability and hence on the advance of
tunnel boring (Norwegian Institute of Technology, 1994). In the NTH model great emphasis is
placed on joint mapping during field investigations. The model applies the following types of
jointing:
- Systematically jointed rock masses:
⋅ parallel-oriented joints 1 (one set),
⋅ parallel-oriented fissures 1 and foliation planes or bedding planes (one set).
⋅ two or more joint sets and/or fissure sets
- Massive rock masses. 1
Thus, by this division some kind of joint openness, roughness and continuity has been included. The
jointing factor (ks) for joints and fissures is shown in Fig. 7-4. Penetration rates are more or less
proportional to the factor (ks), which is adjusted for other values of DRI than 49 as shown in the
upper diagram in Fig. 7- 4 by curve 1 and 2.

As for other types of rock engineering, a well defined description of jointing is important as its
influence is the dominating rock mass property. In the revised model of NTH (Norwegian Institute
of Technology, 1994), the three-dimensional occurrence of jointing is partly included as the value of
(ks) is found from
ks-tot = Σ ksi - (n - 1) 0.36 eq. (7-3)

where ksi is the value of ks for each joint set given in Fig. 7-4.

Although this is a significant improvement from the earlier NTH prognosis models, where only the
spacing of one joint set was included, it seems that this revision does not yet fully include the effect
of the three-dimensional occurrence of joints. As the degree of jointing and the jointing characteris-
tics are not clearly defined by SINTEF, it is hardly possible to correctly convert the NTH
characterization of jointing to three-dimensional block volume measurements.

1
The following definitions are applied in the NTH model:
-RLQWV are defined as pervasive joints, which can be traced around the whole tunnel profile. They can be open
(as in stress relief joints) or clay-coated with weak/smooth minerals (as calcite, chlorite).
)LVVXUHV include discontinuous joints which only partly can be observed around the tunnel profile, in addition
healed joints with low shear strength and foliation or bedding partings (as in mica schist and mica gneiss)
0DVVLYHURFN includes rock masses without joints or fissures, or with healed joints with filling of high shear
strength (for example joints filled with hydrothermal minerals as quartz, epidote, etc.)
7-7

ks > 2.0

1.25 ks > 2.0


,5
'
N 1

0.75

10 30 50 70 90
'5,

IV

fissure class
4

joint class


V
N




U
R
W
NTH Joint/fissure
F 3 b class spacings
D
I III-IV II-III

Q
R
L
W
D
U
0 -
W
H
c
Q 2
0-I 1.6 m
H
S
 d I- 0.8 m
0 III II
% e I 0.4 m
7
II-III I-II I-II 0.3 m
1 f II 0.2 m
II I
II-III 0.15 m
I 0-I
0.36 g III 0.1 m
0 0
IV 0.05 m
0
0 20 40 60 80 100
$QJOHUEHWZHHQWXQQHOD[LVDQGGLVFRQWLQXLW\SODQH

Fig. 7-4 The rating of the jointing factor ks as a function of the spacing of joints and fissures. In the upper figure is
shown the adjustment of ks for other DRI values than 49. (from Norwegian Institute of Technology, 1994)

Eq. (A3-27) [Vb = β × Jv - 3 ] has been applied in the transitions made from the NTH fissure and
joint classes in Fig. 7-4. As the classes here consist of spacings related to one joint set eq. (A3-27)
can be written
Vb = β(1/S) - 3 = β × S 3 eq. (7-4)

where S is the spacing of the joints or fissures in the set.

Many of the tunnels used in the development of the NTH model have mostly one joint set. In such
cases the blocks have flat (tabular) shape; this is especially the case for small joint spacings where β
= 150 - 200 have been used. For large spacings where possibly other joint sets may occur, β = 50
has been applied. This is shown in Table 7-2.

TABLE 7-2. THE BLOCK SHAPE FACTOR USED TO FIND BLOCK VOLUME FROM THE SPACING GIVEN
IN THE NTH DISCONTINUITY CLASS
17+GLVFRQWLQXLW\FODVV , , ,,, ,, ,,,,, ,,, ,,,,9 ,9

Spacing (S) of discontinuities 1.6 0.4 0.3 0.2 0.15 0.1 0.075 0.05
(m)

Equivalent block volume (m3) 200 10 4 1.5 0.6 0.2

Applied block shape factor β 50 150 150 175 175 175 200 200

By applying the ratings for (ks) for the least favourable angle (i.e. α = 0o in Fig. 7-4) in Fig. 7-5 the
following correlation has been found for common characteristics of MRLQWV
7-8

ks = 1.6 co × Vb - 0.33 eq. (7-5)

where Vb is the block volume in m3.


co is a factor representing orientation of the main joint set. The values of co given in
Table 7-3 have been found from Fig. 7-4. The most favourable angles are for joints
intersecting the tunnel at 45 - 70o.

TABLE 7-3 RATINGS OF THE JOINT ORIENTATION FACTOR FOR TBM


DQJOHEHWZHHQWXQQHO
0-15o 15 - 30o 30- 45o 45 - 75o 75 - 90o
D[LVDQGMRLQWVHW
1.5
DYHUDJHYDOXHRIFR 1 1.25 1.5 1.75
(co = 1.75 for Vb < 0.1 m3)

According to NTH the angle between a (horizontal) tunnel axis and joint plane can be found from
δ = arcsin (sin βj sin ( αt - αj)) eq. (7-6)

where αt is the strike of the tunnel,


αj is the strike of the joint, and
βj is the dip of the tunnel.
10

7 fissures
joints
TBM JOINTING FACTOR (kS )

3
a
a Fo
r jo
2 int
s:
ks
b
b =1
Fo .6V
r fiss b -0.33
c ure c
1 s: d
d ks =
0.9
0.7 e V b -0.26 e

0.5
f f
0.36

0.01 0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 m3
BLOCK VOLUME (Vb)

Fig. 7-5 The correlation between the TBM jointing factor ks and the block volume Vb. The ks and the Vb values
for the points a - f have been found from Fig. 7-4 and Table 7-2 respectively.

As described in Chapter 4, Section 4.2 an average joint condition factor jC = 1.75 often represents
common joint characteristics. This gives JP = 0.2 jC × VbD = 0.265 Vb 0.33. Thus eq. (7-5) can be
expressed as
ks = 0.424 co × JP - 1 eq. (7-7)

Similar, for ILVVXUHV in Fig. 7-4 with assumed average joint condition factor jC = 6, the TBM joint-
ing factor is
ks = 0.9 × Vb - 0.26 × co = 0.432 co × JP - 1 eq. (7-8)

which is close to eq. (7-6) considering the degree of accuracy connected to the quality of the input
data.
7-9

The factor ks represents, as mentioned, rocks with drilling rate index DRI = 49. For other values ks
is adjusted by a factor (kDRI) as given in the upper diagram in Fig. 7-4. An average expression for
the two curves, curve 1 and 2, is
kDRI = 0.14 DRI eq. (7-9)

Using eq. (7-7) and (7-8) the ’equivalent TBM jointing factor’ can be expressed as
keq = ks × kDRI = (0.43 co × JP-1 )(0.14 DRI ). As DRI = E × σc - 0.6 (see eq. (7-2)), the
’equivalent TBM jointing factor’ for MRLQWHG rock masses is
0.06 co E
k eq (0.43 co × JP )(0.14(E × σ c ) ) 
-1 -0.6 0.5
eq. (7-10)
JP × σ c0.3

ForPDVVLYH rock masses (Vb > approx. 10 m3) the rock properties, expressed as DRI, have a
relatively stronger influence on the TBM performance. This is expressed in curve 1 in the upper
diagram in Fig. 7-4. By applying eq. (7-2) this curve can mathematically be expressed as
kDRI = 0.06 DRI 0.72 = 0.06 (E × σc - 0.6 )0.72 = 0.06 E 0.72 × σc - 0.43 eq. (7-11)

As ks = 0.36 (see Fig. 7-4) the ’equivalent TBM jointing factor’ for PDVVLYH rock is 2
keq = ks × kDRI = 0.36 (0.06 E 0.72 × σc - 0.43 ) = 0.022 E 0.72 × σc - 0.43 eq. (7-12)

$VVHVVPHQWRIWKHQHWDGYDQFHRIERULQJ

The penetration rate (io) per revolution is found from the equivalent TBM jointing factor (keq) and
the equivalent thrust per cutter (Meq) in Fig. 7-6. This factor is for cutter diameter 483 mm (19
inches) and a mean cutter spacing 70 mm. For diverging cutter dimensions the equivalent thrust is
found from
Meq = MB × kd × ka eq. (7-13)

where MB is the applied thrust per disc (given in kN),


kd is the correction factor for cutter diameter as given in Fig. 7-7, and
ka is the correction factor for cutter spacing as given in Fig. 7-8.

These correction factors can also be found from the expressions


kd = 2.35 - 0.0028 Dc eq. (7-14)
ka = 1.35 - 0.005 Sc eq. (7-15)
(Dc is the cutter diameter, and Sc is the spacing between the cutters.)

The net advance rate (in m/h) is found from


I = io × RPM × 60/1000 eq. (7-16)
where the value of io is found from Fig. 7-6. io can also be calculated using
io = F × keqG (mm/rev.) eq. (7-17)
where F = 0.0015 Meq 1.5
and G = 30 keq × Meq
- 0.5 - 0.8
(for keq < 3.5) 3

2
If JP =1 is applied in eq. (7-9), the following expression is found: keq = 0.022 E × σc - 0.3
Its difference from eq. (7-11) is small.
3
A rough extrapolation of Fig. 7-6 gives io = 0.03 M × keq0.18 for keq ≥ 3.5
7 - 10
14

12
(mm/rev)

utter
00 kN/c
Meq = 3
10


RL




250
V
V 8
D
S

U
H
200
S

Q 6
R
L
W 150
D
U
W
H
Q 4 100
H
3

0
0 1 2 3 4 5 6 7
(TXLYDOHQW7%0MRLQWLQJSDUDPHWHU  NHT
Fig. 7-6 The influence of jointing on the TBM boring (from Norwegian Institute of Technology, 1994).

1.6

1.4

G
N
1.2

1.0

0.9
280 300 340 380 420 460 500 540
&XWWHUGLDPHWHU (mm)

Fig. 7-7 Correction factor kd for the size of the cutters (from Norwegian Institute of Technology, 1994)

1.1

D
N
1.0

0.9

0.85
55 60 65 70 75 80 85 90
$YHUDJHFXWWHUVSDFLQJ PP

Fig. 7-8 Correction factor ka for the mean spacing of the cutters. (from Norwegian Institute of Technology,
1994).
7 - 11

([DPSOH

The following data have been given on rock mass conditions and TBM factors:
--------------------------------------------------------------------------------------------------------------------------------------
5RFNPDVVIDFWRUVPDSSHG (YDOXDWLRQVPDGH

- Rock type: mica schist Assumed compressive strength σc = 80 MPa,


For mica schist the rock factor E = 750 (eq. (7-2)

- Foliation partings spaced: 0.2 m Block volume


Vb = β × Jv-3 = 200 (1/0.2)-3 = 1.6 m3 (eq. (A3-27)

- Joint characteristics: Joint condition factor:


smooth and slightly undulating; jC = jL × jR/jA = 3 ⋅ 2/1 = 6 (eq. (4-8)
fresh walls; discontinuous short joints

- Angle between tunnel and partings: 45o TBM joint orientation factor co ≈ 1.5 (Table 7-3)

7%0PDFKLQHIDFWRUV

- Diameter of TBM: 4.5 m


- Max. gross thrust per cutter: MB = 290 kN Correction factor: ka = 0.975 (Fig. 7-9) or eq. (7-15)
- Cutter spacing: Sc = 75 mm Correction factor: kd = 1 (Fig. 7-10) or eq. (7-14)
- Cutter size: Dc = 483 mm (19 inches)
- Cutterhead RPM: 11.1 rev/min
-------------------------------------------------------------------------------------------------------------------------------------
&DOFXODWLRQV

The jointing parameter in RMi is


JP = 0.2 jC × VbD = 0.55 (eq. (4-4))
where D = 0.26 for jC = 6 (eq. (4-5))

The ’equivalent TBM jointing factor’ for jointed rock masses is


keq = (0.06co E )/(JP × σc 0.3 ) = (0.06 × 1.5 750 ) /(0.55 × 80 0.3 ) = 1.2 (eq. (7-10))

The equivalent thrust per cutter


Meq = Mb × ka × kd = 290 × 1 × 0.975 = 283 kN (eq. (7-13))

Using Meq and keq the penetration is


io = 7.8 mm/rev (Fig. 7-6)
(or io = 7.7 mm/rev when eq. (7-17) is applied )

The net boring rate is then


I = io × RPM × 60/1000 = 7.8 × 11.1 × 60/1000 = 5.2 m/h (eq. (7-16))

As all calculations shown in the example above can be performed using equations given in the
text, a spreadsheet has been developed as shown in Table 7-4. The same input values as used in the
example have been used.
7 - 12

Table 7-4 THE BORING PENETRATION RATE CAN BE CALCULATED APPLYING A SPREADSHEET
USING THE EQUATIONS DEVELOPED

(67,0$7,1*7%03(1(75$7,215$7( (valid for disc cutters)


Tunnel:
,13877%03$5$0(7(56 TBM type: &$/&8/$7,216

TBM diameter = 4,5 m Spacing between cutters Sc = 75 mm 5HIHUHQFH

Number of cutters = RPM = 11,1 Cutter diameter Dc = 483 mm E= 750 eq. (7-2)
fs = 0,760 eq. (4-7)
,13873$5$0(7(56
Location: example js = 1,0 Table 4-3
Rock: mica schist jw = 1,5 Table 4-4
Rock group [1 = non-shistose hard rock, 2 = metamorphic schist, 3 = argillaceous rock] 2 jL = 4,0 Table 4-8
Compressive strength of rock ( MPa ) σc 80 jA = 1,0 Table 4-6

Joint size [very short, short, medium, long] short jC = 6,00 eq. (4-2)
Joint continuity [cont(inuous), discont(inuous)] discont JP = 0,553 eq. (4-4)
Joint surface condition [smooth, slightly rough, rough, very rough] smooth RMi = 44,256 eq. (4-1)
Joint planarity [planar, slightly (undulating), undul(ating), stepped] slightly Db = 0,197 eq. (6-8)
Possible coating on joint wall [none, sand, clay] none co = 1,5 Table 7-3
Possible filling in joint [none, sand, clay, thick clay] none Structure: jointed
Block volume ( m3 ) Vb 1,60 keq massive - eq. (7-12)
Block shape [compact, long, flat, very (flat or long)] very keq jointed 1,2 eq. (7-10)
1)
Orientation of main joint set [fav(ourable); fair; unfav(ourable); very unfav(ourable)] fair Meq = 282 eq. (7-13)

Applied thust per cutter ( kN ) MB 290 kd = 0,998 eq. (7-14)

ka = 0,975 eq. (7-15)


&$/&8/$7,216 F 7,11 eq. (7-17)
Penetration rate per revolution ( mm/rev ) io = 7,70 G 0,45 eq. (7-17)
Penetration rate per hour ( m/h ) I= 5,13

1)
[Orientation: 0-15o = very unfav(ourable); 15-30o = fair; 30-45o and 75 - 90o = fav(ourable); 45-75o = very favourable]

'LVFXVVLRQRIWKH50LPHWKRGIRU7%0SHQHWUDWLRQDVVHVVPHQW

Rock mass conditions and TBM data from two tunnel projects have been compiled in Appendix 8,
and advance rates have been calculated using the method developed in the foregoing. The results
are shown in Fig. 7-9.

Although it can be said that there is a generally fair connection between the calculated and the real
data, there are few calculated advance rates, which are the same as those experienced in the field.
For some locations the calculated results diverge up to approximately 50% from the boring rate
experienced. There may be several reasons for this, the main being:
1. The ’RMi method’ and the combination of data may have limitations,
2. The input data on ground conditions may be inaccurate
3. The registration of measured boring rates and applied thrust may be inaccurate.

Ad. 1.As the RMi method is developed from the NTH model, it has the same structure. Therefore,
they both suffer from possible deficiencies in the selection of parameters and how they are
structured.

There are also uncertainties connected to the development of the RMi method where transition
from one-dimensional spacings applied in the NTH model to three-dimensional block size applied
7 - 13

in RMi. The assumed value of β = 100 - 200 for the block shape factor here may cause additional
inaccuracy in the RMi method. Fig. 7-9 shows, however, that the calculations carried out by the
’RMi method’ generally give more accurate results than the NTH model.

In the NTH model the drilling rate index (DRI) represents the properties of the rock material. The
determination of this feature, which has to be measured in the laboratory, is time-consuming and
costly. Therefore, average values of this parameter have to be applied which do not include
variation in the rock
10
from the NTH model
9
from the RMi model
&DOFXODWHGERULQJDGYDQFH PK

0
0 1 2 3 4 5 6 7 8 9 10
5HDOERULQJDGYDQFH PK

Fig. 7-9 The calculated and real TBM boring advance using the NTH model and the ’RMi method’.

Ad. 2. It is very difficult to characterize and apply the great variations in rock mass properties in a
simple model or method. Often, rapid changes occur in the rock properties as well as in the
jointing features as described in Section 3 in Chapter 3. Hence, simplifications of the real
conditions using average values may introduce errors.

In addition, there may be errors connected to the way descriptions and characterizations are
performed and how they are quantified. The use of block volume as the measure of the
degree of jointing causes a problem where the joints do not delimit defined blocks. This
happens when only one or two joint sets occur, for instance in schistose rocks (such as mica
schist and mica gneiss) without other discontinuities than foliation partings. In such cases an
equivalent block volume has been estimated applying eq. (7-4) with a block shape factor β =
100 - 200. (Using β = 100 instead of β = 200 gives an error in keq of only 14%.) Other
methods to calculate equivalent block volume are described in Appendix 3, Section 3.2.3.

Errors may also be introduced in the laboratory tests. Farmer and Kemeny (1992) write that
apart from a few simple physical property tests, virtually none of the methods used in rock
testing give reliable data. Testing of small samples introduces in addition significant scale
effects. The error from the compressive strength in the TBM advance calculation is reduced
7 - 14

because this rock property is only applied to the power of 0.3 in eq. (7-9) and 0.43 in eq. (7-
11).

Ad. 3 Stang and Aadal (1991) writes that errors may be introduced in the recording of actual
boring advance rate as well as in registration of applied power during boring. Especially the
latter may have important consequences for the calculated boring advance.

The 1994 version of the NTH’s TBM model clearly states that, in addition to the
specifications and construction of the TBM, the jointing generally has the strongest influence
on boring penetration rate. The benefits in applying RMi parameters in assessment of TBM
tunnel boring are mainly connected with how RMi is characterized:
• The RMi characterization of joints and jointing includes their three-dimensional
occurrence. It therefore incorporates the effect of more than one joint set.
• The RMi parameters also include joint characteristics of importance for the shear
strength of the joints, which generally has a marked influence on the TBM boring rate.

The NTH model would be significantly improved by a better joint and jointing charact-
erization. The TBM jointing factor ( ks) may, therefore, be adjusted in the future when better
jointing descriptions are applied. This may cause that eq. (7-9) and eq. (7-11) may be
reworked and changed.

It is generally much easier and less costly to measure the compressive strength or to find the
compressive strength from point load strength than to measure the drilling rate index, DRI.
Another advantage using the compressive strength is that it often forms a part of the required
rock mass description. Thus, this information may be available at an early stage in the
project. As shown in Section 1 in Appendix 3 the compressive strength can be estimated in
several ways. The use of the point load test or the Schmidt hammer may in many cases give
the required accuracy for the rock strength.
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

Chapter 8

POSSIBLE OTHER APPLICATIONS OF THE RMi IN ROCK MECHANICS


AND ROCK ENGINEERING

"The responsibility of the design engineer is not to compute accurately but to judge soundly."
Evert Hoek and Pierre Londe (1974)

The Rock Mass index, RMi, is different from earlier general classifications of rock masses as it is
more numerical. This is a prerequisite for applications in rock mechanics, rock engineering and
design.

RMi can either be applied directly in the engineering as the main input, or only as part of the input
of the ground composition. In other cases it is more appropriate to apply some of the parameters
used in RMi, for example the block volume (Vb), the joint condition factor (jC), or the jointing
parameter (JP).

