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
Dr. thesis Arild Palmström

RMi – A ROCK MASS CHARACTERIZATION

SYSTEM FOR ROCK ENGINEERING PURPOSES

GRAVEL

CO

SA
ND
IA LS BB
LE
ER
AT
BO

.
M
UL
IL

DE
SO

R
S ILT

Compressive strength of rock

0.25 1.0 5 15 50 100 250 MPa

WEAK STRONG
CLAY

ma ssive
ro ck

1.0 10 100 250 kPa

Compressive strength of soil
w te

at
e
al isi

ra h e
d

nt tio r in
er n
g g h ig h ly
d
rat jo in te
i on ro c k S
SE
AS
ROCK M
In memory of prof. Rolf Selmer-Olsen (1919 - 1989)

Many of his ideas have been quantified in this work

 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.
 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

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$7,2154'
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.

'LVFRQWLQXRXVMRLQWHG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.
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

Chapter 1

INTRODUCTION

"The corner-stone of any practical rock mechanics analysis or rock engineering is the geologi-
cal data base upon which the definition of rock types, structural discontinuities and material
properties is based. Even the most sophisticated analysis can become a meaningless exercise if
the geological information upon which it is based is inadequate or inaccurate."
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
"it is extremely important that the quality of the input data matches the design requirements.
Obviously, it must be realized that if incorrect input parameters are employed, incorrect design
information will result. Speaking in computer jargon, the following expression would be
appropriate: 'garbage in, garbage out'."

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: "At present time most geologic and petrographic descrip-
tions of specimens or bodies of rock are qualitative, whereas rock mechanics determination of
mechanical properties of rock are quantitative.
Because many engineering decisions are based on a combination of geologic and rock mechanics
data, it is important that a more systematic means of combining and correlating this information
should be developed.
There is a need for better documentation and correlation of geological and petrographic data, and
corresponding mechanical property data obtained from both laboratory specimens and/or rock
masses, together with operating experience in the same rock mass, or the subsequent performance
of structure in the rock mass created by excavation."

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. "I am more and more amazed about
the blind optimism with which the younger generation invades this field, without paying attention to
the inevitable uncertainties in the data on which their theoretical reasoning is based and without
making serious attempts to evaluate the resulting errors."

Later, Poisel (1990) points out that the same trends have continued also during the last 10 years: "A
look at rock mechanics papers shows that at present it is only important to perfect mathematical
procedures and to investigate rock-like materials. Rock mechanics, however, still needs
investigations in the field - not only testing to get its parameters - geomechanical models, and
intuition.
Regretfully we do not ask ourselves often enough if the results we reach by mathematical
calculations really apply to nature. Rock mechanical models should be determined by what nature
shows, not by what can be done within mechanical-mathematical limits. Thus it should first be
attempted to find out what is going on in nature, which procedures or which mechanisms take place
in the structure, and only then it should be considered how to model them by mechanical-
mathematical means."

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.

1.1 Outline of this work

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 rock mass. 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 Step 1:
AND MAPPING Acquisition of
geo-data
FIELD INVESTIGATIONS
AND TESTING

SELECTED
PARAMETERS Step 2:
combination of the A general
ARE GIVEN ROCK MASS INDEX
selected parameters characterization
NUMERICAL of the rock mass
VALUES
RMi

PROJECT Step 3:
Input of features
RELATED of importance for
FEATURES the actual design

ROCK DESIGN AND Step 4:

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

applied as
input in rock used in Other
mechanics and communication applications
other types of of the RMi
engineering
*) not included in this work

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):
"Rock mechanics has to provide methods of analysis which are realistic compromises between the
best representation of the actual ground conditions and pragmatic engineering. The rock mechanics
practitioner has to face the fact that many geological engineering and rock mechanics problems
may be too complex to allow rigorous analysis but at the same time are deemed satisfactory for the
construction of major structures. Qualitative evaluations, quantitative descriptions of geologic
features as such, and comparisons of specific test results are always of interest but may not suffice
as basis for engineering decisions and designs.
Neither a purely geological nor a completely technical classification of rock masses will answer its
intended purpose but suitable parameters of rock masses should be defined and quantified in rock
engineering terms."