Rock Mass index (RMi)

Co mmun icati on

Ho ek-B rown
fa ilu re
i nput in ro ck mechan ics
mod e ls and cal cu latio ns
u se d in ro ck en gi ne eri ng

F rag men ta tio n c rite rion


c la ss ifi ca tio n sy s te m s

a nd bla s tin g RMR s ys te m


Nu meri cal
inp ut to e xi sti ng

mod e llin g
Stab il ity an d
ro c k s u pp o rt Deformation
Q - s ys te m modulus of
assessments
rock masses

T BM p rog ress Gro un d


NAT M resp on se
e va lu at ion s c urv e s

applications
shown in this work
Fig. 8-1 Various applications of RMi or of the parameters included in the RMi

Some practical applications of RMi are shown in Chapters 6 and 7 for rock support assessment in
underground excavations and TBM boring penetration estimates. This chapter outlines some of the
possibilities in applying RMi in various types of calculations applied in rock mechanics and rock
engineering.
8-2

8.1 APPLYING RMi TO DETERMINE THE CONSTANTS IN THE HOEK-BROWN


FAILURE CRITERION

The Hoek-Brown failure criterion provides engineers and geologists with a means of estimating the
strength of jointed rock masses. After the criterion was presented in 1980, the ratings of its
constants have been adjusted in 1988, 1991 and 1992. A modified failure criterion was published by
Hoek et al. (1992) as is outlined later in this section.

8.1.1 The original Hoek-Brown failure criterion

In its original form the Hoek-Brown criterion is expressed in terms of the major and the minor
principal stresses at failure as
σ1' = σ3' + (m × σc × σ3' + s × σc2)½
B B B B B eq. (8-1)
B B B B PB P P P

where σ1'
B B is the major principal effective stress at failure.
σ3' B B is the minor principal effective stress.
σc B B is the uniaxial compressive strength of the intact rock material from which the rock
mass is composed.
s and m are empirical constants representing inherent properties of jointing conditions and
rock characteristics.

For σ3' = 0, eq. (8-1) expresses the unconfined compressive strength of a rock mass:
B B

σcm = σc × s½ B B B B P P eq. (8-2)

According to Hoek and Brown (1980) the constants m and s depend on the properties of the rock
and the extent to which it has been broken before being subjected to the [failure] stresses. Both
constants are dimensionless. Hoek (1983) explains that they are "very approximately analogous to
the angle of friction, Φi', and the cohesive strength, c', of the conventional Mohr-Coulomb failure
B B

criterion".

To determine m and s Hoek and Brown (1980) adapted the classifications of Bieniawski (1973)
and of Barton et al. (1974). This is shown in Table 8-1. As the structure of RMi is similar to eq. (8-
2) which expresses the uniaxial compressive strength for rock masses, RMi offers a method to
determine the constants m and especially s, as described in the following.

8.1.1.1 The constant s

As described in Section 4.5.2 in Chapter 4, the jointing parameter (JP) is similar to s, though the
understanding is somewhat different regarding the features in a rock mass each of them represents.
From eq. (8-2) and the expression RMi = σc × JP, it is found that B B

JP = s eq. (8-3)

JP can be found directly from the registration of block size (Vb) and joint condition factor (jC),
while s is determined via values found by Q or RMR in Table 8-1. As these classification systems
also include external features such as ground water and stresses, they do not in the best way
characterize the mechanical properties of a rock mass. Another drawback is that they both apply
RQD, which in Appendix 4, Section 6 has been shown to often poorly represent the variation in
jointing.
8-3

TABLE 8-1 THE CONSTANTS s AND m FOR UNDISTURBED AND DISTURBED ROCK MASSES
VARYING WITH THE ROCK TYPE AND THE COMPOSITION OF THE ROCK MASS (from Hoek
and Brown, 1988).
Approximate relationship between rock mass quality and material constants

Disturbed rock mass m and s values undisturbed rock mass m and s

DEVELOPED CRYSTAL CLEAVAGE

DEVELOPED CRYSTAL CLEAVAGE


STRONG CRYSTALS AND POORLY
LITHIFIED ARGILLACEOUS ROCKS

FINE GRAINED POLYMINERALLIC

gneiss, granite, norite, quartz-diorite


mudstone, siltstone, shale and slate
CARBONATE ROCKS WITH WELL

IGNEOUS CRYSTALLINE ROCKS


EMPIRICAL FAILURE CRITERION

METAMORPHIC CRYSTALLINE
POLYMINERALLIC IGNEOUS &
dolomite, limestone and marble

ROCKS – amphibolite, gabbro,


ARENACEOUS ROCKS WITH

andesite, dolerite, diabase and


2

σ'1 = σ '3 + mσcσ '3 + sσc

sandstone and quartzite


(normal to cleavage)

COARSE GRAINED
σ'1 = major principal effective stress
B B

σ'3 = minor principal effective stress


B B

σc = uniaxial compressive strength of intact


B B

rock, and
m and s are empirical constants.

rhyolite
INTACT ROCK SAMPLES
Laboratory size specimens free from m 7.00 10.00 15.00 17.00 25.00
discontinuities s 1.00 1.00 1.00 1.00 1.00
CSIR rating: RMR = 100 m 7.00 10.00 15.00 17.00 25.00
NGI rating: Q = 500 s 1.00 1.00 1.00 1.00 1.00
VERY GOOD QUALITY ROCK MASS
Tightly interlocking undisturbed rock with m 2.40 3.43 5.14 5.82 8.56
unweathered joints at 1 to 3 m s 0.082 0.082 0.082 0.082 0.082
CSIR rating: RMR = 85 m 4.10 5.85 8.78 9.95 14.63
NGI rating: Q = 100 s 0.189 0.189 0.189 0.189 0.189
GOOD QUALITY ROCK MASS
Fresh to slightly weathered rock, slightly m 0.575 0.821 1.231 1.395 2.052
disturbed with joints at 1 to 3 m s 0.00293 0.00293 0.00293 0.00293 0.00293
CSIR rating: RMR = 65 m 2.006 2.865 4.298 4.871 7.163
NGI rating: Q = 10 s 0.0205 0.0205 0.0205 0.0205 0.0205
FAIR QUALITY ROCK MASS
Several sets of moderately weathered m 0.128 0.183 0.275 0.311 0.458
joints spaced at 0.3 to 1 m s 0.00009 0.00009 0.00009 0.00009 0.00009
CSIR rating: RMR = 44 m 0.947 1.353 2.030 2.301 3.383
NGI rating: Q = 1 s 0.00198 0.00198 0.00198 0.00198 0.00198
POOR QUALITY ROCK MASS
Numerous weathered joints at 30-500 m 0.029 0.041 0.061 0.069 0.102
mm, some gouge Clean compacted waste s 0.000003 0.000003 0.000003 0.000003 0.000003
rock
m 0.447 0.639 0.959 1.087 1.598
CSIR rating: RMR = 23
s 0.00019 0.00019 0.00019 0.00019 0.00019
NGI rating: Q = 0.1
VERY POOR QUALITY ROCK MASS
Numerous heavily weathered joints m 0.007 0.010 0.015 0.017 0.025
spaced >50 mm with gouge. Waste rock s 0.0000001 0.0000001 0.0000001 0.0000001 0.0000001
with fines
m 0.219 0.313 0.469 0.532 0.782
CSIR rating: RMR = 3
s 0.00002 0.00002 0.00002 0.00002 0.00002
NGI rating: Q = 0.01

Hoek and Brown worked out their failure criterion mainly from triaxial test data on intact rock
specimens. For jointed rock masses they had very few triaxial test data, in fact only those made on
the Panguna andesite. Therefore, the values of s given by Hoek and Brown for the various jointed
rock masses, are very approximate. The jointing parameter (JP) is based on measured strength in 8
"samples" of rock masses. By applying the defined parameters block volume (Vb) and jointing
parameter (JP) in RMi, the accuracy of the parameter s in Hoek Brown failure criterion can be
considerably improved.
8-4

8.1.1.2 The constant m

In addition to adjustments in the ratings of the constant m, Wood (1991) and Hoek et al. (1992)
have introduced the ratio mb / mi , where mi represents intact rock as given in Table A3-8 (in
B B B B B B

Appendix 3). The constant mb is the same as m in the original criterion. It varies with the jointing
B B

Based on the variation of m in Table 8-1 and of its ratings for disturbed and undisturbed values of
m in Wood (1991), Fig. 8-2 has been worked out.

Fig. 8-2 The variation of mb /mi with the jointing parameter (JP), based on the data in Table 8-1 and Wood
B B B B

(1991).

As shown in Fig. 8-1, the variation of mb can be mathematically expressed as: B B

a) for undisturbed rock masses


mb = mi × JP 0.64
B B B B P P eq. (8-4)
b) for disturbed rock masses
mb = mi × JP 0.857
B B B B P P eq. (8-5)

Applying eqs. (8 - 3) and (8-4) in eq. (8-1), the failure criterion can be written as
σ1' = σ3' + [σc × JP 0.64 (mi × σ3' + σc × JP 1.36 )] ½
B B B B B Beq. (8-6) P P B B B B B B P P P P

where s and m are replaced by JP and mi B B

8.1.2 The modified Hoek-Brown failure criterion

From more than 10 years of experience in using the Hoek-Brown criterion, Hoek et al. (1992) found
a need to modify the criterion to the following form:
σ3 a ’
σ ’1 = σ ’3 + σ c ( mb ) eq. (8-7)
σc

where mb and a are constants which depend on the composition, structure and surface of the
B B

jointed rock mass.

mb is found from the ratio mb /mi in Table 8-3. mb /mi varies between 0.001 in crushed rock
B B B B B B B B B B

masses with highly weathered, very smooth or filled joints to 0.7 in blocky rock masses with rough
8-5

joints. In massive rock mb /mi = 1. The value of a varies between 0.3 and 0.65. It has its highest
B B B B

value for the crushed rock masses with altered, smooth joints and lowest for massive rock masses.

The value of a varies between 0.3 and 0.65. It has its highest value for the crushed rock masses
with altered, smooth joints and lowest for massive rock masses, as shown by Hoek et al. (1992) in
Fig. 8-3.

Unweathered, discontinuous, very tight aperture,


MODIFIED HOEK-BROWN FAILURE CRITERION

Slightly weathered, continuous, tight aperture,

polished / slickensided surfaces, hard infilling


Moderatelyweathered, continuous, extremely

polished / slickensided surfaces, soft infilling


Highly weathered, continuous, very narrow,
rough surface, iron staining to no infilling

narrow, smooth surfaces, hard infilling

Highly weathered, continuous, narrow


σ’3 a
σ’1 = σ’3 + σc ( mb
σc )

very rough surface, no infilling


σ’1 = major principal effective stress at failure

SURFACE CONDITION
σ’3 = minor principal effective stress at failure
σc = uniaxial compressive strength of intact
pieces in the rock mass

VERY GOOD

VERY POOR
mb and a are constants which depend on the
condition of the rock mass

GOOD

POOR
FAIR
STRUCTURE

BLOCKY - well interlocked, undisturbed mb / mi 0.7 0.5 0.3 0.1


rock mass; large to very large block size a 0.3 0.35 0.4 0.45

VERY BLOCKY - interlocked, partially mb / mi 0.3 0.2 0.1 0.04


disturbed rock mass; medium block size a 0.4 0.45 0.5 0.5

BLOCKY / SEAMY - folded and faulted, mb / mi 0.08 0.04 0.01 0.04


many intersecting joints; small blocks a 0.5 0.5 0.55 0.6

CRUSHED - poorly interlocked, highly mb / mi 0.03 0.015 0.003 0.001


broken rock mass; very small blocks a 0.5 0.55 0.6 0.65

Fig. 8-3 Estimation of mb /mi and a based on the degree of jointing (block size) and joint characteristics (from
B B B B

Hoek et al., 1992).

The exponent a may partly be compared with the factor D in the expression for RMi in eq. (4-4)
which varies between 0.2 and 0.6. D has its highest values for smooth, or altered joints large joints,
and lowest values for rough, small joints, see Section 4.2 in Chapter 4. The exponent D does not,
however, include the block volume (Vb) as is the case for a, as Vb has been included directly in
eq. (4-4) to determine the jointing parameter (JP).

8.2 RMi USED TO EVALUATE THE SHEAR STRENGTH OF ROCK MASSES

In the paper by Hoek (1983) the empirical failure criterion has been derived by Dr. J. Bray into the
failure envelope given by
τ = (CotΦi' - CosΦi') (m × σc / 8)
B B B eq.(8-8)
B B B

where τ is the shear stress at failure, and


Φ i'
B B is the instantaneous friction angle.
8-6

The value of the instantaneous friction angle is given by


Φi' = Arctan [4hCos 2 (π/6 + Arcsin h - 3/2) - 1] - 1/2
B B P P P P P P eq. (8-9)

where σ' = the effective stress


h = 1 + 16(m × σ' + s × σc) /3m2σc B B P P B B eq. (8-10)
m = mi × JP 0.64
B (for undisturbed rock masses)
B P P eq. (8-11)
s = JP 2 P P eq. (8-12)

The instantaneous cohesive strength is found as


ci' = τ - σ' TanΦi'
B B B B eq. (8-13)

TABLE 8-2 COMPUTER SPREADSHEAT USED TO CALCULATE THE CONSTANTS s AND m , THE
SHEAR STRESS (τ), THE INSTANTANEOUS FRICTION ANGLE (Φi'), AND THE COHESIVE B B

STRENGTH (ci') FROM INPUT OF RMi PARAMETERS. B B

INPUT DATA example 1 example 2 example 3


ROCK CHARACTERISTICS Type of rock = limestone granite gneiss
Rock compressive strength ( MPa ) σc 50,00 160,0 130,0
H & B's m - factor for intact rock Table A3-8 mi 8,40 32,7 29,2
JOINT CHARACTERISTICS
Joint smoothness factor Table 4-2 js 2,00 3,0 1,0
Joint waviness factor Table 4-3 jw 3,00 1,0 2,0
Joint alteration factor Table 4-5 jA 2,00 2,0 3,0
Joint length and continuity factor Table 4-7 jL 3,00 1,0 2,0
JOINTING DENSITY MEASUREMENTS
Alt. 1: measured joint spacings
Main joint set (min. spacing) (m) S1 0,30
Joint set 2 (m) S2 0,50
Joint set 3 (max. spacing) (m) S3 0,50
Alt. 2: measured block volume ( m3 ) Vb 0,01
Assumed block shape factor (Fig. A3-31) β
Alt. 3: RQD measurement RQD 45
STRESSES
Effective normal stress ( MPa ) σn' 0,10 1,0 10,0

CALCULATIONS
RMi PARAMETERS
Volumetric joint count eq. (A3-22) Jv 7,33
Joint condition factor eq. (4-2) jC 9,00 1,50 1,33
Block volume, eq. (A3-19) or eq. (A3-27) ( m3 ) Vb 0,0750 0,0036 0,0100
Block shape factor eq. (A3-28) β 29,58
Jointing parameter eq. (4-4) JP 0,3235 0,0359 0,0462
Rock Mass index eq. (4-1) RMi 16,18 5,74 6,01
HOEK - BROWN PARAMETERS
s - value (= JP2 ) eq. (8-11) s 0,1047 0,0013 0,0021
m - value eq. (8-10) m 4,08 3,89 4,08
Calculation factor eq. (8-9) h 1,0362 1,0090 1,1012
SHEAR STRENGTH PARAMETERS
Instantaneous friction angle eq. (8-8) ( degree ) φ 56,32 65,61 47,81
Shear stress eq. (8-7) ( MPa ) τ 2,85 3,15 15,58
Instantaneous cohesion eq. (8-12) ( MPa ) c 2,70 0,94 4,55

Though the expression in eq. (8-8) seems complex, it can easily be applied using a spreadsheet on a
desk computer. Table 8-2 shows an example where eqs. (8-8) to (8-13) have been applied in an
Excel spreadsheet.

It should be born in mind that the Hoek-Brown failure criterion is only valid for continuous rock
masses (Hoek and Brown, 1980), i.e. massive rock or highly jointed and crushed rock masses, as is
outlined in Chapter 5, Section 5.1 and in Chapter 6, Section 6.4.
8-7

8.3 RMi USED IN THE INPUT TO GROUND RESPONSE CURVES

"However, in our field, theoretical reasoning alone does not suffice to solve the problems
which we are called upon to tackle. As a matter of fact it can even be misleading unless every
drop of it is diluted by a pint of intelligently digested experience."
Karl Terzaghi (1953)

Ground-response interaction diagrams are well established aids to the understanding of rock mass
behaviour and tunnel support mechanics. They are limited to continuous materials, i.e. massive rock
or highly jointed and crushed (particulate) rock masses (see Chapter 5, Section 5.1). According to
several authors (Rabcewicz, 1964; Ward, 1978; Muir Wood, 1979; Hoek and Brown, 1980; Brown
et al. 1983) they may also be used quantitatively in designing tunnel support. For this use it is
essential to be able, from the field observations and assessment of the stresses and moduli, to
predict the ground response curve for a particular rock mass, stress regime, and tunnel geometry.

Many approaches to the calculation of ground response curves have been reported in the literature.
Most use closed-form solutions to problems involving simple tunnel geometry and hydrostatic in-
situ stresses, but some use numerical methods for more complex excavation geometries and stress
fields. However, with improved knowledge of the engineering behaviour of rock masses and the use
of desk computers it is now possible to incorporate more complex and realistic models of rock mass
behaviour into the solutions.

Two solutions of the ground-support interaction diagrams using a simple axisymmetric tunnel
problem were presented by Brown et al. (1983). Both analyses incorporate the Hoek- Brown failure
criterion for rock masses. The material behaviour applied in the closed-form solution is shown in
Fig. 8-4.

σ3 = constant
σ1 − σ3

po A
ε1e ε1
Support pressure, pi

-
B
ε3
Ground response curve
ε3p Gradient = f

ε1p Support C
reaction
ε1e ε1 line
+
D

0 p
-
δio
v
- Gradient = F
p
v
ε1
+ ε1
p

Fig. 8-4 Left: The material behaviour used by Brown et. al.(1983) in the closed-form solution.
Right: The ground response curve.
8-8

The input data used in the closed-form solution are:


ri the internal tunnel radius B B

σc the compressive strength of intact rock B B

po the in situ hydrostatic rock stress B B

f the gradient of line in the -ε3p , ε1p diagram (Fig. 8-4) B PB P B PB P

Data for original non-disturbed rock mass:


m and s are material constants in the Hoek-Brown failure criterion
E and ν are Young's modulus and Poisson's ratio
Data for broken rock mass in the 'plastic zone':
mr and sr are material constants in the Hoek-Brown failure criterion.
B B B B

The following calculation sequence is given by Brown et al. (1983):


1. M = ½ [(m/4)2 + (mpo/σc) + s] ½ - m/8. P P B B B B P P

2. G = E/[2(1 + ν)].
3. For pi ≥ po – Mσc, deformation around the tunnel is elastic: δi/ri = (po – pi /2G.
B B B B B B B B B B B B B B

4. For pi < po – Mσc, plastic deformation occurs around the tunnel: ue /re = Mσc /2G.
B B B B B B B B B B B B

5. N = 2{[(po – Mσc )/mr σc] + sr /mr2)}½. B B B B B B B B B B B PB P P P

6. re /ri = exp{N – 2[pi/mr σc) + (sr /mr2 )]½}.


B B B B B B B B B B B PB P P P

7. δi/ri = Mσc /[G(f + 1)]{[(f – 1)/2] + (re /ri )f+1}.


B B B B B B B B B B P P

Brown et al. (1983) indicate that where appropriate for a given rock mass, the constant f can, in
place of an experimentally determined or back-calculated value, be calculated from
f=1+F eq. (8-13)

m
where F = eq. (8-14)
σ
2(m re + s)1/2
σc

and σre = po - M × σc B B B B B B eq. (8-15)

s = JP2 can be found from eq. 4-4 or from Fig. 4-4 based on field characterization of the block
P P

size (Vb) and joint condition (jC) as described in Chapter 4, while m can be found from Table A3-
8 and Fig. 8-2.

For the broken, (plastic) zone the corresponding sr and mr values have to be estimated from B B B B

experience. It is known that the rock mass breaks up during the deformation (squeezing) process,
which is gradually reduced towards the boundary between the plastic and elastic zone. Applying the
'common' joint condition (joint condition factor jC = 1.75) for the new breaks, eq. 4-5 (JP = 0.25
Vb1/3) can be applied to find
P P

sr = JP2 = 0.06 Vb2/3 eq. (8-16) B B P P P P

The calculations can be readily carried out using a desk computer. If the actual case is not
axisymmetric, because the tunnel cross section is not circular or the in situ stress field is not
hydrostatic, it will be necessary to use numerical method to calculate the stresses, strains and
displacements in the rock masses surrounding the tunnel.

Another method of finding the ground response curve has been shown by Hoek and Brown (1980),
where also data to determine reaction from the support is given.
8-9

8.4 RMi USED FOR NUMERICAL GROUND CHARACTERIZATION IN THE


NATM

The principles of the new Austrian tunnelling method (NATM) is outlined in Fig. 8-5.

–Key features of the New Austrian Tunnelling Method –


Professor Rabcevicz, of Salzburg, followed by final collapse – unless NATM aims at a temporary semi-rigid
Austria, has been one of the chief de- resisted in time by a lining. lining stressed by a moderate rock load pr B B

velopers of the New Austrian Tun- The tunnel lining must be neither that will be just above the theoretical
nelling Method. Goal of NATM: to too stiff nor too flexible. If stiff, pr B B
minimum value.
provide safe and economic support in will remain unnecessarily large – the (Our thanks to Mr. Herbert
tunnels excavated in materials inca- lining will be uneconomic. With Nussbaum, an engineering consultant and
pable of supporting themselves – e.g. increased ∆r, however, the pressure expert on tunneling from Weat
crushed rock, debris, even soil. Sup- pr decreases. With a lining allowing
B B
Vancouver, B.C., Canada. Based on
port is achieved by mobilizing what- too much yield, pr will be big and the B B
phone interviews with him, we were able
ever humble strength the rock or earth lining uneconomic and unsafe. to write this box and captions for all
possesses. figures – GD).
The New Austrian Tunnelling
Method has several features:
• It relies on strength of surrounding Before excavation After excavation

rock to provide tunnel support. This is


done by inhibiting rock deterioration,
joint opening, and loosening due to pr 0 pr When
excessive rock movements. ∆r pt
surrounding
rock moves into
• It uses protective measures like tunnel cavity,
r
pt 0 -∆
stresses pr at
lining tunnel walls with shotcrete and
r0
r=

cavity-rock
driving anchors into unstable rock. In interface
many cases, a second, inner linings decrease
r0 Cavity Plastic zone Elastic zone dramatically,
not needed – e.g. for water conduits, making possible
short highway tunnels. the use of
R

pr
R tunnel linings
• It involves installation of sophis-
-r
that are much
less thick, much
ticated instrumentation at the time less costly. But
initial shotcrete lining is placed, to since total
weight of
provide info for designing a second mountain is
inner lining. constant,
stresses in rock
• It completely eliminates costly 100%
pr 0 ,
ng e
ni lur must increase
se fai
Too

interior supports for tunnel walls, such o n elsewhere to


Pressure on tunnel lining, (pr )

Lo de
stiff

d carry this
80% su
as heavy steel arches. ff
, sa

sti constant load.


oo
fe

,t Note stresses pt
Tunneling into hard or inferior fe
B B

60%
nsa increase to max
U
rock disturbs the existing equilibrium Au
at a short
str distance from
of forces. A rearrangement of stresses 40%
ian
me cavity.
thod afe
within the rock surrounding the cavity 20%
, safe Uns
Incompetent rock
follows (See Fig.). As time Competent rock

progresses, the freshly excavated 0%


1 2 3

tunnel radius ro decreases to (ro – ∆r).


B B B B
Rock movement, ∆r

If tunneling in competent rock, further


deformation is
Fig. 8-5 The main ideas and principles of NATM (from Rabcewicz, 1975).

Brosch (1986) recommends that "informative geological parameters lending themselves to


quantification be used for describing rock mass in future tunnel projects in Austria. This calls for
characterization based on verifiable parameters to provide numerical geo-data for rock
engineering and design to be used in rock construction".

From this statement it is obvious that RMi offers an excellent opportunity to improve the input
parameters used in design works of NATM projects.
8 - 10

NATM has its own classification, mainly based on the behaviour in the excavated tunnel. The
various classes can also be assessed from field observations of the rock mass composition and
assessment of the rock stresses. There does not seem to exist any numerical system for classifying
the important parameters of the rock mass. The ground is mainly characterized on an individual
basis, based on personal experience (Kleeberger, 1992).

TABLE 8-3 THE CLASSIFICATION OF GROUND BEHAVIOUR APPLIED IN ÖNORM B 2203


----------------------------------------------------------------------------------------------------------------------------------------------
NATM class ROCK MASS BEHAVIOUR
------------------------------ -----------------------------------------------------------------------------------------------------------
1 Stable Elastic behaviour. Small, quick declining deformations. No relief features after scaling.
The rock masses are long-term stable.

2 Slightly ravelling Elastic behaviour, with small deformations which quickly decline. Some few small
structural relief surfaces from gravity occur in the roof.

3 Ravelling Far-reaching elastic behaviour. Small deformations that quickly decrease. Jointing causes
reduced rock mass strength, as well as limited stand-up time and active span*) . This P P

results in relief and loosening along joints and weakness planes, mainly in the roof and
upper part of walls.

4 Strongly ravelling Deep, non-elastic zone of rock mass. The deformations will be small and quickly reduced
when the rock support is quickly installed. Low strength of rock mass results in possible
loosening effects to considerable depth followed by gravity loads. Stand-up time and
active span are small with increasing danger for quick and deep loosing from roof and
working face.

5 Squeezing or "Plastic" zone of considerable size with detrimental structural defects such as joints,
swelling seams, shears. Plastic squeezing as well as rock spalling (rock burst) phenomena.
Moderate, but clear time-dependent squeezing with only slow reduction of deformations
(except for rock burst). The total and rate of displacements of the opening surface is
moderate. The rock support can sometimes be overloaded.

6 Strongly squeezing Development of a deep squeezing zone with severe inwards movement and slow decrease
or swelling of the large deformations. Rock support can often be overloaded.
-----------------------------------------------------------------------------------------------------------------------------------------------
*)
P P Active span is the width of the tunnel or the distance from support to face in case this is less than the width of the tunnel

The NATM uses the Fenner-Pacher diagram, which is similar to the ground reaction curve, for
calculation of the ground behaviour and rock support determination. A comparison between
terms applied in NATM and by Terzaghi is presented in Table 6-1 in Chapter 6.

8.4.1 The use of RMi in NATM classification

Seeber et al. (1978) have made an interesting contribution to quantify the behaviouristic
classification in the NATM by dividing the ground into two main groups:
1. The¨'Gebirgsfestigkeitsklassen' ('rock mass strength classes') based on the shear strength
properties of the rock mass.
2. The 'Gebirgsgüteklassen' ('rock mass quality classes') determined from the 'rock mass
strength classes' and the rock stresses acting. These are the same classes as applied in the
NATM classification in Table 8-3 (see also Table 6-1 in Chapter 6).
The first group can be compared to RMi, but the input parameters are different. Fig 8-6 shows
that it is possible to use the shear strength parameters found in Section 2 to determine these data,
as they consist of rock mechanics data characterized by one of the following parameters:
8 - 11

- friction angle of rock mass (Φ), found from eq. (8-8) using very low normal stress,
- cohesion of rock mass (c), which can be found by applying eq. (8-12), and/or
- modulus of elasticity (E) and modulus of deformation (V).

WORKLINE Eel
ROCK φ cel
2
STRENGTH Vel (N/cm ) cpl
v CLASS
σ vpl 2
(Grad) (N/cm )
σv
10,00 000 55
E 1000
1 10,00 000 50
10
5,00 000 45

800 000 50
2 800
800 000 45
ε 10
400 000 40

V 1 000 000 45
σ 3 500 000 40
500
σv 250 000 35
10

800 000 40
E 400
4 400 000 35
10
200 000 30

600 000 35
5 300
ε 300 000 30
10
150 000 25

450 000 35
σ 6 150 000 30
150
10
75 000 25
V

σv
300 000 30
7 100
100 000 25
10
E 50 000 20

ε 150 000 25
50
8 50 000 20
10
25 000 15

Fig. 8-6 Rock mass strength classes ('Gebirgsfestigkeitsklassen') applied by Seeber et al. (1978)

The value of the shear strength parameters can be determined from the defined parameters in
RMi as shown in Section 8.2. In this way, the NATM classes can be defined and determined also
from numerical rock mass characterizations. NATM may effectively benefit from this
contribution, especially when it is applied in the planning stage of tunnelling projects.

Suggested RMi parameters to characterize the various NATM classes are shown in Table 8-4.
The competency factor is further described in Chapter 6, Section 4.1.

TABLE 8-4 SUGGESTED NUMERICAL DIVISION OF GROUND ACCORDING TO NATM CLASSIFICATION


NATM class Rock mass properties Competency factor
( JP = jointing parameter) ( Cg = RMi/σθ ) B B

1 Stable Massive ground (JP > approx. 0.5) Cg > 2


2 Slightly ravelling 0.2 < JP < 0.6 Cg > 1
3 Ravelling 0.05 < JP < 0.2 Cg > 1
4 Strongly ravelling JP < 0.05 0.7 < Cg < 2
5 Squeezing Continuous ground *) P P 0.35 < Cg < 0.7
6 Strongly squeezing Continuous ground *) P P Cg < 0.35
*)
P Continuous ground is where CF < approx. 5 or CF > approx. 100
P (CF = tunnel diam./block diam.)
8 - 12

8.4.2 RMi used for input to Fenner-Pacher ground response diagrams

The Fenner-Pacher curves are, as mentioned, similar to the ground response curves described in
Section 8.1. These curves can, therefore, be applied also for NATM support evaluations.

The benefit in applying RMi to characterize the ground is that the curves can then be based on
appropriate numerical strength parameters. As RMi can be estimated from simple pre-
investigations, the curves can be worked out at an early phase of the project.

By combining the rock mass strength classes ('Gebirgsfestigkeitsklassen') in Fig. 8-6 with rock
stresses from overburden Seeber et al. (1978) have worked out characteristic ground response
curves for the 8 typical rock behaviour classes in the NATM, as shown for class 3 - 7 in Fig. 8-7.
These curves can be applied for the purpose of dimensioning or controlling rock support. They
enable, theoretically in a simple manner, to assess the effect of bolt length and also to find the
connection between deformation and load on rock support.

Rock mass strength classes This table presents only the qualitative relation between both rock mass classes. It is not suitable for use
with “Kennlinien - Bemessungsverfahren” (characteristics - calculation method), but replaces
Eel cel (in arrow direction) approximately the corresponding localized rock mass quality class.
No. φ
vel 2
(N/cm )
1,000 000 o
1 1,000 000 1000 50
o 500
2 800 000 800 45 0
800 000
300
1,000 000 o 0
3 500 40 200
500 000 0
800 000 o H=
4 400 35 100
400 000 0 m
o
5
600 000 300 30 50
300 000 0
o
450 000 150 30
6 150 000
20
300 000
o 0
7 100 25
100 000
10
0
o
150 000 50 20
50 000 50

150 000 o
8 50 15
50 000

PA = 0
Convergence .1 0.2 cm 1.0 2 3 5 10 20 50 100 200 300 1000 2000 cm
5 Squeezing 6 Strongly squeezing
Rock mass quality classes 3 Ravelling 4 Strongly ravelling 7 Flowing
or swelling or swelling
Classes 1 and 2 are to the left outside the diagram

Fig. 8-7 The characterization of the ground into NATM classes as applied by Seeber et al. (1978). Classes 1 and 2
are not covered.

The practical use of the 'standard characteristic (ground response) curves' is shown in Figs. 8-8
and 8-9. Both figures are for the same type of NATM class 5 ('Gebirgsfestigkeitsklasse' 8, φ = 15
- 25o, and overburden 400 m), i.e. 'sehr grebräch' or 'squeezing or swelling'. Lines for bolt
P P

lengths and concrete lining are shown in both figures. Fig. 8-9 shows how the curves in Fig. 8-8
can be used to determine the support pressure and the corresponding displacement, which
depends on the type of rock support.

The characteristic ground response curves are for circular tunnels with 6 m radius. As the
displacements are approximately proportional to the excavation radius (Seeber et al., 1978), they
can easily be estimated also for other tunnel sizes. Fig. 8-10 shows the displacements in circular
tunnels of various sizes located in the same NATM class.
8 - 13

2 3 2
Radius of excavation = 6 m γr = 28 kN/m Eel = 150 x 10 N/cm
3 2
Overburden = 400 m ν = 0.3 Epl = 150 x 10 N/cm
2 3 2
Displacement cel = 50 N/cm Vel = 50 x 10 N/cm
2 3 2
Plastic radius cpl = 10 N/cm Vpl = 25 x 10 N/cm

250 φ = 25

φ = 25
φ= o
φ=2

φ=

φ=
15

15
20
o

o
0
o

o
12 3
6

200

6
12

150 bo
lt l
Support pressure pA (N/cm2 )

en
gt
h 3m
bo
lt l
e
ng
th
bo 6m
lt
l en
g
bo

th 3
12
l

en
tl

gt m
h
100 24
m

12
6

12

50

0
0 10 20 30 40 50 60 70 80 90 100
Displacement U (cm)

0 5 RA 10 15 20 25 30
Plastic radius Ro (m)
RA = excavated radius

Fig. 8-8 One example of the 96 standard ground response curves worked out by Seeber et al. (1978) for circular
tunnels with 12 m diameter.
8 - 14

2 3 2
Radius of excavation = 6 m γr = 28 kN/m Eel = 150 x 10 N/cm
3 2
Overburden = 400 m ν = 0.3 Epl = 150 x 10 N/cm
2 3 2
Displacement cel = 50 N/cm Vel = 50 x 10 N/cm
2 3 2
Plastic radius cpl = 10 N/cm Vpl = 25 x 10 N/cm

250
INFLUENCE FROM BOLT LENGTH

φ=2

φ=2
0

0
o
o
12 3
6 o
Ro for φ = 20

200

6
12

R3 = +9 m
151 for L = 3 m
150 bo
RA = 6 m 3m lt l
Support pressure pA (N/cm2 )

en
gth
3m
bo
lt l
e
ng
th
6m
R5 = + 10 m
111 for L = 5 m
bo

bo 3
6m
lt

en lt le
l

100 for concrete lining g th ng


100 24 th
12
m m
R12 = + 8 m
86 for L = 12 m 12
6
12 m R24 = -1 m
70 for L = 24 m
12
24 m

RA = 6 m
59 cm for concrete lining

50
48 cm for L = 12 m
36 cm for L = 24 m

107 cm for L = 3 m
68 cm for L = 6 m

0
0 10 20 30 40 50 60 70 80 90 100
Displacement U (cm)

0 5 RA 10 15 20 25 30
Plastic radius Ro (m)

Bolt length and required support pressure Displacement and bolt lengths required for
2
for constant displacement U = 59 cm constant support pressure pA = 100 N/cm
Rock Rock bolts Concrete Rock Rock bolts Concrete
support LAN = 3 m 6 m 12 m 24 m lining support LAN = 3 m 6 m 12 m 24 m lining
2 2
pA (N/cm ) 151 111 86 70 100 pA (N/cm ) 107 68 48 36 59

LAN = length of rock bolts pA = support pressure U = displacement

Fig. 8-9 The influence of bolt length on the support pressure and displacement in the tunnel. The response curve
is the same as shown in Fig. 8-8 (revised from Seeber et al., 1978).
8 - 15

2
pA (N/cm )

250 H = 500 m
2
Eel = EPL = 600 000 N/cm
2
Vel = 300 000 N/cm
2
200 Vpl = 150 000 N/cm
2
Cel = 300 N/cm
2
Cpl = 10 N/cm
3
150 γ = 28 N/cm
RA = 8.0 m ν = 0.25
φEL = 30o
RA = 6.0 m φPL = 35o
100
RA = 4.0 m
RA = 2.0 m
50

0 u
0 2 4 6 8 10 12 14 16 18 20 22 24 (cm)

U = displacement pA = support pressure RA = excavation radius


B B B B

Fig. 8-10 The displacements varying with the size of the tunnel within the same ground class (from
Seeber et al., 1978).

It is obvious that the accuracy of the procedure depends in particular on the accuracy of the input
parameters. As they, according to Seeber et al. (1978), generally present a scatter of approx.
100%, a computation, which bases itself on these data cannot possibly results in a better
accuracy. If, however, convergence measurements are available at a somewhat later date, the
results can then be used to improve the accuracy of the input parameters considerably.

8.5 THE USE OF RMi PARAMETERS IN CLASSIFICATION SYSTEMS

"A fundamental requirement for any classification is the need for established criteria in order
to arrange the rock being classified systematically into significant groups and categories."
Williamson and Kuhn (1988)

RMi is not directly applicable in the main classification systems, as they often are completed
systems of 'their own'. Some of the parameters involved in RMi may, however, be used, which
can be of interest where they are considered more accurate or if they are easier measured.

The existing two main classification systems are the RMR (or Geomechanics) system developed
by Bieniawski (1973) and the NGI Q-system by Barton et al. (1974). The systems partly apply
different parameters in different modes; consequently, the established mathematical connections
between them are generally empirical and approximate.
8 - 16

TABLE 8-5 THE RMR CLASSIFICATION SYSTEM OF ROCK MASSES. THE RATINGS FOR EACH
PARAMETER ARE SUMMED UP TO ARRIVE AT THE RMR VALUE FOR THE ACTUAL ROCK
MASS (from Bieniawski, 1984).
A. Classification parameters and their ratings
PARAMETER Range of values // RATINGS
Strength Point-load strength For this low range: uniaxial
> 10 MPa 4 - 10 MPa 2 - 4 MPa 1 - 2 MPa compr. strength is preferred
of intact index
rock Uniaxial com- 5 - 25 1-5 <1
1 material > 250 MPa 100 - 250 MPa 50 - 100 MPa 25 - 50 MPa
pressive strength MPa MPa MPa
RATING 15 12 7 4 2 1 0
Drill core quality RQD 90 - 100% 75 - 90% 50 - 75% 25 - 50% < 25%
2
RATING 20 17 13 8 5
Spacing of discontinuities >2m 0.6 - 2 m 200 - 600 mm 60 - 200 mm < 60 mm
3
RATING 20 15 10 8 5
Length, persistence <1m 1-3m 3 - 10 m 10 - 20 m > 20 m
Rating 6 4 2 1 0
Separation none < 0.1 mm 0.1 - 1 mm 1 - 5 mm > 5 mm
Rating 6 5 4 1 0
Conditio
4 n of
Roughness very rough rough slightly rough smooth slickensided
discon- Rating 6 5 3 1 0
tinuities none Hard filling Soft filling
Infilling (gouge)
- < 5 mm > 5 mm < 5 mm > 5 mm
Rating 6 4 2 2 0
Weathering unweathered slightly w. moderately w. highly w. decomposed
Rating 6 5 3 1 0
Inflow per 10 m
none < 10 litres/min 10 - 25 litres/min 25 - 125 litres/min > 125 litres /min
Ground tunnel length
5 water pw / σ1 0 0 - 0.1 0.1 - 0.2 0.2 - 0.5 > 0.5
General conditions completely dry damp wet dripping flowing
RATING 15 10 7 4 0
pw = joint water pressure; σ1 = major principal stress

B. Rating adjustment for discontinuity orientations


Strike and dip
Very favourable Favourable Fair Unfavourable Very unfavourable
orientation of joints
Tunnels 0 -2 -5 -10 -12
RATINGS Foundations 0 -2 -7 -15 -25
Slopes 0 -5 -25 -50 -60

C. Rock mass classes determined from total ratings


Rating 100 - 81 80 - 61 60 - 41 40 - 21 < 20
Class No. I II III IV V
Description VERY GOOD GOOD FAIR POOR VERY POOR

D. Meaning of rock mass classes


Class No. I II III IV V
10 years for 6 months for 1 week for 10 hours for 30 minutes for
Average stand-up time
15 m span 8 m span 5 m span 2.5 m span 1 m span
Cohesion of the rock mass > 400 kPa 300 - 400 kPa 200 - 300 kPa 100 - 200 kPa < 100 kPa
Friction angle of the rock mass < 45o 35 - 45o 25 - 35o 15 - 25o < 15o

8.5.1 Input to the RMR (Geomechanics) system

As the RMR is based on the sum of several parameters, while RMi and partly also the
parameters involved in are expressed exponentially, it is difficult to directly apply RMi in RMR.
An exception is the compressive strength, σc , which is the same in both systems. Also the joint
B B

condition factor (jC) has similarities with the joint condition applied in RMR.