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

Describe 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.
Characterize is to report the particular qualities, features, or traits of. Rock mass characterization 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.
Classify is to arrange or group in classes according to some system or principle. Rock mass
classification 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 characterize 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.
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

Chapter 2

ROCK MASSES AS CONSTRUCTION MATERIALS

"Rock masses are so variable in nature that the chance for ever finding a common set of
parameters and a common set of constitutive equations valid for all rock masses is quite
remote."
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
"inadequacies in site characterization of geo-data probably present the major impediment to the
design, construction and operation of excavations in rock. Improvements in site characterization
methodology and techniques, and in the interpretation of the data are of primary research require-
ments, not only for large rock caverns, but for all forms of rock engineering."

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

BASIC ELEMENT SIZE RANGE STRUCTURES CONSIDERED AREA

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

INVESTIGATIONS
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 size or volume of the material involved, see Fig. 2-2 and Table 2-1,
- the structure and composition of the material,
- the many construction and utilization purposes of it, see Table 2-2, and
- the difficulties in measuring the quality of the material (see also Appendix 4).

TABLE 2-2 MAIN TYPES OF WORKS CONNECTED TO ROCKS AND ROCK MASSES
TYPE ACTUAL PROCESS OR USE
- 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.

2.1 ROCKS AND THEIR MAIN FEATURES

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

2.1.1 Fresh rocks

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

Igneous rocks 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.

Sedimentary rocks 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.

Metamorphic rocks 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.

2.1.2 The influence from some minerals

Certain elastic and anisotropic minerals 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 swelling minerals of 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

rocks with rocks with

isotropic strongly
or slightly r o c k s w i t h r e d u c e d s t r e n g t h a n d d ur a b i li t y
anisotropic
anisotropic properties
properties

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

2.1.3 The effect of alteration and weathering

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 weathered near the surface, and are sometimes altered by hydrothermal
processes. Both processes generally first affect the walls of the discontinuities 1. The main results of
rock weathering and alteration are:
1. Mechanical disintegration 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 decomposition, 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 disintegration leads mainly to a greater number of joints in rock masses located in the upper
zone of weathering, while decomposition influences the joint condition as well as the rock material.

2.1.4 Geological names and mechanical properties of rocks

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).

2.2 DISCONTINUITIES IN ROCK

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.
ROCK DEFECTS JOINTS WEAKNESS ZONES
faults

joints
partings
cracks
fissures
bedding planes
seams / shears

0.01 0.1 1 10 100 1000 10 000

LENGTH (m)
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).

2.2.1 Faults

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.

A B C

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

2.2.2 Joints and their main features

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.

joint
ity of the
ontinu
and c
length

int
of jo
on dit ion ce:
c sur fa
wall thness ting
oo a
- sm ssible coalte ration
- po ssible k
- po ll c
ro
of wa

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.

2.2.3 The main jointing characteristics

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

2.2.4 The rock mass

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.

A
2
3
1
2

4
3
3 B

A 4
2

2 50
- 15
0m
3
500
- 15
00
m 2
5 4
B

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.

2.3 ROCK MASS CHARACTERIZATION FOR DESIGN AND CONSTRUCTION

PURPOSES

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
Name of classification Form and Type*) Main applications Reference
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
*)
Definition of the following expressions:
Descriptive form: the input to the system is mainly based on descriptions
Numerical form: the input parameters are given numerical ratings according to their character
Behaviouristic form: the input is based on the behaviour of the rock mass in a tunnel
General type: the system is worked out to serve as a general characterization
Functional type: 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
CLASSIFICATION SYSTEM NO.  1 2 3 4 5 6 7 8 9 10 11 12 13
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

CLASSIFICATION SYSTEM NO.  1 2 3 4 5 6 7 8 9 10 11 12 13

Legend:
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)
Classification system no.:
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)
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

Chapter 3

COLLECTION OF GEO-DATA - LIMITATIONS AND

UNCERTAINTIES

"The geotechnical engineer should apply theory and experimentation but temper them by
putting them into the context of the uncertainty of nature. Judgement enters through
engineering geology."
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

3.1 GEO-DATA FOUND BEFORE, DURING AND AFTER EXCAVATION

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 before
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 before 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,
during and after 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.

3.2 SOME METHODS USED IN GEO-DATA COLLECTION

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.

3.2.1 Geological observations and mapping

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)

TERM DESCRIPTION

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

collection of geo-data

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 statistical 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 sampled 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.

3.2.3.1 Limitations and deficiencies in core logging

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. "Post construction analyses of boring logs at
power plant sites indicate that the likelihood of recognizing even many inch wide faults in boring
logs is extremely small."