In the RMR system the jointing is characterized by the RQD and by the spacing of joints. As
shown in Appendix 4, RQD generally is an inaccurate measure of the block size or discontinuity
intensity. Also discontinuity spacing - though Bieniawski (1989) has made attempts to define it
better - is often insufficiently defined (refer to Section 3.7 in Appendix 3). These two parameters
8 - 17

for block size in RMR may be considerably better characterized by the block size (Vb). This
would in addition make RMR simpler to use.

The characterization of the parameter for 'condition of discontinuities' has been significantly
improved in the 1989 version of RMR as can be seen by comparing Tables 8-5 and 8-6. Still,
the joint condition factor (jC) in RMi may be considered as an improvement compared to the
corresponding RMR.

TABLE 8-6 THE IMPROVED DIVISION AND RATING OF DISCONTINUITY CONDITIONS (from Bieniawski,
1989).
Guidelines for Classification of Discontinuity Conditions *)
Parameter Ratings
Discontinuity length <1m 1-3 m 3-10 m 10-20 m >20 m
(persistence/continuity) 6 4 2 1 0
None <0.1 mm 0.1-1.0 mm 1-5 mm >5 mm
Separation (aperture)
6 5 4 1 0
Very rough Rough Slightly rough Smooth Slickensided
Roughness
6 5 3 1 0
Hard filling Soft filling
Infilling (gouge) None <5 mm >5 mm <5 mm >5 mm
6 4 2 2 0
Unweathered Slightly weathered Moderately weathered Highly weathered Decomposed
Weathering
6 5 3 1 0
*) Note: Some conditions are mutually exclusive: For example, if infilling is present, it is irrelevant what the roughness may be, since its
effect will be overshadowed by the influence of the gouge. In such cases, use the main classification table directly.

8.5.2 Input to the Q-system

The Q-system for classification of rock masses is defined as


Q = (RQD/Jn) × (Jr/Ja) × (Jw/SRF) eq. (8-17)

where RQD is the rock quality designation (Deere, 1966)


Jn is the joint set number,
Jr is the joint roughness number,
Ja is the joint alteration number,
Jw is the joint water reduction number, and
SRF is the stress reduction factor.

The Q system and the RMi system have a similar structure and also some parameters are similar.
It is probably the classification system in which the parameters in RMi best can be utilized.

As the Q-system includes external features (stresses and water pressure) acting, only the first
four parameters (RQD/Jn)⋅(Jr/Ja) can be compared with RMi. The Q-system does not directly
include a parameter for the rock material; therefore these four parameters express the jointing in
the rock mass similar to the jointing parameter (JP) in RMi. A comparison has been discussed in
Chapter 9.

The ratio (RQD/Jn) in the Q-system is an expression for the block size (Barton et al., 1974) and
can be compared to the block volume (Vb) in RMi. Appendix 4 concludes, however, that this
ratio very poorly represents the block size. Using the block volume (Vb) instead of (RQD/Jn)
would improve the quality of the Q-system.
8 - 18

The values of the factors for joint roughness (Jr with jR) and joint alteration (Ja with jA) are
similar in both systems. As the RMi also includes a factor for the joint size (jL) in its joint
condition factor (jC), a better characterization of the shear strength of joints may be achieved.
Also here the Q-system may benefit from applying this RMi parameter.

8.5.3 Input to other classification systems

From the simple structure of RMi it is easy to determine the effect of the various parameters, and
consequently how RMi can be developed for other purposes.

In the literature 'new' classification systems are often developed for a new project, based on the
requirement to 'tailor' the classification to the actual rock masses found in the specific area. With
its simple structure, RMi is suited for such developments adapting it to local ground conditions.

8.6 A CONTRIBUTION TO IMPROVED COMMUNICATION

In engineering geology and civil engineering, as in other areas, there is need for clear and
effective communication between individuals involved. The geologist and the engineering
geologist provide the basic data of the ground on which the engineering calculations are based.
Generally, interpretations and correlations between geological and geotechnical data have been
made by individuals, based on their personal experience rather than on any collective basis. For
successful results, close association must exist between geologist and engineer, with full
appreciation and understanding of the parts played by each. "The accuracy of the final answer
can only be as accurate as the geological data at hand." (Piteau, 1970).

Communication problems are compounded by the fact that the engineering geologist is dealing
with a material that is difficult to define due to its complex nature. Williamson and Kuhn (1988)
are of the opinion that "The use of subjective geologic terminology has proven to be less than
helpful in resolving this problem with such terms as 'slightly weathered', 'moderately hard' and
'highly fractured'. These terms do not communicate the true picture even from one geologist to
another, because each has a different perception of the meaning". Thus, there is still a demand
for improving the applied terms in engineering geology and rock mechanics.

8.6.1 Identification chart for geological materials

As a part of the contribution for improved communication a general identification chart for
geological materials has been developed. It is a further development of the chart presented by
Palmström (1986). The chart is in part similar to the 'unified classification chart' developed by
Deere et. al (1969) as shown in Fig. 8-11.

Deere et al. used a combination of the following geological features:


- the particle or block size; and
- the continuity of the geologic materials.
8 - 19

Intact Rock Properties


Shear Strength / Stress Level
Strength / Stress Level
Deformability
Volumetric Stability
Volumetric Stability

30
Spacing, Directions &
i ng

0
D int Shear Strength / Stress Level

0
Character of Dis- Cla
RQ 200 e Jo ock Continuous y Clay Content
continuities 0 id e R 2
W iv coherent 00
v ass Ground Water
M
Level
10 Structural
00

nt i-
co Sem
.
Jo ide
ing

98 uum
W
int

Contin ics

Co
S e r ent
al
n
Strength of 500 mecha

he
n

mi-
450 gy)

io
Rock Mass Rheolo
s

t at
95
uou

n a l and gr av i
Mod. Close

Discontin

300 Rock
Jointing

90 mecha

Silt
Tectonic

n de r
nics
200 02

it h b i
75 150 s
nic
cha

t io
il me
Se ont.

10 0 S o

osi

tw
C

50
mi-

ep
Clo ting

06
Joi

Strength and

en
n

er
se

Coherence of

D
25 M
S

Rock Mass ay
50
o

h
ha
10
m

C ve

o
e

oh C
An i Clay and Size
Volumetric 2
er
e s o t r o py
vC n Content, Cementation
Stability 25 los ce 2
Cr e Jo Non-Coherent
nd
us
20

he intin Continuous Sa
d g Density (Friction Angle
and Dilatancy)
Ground Water Gravel
Level Fr Ground Water Level
ac
t u 2
reS
p a c ing , m m , D 1 0

Maj
o r C on
t r o llin g Fact ors

Fig. 8-11 The unified classification chart (from Deere et al., 1969)

There are many overlapping characteristics of soils and rocks in rock engineering; stiff clays
extend naturally into shales and slates; varying degrees of alteration or cementation can form any
intermediate characteristic between solid rock and indurated soil. It is significant that major
problems in rock tunnelling are frequently associated with weakness zones where the rock
approaches the character of soil. Often, the most important characteristic separating soil from
rock is the relative importance of the discontinuities in rocks. Other differences are:
∗ A soil mass consists of an assemblage of uncemented angular to round particles randomly
located. The voids in between the particles may or may not be filled with water (or more
fine-grained materials in the case of moraines and coarse-grained materials (scree)). It is
essentially a continuous material.
∗ A rock mass, on the other hand, can sometimes be considered as a continuous, sometimes
as a discontinuous material made up of an aggregate of blocks or particles properly
organized or piled like the bricks in a wall, more or less separated by planes of weakness.
These blocks generally fit tightly. The spaces between the blocks may or may not contain
water and soft and/or hard infilling materials.

In soils there will be a tendency for failure to occur arbitrarily, but in rock masses the tendency is
for the failure to follow pre-existing planes of weakness. A second important difference between
soil and rock behaviour is, that in rock masses, the shear strength will be determined largely by
the shear strength of the discontinuities and not by the rock strength.
8 - 20

TURE OF
UC A
TR
SO
S

IL D
2 m m3 ) (V b b = 60
Db = 5 mm e =
125 m m
b) (V b e
= GRAVEL cm 3
(D )
s
le CO
(V Db
ic BB
LE b =
ND
t

e = 6
r

SA 0. 00
pa

12 m
5 m
il

)
m mm

m3
3
be = 0 so
4 m

)
10 .-06

BO
.

UL
DE
( V Db
=

R
SILT

Unia xia l compressive strength of intact rock


Db = 2 µm

0.25 0.5 1.0 5 15 50 100 250 MPa

STRONG
WEAK
INTACT

VERY
STRONG
MEDIUM
VERY
WEAK

EXTRA
STRONG

STRONG
EXTRA
WEAK

ROCK
CLAY

STIFF
VERY
SOFT

SOFT
CLAY

CLAY

CLAY

CLAY
FIRM

ROCK
1.0 10 25 50 100 250 kPa
Unia xia l compressive strength of soil ma ss

1 ,00 0 3
σc =

m
1 2 m)
SIVE
1

inted
MAS

(D b =
y jo

Vb =
e
σc
=5

akl
we
E qu

σc =
i v

R
15
DE
ale

ed
L
nt
OU
joi
n

σc =
ts

ly

50
1. m 3
o

te

)
m
il p

σc =
ra
ar

1
2
e

100
tic

od

b =
le

(D Vb
e =

σc =
250
in t ed BB
LE
)

y jo CO
b

SA ngl (
V

ND
stro
V crushed
3 e
(D b b = 1 dm ) m
m GRAVEL =1 2m lu OF A R
e =
12 m 3 Vb = 0 .1 TU
RE OC
mm (D b vo
C

e
k
U

)
MA

c
ST R

3
Vb = 1 cm
blo
SS

(Dbe = 0.012 m)

Fig. 8-12 A general identification chart for rock masses and soil materials
8 - 21

OF A SOIL
RE
U
CT
STRU

)
(Db 2m m Db =
c le s Db = GRAVEL 60 mm
art i
lp CO
oi BB
LE D
b
ND
s

SA =
60
0
m
m
mm

BO
06
0.

UL
=

DE
Db

R
SILT

Unia xia l compressive strength of intact rock


Db = 2 µm

0.25 0.5 1.0 5 15 50 100 250 MPa

STRONG
WEAK

VERY

EXTRA
STRONG

STRONG
MEDIUM
EXTRA
WEAK

VERY
WEAK

STRONG
INTACT
CLAY

ROCK
STIFF
VERY
SOFT

SOFT
CLAY

CLAY

CLAY

CLAY
FIRM

ROCK
1.0 10 25 50 100 250 kPa
Unia xia l compressive strength of soil mass
σc =

0 m3
SIVE
1

1,00
MAS
σc =

Vb =
5
Eq

250
ui

σc =
R
v al

15
DE
e

UL
100

1
BO

0
n

σc = 0.
ts

50 i= A
o

RM
il p

50

σc =
ar

100
t ic

=
le

Vb

σc =
15

250 LE
05 BB
)

0. CO
b
5

SA
(
V

ND
0.1 .25
Vb 0 1
d m
3
e
=1 =1 m
mm 3 GRAVEL
Vb lu
v o
ck
3
Vb = 1 cm O F A R O CK
blo RE
M
U
CT

AS
S TR U

Vb = block volume
Db = particle diameter

Fig. 8-13 The RMi values for 'normal joint condition' (jC = 1) plotted on the identification chart. The RMi value is
found from the intersection between uniaxial strength (following the circles) and block size.
Example: uniaxial compressive strength of rock σc = 50 MPa and block volume Vb = 1 dm3 gives a B B P P

location in 'A'. Here RMi ≈ 0.7

The identification chart can be used to show the position of and the relation between actual
materials. It may help
- to identify the difference/similarities between various geologic materials,
- to identify and compare the different types of behaviour for various materials, and
- to improve communication.
8 - 22

Thus, it is hoped that the identification chart presented in Fig. 8-12 for rock masses will shorten
the gap between various rock classification systems and will contribute to a better
communication between soil and rock mechanics people. In Fig. 8-13 the RMi values may be
found roughly for 'normal joint condition' i.e. the joint condition factor jC = 1 - 2.

8.7 POSSIBLE USE OF RMi IN NUMERICAL MODELS

Numerous authors have demonstrated the use of numerical models in tunnel design. They have
produced a wealth of information, much of which is of considerable general interest, concerning
the two-dimensional stress and deformation patterns around tunnels. In using these powerful
numerical tools, it is necessary to be constantly aware of the fact that the answers produced are
only as good as the input information. In view of this limitation, the potential of numerical
models can today rarely be fully utilized in the practice of engineering design.

However, sensitivity studies made possible by computers can explore the influence of variations
in the value of each input parameter making a contribution to engineering judgement of the
accuracy in the calculation.

No attempts have been made in this work to apply RMi or its input parameters in numerical
models. Block size, block type and shape in the RMi system can, however, be valuable as input
to numerical models. Due to the fact that the input parameters to RMi are well defined, their
possible use in numerical models may consequently result in improved numerical predictions.
)URP3DOPVWU|P$50L±DURFNPDVVFKDUDFWHUL]DWLRQV\VWHPIRUURFNHQJLQHHULQJSXUSRVHV
3K'WKHVLV2VOR8QLYHUVLW\1RUZD\S

&KDSWHU

',6&866,21$1'&21&/86,216

$FRPSOHWHXQGHUVWDQGLQJRIWKHIHDWXUHVDQGPHFKDQLVPVLQYROYHGLQWKHVWUHQJWKRIURFN
PDVVHVSUHVHQWVIRUPLGDEOHWKHRUHWLFDODQGH[SHULPHQWDOSUREOHPVDQGKHQFHVLPSOLI\LQJ
DVVXPSWLRQVDUHUHTXLUHGLQRUGHUWRSURYLGHDUHDVRQDEOHEDVLVIRUHVWLPDWLQJWKHVWUHQJWKRI
MRLQWHGURFNPDVVHVIRUHQJLQHHULQJGHVLJQSXUSRVHV
Evert Hoek (1986)

The steadily increasing trend in the field of engineering geology and rock mechanics to substitute
the geological reality with mathematical idealizations often causes reduced interest in the quality of
the input parameters representing reality. As mentioned in the introduction the main objective in
this work has been to involve and represent geological parameters better in rock engineering and
design. In addition important goals have been to:
- improve the measurements and descriptions of important rock mass parameters used in a
general rock mass index, RMi, which characterizes the relative strength of a rock mass.
- develop well defined characterizations which will improve quality of geological input data
and thus contribute to improved communication between people involved in rock engineering.

In addition to definitions the methods presented are backed by expressions and equations; many of
them are supported by diagrams and tables. As this work relates to the field of rock construction,
civil engineering and design, it involves, as Bieniawski et al. (1993) point out, three main
disciplines, namely:
- engineering geology, which provides the framework for collecting and developing facts of the
nature of the rock mass,
- rock mechanics, which provides the theory and empirical rules of rock deformation and
failure, and
- rock engineering, which is the practical application of engineering geology, rock mechanics,
and geohydrology to design of rock structures.

With so many topics covered a detailed discussion would tend to be comprehensive; therefore,
’local’ discussions, comments and summing-ups have been outlined in many of the chapters and
sections.

 217+(/$<2872)7+(50L6<67(0

Central to the process of developing the RMi system and its use in practical rock engineering has
been to arrive at:
- a simple structure for the RMi to characterize rock masses; such that its
- parameters can be applied to several engineering purposes; and that it can be
- applied both for preliminary estimates as well as for more accurate calculations and
assessments based on the quality of its input parameters.
9-2

This is well in accordance with Lane (1948) who, in his critique of Casagrande’s original published
proposal of the Unified Soil Classification, listed four requirements that should be satisfied in any
classification system of natural materials 1 for engineering purposes:
L ,WVKRXOGGHVFULEHWKH
PDWHULDO
LQZHOOXQGHUVWRRGWHUPVWKDWFRQYH\DQLGHDRILWVW\SHDQG
EHKDYLRXU
LL 7KHV\VWHPVKRXOGIXUQLVKDQLQGLFDWLRQRI
PDWHULDO
SURSHUWLHVDQGSHUIRUPDQFH
LLL ,WVKRXOGEHDSSOLFDEOHIURPYLVXDOH[DPLQDWLRQERWKLQDVLPSOLILHGIRUPDQGZLWKH[SHULHQFH
LQDPRUHUHILQHGIRUP
LY ,WVKRXOGHPSOR\DVLPSOHV\VWHPRIQRWDWLRQIRUJUDSKLFDEVWUDFWVRIERULQJORJV VHFWLRQV
SODQPDSVFRQWUDFWVSHFLILFDWLRQVUHSRUWV RQGUDZLQJVUHFRUGLQJERULQJLQIRUPDWLRQ
The fourth requirement is not outlined in this work, but the methods presented may well be
further developed to also cover such purposes.

RMi is a numerical, general characterization of rock masses. As a measure of the rock mass
properties, it includes only inherent parameters of the rock mass. When applied in practical rock
engineering RMi is adjusted for the local features of significant importance for the actual use, work
or utility. It is thus a flexible system applicable to many different purposes related to rock
construction as indicated in Fig. 9-1.

To avoid confusion in a field, which is already crowded with classification methods and systems,
RMi is tied to basic characteristics of geologic materials. Thus, few new terms or parameters have
been introduced.

assessments output and applications


input parameters
and/or tests

-2,17,1*,17(16,7<
MRLQWLQJ

-2,17&+$5$&7(5,67,&6 SDUDPHWHU

5RFN0DVVLQGH[ 50L
XQLD[LDO
FRPSUHVVLYH
VWUHQJWK

/2&$7,213$5$0(7(56

&RPPXQLFDWLRQ
&216758&7,213$5$0(7(56

5(48,5(0(176

V V
V F Q
L
P Q R
J L
H D W
N Q W
L K D
F V O
J W F
R Q \ X
U V H
 L V F
U [  O
Q L P D
L H [ Q 
 H N F
G H R 
Q  L F
H L W G
L R D R Q
O J W U
S Q  F  D
W L 
S H X I
L Q
L V
O
$ S V 
W H
Q V X G
,
D S R
O
F Q P
,

Fig. 9-1 Geological properties/parameters are quantified and combined to a general rock mass index RMi. RMi
and/or its parameters can be applied as input in various types of rock engineering.

1
In the statements below, the word ’material’ has been applied instead of ’soil’.
9-3

&RPSDULVRQVZLWKWKHSULQFLSOHVLQRWKHUURFNHQJLQHHULQJV\VWHPV

Chapters 6 - 8 show that RMi and/or its parameters have many applications in rock mechanics and
rock engineering. The same principle has been presented both by Castelli (1992) using a ’basic rock
mass quality factor’ as the general factor, and by Hack and Price (1993) introducing the term
’reference rock mass’ for the same. This is shown in Figs. 9-2 and 9-3 respectively.

EVALUATION
ROCK MASS
OF ROCK
PROPERTIES
MASS

ADJUSTMENT
ENGINEERING
BY ORIENTATION
PARAMETERS ENGINEERING

EVALUATION SUITABILITY
LOCATION FOR
PARAMETERS OF LOAD
CONDITION PROJECT

(PARAMETER
SELECTION)
WEIGHT PROJECT WEIGHT
FACTOR

Fig. 9-2 Layout of the classification system by Castelli (1992).

other
REFERENCE ROCK exposures
MASS
exposure REFERENCE ROCK MASS
REFERENCE ROCK-MASS
method of excavation RATING
degree of weathering (RFR)
spacing (SPA)
spacing condition (CD)
discontinuities
condition orientation (SPA + CD + IRS) +/- SW =
discontinuitites persistence RFR
orientation
persistence

intact rock strength intact rock strength (IRS)


susceptibility to weathering susceptibility to weathering (SW)

SLOPE STABILITY SLOPE


correction factor for slope
ASSESSMENT method of excavation orientation and dip with
degree of weathering respect to orientation of
orientation and dip slope discontinuities (SLO)
REFERENCE ROCK MASS permanent water
height slope
orientation

discontinuities spacing (SPA) SLOPE ROCK-MASS


condition (CD) STABILITY RATING
persistence (SSR)
correction factors for (ASPA + ACD + AIRS)
persistence in relation AWA + SW +SLO =
intact rock strength (IRS) with height of slope SSR
susceptibility to weathering (SW)

Fig. 9-3 Layout of the classification system by Hack and Price (1993).
9-4

Bieniawski (1984, 1989) applies also a similar principle when the RMR system is used in rock
support recommendations in mining. The basic (common) RMR value is adjusted for stresses,
blasting damage and for faults as shown in Fig. 9-4.