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 tests 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 "a surprising number of in situ testing pro-
grammes in intrusive igneous rock and high-grade metamorphics give values of static modulus
ranging between 7 and 14 GPa using bearing plates, and between 41 and 55 GPa using flat
jacks".
 3 - 11

3.3 UNCERTAINTIES AND ERRORS IN GEO-DATA COLLECTION

"In thinking about sources of uncertainty in engineering geology, one is left with the fact
that uncertainty is inevitable. One attempts to reduce it as much as possible, but it must
ultimately be faced. It is a well recognized part of life for the engineer. The question is not
whether to deal with uncertainty, but how?"
Herbert H. Einstein and Gregory B. Baecher (1982)

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

Uncertainty 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.
Error is defined as the difference between computed or estimated result and the actual value.
A bias 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 omissions. 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:
Gross errors, 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.
Systematic errors 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)
Errors of method 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 unpredictable
deviations or random errors, 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).

3.4 SUMMARY

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
FEATURES THAT MAY REDUCE MEASURE- CONSEQUENCES FOR THE MEASUREMENTS
MENT OR INTERPRETATION QUALITY

Observations in outcrops:
- 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

Observations in excavated cuttings, trenches,

adits, etc:
- weathered rocks occur in the surface (but not - the rock conditions observed are different from the
deeper) conditions in deeper located rock masses

Observations in deep located underground

openings:
- 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.

DATA COLLECTED FROM

ROCK MASS DRILL UNDER- REFRACTION
PARAMETER CORES GROUND OUTCROPS SEISMIC
ADITS
OPENINGS PROFILES
Rocks
- distribution of rocks x x x x/(x) -
- sample for strength tests x x x x/(x) -

Joints and jointing

- 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)/- -

Faults and weakness zones

- 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.
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

Chapter 4

THE COMBINATION OF GEO-DATA INTO A ROCK MASS INDEX

"Try always to combine theory and practice and to confront ideas and experience."
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 strength characterization of rock masses. Williamson and Kuhn (1988) are of the
opinion that "no classification system can be devised that deals with all the characteristics of all
possible rock materials or rock masses. What we are aiming for, therefore, is a system that would
generally group the rocks in such a way that those parameters which are of most universe concern
are clearly dealt with and the number of symbols are kept to a minimum."

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
"which are the same irrespective of place or circumstances. For this reason it was considered
necessary to omit factors related to environment from the classification, although stress
applications, pore-water and other influences have a pronounced effect on the behaviour of a rock
in any given situation. Just as a structural engineer who is designing a steel structure will establish
the stress distributions of the structure separately from the specifications of the steel, so in any
specific problem in rock mechanics the environmental factors will be considered and established
for that problem in addition to the determination of the nature or classification of the rock." 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.
4.1 THE STRUCTURE OF THE ROCK MASS INDEX

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.

4.1.1 The input parameters selected

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 main 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, the block shape, 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.

4.1.2 The Rock Mass index (RMi)

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 JO INTING
TERMINATION PARAM ETER
JP
DENSITY BLOCK VOLUME
OF JOINTS Vb
ROCK MASS INDEX
UNIAXIAL RMi
ROCK CO MPRESSIVE
MATERIAL STRENGTH
σc

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.
4.1.3 The combination of the input parameters

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 shear strength properties 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

4.2 CALIBRATION OF RMi FROM KNOWN ROCK MASS STRENGTH DATA

"The purpose of science is to simplify, not to complicate. The function of an engineering

geologist, geotechnical or rock engineer is to examine and observe the complex variables of an
area or project site and from this effort arrive at a set of simple, significant generalizations".
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.
Sample Location Rock type σc jC Vb JP
no 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

100
5

10

m3
2

1
5

Vb
2
SAMPLE 3: jC = 0.75 - 1.5
0.1

volume
SAMPLE 7: jC = 3.4 - 4.5 5

2
SAMPLE 4: jC = 0.2 - 0.3
10
Block

5
SAMPLE 5: jC = 2 - 2.5
2
SAMPLE 6: jC = 5 - 10?
3
dm

SAMPLE 2: jC = 1.5 - 2.5 5

1
0.

2
=

2
0.
3
jC

0.

SAMPLE 1: jC = 4 - 6
0 .5

0.1
1

5
2

2
8
16

30

10
jC =

1
5

0.1
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
Jointing parameter JP

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

Ave ra g e j oi n t spa c in g ( m )
0.8 1 approx. scale effect of 5

m a ss i v e ro c k
0.6 compressive strength ( σ c )
Volume tri c joint c ount (join ts /m )

(Vb)
3

1
1.5 0.5 2
1.5
2 0.4
1.5
2

VOLUME
0.3 5
2
3
3 2 3 2
4 0.2
3
4
5
4 5
5 6

BLOCK
RQD
5 6
2
8 0.1
6
8 0.08
10
8 10 5
0.06
10
15 0.05 2
15
20 0.04
15
20
0.03 4 5
20
30
30 2
0.02
30
50
50 5
60
50 60
2
80 0.01
60
80
100
80 100 1 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:

1: For Vb = 0.00005 m3 (50 cm3 ) and jC = 0.2 the JP = 0.0006;
2: For two joint sets with jC = 1.5 and the volumetric joint count Jv = 3.3 the JP = 0.3;
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)
4: 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

CHARACTERIZATION
RMi VALUE
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 scale effects 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 FOR 'COMMO N' JO INT CONDITION (jC = 1 - 2)

5

2
RM
10 i=
10
3
m

5 0
60
2 40
1 20
5 15
Block volume Vb

10
2
6
100 4
5
2
2
dm3

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
Uniaxial compressive strength of intact rock σ c

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
βo 3 27 3
Db = Vb = Vb (eq. (6-8))
β β
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.

4.3 NUMERICAL VALUES OF THE INPUT PARAMETERS TO RMi

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.

4.3.1 The compressive strength of intact rock (σc )

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).

INFLUENCE
FROM
MOISTURE
laboratory

UNIAXIAL
test

COMPRESSIVE
TEST
COMPRESSIVE STRENGTH OF ROCK
field or laboratory test

POINT LOAD TEST

SCHMIDT
HAMMER
TEST

SIMPLE FIELD
(HAMMER) TEST

ROCK NAME
AND STRUCTURE
observations and
strength tables

ANISOTROPY

WEATHERING
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, the lowest compressive strength should be applied 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.

4.3.2 The block volume (Vb)

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
% Smaller

60

3
Vbmin = 0.01 m
40 Vb25 = 0.07 m
3

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
Block volume Vb

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

BLOCK VOLUME
MEASUREMENT

REFRACTION
SEISMIC

Vb
VOLUMETRIC JOINT COUNT
MEASUREMENT
in surfaces

drill cores
or from

WEIGHTED
JOINT DENSITY
MEASUREMENT
Jv

1-D FREQUENCY
from drill cores

MEASUREMENT
ASSUMED

FACTOR
BLOCK
SHAPE

ROCK QUALITY
DESIGNATION (RQD)

= transition / correlation

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

4.3.3 The joint condition factor (jC)

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 asperities, particularly those of
roughness and hardness, are likely to be dependent on the mineralogical and lithological make-up of
the rock. Mineral coatings 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 condition of the joint, i.e.
− The strength (hardness) of the wall rock material in clean 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 filled joints.
− Size (persistence) and termination of the joint.
2. The external features, 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. The roughness factor (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. The alteration factor (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. The size factor (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 smoothness or 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).
TERM DESCRIPTION factor js
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.

Waviness 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).

TERM undulation (u) waviness factor (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.
waviness*)
smoothness *) 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 weathering or alteration 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

A. CONTACT BETWEEN THE TWO ROCK WALL SURFACES

TERM DESCRIPTION jA
--------------------------------- ---------------------------------------------------------------------------------- --------------
Clean joints
-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
Coating or thin filling
-Sand, silt, calcite etc. ....... Coating of friction materials without clay ......................................... 3
-Clay, chlorite, talc etc. ...... Coating of softening and cohesive minerals ...................................... 4

B. FILLED JOINTS WITH PARTLY OR NO CONTACT BETWEEN THE ROCK WALL SURFACES

Partly wall No wall

contact contact
TYPE OF FILLING DESCRIPTION thin fillings thick filling
MATERIAL (< 5 mm*)) or gouge
jA jA
--------------------------------- -------------------------------------------------------------- -------------- --------------
-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)
TERM FOR
GRADE WEATHERING DESCRIPTION
OR ALTERATION

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

II Slightly 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.

III Moderately 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.

IV Highly 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.

V Completely 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.

VI Residual soil 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 joint length can 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. Joint continuity 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).