Strength of
intact rock
Blasting damage
Rating: 0-15 adjustment Ab
0.8-1.0

Discontinuity
Discontinuity orientation
density adjustment
In-situ stress &
RQD: 0-20 change of stress
Spacing: 0-20 adjustment
As
Rating: 0-40
Basic RMR
0.6-1.2
0-100

Discontinuity
condition Major faults &
fractures
Rating: 0-30 S
0.7-1.0

Adjusted RMR
Groundwater
condition RMR x Ab x As x S
Rating: 0-15 max. 0.5

Support recommendations

Fig. 9-4 Layout of the RMR classification when applied in mining (from Bieniawski, 1984)

Similarly, in the draft of 1993 for "The Chinese National Standard for Engineering Classification of
a Rock Mass" a ’rock mass basic quality index’ is suggested (Xuecheng, 1993).

It is important that when a general index or basic factor is chosen, it characterizes those properties
of the material which are significant in engineering. In this way the index or factor is useful in
communication as well as where comparisons are made between various localities. As mentioned in
Chapter 4 there is a need, also in the field of rock engineering, to report the properties of the
material (rock mass) used in the construction.

 217+(6758&785(2)7+(50L

The rock mass index, given as RMi = σc × JP, is used to estimate the strength properties of a rock
mass based on combinations of important rock mass features 2. As outlined in Chapter 4, RMi can
be compared with the ’unconfined compressive strength of rock masses’ in the Hoek-Brown failure
criterion for rock masses given as σcm = σc × V½ .

Hoek and Brown derived their failure criterion mainly from triaxial test data on intact rock
specimens. For jointed rock masses they had only triaxial test data from the Panguna andesite.
Therefore, the values of V for jointed rock masses, have been considered YHU\DSSUR[LPDWH as
stated by Ward (1991).

2
σc is the compressive strength of intact rock material, JP is the ’jointing parameter’ as outlined in Chapter 4.
9-5

In this work it has been possible to find another 7 sets of compressive strength data for rock masses
used to determine the value of JP for various types of jointing. As discussed in Chapter 4, this is
still considered insufficient as a basis for an expression to characterize the strength properties of
rock masses. Although relatively good correlations were found between the strength data used, it is
probable that the combination of block volume and joint characteristics in JP is relatively rough.
With the variability in the rock masses it is hardly possible to find an accurate, ’realistic’ and at the
same time simple combination of the features and parameters acting. It is, therefore, important to
stress that RMi characterizes the UHODWLYH compressive strength of different rock masses.

After introducing it in 1980, Hoek and Brown found from practical applications of their failure
criterion that the values of the constant V were conservative. Therefore, they adjusted the V-values in
1988. A general comparison between V and JP is difficult to carry out because different definitions
of the two parameters are applied. As seen in Fig. 9-5, the values of V (= JP 2), are mainly lower
than JP2 for the ’original’ V, while they are mostly higher for the ’revised’ V. Thus, it seems from
Fig. 9-5 that there is a fair connection between the ’original’ s and JP for small values and
between the ’revised’ V and JP for the higher values.

0,1

0,01

0,001
’original’ values of s
revis ed values of s

0,0001
0,0001 0,001 0,01 0,1 1
s

Fig. 9-5 Comparison between the jointing parameter (JP) and the parameter V in the Hoek-Brown failure
criterion. Values of V have been found via the Q-system.

As for the Hoek-Brown failure criterion, also RMi - when applied directly in calculations - is
restricted to FRQWLQXRXV rock masses. In discontinuous rock masses the use of RMi must be adjusted
to the local conditions. This has been shown in the application of RMi in design of rock support
(Chapter 6) where discontinuous and continuous rock masses have been treated separately in the
assessment of stability and rock support.

 217+(,13873$5$0(7(567250L

Basically, the properties of the URFNPDWHULDO can be found from laboratory tests. -RLQWV, however,
have generally so large dimensions that their characteristics can hardly be tested mechanically. An
additional difficulty in working out a system for characterizing a rock mass is its three-dimensional
structure. Most characterizations are performed as rough, inaccurate one-dimensional measurements
of joint spacing or joint frequency based mainly of the dominating joint set. Few rules or guidelines
9-6

exist how to include the effect of the other joint sets and joints. An important consideration in this
work has been, in a simple way, to include the 3-dimensional composition of rock masses. For this
the block volume has been selected; in addition, it is possible to include the effect of variations in
the joint condition factor (jC) as shown in Chapter 5.

To help the user, several methods have been presented in Appendix 3 on how to compare and select
the results from various kinds of jointing descriptions and measurements.

7KHXQLD[LDOFRPSUHVVLYHVWUHQJWKRIWKHURFN

As the uniaxial compressive strength of rocks is well defined, this measurement should in most
cases be easy to apply in the system. This parameter makes it possible also to include clay materials
in the characterization by using their unconfined, undrained compression strength. Also clay fillings
in faults and weakness zones where the blocks do not have contact, can in this way be characterized
by the strength of the clay material. For granular soil materials like silt, sand and gravel the
compressive strength is, however, difficult to measure; and has to be roughly estimated where such
input is required.

Uniaxial compressive strength test is most frequently used to define rock behaviour. The test is open
to misinterpretation (Farmer and Kemeny, 1992). The reason for this lies principally in the structure
and breakdown mechanism of the rock. There are also possible sources of error in the sampling and
sample preparation.

Appendix 3 shows various other methods for determining the uniaxial compressive strength. For
example, the point load strength is a quick and easy method for determining the strength of rocks,
and it may sometimes give more reliable results than compression tests on machined cylindrical
samples.

The lowest value of σc should be applied in RMi which means that for anisotropic rocks the test
should be made at approximately 45o to the schistosity or layering, as shown in Appendix 3, Section
1.

-RLQWLQJ

9.3.2.1 Joint characteristics

Ideally, all the characteristics (smoothness, waviness, length) included in the joint condition factor
jC should be measured accurately. Such measurements of the joints would generally be either
extremely time-consuming or in most cases, practically impossible to carry out. Only a few joints in
the rock mass can be observed and measured and extrapolations have to be made.

Methods given in Chapter 5 show how to calculate jC, also where the input parameters vary. For
rough estimates, where limited information on the joint characteristics exists, a ’common’ value jC
= 1 - 2 has been recommended. Using this it is possible to give preliminary estimates of RMi, also
where input data on joint properties are lacking.

9.3.2.2 Block volume


9-7

The intensity of jointing has in most cases the greatest influence on the engineering properties of
rock masses. An important feature in the RMi system is the selection of block volume for this
parameter. By this it is possible to describe the whole range of degree of jointing from spacing less
than 1 cm to several meters.

The measurement of block volume is especially useful where small blocks or irregular jointing
occur. Crushed rock masses both in excavations and in drill cores are examples where the use of
block size can give a more accurate description than traditional joint measurements.

Not all types of rock masses are, however, made up of separate blocks. As soon as many of the
joints are discontinuous, or less than three joint sets occur the joints do not delimit defined blocks.
This feature has been dealt with in Chapter 5 and in Appendix 3, Section 3 where guidelines are
worked out for how to determine an equivalent block volume. In fact, this is not a feature specific to
the measurement of block size. It is apparently "hidden" in jointing measurements like RQD and
joint spacing. The presence of discontinuous joints and/or few joint sets has, however, a significant
effect on the properties and behaviour of rock masses and should be included in the engineering
characterization.

 217+(9$5,$7,216$1'81&(57$,17,(6,152&.0$66(6

:KHQWKHVWDUWLQJSRLQWLVOHDVWVRXQGPRVWVXUSULVLQJUHVXOWVPD\EHIRXQG
Henrik Ibsen (1867)

The variations in the structure and composition of a rock mass often result in problems and
uncertainties when its features or properties are to be described and characterized.
According to Bieniawski et al. (1993) the hardest challenges to the designer are the variability of the
rock conditions and the lack of sufficient information when rigorous analyses are being made.

Where the rock mass characterization is carried out EHIRUH construction, uncertainties are introduced
from the LQWHUSRODWLRQV made between more or less known conditions at the surface and from
various forms of H[WUDSRODWLRQV carried out from these (known) conditions to areas with unknown
information. Except for wrong interpretations, improved characterization of the rock mass by RMi
will generally increase the quality of the geological input data to be applied in evaluation,
assessments or calculations. This will in turn lead to better designs.

Where the rock mass characterization is carried out in the tunnel or cavern, the quality of the input
data depends mainly on how the description and characterization of the input parameters are
performed. In these cases improved methods for characterizations have a direct impact on the
engineering quality.

As there often are wide variations in the rock masses and their behaviour, even within limited
areas/volumes, it is often wise to use a variation range of the geological parameters used in RMi.
The great variations in structure/composition result in that the relative strength properties of rock
masses almost always will vary within certain ranges, refer to Chapter 5.
9-8

 &203$5,621%(7:((150L$1'27+(50(7+2'686(',152&.
(1*,1((5,1*

Chapters 6 - 8 show that RMi can be applied in various types of analyses for rock engineering
purposes included underground stability and rock support determination and TBM penetration
progress. In addition, some of the parameters in RMi can be used individually in classification
systems where they may improve the input because they represent reality better and/or they may be
found easier or more relevant to use.

7KHURFNTXDOLW\GHVLJQDWLRQ 54'

RQD is probably most commonly parameter in drill core logging, as it is rapid and easy to learn.
Today, RQD is applied in the main classification systems as an input parameter for the block size or
the jointing density. RQD is one-dimensional; therefore it is strongly directional. This was stressed
among others by Bjerrum (1965). Therefore, Hudson and Priest (1983) and several other authors
recommend carrying out core drillings in three directions to obtain reliable results. This is, however,
an expensive solution to obtain information on the jointing.

52&. 48$/,7<

0 30 60 90 100 '(6,*1$7,21 54'

92/80(75,&
-2,17 &2817
3
100 50 20 10 5 2 1 0.5 0.2 joints/m

%/2&.
1 10 0.1 1 10 0.1 1 10 100 1000 10 000 100 000
92/80(
3 3 3
cm dm m

Fig. 9-6 Range of jointing covered by RQD, block volume (Vb), and volumetric joint count (Jv).

Fig. 9-6 shows that RQD only covers a small part of the range of the block sizes in a rock mass.
This is further shown in Fig. 4-4 (in Chapter 4) and in Appendix 4. Thus RQD does neither express
the variations for a high degree nor for a low degree of jointing. Therefore, several authors have
suggested improvements in the registration of RQD, such as:
• Sen and Eissa (1991) have published a method for more accurate determination of the block
volume from core drillings and RQD measures. It requires, however, considerable amount of
information on jointing to establish the statistical jointing distribution required.
• Barton et al. (1974) have improved the application of RQD in the Q-system by dividing it with a
factor (Jn) for the number of joint sets.
• Also Bieniawski (1973) has modified the use of RQD in the RMR system by adding a rating for
the spacing of the joints.

The extensive use of RQD comes partly from the fact that it is a simple, cheap, and rapid parameter
for characterizing drill cores. Considering the costs of coring a hole, it is, however, remarkable that
so little work has been performed on improving core logging. The weighted jointing density method
described in Appendix 3, which is fairly rapid and simple, offers an attractive improvement of core
characterization.
9-9

5RFNVXSSRUWGHVLJQV\VWHPV

The two most used classification systems - the rock mass rating (RMR) of Bieniawski (1973) and
the Q-value of Barton et al. (1974) - both arrive directly at their quality value related to stability and
rock support from various parameters in the rock mass.
These empirical systems are models for support practice and not tunnel mechanics or analysis of
rock masses. As noted by Dowding et al. (1975), the selection of initial supports is often governed
by factors having nothing to do with required capacity, i.e. material availability. Also, there is an
understandable tendency by the tunnellers to be safe, which leads to overdesigned supports. Final
supports are usually overdesigned again by being conservative and by not often considering the
effect of initial support. All of this would be unimportant if the degree of overdesign were known,
but it is not (Einstein et al., 1979).

9.5.2.1 Comparison between RMi and the classification systems of RMR and Q

By excluding the parameters related to the location (rock stresses, water pressure and orientation of
joints), both classification systems can indicate the general conditions of the rock mass and thus be
compared with RMi. The results from an investigation of the connection between RMi - RMR and
between RMi - Q are shown in Figs. 9-7 and 9-8. The values have been found using various input
data on:
- uniaxial compressive strength of rock material;
- joint set spacings and number of joint sets; and
- joint roughness and alteration.
The corresponding values of RQD, block volume and spacing have been found using equations
derived in Appendix 3.

1000
RMi = 10 (RMR-40)/15

100

10

0,1
20 30 40 50 60 70 80

R MR

Fig. 9-7 Comparison between RMR and RMi

As shown a general correlation3 can be expressed as


RMi ≈ 10 (RMR - 40) /15 eq. (9-1) 4

3
The correlation is best for RMR < approx. 70
4
Applying this correlation and Vel = 10 (RMR-10) / 40 (Serafim and Pereira, 1983) the deformation modulus for rock
masses can be expressed as Vel ≈ 5.6 RMi 0.375 . This equation should, however, be further documented from in situ
measurements.
9-10

For some conditions there is, however, considerable deviation between the two corresponding
values. A main reason may be the different ways the two systems apply jointing intensity (or block
size). In the RMR system the jointing is characterized by RQD and by spacing of joints. The
insufficient definition of spacing, and in addition the use of RQD causes that the jointing parameter
in the RMR system generally is an inaccurate parameter. This is further documented in Appendix 4.

1000

100

10

0,1
1 10 100
Q = R QD /J n x J r/J a

Fig. 9-8 Comparison between values of Q and RMi. The parameters for Jw and SRF have been given value one.

There is poor correlation between values of Q and RMi in Fig. 9-8. The main reason is that the Q-
system does not include a strength parameter for the rock. The input of RQD to the value of Q may
also increase the inaccuracy of the correlation. In Fig. 9-9 the jointing parameter, JP, has been
applied instead of RMi, thus avoiding the rock strength parameter such that the two systems contain
the same parameters. For Q > 1 the expression
JP = approx. 0.01 Q
gives as shown a rough correlation. For lower Q-values it does not seem to be any correlation. This
is probably caused mainly by RQD’s incapability characterizing highly jointed rock (see Appendix
4).

0,1

JP = 0.01 Q

0,01

0,001
0,1 1 10 100
Q = R QD /J n x J r/J a
9-11

Fig. 9-9 Comparison between values of Q and the jointing parameter, JP. In the Q value the parameters for Jw and
SRF have been excluded.
The various parameters used in the two systems and in RMi are shown in Table 9-1. Here, the
parameters in the RMi system applied in assessment of rock support as presented in Chapter 6, have
been used.

It is clear from this that the RMi-system includes mainly the same input parameters as the Q-system
and the RMR system. RMi can, however, more easily be adjusted for local features of importance.

Both the RMR and Q systems "mix" the input parameters into one value which is used to
recommend the rock support. They are both fairly simple so that relatively inexperienced people can
carry out the collection of input data and perform the support evaluations. RMi, being a stepwise
system, requires more experience to be fully utilized. It has a structure which makes it easier to
understand the interaction and significance of the various features and parameters used; this makes
the use of judgement easier and more attractive.

TABLE 9-1 COMPARISON OF THE RMR AND Q SYSTEMS WITH THE RMi USED FOR ROCK SUPPORT
-----------------------------------------------------------------------------------------------------------------------------------------------
52&.0$66 3$5$0(7(565(35(6(17(',1
3$5$0(7(5 7KH505V\VWHP 7KH4V\VWHP 6XSSRUWDSSOLFRIWKH50LV\VWHP
-----------------------------------------------------------------------------------------------------------------------------------------------
JOINTING RQD RQD Block volume *)
INTENSITY joint spacing number of joint sets

JOINT CHARAC- joint roughness 2) joint roughness joint roughness *)


TERISTICS joint alteration 2) joint alteration joint alteration *)
joint filling 2) joint filling/coating joint filling/coating *)
joint thickness joint thickness joint thickness *)
joint length *)

STRENGTH OF compressive strength compressive strength *)


ROCK MATERIAL

ROCK STRESSES stress adjustment 1) rock stresses rock stresses

SWELLING swelling pressure 5)

GROUND WATER leakage and pressure leakage and pressure 4)

ORIENTATION OF strike & dip of joints strike & dip of joints


DISCONTINUITIES strike & dip of zones

OTHER FEATURES blasting damage 1) 4)


major faults/fracture 1) weakness zones weakness zones

DIMENSION OF related to 10 m span span and/or height span and/or height


EXCAVATION

USE OF THE excavation support ratio 6)


EXCAVATION
------------------------------------------------------------------------------------------------------------------------------------------------
*)
Applied in the general (basic) rock mass characterization (RMi).
1)
Parameters used to adjust the basic RMR value for mining purposes.
2)
A better division has been presented by Bieniawski, 1989. (See Table 8-6)
4) May be included where it has important influence.
5) The effect from swelling often requires special analysis to be carried out; therefore swelling is not included in the
support method of RMi.
9-12

6) Support charts for various applications of the excavation remain to be worked out.

Castelli (1992) has shown another way to compare classification systems by finding how many
arrangements for all probable combinations they offer. He found that
- the Rock Mass Rating has totally 21,875 combinations, while
- the Q-system has a total of 2,363,904 combinations, or
- the Q system excluding the SRF factor has 147,744 combinations.

As Castelli points out, the number of combinations will in practice be reduced because it is unlikely
that some combinations occur simultaneously. For example, Castelli mentions that in the RMR
system, RQD < 25% and discontinuity spacing > 2 m is extremely unlikely. In spite of that, each
system characterizes a large number of different rock masses. Especially, the Q-system is
considerably subdivided. The RMi system, includes a wider variation in jointing by applying block
volume and in addition includes joint size and the compressive strength of the rock, incorporates
probably a higher number of combinations than both systems.

 7+(1((')25$
/$1*8$*(
,152&.0(&+$1,&6$1'52&.
(1*,1((5,1*

Because many engineering decisions are based on a combination of geologic, geotechnical and rock
mechanics data, ISRM (1971) finds it important that a more systematic means of combining and
correlating this information should be developed.

Similarly, a common deficiency in both geological and geotechnical literature has been the lack of an
adequate and generally accepted means of transmitting an overall assessment of the nature of rock
masses to those who have not had the opportunity of observing them. A language common to rock
mechanics specialists and experts from related fields should, according to ISRM (1980), be available.
This has also been mentioned by Matula (1969).

Such a ’language’ or guidelines for descriptions using well defined terms will improve the
communication between the field geologist and rock engineer (Hoek and Brown, 1980). It will also
help in a meaningful interpretation so that field data should be intelligible to other engineers and
geologists who may become involved in the project. Deere et al. (1969) mention that a better
language will also help in the accumulation of experience associated with various classification
systems, when the description of the parameters are quantitative and can be ’translated’ from one
system to another.

Müller (1982) finds that the geological descriptions as well as the geomechanical testing and
presentation of data generally are good. But the transfer to application, the technical interpretation, is
missing or is poor. The engineering geologist should act like an interpreter between the geologist and
the rock engineer. 1HYHUVKRXOGURFNPHFKDQLFVEHXVHGZLWKRXWWKHEDFNJURXQGRIHQJLQHHULQJ
JHRORJ\6XFKDXVHRUPLVXVHZRXOGEHZRUVHWKDQWKHXVHRIHQJLQHHULQJJHRORJ\ZLWKRXWURFN
PHFKDQLFV

The following features in this work can possibly improve these draw-backs and problems in com-
munication and exchange of geo-data:
- The defined characterizations of the various rock mass parameters included in the RMi system
in Chapter 4 as well as in Appendix 3.
9-13

- A list of definitions for common rock mass features in rock engineering, given in Appendix 3
useful for a more concise use of descriptive terms.

- Guidelines on how to measure and combine many of the parameters used and apply them in the
characterization as shown in Chapter 5 and Appendix 3.
- The guidelines presented for ’translation’ of qualitative descriptions into quantitative numbers in
Appendix 3.

The use of RMi as a general, numerical system of rock masses, connected to descriptions will,
therefore hopefully, strengthen the communication between those engaged in rock construction, rock
engineering and design.