JOINT RATING OF jL FOR
LENGTH TERM AND TYPE continuous discontinuous
INTERVAL 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.

4.4 POSSIBLE AREAS OF APPLICATION OF THE RMi

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.
RELATIVE IMPORTANCE OF:
UTILITY ACTIVITY 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

Legend: + 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.

4.5 DISCUSSION

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.

4.5.1 Limitations of the RMi

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. The range and types of rock masses covered by the RMi.
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. The accuracy in the expression of the RMi.
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. The effect of combining parameters that vary in range.

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 relative 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.

4.5.2 Other similar rock mass characterization methods

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 compressive 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
"representative for the jointed effect of a rock mass".

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 compressive strength of rock masses (σcm), which is a part of the Hoek-Brown failure
criterion for rock masses, and is expressed as
σcm = σc × s ½ eq. (4-11)

Here σc = the uniaxial compressive strength of the intact rock material, and
s = an empirical constant. The value of s 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 s 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’ + (m σc σ3’ + s σc )

2.0

σ1’
1.5
Uniaxial compression
β 2 1/2
σcm = (s 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).
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

Chapter 5

ROCK MASSES CHARACTERIZED BY THE RMi

"How convinced are we that the rock properties we feed into the design models actually
describe the way the rock mass reacts to a perturbation like a tunnel or a cavern?"
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.

4.1 ON CONTINUOUS AND DISCONTINUOUS ROCK MASSES

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 discontinuous. 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 continuous (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 "when the ratio of fracture
spacing to a tunnel diameter is between the approximate limits of 1/5 and 1/100. For a
range outside these limits, the rock may be considered continuous, though possibly
anisotropic." (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 x 5 x 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).

5.2 ZONING OF ROCK MASSES INTO STRUCTURAL REGIONS

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 structural regions
(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 zoning of rock masses 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 detailed jointing (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
ce
surfa
und
Gro CLASS5 (b)

CLASS 5 (b)
DI
VE
RS
ION
ROCHES FAU

CLASS 3
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.

5.3 PRINCIPLES IN CHARACTERIZING THE VARIATIONS IN ROCK MASSES

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 have 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.

5.3.1 Variations in the rock material

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. In rocks where the strength varies with direction of the testing, the lowest
value should be applied for σc as input to RMi.

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.

5.3.2 Variations in the jointing

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

different joint alteration

factor and joint length
between the joint sets

short, smooth & planar

and fresh joints

long smooth & slightly

undulating joints

different joint size,

termination and roughness
for the joint sets

discontinuous,
rough & planar joints

discontinuous, short foliation

partings; fresh, smooth & undulating

different joint size and

termination for the
two joint sets

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.

A . Block volume found from joint spacing or joint density measurements

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 app-
roximate 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).

B. Probability calculations to determine the variation in block volume

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 different conditions for the various joint sets, some simplifications have
to be made to combine them as shown in the following:
 5 - 10

A. jC determined for one joint or joint set where the parameters involved in it vary

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 probability calculations 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

B. The resulting jC or JP for the rock mass when jC varies for each joint or joint set

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 that "when
joint sets show different shear strength, the set which shows the smallest mean
angle of friction should be adopted, unless specific circumstances warrant
otherwise. A record of the angles of friction corresponding to other fracture sets
may prove of interest."

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:

Case 1
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
1m 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' Approx. Average Adjusted Assumed joint
joints joint joint number condition
length length of joints factor
(na) ( Lj) Na* = na⋅Lj/√A (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)

Case 2
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 volume 1 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 bedding joints 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.

5.3.3 Singularities and weakness zones

Singularities, 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
TYPE OF WEAKNESS ZONE
JP RMi
Zones of weak materials
• Layers of soft or weak minerals, 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
• Zones of weak rocks or brecciated rocks, 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 *
• Weathered surface or near surface deposits 0.005 - 0.05 0.05 - 3

Faults and fault zones

• Tension fault zones
- feather joints and filled zones, such as:
> clay-filled zones 1) ** 0.01 - 0.05
> calcite-filled zones 2) ** 0.5 - 5
• Shear fault zones
- 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)
• Altered faults
- 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

It is difficult to assume general numerical

Recrystallized and cemented/welded zones
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

Fig. 5-10 Example of the influence from a singularity on stability (from Cecil, 1970)
 5 - 16

Large and moderate weakness zones 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.

5.3.4 Summary of the possibilities and methods to determine the block volume or the
jointing parameter where the jointing characteristics vary

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
J O I N T C O N D I T I O N FACTOR (jC)
DIFFERENT jC
TYPE OF JOINTING SAME jC
- consisting of FOR ALL between the between the between the
JOINTS sets only joints in the sets single joints

Regular jointing
- mainly of joint sets x x (x) -
- columnar jointing x (x) ? -

Mixed jointing
- joint sets + random joints x x : :

Irregular jointing
- mainly random or irregular joints x - - :

Foliation jointing
- long joints + short partings x x (x) ?

Bedding jointing
- 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
Variations in the block size (Vb)
A. The block volume found from joint spacing or joint density measurements
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.

B. Probability calculations to determine the variation in block volume

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

Variations in the joint condition factor (jC)

A. The jC determined for one joint or joint set where its parameters vary
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 probability calculation similar to example 5-3.

B. The resulting jC or JP for the rock mass when jC varies for each joint or joint set
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:
Case 1 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.
Case 2 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.
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

Chapter 6

THE USE OF RMi IN DESIGN OF ROCK SUPPORT IN UNDERGROUND

OPENINGS

"The basic aim of any underground excavation design should be to utilize the rock itself as the
principal structural material, creating as little disturbance as possible during the excavation
process and adding as little as possible in the way of concrete or steel support. In their intact
state and when subjected to compressive stresses, most hard rocks are far stronger than concrete
and many are of the same order of strength as steel. Consequently, it does not make economic
sense to replace a material which may be perfectly adequate with one which may be no better."
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:
Stability is here used to express the behaviour of rock masses related to their "likelihood of
being fixed in position" (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.

Failure, '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 ground is frequently used in this chapter. By ground is here meant 'the in situ rock
mass subjected to stress and water'.

Rock mass 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 competency of the ground. 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 "a rock layer which, during folding, flexes
without appreciable flow or internal shear" (Dictionary of geology, 1972), competent ground is a
rock mass or soil having higher strength than the stresses acting on it.

6.1 STABILITY ANALYSES AND ROCK SUPPORT DESIGN

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. "It is seldom possible, neither
to acquire the accurate mechanical data of the ground and forces acting, nor to theoretically
determine the exact interaction of these" (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: "The result of an analysis depends very much on the assumed model and the values of the
significant parameters. The main purposes of the structural analysis are to provide the design
engineer with:
1) A better understanding of the ground-structure interaction induced by the tunnelling
process.
2) Knowledge of what kinds of principal risks are involved and where they are located.
3) A tool for interpreting the site observations and in-situ measurements.
The available mathematical methods of analysis are much more refined than are the properties that
constitute the structural model. Hence, in most cases it is more appropriate to investigate
alternative possible properties of the model, or even different models, than to aim for a more
refined model."

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

6.2 INSTABILITY AND FAILURE MODES IN UNDERGROUND EXCAVATIONS

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)
Terms applied by Terzaghi
Firm ground is a material which will stand unsupported
in a tunnel for several days or longer. The term includes
a great variety of materials: sands and sand-gravels with
clay binder, stiff unfissured clays at moderate depths,
and massive rocks.
Ravelling ground indicates a material which gradually
breaks up into pieces, flakes, or fragments. The process
is time-dependent and materials may be classified by the
rate of disintegration as slowly or rapidly ravelling. For a
material to be ravelling it must be moderately coherent
and friable or discontinuous. Examples are jointed rocks,
fine moist sands, sands and sand-gravel with some
binder, and stiff fissured clays.
Squeezing ground.
Squeezing rock slowly advances into the tunnel without
perceptible volume increase. It is merely due to a slow
flow of the material towards the tunnel, at almost
constant water content. The manifestations and the
causes of the squeeze can be very different for different
clays and decomposed rocks.
Prerequisite for squeeze is a high percentage of
microscopic and sub-microscopic particles of micaceous
minerals or of clay minerals with a low swelling
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
expansion from water adsorption. The capacity to swell
seems to be limited to those rocks which contain clay
minerals such as montmorillonite, and to rocks with
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)
Term used in The new Austrian tunnelling
method (NATM)
Stable
Elastic behaviour of the surrounding rock mass.
Small, quickly declining deformations. No relief
features after scaling.
The rock masses are long-term stable.
Loosening
Elastic behaviour of the rock mass, with small
deformations which quickly decline. Some few small
structural relief surfaces from gravity occur in the
roof.
Ravelling
Far-reaching elastic behaviour of the rock mass with
small deformations that quickly decrease. Jointing
causes reduced rock mass strength, as well as limited
stand-up time and active span*) . This results in relief
and loosening along joints and weakness planes,
mainly in the roof and upper part of walls.
Strongly ravelling
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 or swelling
The "plastic" zone of considerable size with
detrimental structural defects such as joints, seams,
shears results in plastic squeezing as well as 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.
Strongly squeezing or swelling
Development of a deep squeezing zone with severe
inwards movement and slow decrease of the large
deformations. Rock support can often be overloaded.
 6-5