 %(1(),76$1'/,0,7$7,216,1$33/,&$7,212)7+(50L6<67(0

Some of the EHQHILWV using the RMi system are:

• In the author’s opinion RMi will give a significant improvement in the geological data applied in
connection with rock constructions:
- by its more systematic use of the input of rock mass characterization; and
- by the way its parameters are determined, including the possibilities applying results
from different types of measurements and descriptions.
• It can also be used when limited information on the ground conditions is available; for example,
in early stages of a project or where rough estimates are sufficient.
• It can easily be used in comparisons and exchange of knowledge between different locations, as
well as in general communication.
• It is a stepwise system suited for engineering judgement.
• It is much easier to find the values of V (= JP 2 ) using the RMi system than the methods outlined
by Hoek and Brown (1980) which incorporate the use of classification systems.
• It covers a wide spectrum of rock mass variations and therefore has possibilities for wider
application than most other rock mass classification and characterization systems. This has for
example:
- the system developed for assessment of rock support; and
- the system developed for assessing the boring rate of TBMs.
• The use of parameters in RMi can improve the input in other rock mass classification systems and
in the NATM.

It is, however, important to also realize the following OLPLWDWLRQV of the RMi:

• As a result of the often large volumes involved plus the three-dimensional and complex structure
of rock masses combined with the inaccessibility for seeing or observing the real conditions, it is
not possible to:
- apply all the various parameters in a rock mass in a simple system;
- collect exact information on the rock mass structure as measurement and description of rock
masses generally is based on extrapolation;
- obtain/record the exact data on the rock mass conditions even with the most sophisticated
investigation program;
9-14

The assessments in rock design and engineering must, therefore, often be based on data found
from simplified descriptions. This is, however, a limitation valid also for other methods used in
the collection of geo-data.

• We have to accept that the data we use in our calculations, evaluations and assessments often have
limited accuracy. A part of this stems from the descriptions of rock masses and how their features
are combined. The results from design and engineering can never be better than the quality of the
geo-data used, however sophisticated the calculation methods and models applied may be.

• Considering the uncertainties connected with rock masses and the simplifications made in the
expression of RMi, it should be stressed that it expresses only the index strength of a rock mass.

• RMi basically expresses the general features of rock masses and should not be uncritically applied
where more specific analyses need to be carried out.

 620(&21&/8',1*5(0$5.6

By using mostly standard geological and rock mechanical descriptions and classifications, the RMi
system should be relatively easy to adapt and apply. Most users should soon be familiar with finding
the value or ratings of the parameters involved. Simple features and expressions have been preferred
here rather than more sophisticated and complicated solutions, even if the latter could give more
accurate results.

As most of the evaluations, assessments and engineering are based on observations, the results are
wholly dependent on the quality of the input data found from visual description or from
measurements. The well defined parameters used in RMi and the way they are structured may, as
mentioned, improve the quality of such data. Prerequisite for this is, however, that experienced people
are responsible for the interpretation and acquisition of the geo-data.

It is not possible to exactly characterize all the varieties of such complex a material as a rock mass is
by any simple system. Though RMi offers a stepwise method where variations in the rock mass
possibly may be better characterized than in most other systems, the results are generally rough. For
practical purposes, however, they should in most cases be sufficient as input in the assessments,
keeping in mind that considerable efforts have to be made to obtain better structure and strength
information from the rock mass.

RMi is a tool - a help - during the rock engineering design process. It can never replace geological
feeling and/or practical experience. Further, it can never cover all of the innumerable
cases/types/occurrences of rock masses. Practical judgement always has to be the main factor when
assessments and evaluations are made in this field.

At the end of this chapter it is, therefore, appropriate to emphasize the role of engineering judgement.
Prerequisite for this is a sound XQGHUVWDQGLQJof the site conditions and the impact caused by the
construction works. Any attempt to replace geology by classifications or numerical values may lead to
loss of geological understanding. The aim in the development of RMi has been that its simple
structure largely can maintain ’the geological feeling’ of the experienced rock engineer and the
engineering geologist. This can be achieved, as mentioned earlier, by connecting RMi to an additional
description of the rock mass using well defined terms.
9-15

)XWXUHGHYHORSPHQWV

It has, during development of this work, been important to show how RMi can be used in rock
engineering and design in rock support evaluation, and in assessment of TBM penetration rate. A lot
of work remains, including the collection of field data on the ground conditions and the experience
gained from construction case studies, to refine these methods.

So many aspects are involved in acquisition of geological data and their use in rock engineering, that it
has not been possible within this work to fully develop definitions, methods and expressions which
cover all possible applications. The expressions, diagrams and tables developed may possibly be
refined when more experience and data from constructions combined with rock mass characteristics
are available. More accurate expressions, can then be developed.

Some of the most interesting additional utilizations of RMi and/or its parameters include:
- a method to determine the deformation modulus of rock masses;
- input to design of rock blasting and fragmentation;
- input to numerical models; and
- a method to improve the interpretation of refraction seismic results to characterize
jointing shown in Appendix 5.

The use of the RMi system by the author in engineering practice for a couple of years has shown
promising results as well as interesting possibilities for development in rock mechanics and rock
engineering.
)URP3DOPVWU|P$50L±DURFNPDVVFKDUDFWHUL]DWLRQV\VWHPIRUURFNHQJLQHHULQJSXUSRVHV
3K'WKHVLV2VOR8QLYHUVLW\1RUZD\S

&KDSWHU

5()(5(1&(6

Amadei B. (1988):
Strength of a regularly jointed rock mass under biaxial and axisymmetric loading conditions.
Int. J. Rock Mech. Min. Sci. & Geomech. Abstr., Vol. 25, No. 1, pp. 3 - 13.

Amberg W. and Christini F. (1986):


The new Austrian tunnelling method in railway tunnel construction.
Rasegna dei lavori pubblici, No 5, pp. 241/1 - 252/1.

American Geological Institute (1962):


Dictionary of geological terms.
Dolphin Reference Books, 545 pp.

Aydan Ö., Akagi T. and Kawamoto T. (1993):


The squeezing potential of rocks around tunnels; teory and prediction.
Rock Mech. Rock Engn., No. 26, pp 137 - 163.

Baecher G.B., Lanney N.A. and Einstein H.H. (1977):


Rock joint properties and sampling.
Proc. 19th U.S. Symp. on Rock Mechanics, Keystone.

Baecher G.S. and Lanney N.A. (1978):


Trace length biases in joint surveys.
19th US Symp. on Rock Mechanics, Stateline, Nevada, pp. 56-65.

Barton N. (1973):
A review of the shear strength of filled discontinuities.
Proc. Conf. on Fjellsprengningsteknikk/Bergmekanikk, Tapir, Trondheim, 38 pp.
(also in Norwegian Geotechnical Institute, Publ. No. 105)

Barton, N., Lien, R. and Lunde, J. (1974):


Engineering classification of rock masses for the design of rock support.
Rock Mechanics 6, 1974, pp. 189-236.

Barton N., Lien R. and Lunde J. (1975):


Estimation of support requirements for underground excavations.
Proc. Sixteenth Symp. on Rock Mechanics, Minneapolis, pp. 163-177.

Barton N. (1976):
10-2

The shear strength of rock and rock joints.


Int. J. Rock Mech. Min. Sci. & Geomech. Abstr., Vol. 13, No. 9, pp. 255-279.

Barton N. and Choubey V. (1977):


The shear strength of rock joints in theory and practice.
Rock Mechanics, No. 1/2, pp. 1-54, (also in Norwegian Geotechnical Institute,
Publ. No. 119)

Barton, N., Lien, R. and Lunde, J. (1980):


Application of Q-system in design decisions concerning dimensions and appropriate support for
underground installations.
Proc. Int. Conf. Subsurface Space, Pergamon Press, pp. 553-561.

Barton N. and Bandis S. (1980):


Some effects of Scale on the shear strength of joints. Technical note.
Int. J. Rock Mech. Min. Sci. & Geomech. Abstr., Vol 17, pp. 69-73.

Barton N. (1987):
Predicting the behaviour of underground openings in rock.
4th Manual Rocha Memorial Lecture, Lisbon (in Norwegian Geotechnical Institute, Publ. No. 172)
21 pp.

Barton N. (1990a):
Cavern design for Hong Kong rocks.
Norwegian Geotechnical Institute, Publ.no. 180, pp. 1-24.

Barton N. (1990b):
Scale effects or sampling bias?
Proc. Int. Workshop Scale Effects in Rock Masses, Balkema Publ., Rotterdam, pp. 31-55.

Barton N. and Bandis S. (1990):


Review of predictive capabilities of JRC-JCS model in engineering practise.
Proc. Int. Conf. Rock Joints, Balkema Publ., Rotterdam, pp 603-610.

Barton N. (1993):
Physical and discrete element models of excavation and failure in jointed rock.
Keynote lecture presented at ISRM Int. Symp. on Assessment and Prevention of failure Phenomena
in Rock Engineering, Istanbul, Turkey.

Barton N. (1991 - 1994):


Private communication.

Bates R.L., Jackson J.A. (1980):


Glossary of geology.
American Geological Institute, Fall Church, Virginia, second edition 1980.

Bergh-Christensen J. (1968):
On the blastability of rocks. (in Norwegian)
Lic.Techn. Thesis, Geological Inst., Techn. Univ. Norway, Trondheim.
10-3

Bergh-Christensen J. and Selmer-Olsen R. (1970):


On the resistance to blasting in tunnelling.
Proc. 2nd ISRM Congr., Belgrade, Vol. 3, paper 5-7.

Bergman M. (1975):
Borehole investigations in rock; evaluation of the reliability of methods (in Swedish).
National Swedish Building Research, report no R17:1975, 69 pp.

Bergman S.G.A. (1956):


Functional rock classification. (in Swedish).
IVA Publ. 142, Stockholm.

Bhawani Singh, Jethwa J.L., Dube A.K. and Singh B. (1992):


Correlation between observed support pressure and rock mass quality.
Tunnelling and Underground Space Technology, Vol. 7, No. 1, pp. 59-74.

Bieniawski, Z.T. (1973):


Engineering classification of jointed rock masses.
Trans. S. African Instn. Civ. Engrs., Vol 15, No 12, Dec. 1973, pp 335 - 344.

Bieniawski, Z.T. (1974):


Geomechanics classification of rock masses and its application in tunneling.
Proc. Third Int. Congress on Rock Mechanics, ISRM, Denver 1974, pp.27-32.

Bieniawski Z.T. (1984):


Rock mechanics design in mining and tunneling.
A.A. Balkema, Rotterdam, 272 pp.

Bieniawski, Z.T. (1988):


Rock mass classification as a design aid in tunnelling.
Tunnels & Tunnelling, July 1988

Bieniawski Z.T. (1989):


Engineering rock mass classifications.
John Wiley & Sons, New York, 251 pp.

Bieniawski Z.T., Bauer S.J. and Costin L.S. (1993):


Geotechnical design methodology workshop.
News Journal of International Society for Rock Mechanics, Vol. 1, No. 4, pp. 42-45.

Bjerrum L.B. (1965):


Discussion of paper: Functional rock classification by S.G.A. Bergman.
IVA report 142, pp. 124-125.

Blindheim O.T., Boniface A. and Richards L.A. (1991):


Boreability assessments for the Lesotho Highlands water project.
Tunnels & Tunnelling, June 1991, pp. 55-58.

Braun W.M. (1980):


Application of the NATM in deep tunnels and difficult formations.
10-4

Tunnels and Tunnellin, March 1980, pp. 17-20.

Brekke T.L. (1965):


On the measurement of the relative potential swellability of hydrotermal motmorillonite clay from
joints and faults in Pre-Cambrian and Paleozoic rocks in Norway.
Int. J. Rock Mech. Mining Sci., Vol 2, pp 155-165.

Brekke T.L. and Selmer-Olsen R. (1965):


Stability problems in underground construction caused by montmorillonite carrying joints and
faults.
J. of Engineering Geology, Vol. 1, No. 1, pp 3-19.

Brekke T.L. and Howard T.R. (1972):


Stability problems caused by seams and faults.
Rapid Tunneling & Excavation Conference, 1972, pp. 25-41.

Bridges M.C. (1976):


Presentation of fracture data for rock mechanics.
Proc. 2nd Australia-New Zealand Conf. on Geomechanics, Brisbane

British Standard Institution (1981):


Code of practice for site investigations.
BS 5930: 1981, 147 pp.

Broch E. and Franklin J.A. (1972):


The point-load strength test.
Int. J. Rock Mech. Min. Sci., Vol. 9, pp. 669-697.

Broch E. and Leivestad S.I. (1973):


On the influence of moisture and anisotropy upon shear strength of rocks. (in Norwegian)
Publ. no.9, Geol. Inst., The Technical University of Norway, 30 pp.

Broch E. (1977):
The point-load test and its use in engineering geology. (in Norwegian)
Rep. no. 2, Geological Inst., Techn. Univ. Norway, Trondheim, 148 pp.

Broch E. (1979):
Changes in rock strength caused by water.
Proc. 4th ISRM Congr., Montreux, Vol. 1; Balkema, Rotterdam. pp. 71-75.

Broch E. (1983):
Estimation of strength anisotropy using the point-load test.
Int. J. Rock Mech. Min. Sci. & Geomech. Abstr., Vol. 20, No. 4, pp. 181 - 187.

Broch E. (1988):
Site investigations.
Norwegian Tunnelling Today, Tapir publ. Trondheim, Norway, pp. 49 - 52.
10-5

Brook N. (1985):
The ekvivalent core diameter method of size and shape correction in point load testing.
Int. J. Rock Mech. Sci. & Geomech. Abstr., Vol. 22, No. 2, pp. 61 - 70.

Brosch F.J. (1986):


Geology and the classification of rock masses - examples from Austrian tunnels.
Bull. IAEG no 33, 1986, pp 31 - 37.

Brown and Hoek (1978):


Trends in relationships between measured in situ stresses and depth.
Int. J. Rock Mech. Min. Sci.& Geomech. Abstr., Vol. 15, pp 211-215.

Brown E.T. (1981):


Putting the NATM into perspective.
Tunnels and Tunnelling, Nov. 1981, pp. 13-17.

Brown E.T., Bray J.W., Ladanyi B. and Hoek E. (1983):


Ground response curves for rock tunnels.
J. Geot. Engn., Vol. 109, No. 1, pp. 15 - 39.

Brown E.T. (1985):


From theory to practice in rock engineering.
Trans. Instn.. Min. Metall. (Sect. A: Min.industry), 94, 1985, pp A67-83.

Brown E.T. (1986):


Research and development for design and construction of large rock caverns.
Proc. Int. Symp. on Large Rock Caverns. Helsinki, Finland, pp. 1937-1948.

Brown E.T.and Hoek E. (1988):


Determination of shear failure envelope in rock masses.
J. Geot. Engn., Vol 114, No. 3, pp. 371 - 373.

Brown E.T. and Hoek E. (1988):


Discussion on "Determination of shear failure envelope in rock masses. by Ucar R."
J. Geot. Engn., Vol. 114, No. 3, pp. 371 - 373.

Burton A.N. (1965):


Classification of rocks for rock mechanics.
Letter to the editor, Int. J. Rock Mech. Mining Sci., Vol. 2, p. 105.

Burwell E.F.jr. and Roberts G.D. (1950):


The geologist in the engineering organization.
Chapter 1 in Application of Geology to Engineering Practice (Berkey Volume): Geological Society
of America, 327 pp.

Call R.D., Savely J.P. and Nicholas D.E. (1976):


Estimation of joint set characteristics from surface mapping data.
Monograph on Rock Mechanics Applications in Mining. SME-AIME, pp 65-73.

Cameron-Clarke I.S. and Budavari S. (1981):


10-6

Correlation of rock mass classification parameters obtained from borecore and in-situ observations.
J. Engn. Geol., Vol 17, pp. 19-53.

Carmichael T.J. and Lee C.F. (1977):


Rock mass characterization for location and design of an underground oil storage facility.
18th US Symp. on Rock Mechanics, Keystone, Colorado, 1977, pp. 5A4-1 - 5A4-8.

Carmichael R.S., Ed. (1989):


Handbook of physical properties of rocks and minerals.
CRC Press, Florida

Castellani V. and Dragoni W. (1991):


Italian tunnels in antiquity.
Tunnels & Tunnelling, March 1991, pp. 55-57.

Castelli E. (1992)
Geomechanics characterization methologies: a matrix approach.
Periodico della Societa Italiana Gallerie 38, pp.13-25.

Cecil O.S. (1970):


Correlations of rock bolt - shotcrete support and rock quality parameters in Scandinavian tunnels.
Ph.d. thesis Univ. of Illinois 1970; also in Proc. Swedish Geotech. Institute, No 27, 1975, 275 pp.

Cecil O.S. (1971):


Correlation of seismic refraction velocities and rock support requirements in Swedish tunnels.
Reprints and preliminary reports, No. 40, Swedish Geotechnical Institute, Stockholm.

Chappell B.A. (1987):


Predicted and measured rock mass moduli.
J. of Mining Sci. and Tech., 6 (1), pp 89-104.

Chappell B.A. (1990):


Rock mass characterization for dam foundations.
J. Geotechn. Engn., Vol 116, No. 4, pp 625 - 646.

Chen J.F. and Vogler U.W. (1992):


Rock cuttability/boreability assessment research at the CSIR
TUNCON ’92, Design and Construction of Tunnels, Maseru, Lesotho, pp. 91-97

Clerici A. (1993):
Indirect determination of the modulus of deformation of rock masses - Case histories.
Proc. Int. Conf. Eurock’93, pp. 509 - 517.

Coates D.F. (1964):


Classification of rocks for rock mechanics.
Rock Mech. and Mining Sci., Vol 1, pp. 421-429.

Colback P.S. and Wiid B.L. (1965):


The influence of moisture content on the compressive strength of rock.
Proc. Symp. on Rock Mechanics, Toronto, pp. 65-83.
10-7

Coon R.F. and Merritt A.H. (1970):


Predicting the modulus of deformation using rock quality indexes.
ASTM Special Publication 477, ASTM, Philadelphia, pp. 154 - 173.

Cording E.J. and Deere D.U. (1972):


Rock tunnel supports and field measurements.
Proc. Rapid Excavation and Tunnelling Conf. 1972, pp. 601-622.

Cording E.J., Hendron A.J. and Deere D.U. (1972):


Rock engineering for underground caverns.
Proc. Symp. Underground Rock Chambers, ASCE, pp. 567-600.

Cording E.J. and Mahar J.M. (1974):


The effect of natural geologic discontinuities on behavior of rock in tunnels.
Proc. Rapid Exc. & Tunn. Conf., AIME., pp. 107-138.

Cottiss G.I., Dowel R.W. and Franklin J.A. (1971):


A rock classification system applied in civil engineering, part 1.
Civil Engn. and Public Works Review, June 1971, pp. 611-614.

Cottiss G.I., Dowel R.W. and Franklin J.A. (1971):


A rock classification system applied in civil engineering, part 2.
Civil Engn. and Public Works Review, July 1971, pp. 736-743.

Cowie P.A. and Scholz C.H. (1992):


Displacement-length scaling relationship for faults; data synthesis and discussion.
J. of Struct. Geol., Vol. 14, No. 10, pp. 1149 - 1156.

Cruden D.M. (1977):


Describing the size of discontinuities.
Int. J. Rock Mech. and Min. Sci, No 14, pp. 133-137

Darling P. (1991):
Tunnelling through trouble in the Rockies.
Tunnels & Tunnelling, March 1991, pp. 36-38.

Dearman W.R. and Fookes P.G. (1974):


Engineering geological mapping for civil engineering practice.
Quart. J. Engn. Geol., Vol 7, pp. 223-256.

Dearman W.R. (1991):


Engineering geological mapping.
Butterworth - Heinemann Ltd., Oxford.

Deere D.U. (1963):


Technical description of rock cores for engineering purposes.
Felsmechanik un Ingenieurgeologie, Vol. 1, No 1, pp. 16-22.

Deere D. and Miller R.D. (1966):


10-8

Engineering classification and index properties for intact rock.


Univ. of Illinois, Tech. Rept. No. AFWL-TR-65-116, 1966.

Deere D.U., Peck R.B., Monsees J.E. and Schmidt B. (1969):


Design of tunnel liners and support system.
Office of high speed ground transportation, U.S. Department of transportation. PB 183799.

Deere D.U. (1971):


The foliation shear zone - an adverse engineering geologic feature of metamorphic rocks.
Boston Soc. Civ. Engn., Vol 60, No. 4, pp. 163-176.

Deere D.U., Merritt A.H. and Cording E.J. (1974):


Engineering geology and underground construction.
General report, 2nd Intn. Congr. of Int. Assoc. of Engn. Geol., Sao Paulo, Brazil, 1974, pp VII-GR.
1-26.

Denkhaus H.G. (1965):


Strength of rock material and rock systems.
Int. J. Rock Mech. Mining Sci., Vol 2, pp 111-126. 1965

Dershowitz W.S., Einstein H.H. (1988):


Characterizing rock joint geometry with joint system models.
Rock Mech. and Rock Engn., Vol 21, 1988, pp 21 - 51.

Dowding C.H. and Miller J.B. (1975):


Comparison of predicted and encountered geology for seven Colorado tunnels.
Report prepared for NSF, MIT Department of Civil Engineering, R 75-6, 1975.

Dowding C.H. (1978):


Future challenges in site characterization.
Site Characterization and Exploration, NSF Site Characterization Workshop, ASCE.

Duddeck H. (ed.) (1988):


Guidelines for the design of tunnels.
ITA Working Group on General Approaches to the Design of Tunnels.
Tunneling and Underground Space Technology 3, pp 237-249.

Einstein H., Steiner W. and Baecher G.B. (1979):


Assessment of empirical design methods for tunnels in rock.
RETC 1979, pp. 683-705.

Einstein H.H. and Baecher G.S. (1982):


Probabilistic and statistical methods in engineering geology. I. Problem statement and introduction
to solution.
Rock Mechanics, Supp. 12, pp. 47-61.

Einstein H.H. and Baecher G.B. (1983):


Probabilistic and statistical methods in engineering geology.
Rock Mechanics, Vol. , pp 3972.

Einstein H.H. (1991):


10-9

Observation, quantification and judgement: Terzaghi and engineering geology.


J. Geotech. Engn., Vol. 117, No. 11, pp. 1772-1778.

Einstein H.H. (1993):


Swelling rock.
ISRM News, No. 2, pp. 57-60.

Eriksson C. and Krauland N. (1975):


Rock mechanical views regarding the slide in the hanging wall in the Långsele mine.
Internal note Boliden, Sweden, 15 pp.

Eshwaraiah H.V. and Upadhyaya V.S (1990):


Influence of rock joints in performance of major civil engineering structures.
Proc. of Mechanics of Jointed and Faulted Rock. Balkema publ. pp. 951-968.

Ewan V.J., West G., Temporal J. (1983):


Variation in measuring rock joints for tunnelling.
Tunnels & Tunnelling, April 1983, pp 15 -18.

Fairhurst C. (1988):
Foreword.
Int. J. Rock Mech. Min. Sci & Geomech. Abstr., Vol. 25, No. 3. pp v - viii, 1988.

Faria Santos C. and Bieniawski Z.T. (1989):


Floor design in underground coal mines.
Rock Mechanics and Rock Engineering 22, 1989, 249-271.

Farmer I.W. (1977):


Rock engineering - applications in tunnelling.
Tunnels & Tunnelling, July 1977, pp. 84-88.

Farmer I.W. and Kemeny J.M. (1992):


Deficiencies in rock test data.
Proc. Int. Conf. Eurock '92, Thomas Telford, London, pp. 298-303.

Fisher P. and Banks D. (1978):


Influence of the regional geologic setting on site geologic features.
Proc. Site Charact. and Expl. ASCE, New York, pp. 163-185.

Franklin J.A. (1970):


Observations and tests for engineering description and mapping of rocks.
Proc. 4th Congress ISRM, Beograd 1970, 6 pp.

Franklin J.A., Broch E. and Walton G. (1971):


Logging the mechanical character of rock.
Tran. Inst. Min. Metall. A80, A1-A9.

Franklin J.A. (1975):


Safety and economy of tunneling.
Proc. Tenth Canadian Rock Mech. Symp., Kingstone, pp. 27-53.
10-10

Gardener R. (1992):
Seismic refraction as a tool in the evaliation of rock quality for dredging and engineering purposes:
case studies.
Proc. Intn. Symp. Eurock’92, Thomas Telford, London, pp. 153 - 158.

Geological Society Engineering Group Working Party (1971):


Report on the logging of rock cores for engineering purposes.
Q. J. Engn. Geol., 3, pp. 1-24.

Ghosh D.K. and Srivastava M. (1991):


Point-load strength: An index for classification of rock material.
Bull. Int. Ass. Engn. Geol., No. 44, pp. 27 - 33.

Gillespie P.A., Walsh J.J. and Watterson J. (1992):


Limitations of dimension and displacement data from single faults and the consequences for data
analysis and interpretation.
J. of Struct. Geol., Vol. 14, No. 10, pp. 1157 - 1172.

Golser J. (1979):
Another view of the NATM.
Tunnels and Tunnelling, March 1979, pp. 41.

Goodman R.E. (1970):


The deformability of joints.
Determination of the in-situ modulus of deformation of rock; Amer. Soc. Test & Mats., STP 477,
pp. 174-196.

Goodman R.E. and Shi G.H. (1985):


Block theory and its application to rock engineering.
Prentice-Hall, Englewood Cliffs, NJ

Goodman R.E. (1989):


Introduction to rock mechanics.
John Wiley & Sons, New York, 561 pp.

Graham P.C. (1976):


Rock exploration for machine manufacturers.
Proc. Symp. on Exploration for Rock Engineering, Johannesburg, pp. 173 -180.

Greminger M. (1982):
Experimental studies of the influence of rock anisotropy on size and shape effects in point-load
testing.
Int. J. Rock Mech. Min. Sci. & Geomech. Abstr. 19, pp. 241-246.

Grimstad E. (1993):
Private communication.

Grimstad E. and Barton N. (1993):


Updating the Q-system for NMT.
Proc. Int. Symp. on Sprayed Concrete, Fagernes, Norway 1993, Norwegian Concrete Association,
Oslo, 20 pp.
10-11

Grujic N. (1974):
Ultrasonic testing of foundation rock.
Proc. 4th ISRM Intn. Congr., Denver

Guralnik D.B. (1980):


Webster’s new world dictionary.
Second college edition, Simon and Schuster, New York, 1690 pp.

Gu D. and Wang S. (1981):


On the engineering geomechanics of rock mass structure.
Bull. IAEG, No. 23, pp. 109-111.

Haak A. (1987):
Where are the limits if the new Austrian tunnelling method?
Extracts from a discussion involving representatrives from the National committees of Austria,
Switzerland and the Federal Republic of Germany. Tunnel 3/87, pp. 126-128.

Habenicht H. (1979):
Description of joint systems taken fron drill cores. (in German)
Rock Mechanics, No. 11, pp. 217-242.

Hack H.R.G.K. and Price D.G. (1993):


A rock mass classification system for the design and safety analysis of slopes.
Proc. Int. Conf. Eurock ’93, Balkema, Rotterdam, pp.803-810.

Hagenhofer F. (1990):
NATM for tunnels with high overburden.
Tunnels and Tunnelling, May 1990, 2 pp.

Hagerman T.H. (1966):


Different types of rock masses from rock mechanics point of view.
Rock Mechanics and Engineering Geology, Vol IV, No. 3, 1966, pp. 183-198.

Haimson B.C. (1978):


The hydrofracturing stress measuring method and recent field results.
Int. J. Rock Mech. Mi. Sci. & Geomech. Abstr., Vol. 15, pp. 167 - 178.

Haimson B.C. (1993):


Scale effects in rock stress measurements.
Proc. Int. Conf. on Scale Effects in Rock Masses, pp. 89 - 101.

Hansagi I. (1965):
Numerical determination of mechanical properties of rock and of rock masses.
Int. J.Rock Mech.Mining Sci., Vol 2, pp. 219-223

Hansagi I. (1965b):
The strength properties of rocks in Kiruna and their measurements (in Swedish).
Ingeniörsvetenskapsakademiens meddelande, IVA no.142, Stockholm, pp. 128-143.
10-12

Hansen T.H. (1988):


Rock properties.
Norwegian Rock and Soil Assoc., Publ. no 5, 3 pp.

Helfrich H.K., Hasselström B. and Sjögren B. (1970):


Complex geoscience investigation programmes for siting and control of tunnel projects.
The Technology and Potential of Tunnelling, Vol. 1, N.G.W. Cook, editor; Johannesburg.

Helfrich H.K. (1971):


Mapping of the rock strength by refraction seismic measurements (in Swedish).
IVA Report 38, pp. 25-35.

Herget G. (1982):
Probabilistic slope design for open pit mines.
Rock Mechanics, Suppl. 12, 163-178, 1982.

Hodgson R.A. (1961):


Classification of structures on joint surfaces.
American J. of Science, Vol 259, 1961, pp 493 - 502.

Hoek E. (1965):
Rock fracture under static stress conditions.
Nat. Mech. Engg. Res. Inst. Report MEG 383, CSIR; S. Africa, 1965, 200 pp.

Hoek E. and Londe P. (1974):


General report, surface workings in rock.
Proc. Third Int. Congr. on Rock Mech., Denver.

Hoek E. and Brown E.T. (1980):


Underground excavations in rock.
Institution of Mining and Metallurgy, London 1980, 527 pp.

Hoek E.: (1981):


Geotechnical design of large openings at depth.
Rapid Exc. & Tunn. Conf. AIME 1981.

Hoek E. (1982):
Geotechnical considerations in tunnel design and contract preparation.
Trans. Inst. Min. Metall., London, Vol. 91, pp. A101-119.

Hoek, E. (1983):
Strength of jointed rock masses.
The Rankine Lecture 1983, Geotechnique 33, no 3 pp 187-223

Hoek E.(1986):
Practical rock mechanics - development over the past 25 years
Keynote address delivered 24.2.1986

Hoek E. and Brown E.T. (1988):


The Hoek-Brown failure criterion - a 1988 update.
10-13

Proc. 15th Canadian Rock Mechanics Symp. 1988, pp. 31-38.

Hoek E., Wood D. and Shah S. (1992):


A modified Hoek-Brown failure criterion for jointed rock masses.
Proc. Int. Conf. Eurock ’92, Chester, England, pp. 209-214.

Hoek E. (1994):
Strength of rock masses.
News Journal of ISRM, Vol. 2, No. 2, pp. 4-16.

Hoek E. (1994):
The challenge of input data for rock engineering.
Letter to the editor. ISRM, News Journal, Vol. 2, No. 2, 2 pp.

Houghton D.A. (1976):


The role of rock quality indices in the assessment of rock masses.
Proc. of the Symp. on Exploration for rock engineering, Johannesburg, South Africa,
pp. 129-135.

Hudson J.A., Priest S.D. (1979):


Discontinuities and rock mass geometry.
Int. J. Rock Mech. Min. Sci & Geomech. Abstr., Vol 16, 1979, pp 339 - 362.

Hudson J.A. and Priest S.D. (1983):


Discontinuity frequency in rock masses.
Int. J. Rock Mec. Min. Sci. & Geomech. Abstr., Vol 20, No 2, pp. 73-89, 1983.

Hudson J.A. (1989):


Rock mechanics principles in engineering practice.
CIRIA Ground Engineering report, 72 pp.

Hustrulid W.A. (1971):


A comparison of laboratory cutting results and actual tunnel boring performance.
Mining Dept., Colorado School of Mines, Golden, Colorado.

Ikeda K. (1970):
A classification of rock conditions for tunneling.
Quarterly Reports, Vol. 11, No. 2, pp. 71-74.

Ilsley R.C., Costello M.J. (1983):


Discontinuity characterization for underground openings in the Milwaukee water pollution abatement
program.
Underground Space, Vol 7, 1983, pp 214 - 220.

International Society for Rock Mechanics (ISRM), Commission on standardization of laboratory and
field tests (1971):
Suggested methods for determining the slaking, swelling, porosity, density and related rock index
properties.
Int. Soc. Rock Mech. secretary, Lisbon.
10-14

International Society for Rock Mechanics (ISRM), Commission on "Definition of the most promising
lines of research" (1971):
Final report. Int. Soc. Rock Mech. secretary, Lisbon.

International Society for Rock Mechanics (ISRM), Commission on standardization of laboratory and
field tests (1972):
Suggested methods for determining the uniaxial compressive strength of rock materials and the point
load strength index.
Committee on laboratory tests. Int. Soc. Rock Mech., Lisbon.

International Society for Rock Mechanics (ISRM), Commission on Terminology, Symbols and
Graphic Representation.(1975):
Terminology.
Int. Soc. Rock Mech. secretary, Lisbon.

International Society for Rock Mechanics (ISRM), Commission on standardization of laboratory and
field tests (1978):
Suggested methods for the quantitative description of discontinuities in rock masses.
Int. J. Rock Mech. Min. Sci. & Geomech. Abstr., Vol. 15, No. 6, pp. 319-368.

International Society for Rock Mechanics (ISRM), Commission on classification of rocks and rock
masses (1980):
Basic geotechnical description of rock masses.
Int. J. Rock Mech. Min. Sci. & Geomech. Abstr., Vol. 18, pp. 85-110.

International Society for Rock Mechanics (ISRM) working groups (1981):


Rock characterization, testing and monitoring.
Brown E.T., editor, Pergamon Press, New York, 211 pp.

International Society for Rock Mechanics (ISRM), Commision on Testing Methods (1981):
Suggested methods for determining the uniaxial compressive strength and deformability of rock
materials.
Int. Soc. Rock Mech., secretary, Lisbon., 5 pp.

International Society for Rock Mechanics (ISRM), Commssion on swelling rock (1983):
Characterization of swelling rock.
Int. Soc. Rock Mech. secretary, Lisbon.

International Society for Rock Mechanics (ISRM), Commision on Testing Methods (1985):
Suggested method for determining point load strength.
Int. J. Rock Mech. Min. Sci. & Geomech. Abstr., Vol. 22, pp. 51-60.

International Society for Rock Mechanics, Commission on testing methods (1989):


Suggested method for large scale sampling and triaxial testing of jointed rock.
Int. J. Rock Mech. Min. Sci. & Geomech Abstr., Vol. 26, No. 5, pp. 427- 434.

International Society for Rock Mechanics (ISRM), Commission on failure mechanisms around
underground excavations, First report (1989):
Observations, researches and recent results about failure mechanisms around single
galleries.
Int. Soc. Rock Mech. secretary, Lisbon.
10-15

International Tunnelling Society (ITA) (1990):


ITA Recommendations on contractual sharing of risks.
Tunnels and Underground Space Technology, Vol. 5, No. 4, 1990

Isaksen V. and Solberg E. (1990):


Fullface boring at Svartisen power plant (in Norwegian)
Final dissertation work University of Trondhein, Norway

Jaeger J.C. (1969):


Behavior of closely jointed rock.
Proc. 11th Symp. Rock Mech., pp. 57 - 68.

Janelid I. (1965):
Rock mechanics and its significance in mine and rock excavation design (in Swedish).
Royal Academy of Engineering Sciences, Report 142, pp 7-12. 1965

Jennings J.E. and Robertson A.MacG. (1969):


The stability of slopes cut into natural rock.
Conf. on Soil Mechanics and Foundation Engineering, Mexico, Vol. II, pp. 585-590.

John K.W.: (1969):


Civil engineering approach to evaluate strength and deformability of regularly jointed rock
11th int. symp. on rock mech. pp. 69-80

John K.W. and Baudendistel M. (1981):


A compromise approach to tunnel design.
22nd US Symp. on Rock Mechanics, 1981, pp. 333-341.

Judd W.R. and Huber C. (1961):


Correlation of rock properties by statistical methods.
Inc. Symp. on Mining Research, Missouri, 1961, pp. 621-648.

Kaiser P.K., MacKay C. and Gale A.D. (1986):


Evaluation of rock classifications at B. C. Rail Tumbler Ridge Tunnels.
J. Rock Mech. and Rock Engn., Vol. 19, pp. 205-234.

Karmis M. (ed.) (1986):


Application of rock characterization in mine design.
SME Publication, Littletown Co.

Kikuchi K., Kobayashi T., Inoue M. and Izumiya Y. (1985):


A study on the quantitative estimation of joint distribution and the modelling of jointed rock masses.
Tokyo Electric Power Services Co., Engineering geological department, Civil operation center. 10 pp.

Kirkaldie L. (1988):
Rock classification systems for engineering purposes.
STP 984, Amer. Society for Testing Materials, 167 pp.

Kleeberger J. (1992):
Private communication.
10-16

Knill J.L. (1969):


The application of seismic methods in the prediction of grout take in rock.
Proc. Int. Conf. In Situ Investigations in Soil and Rocks, pp. 63-70.

Koerner U. (1971):
Critical notes on rock classification in underground construction from a geological point of view.
Die Bautechnik, No. 9, pp. 318-319.

Krauland N., Söder P. and Agmalm G. (1989):


Determination of rock mass strength by rock mass classification - Some experiences and questions
from Boliden mines.
Int. J. Rock Mech. Min. Sci & Geomech. Abstr., Vol. 26, No 1, pp 115 - 123.

Krauland N. (1992):
Private communication.

Krumbein W.C. and Greybill F.H. (1965):


An introduction to statistical models in geology.
McGraw-Hill, Inc. New York, 475 pp.

Lama, R.D. and Vutukuri, V.S. (1978):


Handbook on mechanical properties of rocks. Trans Tech Publications, Clausthal, Germany, 1978,
1650 p.

Lane K.S. (1948):


Discussion on A.M. Casagrande: ' Classification and identification of soils'.
Trans. of Am. Soc. of Civil Engn., Vol. 113, pp. 950 - 951.

LaPointe P.R. (1988):


A method to characterize fracture density and connectivity through fractal geometry.
Int. J. Rock Mech. Min. Sci. & Geomech. Abstr., Vol 25, No. 6., pp. 421-429.

Lardelli T. (1992):
Private communication.

Lauffer H. (1958):
Classification for tunnel construction (in German)
Beologie und Bauwesen, Vol. 24, No. 1, pp 46-51.

Lauffer-Innsbruck H. (1988):
On rock classification regarding cutting. (in German)
Felsbau, Vol. 6, pp. 137-149.

Lindblom U.E. (1986):


Developments in design methods for large rock caverns.
General report, Proc. Symp. of Large Rock Caverns, Helsinki, 1986, pp 1937-1948.

Lislerud A. (1988):
10-17

Hard rock tunnel boring: Prognosis and costs.


Tunnelling and Underground Space Technology, Vol. 3, No. 1, pp. 9 - 17.

Louis C. and Perrot M. (1972):


Three dimensional investigation of flow conditions at Grand Maison Damsite.
Proc. Symp. on Percolation Through Fissured Rock, Stuttgart

Löset F. (1990):
Use of the Q-method for securing small weakness zones and temporary support.(in Norwegian)
Norwegian Geotechnical Institute, internal report No. 548140-1, 40 pp.

Löset F. (1992):
Support needs compared at the Svartisen road tunnel.
Tunnels & Tunnelling, June 1992, 3 pp.

Madan M.M. (1991):


An analytical approach to tunnel construction.
Tunnels & Tunnelling, May 1991, pp. 71-74.

Mahtab M.A., and Yegulalp T.M. (1986):


Characterizing jointed rock for tunnel design.
Proc. of Conf. on Large Underground Openings, Florence, Italy, pp. 613-619.

Martin D. (1988):
TBM tunnelling in poor and very poor rock conditions.
Tunnels and Tunnelling, March 1988, pp. 22-28.

Matula M. (1969):
Engineering geologic investigations of rock heterogeneity.
11th US Symp. on Rock Mech. pp 25-42.

Matula M. and Holzer R. (1978):


Engineering topology of rock masses.
Proc. of Felsmekanik Kolloquium , Grundlagen ung Andwendung der Felsmekanik, Karls-
ruhe, Germany, 1978, pp. 107-121.

Maury V. (1976):
An example of underground storage in soft rock (chalk).
Proc. Int. Symp. Rock Store, pp. 681-689.

McFeat-Smith I., Nieuwenhuijs G.K. and Lai W.C. (1986):


Application of seismic surveying, orientated drilling and rock classification for site investigation of
rock tunnels.
Proc. Int. Conf. Rock Engineering and Excavation in an Urban Environment, MIT., pp. 249-261.

Merritt A.H. and Baecher G.B. (1981):


Site characterization in rock engineering.
22nd U.S. Symp. on Rock Mechanics, pp. 49-66

Milne D., Germain P. and Potvin Y. (1992):


Measurement of rock mass properties for mine design.
10-18

Proc. Int. Conf. Eurock ’92, Thomas Telford, London, pp.245-250.

Moreno Tallon E. (1980):


Application of the geomechanical classification for the tunnels at Pajares. (in Spanish).
Curso de Sostenimientos activos en Galerias y Tuneles, Madrid, Fundacion Gomez-Pardo.

Moreno Tallon, E. (1987):


Rock-masses characterization for underground excavation purposes
VI Australian Tunnelling Conference, Melbourne 1987.

Morfeldt C.-O., Bergman M. and Lundström L. (1973):


Rock mass investigations; evaluation of methods (in Swedish).
Swedish Building Research, report R34:1973, 116 pp.

Morfeldt C.-O. (1976):


Caverns and tunnels in rock; engineering and geological follow-up (in Swedish).
Swedish Building Research, report R15:1976, 106 pp.

Movinkel T. and Johannessen O. (1986):


Geologic parameters for hard rock tunnel boring.
Tunnels & Tunnelling, April 1986, pp. 45-48.

Muir Wood A.M. (1979):


Ground behaviour and support for mining and tunnelling.
Tunnels and Tunnelling, Part 1 in May 1979 pp. 43-48, and Part 2 in June 1979, pp. 47-51.

Müller L. (1963):
Rock construction (in German).
Ferdinand-Enke-Verlag, Stuttgart, 624 pp.

Müller L., Bock H. and Müller K. (1970):


Structural geology of rocks - rock mechanics in construction (in German).
Wilhelm Ernst & Sohn Verlag, Berlin

Müller L. (1978):
Removing misconceptions on the new Austrian tunnelling method.
Tunnels and Tunnelling, Oct. 1978, pp. 29-32.

Müller L. (1982):
The influence of engineering geology and rock mechanics in tunnelling.
Proc. of IV Congr. Intn. Ass. of Engn. Geol., New Delhi, 1982, pp. ix 177 - 186.

Mutschler T. and Natau O. (1991):


Further developments for the determination of the stress-strain behaviour of jointed rock mass by large
scale tests.
Proc. 7th. Congr. of ISRM, Aachen, pp. 1557-1560.

Mutschler T. (1993):
Private communication.

Nakano R. (1979):
10-19

Geotechnical properties of mudstone of Neogene Tertiary in Japan.


Proc. Int. Symp. Soil Mechanics, Oaxaca, pp. 75 - 92.

Narr, W., Suppe, J. (1991):


Joint spacing in sedimentary rocks.
J. of Structural Geol., Vol 13, No. 9, pp 1037 - 1048.

Natau O.P., Frölich B.O. and Mutschler T. (1983):


Recent developments of the large-scale triaxial test.
Proc. 7th. Congr. of ISRM, Melbourne, pp. 1557-1560.

Natau O. (1990):
Scale effects in determination of the deformability and strength of rock masses.
Proc. Intn. Conf. on Scale Effects in Rock Masses, pp. 77 - 88.

Natau O., Bühler, Keller S. and Mutschler T. (1995):


Large scale Triaxial testss in connection with a FEM analysis for the determination of the properties of
a transversal isotropic rock mass.
Proc. 8th Int. Congress on Rock Mechanics, Tokyo, 9 pp.

Nieto A.S. (1983):


Some geologic factors in the location design and construction of large underground chambers in rock.
Proc. Rapid Exc. & Tunn.Conf. AIME 1983, pp 569-596.

Nilsen B. and Thidemann A. (1993):


Rock engineering.
Hydropower Development, publ. no. 9, Norwegian Institute of Technology, Division of Hydraulic
Engineering, 156 pp.

Nilsen B. and Ozdemir L. (1993):


Hard rock tunnel boring prediction and field performance.
Rapid Excavation and tunneling Conference, 20 pp.

Nord G., Olsson P. and By T.L. (1992):


Probing ahead of TBMs by geophysical means.
Tunnelling and Underground Space Technology, Vol. 7, No. 3, pp. 237-242.

Norsk Bergmekanikkgruppe (1985):


Handbook in engineering geology - rock. (in Norwegian)
Tapir, Trondheim, Norway, 140 pp.

Norwegian Institute of Technology (1994):


Fullface boring of tunnels (in Norwegian).
PR 1-94, Trondheim Norway, Norwegian Institute of Technology, 159 pp.

Nystuen, J.P., (1989):


Naming geological units in Norway,
Norsk Geologisk Tidsskrift, Vol. 69, Suppl.

Obermeier S.E. (1974):


Evaluation of laboratory techniques for measurement of swell potential of clays.
10-20

Bull. Ass. of Engn. Geol., Vol XI, No. 4, pp. 293-314.

Olivier H.J. (1976)


Importance of rock durability in the engineering classification of Karoo rock masses for tunnelling.
Proc. symp. on Exploration for Rock Engineering, Johannesburg 1976, pp 137 - 144.

Pacher F. (1975):
The development of the New Austrian Tunnelling Method and the main features in design work and
construction.
16th Symp. on Rock Mechanics, Minneapolis, pp. 223-232.

Pacher F. (1978):
The conseption of safety in special cases of rock construction. (in German)
Proc. Felsmechanik Kolloquium Karlsruhe 1978, Trans Tech Publ. pp. 27-44.

Palmström, A. (1974):
Characterization of the degree of jointing and the quality of rock masses (in Norwegian). Internal
report Ing. A.B. Berdal, Hövik, Norway, 26 pp.

Palmström, A. (1982):
The volumetric joint count - a useful and simple measure of the degree of jointing.
Proc. IV Int. Congr. IAEG, New Delhi, 1982, pp V.221-V.228.

Palmström A. (1984):
Geo-investigation and advanced tunnel excavation technique important for the Vardö subsea road
tunnel.
Proc. Int. Symp. Low Cost Road Tunnels, Oslo, Norway, 1985, 16 pp.

Palmström A. (1985):
Application of the volumetric joint count as a measure of rock mass jointing.
Proc. Int. Symp. Fundamentals of Rock Joints, Björkliden, Sweden, 1985, pp 103-110.

Palmström A. (1986):
The volumetric joint count as a measure of rock mass jointing.
Presented at the Conference on Fracture, Fragmentation and Flow, Jerusalem 1986, 19 pp.

Palmström A. (1986):
A general, practical method for identification of rock masses to be applied in evaluation of rock mass
stability conditions and TBM boring progress. (in Norwegian)
Proc. Conf. on Fjellsprengningsteknikk, Bergmekanikk, Geoteknikk, Oslo Norway,
pp. 31.1-31.31.

Palmström A. and Berthelsen O. (1988):


The significance of weakness zones in rock tunnelling.
Proc. Int. Conf. Rock Mechanics and Power Plants, Madrid 1988, 8 pp.

Papadopoulos Z. and Marinos P. (1992):


On the anisotropy of the Athenian schist and its relation to weathering.
Bull. Int. Ass. Engn. Geol., No. 45, pp. 111 - 116.
10-21

Patching T.H. and Coates D.F. (1968):


A recommended rock classification for rock mechanics purposes.
CIM Bull., Oct. 1968, pp 1195-1197.

Patton F.D. (1966):


Multiple modes of shear failure in rock.
Proc. 1st Congr. ISRM, Lisbon, 1966, I: pp. 509-513.

Patton F.D., Deere D.U. (1970a):


Significant geologic factors in slope stability.
Proc. Symp. Planning Open Pit Mines, Johannesburg, pp 143-151.

Pearson J.R.A. (1988):


Key questions in rock mechanics.
Proc. 29th U.S. Rock Mechs. Symp., Minneapolis, pp. 7-15.

Peck R.B. (1980):


Where has all the judgement gone?
Laurits Bjerrum memorial lesson no 5, Norwegian Geotechnical Institute, Oslo, Norway

Peters C.M.F. (1972):


A structural interpretation of the Garlock fault at the Tehachapi crossing.
Proc. Rapid Excav. & Tunn. Conf., AIME., pp. 133-155.

Piteau D.R. (1970):


Geological factors significant to the stability of slopes cut in rock.
Proc. Symp. on Planning Open Pit Mines, Johannesburg, South Africa, 1970, pp. 33-53.

Piteau D.R. (1973):


Characterizing and extrapolating rock joint properties in engineering practice.
Rock Mechanics, Suppl. 2, pp. 5-31.

Poisel R. (1990):
The dualism-continuum of jointed rock.
Proc. Mech. of Jointed and Faulted Rock, 1990, Balkema publ. pp 41-50.

Pollard,D. D., Aydin, A. (1988):


Progress in understanding jointing over the past century.
Bull. Geol. Society of America, Vol 100, pp 1181 - 1204.

Price N.J. (1969):


Laws of rock behavior in the earth’s crust.
11th Symp. on Rock Mech. Berkley, pp. 3-23.

Price N.J. (1981):


Fault and joint development in brittle and semi-brittle rock.
Pergamon Press, 1981, 176 pp.

Proctor R.J. (1971):


Mapping geological conditions in tunnels
10-22

Bull. Ass.Engn.Geol., Vol VIII, No. 1, PP. 1-31.

Pusch R. and Morfeldt C.-O. (1993):


Characterization of rock masses for construction of underground openings from numeric calculation of
stresses, deformations and ground water flow. (in Swedish)
Väg och vattenbyggaren, no. 4/93, pp. 13 - 18.

Rabcewicz L.v. (1964/65):


The new Austrian tunnelling method.
Water Power, part 1 November 1964 pp. 511-515, Part 2 January 1965 pp. 19-24.

Rabcewicz L.v. (1975):


Tunnel under Alps uses new, cost-saving lining method.
Civil Engineering-ASCE, October 1975, pp.66-68.

Rabcewicz L.v. and Golser J. (1973):


Principles dimensioning the support system for the new Austrian tunnelling method.
Water Power, March 1973, pp. 88-93

Ramamurthy T., Venkatappa Rao G. and Singh J. (1993):


Engineering behaviour of phyllites.
Engineering Geology, 33, pp. 209 - 225.

Robbins R.J. (1980):


Present trends and future directions in tunnelling.
The Yugoslav Symp. on Rock Mechanics and Underground Actions, 11 pp.

Robertson A.MacG. (1970):


The interpretation of geological factors for use in slope theory.
Proc.Symp. Planninmh Open Pit Mines, Johannesburg, South Africa, 1970. pp. 55-71.

Robinson C.S. (1972):


Prediction of geology for tunnel design and construction.
Rapid Tunneling and Excavation Conf., pp 105-114.

Rokoengen K. (1973):
Classification of clay zones in rock. (in Norwegian)
Report no. 11, The Technical University in Norway, Trondheim, 46 pp.
(Extract from dr. thesis on 'Swelling properties of clay zones in rock')

Rostami J. (1992):
Design optimization, performance prediction and economic analysis of tunnel boring machines for the
construction of the proposed Yucca Mountain nuclear waste repository.
Dr. thesis, Colorado School of mines, 195 pp.

Ruiz M.D. (1966):


Some technological characteristics of twenty-six Brazilian rock types.
Proc. 1st Congr. Intn. Rock Mech., Lisbon, Vol. 1, pp. 115 - 119.

Russenes B.F. (1974):


Analysis of rock spalling for tunnels in steep valley sides (in Norwegian).
10-23

M.Sc. thesis, Norwegian Institute of Technology, Dept. of Geology, 247 pp.

Rutledge J.C. and Preston R.L. (1978):


New Zealand experience with engineering classifications of rock for the prediction of rock support.
Proc. Int. Tunnelling Symposium, Tokyo, 1978, pp. A3 1-7.

Sadagah B.H., Sen Z. and Freitas M.H.D. (1990):


A mathematical representation of jointed rock masses and its application.
Proc. of Symp. on Mechanics of Jointed and Faulted Rock, 1990, pp. 65-70.

Salustowicz A. (1965):
Zarys mekaniki gorotworu Katowice.
Wydawnictwo "Slask".

Schneider H.J. (1976):


The friction and deformation behaviour of rock joints.
Rock Mechanics No. 8, pp 169-184.

Scholz C.H. (1990):


The mechanics of earthquakes and faulting.
Cambridge University Press, Cambridge, USA, 438 pp.

Seeber G., Keller S., Enzenberg A., Tagwerker J., Schletter R., Schreyer F. and Coleselli A. (1978):
Methods of measurements for rock support and installations in road tunnels using the new Austrian
tunnelling method. (in Germain)
Bundesministerium f. Bauten u. Technik, Strassenforschung Heft 133, 200 pp.

Selmer-Olsen R. (1950):
On faulting and crushed zones in the Bamble formation. (In Norwegian).
Norsk geologisk tidsskrift, No 25, 1950, pp. 171-191.

Selmer-Olsen R. (1964):
Geology and engineering geology. (in Norwegian)
Tapir, Trondheim, Norway, 409 pp.

Selmer-Olsen R. (1971):
Engineering geology. Part 1. (in Norwegian)
Tapir, Trondheim, Norway, 230 pp.

Selmer-Olsen R. (1988):
General engineering design procedures.
Norwegian Tunnelling Today, Tapir 1988, pp. 53-58

Selmer-Olsen R. and Palmström A. (1989):


Tunnel collapses in swelling clay zones, part 1.
Tunnels & Tunnelling, November 1989, pp.

Selmer-Olsen R. and Palmström A. (1990):


Tunnel collapses in swelling clay zones, part 2.
Tunnels & Tunnelling, January 1990, pp.
10-24

Sen Z., Eissa E.A. (1991):


Volumetric rock quality designation
J Geotech. Engn., Vol 117, No 9, 1991, pp 1331 - 1346.

Sen Z. and Eissa E.A. (1992):


Rock quality charts for log-normally distributed block sizes.
Int. J. Rock Mech. Min. Sci. & Geomech. Abstr., Vol. 29, No. 1, pp. 1-12.

Serafim J.L. and Pereira J.P. (1983):


Consideration of the geomechanics classification of Bieniawski.
Proc. Int. Symp. on Engeneering Geology and Undeground constructions, pp. 1133 - 1144.

Seshagiri Rao K., Venkatappa Rao G. and Ramamurthu T. (1987):


Discussion of paper by K.L. Gunsallus and F.H. Kulhawy, "A comparative evaluation of rock strength
measures".
Int. J. Rock Mech. Min. Sci. & Geomech. Abstr. Vol. 24, No. 3, pp. 193-196.

Singh J., Ramamurthy T. and Venkatappa Rao G. (1989):


Strength anisotropies in rocks.
Ind. Geotech. J., 19(2), pp. 147-166.

SINTEF (1990):
Deformation and rock stress measurements at Svartisen power plant, Efjord and Stetind road tunnels
(in Norwegian).
SINTEF Report no. STF36 F90059

SINTEF (1993):
Fullface boring Dalåa - Torsbjørka (in Norwegian)
Final report made for Merkraft A/S (courtesy Merkraft A/S)

Sjögren B., Övsthus A. and Sandberg J. (1979):


Seismic classification of rock mass qualities.
Geophysical Prospecting, Vol. 27, No. 2, pp. 409-442.

Sjögren B. (1984):
Shallow refraction seismics.
Chapman and Hall, London, 270 pp.

Sjögren B. (1993):
Private communication.

Snow D. (1966):
Disc'n.
Theme III, Int. Congr. Rock Mech., Lisbon.

Snow D. (1968):
Fracture deformation and changes of permeability and storage upon changes of fluid pressure.
Quarterly Colorador School of Mines, 63, pp. 201-244
10-25

Spaun G. (1974):
Zur Frage der Sohlhebungen in Tunneln des Gipskeupers.
In: Festschrift Leopold Müller-Salzburg zum 65. Geburtstag, pp. 245-260.

Stang T. and Aadal T. (1991):


Fullface boring at Svartisen power plant (in Norwegian)
Final dissertation work University of Trondhein, Norway

Stephenson D.E. and Triandafilidis G.E. (1974):


Influence of specimen size and geometry on uniaxial compressive strength of rock.
Bull. Ass. of Engn. Geol., Vol Xi, No. 1, 1974, pp. 29-47.

Stimpson B.:(1982)
A rapid field method for recording joint roughness profiles. Technical note.
Int. J. Rock Mech. Min. Sci & Geomech. Abstr., Vol 19 pp 345-346, 1982

Stini I. (1950):
Geology in tunnel construction (in German).
Springer publishers, Vienna, 366 pp.

Swan G. (1981):
Tribology and the characterization of rock joints.
Proc. 22nd US Symp.on Rock Mechanics, MIT. 1981, pp. 432-437.

Söder P.-E. and Krauland N. (1990):


Determination of pillar strength by full scale pillar tests in the Laisvall mine.
Proc. 11th World Mining Congr. on Strata Control in Deep Mines, pp. 39 - 59.

Terzaghi K. (1946):
Introduction to tunnel geology.
In Rock tunnelling with steel supports, by Proctor and White, pp 5 - 153.

Terzaghi K. (1953):
Address.
Proc. Int. Conf. on Soil Mech. and Foundation Engn., Vol. 3, p. 76.

Terzaghi R. (1965):
Sources of error in joint surveys.
Geotechnique, Vol 15, 1965, pp 287-304.

Thorgrimsson S., Loftsson M. and Jensson O. (1991):


Iceland's Blanda hydroelectrical project: Monitoring of deformations, rock support and testing of rock
anchors in the powerhouse cavern.
Tunnelling. and Underground Space Technology, Vol. 6, No. 2, pp.235-239.

Thorpe R., Watkins D.J., Ralph W.E., Hsu R. and Flexser S. (1980):
Strength and permeability tests on ultra-large Stripa granite core.
Technical Information Report No.31, Lawrence Berkeley Laboratory, University of California, 200 pp.

Tirén S.A. and Beckholmen M. (1992):


10-26

Rock block map analysis of southern Sweden.


Geologiska Föreningens i Stockholm Förhandlingar, Vol. 114, Pt. 3, pp 253 - 269.

Tourtelot H.A. (1974):


Geologic origin and distribution of swelling clays.
Bull. Ass. of Engn. Geol., Vol XI, No. 4, 1974, pp. 259-275.

Tsidzi K.E.N. (1986):


A quantitative petrofabric characterization of metamorphic rocks.
Bull. Int.Assoc. Engng. Geol. no 33, pp. 3 - 12.

Tsidzi K.E.N. (1987):


Foliation index determination for fine-grained metamorphic rocks.
Bull. Int.Assoc. Engng. Geol. no 36, pp. 27 - 33.

Tsidzi K.E.N. (1987):


Compressive strength anisotropy of foliated rocks.
Proc. symp. on Mechanics of Jointed and Faulted Rock; Balkema, Rotterdam,
pp. 421 - 428.

Tsoutrelis C.E., Exadactylos G.E. and Kapenis A.P. (1990):


Study of the rock mass discontinuity system using photoanalysis.
Proc. of Symp. on Mechanics of Jointed and Faulted Rock, 1990, pp. 103-112.

Turk N., Dearman W.R. (1985):


Investigation of some rock joint properties: Roughness angle determination and joint closure.
Proc. of Int. Symp. on Fundamentals of Rock Joints, Björkliden Sweden 1985, pp 197 -204.

Unal E. (1983):
Design guidelines and roof control standards for coal mine roofs.
Ph.D. Thesis, The Pennsylvania State University, 1983.

Wagner H. (1987):
Design and support of underground excavations in highly stressed rock.
Proc. of 6th ISRM Congr., Montreal; Keynote paper Vol. 3.

Wahlstrom, E.E. (1973):


Tunnelling in rock.
Amsterdam Elsevier, 250 p.

Wallis P.F., King M.S. (1980):


Discontinuity spacings in a crystalline rock.
Technical note, Int. J. Rock Mech. Min. Sci & Geomech Abstr., Vol 17, 1980, pp 63 - 66.

Ward W.H. (1978):


Ground supports for tunnels in weak rocks.
The Rankine Lecture. Geotechnique 28, No. 2, pp. 133-171.

Watkins M.D. (1971):


10-27

Terminology for describing the spacing of discontinuities of rock masses.


Q.J. Engn. Geol., Vol 3, 1971, pp 193 - 195.

Wickham G.E., Tiedeman H.R. and Skinner E.H. (1972):


Support determinationd based on geologic predictions.
Proc. Rapid Exc. & Tunn. Conf., 1972, PP. 43-64.

Williamson D.A. (1980):


Uniform rock classification for geotechnical engineering purposes.
Transportation Research Record 783, National Academy of Sciences, Washington D.C,
pp. 9-14

Williamson D.A. and Kuhn C.R (1988):


The unified classification system.
Rock Engineneering Systems for Engineering Purposes, ASTM STP 984, American Society for
Testing Materials, Philadelphia, pp. 7 - 16.

Wittke N. and Louis C. (1969):


Several quick tests for determining the mechanical character of rocks.
Geotechnical Colloq., Toulouse, France

Wood D. (1991):
Estimating Hoek-Brown rock mass strength parameters from rock mass classifications.
Transportation Research Record 1330, pp. 22-29.

Xuecheng D. (1993):
Rock mechanics investigations related to the three gorges dam project.
News Journal of International Society for Rock Mechanics, Vol. 1, No. 4, pp. 6-15.

You might also like