10,000
)

(heavily jointed r ock s)

opening diameter
block diameter
CONTINUOUS

COLLAPSE FLOWING GROUND 3)

2)
RUNNING GROUND
1000 1)
SQUEEZING 1)
RAVELLING
1)
RAVELLING
1) LOOSENING
Continuity of ground (

100
0.1 10
INCOM PETENT GROUND COMPETENT GROUND
(overstressed rock mass es)
100
D ISC ON T IN U OUS
(j oi n te d ro cks )

R A V E L L I N G 1)

L O O S E N I N G 1)

10
B LO C K FALLS
STABLE

5
0.01 poor conditi ons 1 good condition s 1000

5
(m as siv e roc ks)
C ON TIN U OUS

4)
SPALLING
1 POPPING 4)
4)
ROCK BURST
STABLE
CREEP
SQUEEZING

0.1
0.1 1 10
INCOM PETENT GROUND COMPETENT GROUND
(overstressed rock mass es)

Quality of ground

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.

6.2.1 Special modes of instability and behaviour related to faults and weakness
zones

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 gouge or filling material.
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)
Dominant material in POTENTIAL BEHAVIOUR CAUSED BY THE GOUGE MATERIAL
fault gouge/filling At face Later
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).

6.2.2 Main types of failure development

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: "Experience shows that the ground does
not commonly react at once to the change of stress produced by the blast. The round blast creates
an unsupported section of roof located between the new face and the last support of the tunnel. As
soon as the natural supporting rock is removed by blasting, some blocks drop out of the roof,
leaving a small gap in the half-dome. If the newly exposed roof section is left without support, some
more blocks drop out after a while, thus widening the gap. Finally the entire mass of rock
constituting the half-dome drops into the tunnel and a new half-dome is formed. The new half-dome
also starts to disintegrate and the process continues until the tunnel section adjoining the working
face is filled with rock debris."
"The rate at which the progressive deterioration or ravelling of the half-dome takes place depends
on the shape and size of the blocks between joints, on the width of the joints, on the joint filling and
the active span."

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.

6.3 THE MAIN FEATURES INFLUENCING UNDERGROUND STABILITY

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 behaviour 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 inherent properties of the material (rock masses) 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 external forces acting in the ground:
a. Magnitude and anisotropy of horizontal and vertical stresses in undisturbed rock.
b. Ground water.
3. The excavation features, 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 time-dependent features, 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.

6.3.1 The inherent properties of the rock mass

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).
STRIKE Dip = 0 - 20o Dip = 20 - 45o Dip = 45 - 90o
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.

6.3.2 The external ground features

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.

6.3.3 The excavation features

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 perimeter blasting. 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.

6.3.4 The time-dependent features

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). "The bridge-action period for a given material increases very rapidly with
decreasing distance between supports. Thus, for instance, a very fine, moist and dense sand can
bridge a span one foot wide for several hours. Yet, the same sand would almost instantaneously
drop through a gap between supports with a width of five feet."
 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.

6.3.5 Summary of Section 6.3

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 swelling 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 long-term effects 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 stand-
up time 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
THE GROUND CONDITIONS CHARACTERIZED BY
The inherent properties of the rock
mass:
- 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

The external forces acting:

- 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
The excavation features:
- The shape and size of the opening
(*) 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
* 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.

6.4 RMi APPLIED TO ASSESS ROCK SUPPORT

"It is essential to know whether the problem is that of maintaining stability with the pre-existing
jointing pattern or whether it is the very different problem of a yielding rock mass. The stress
situation is therefore one of the main parameters in stability and rock support evaluations."
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 continuity factor and
the ground condition factor:
 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
highly jointed
or crushed
tunnel dimension
CONTINUITY FACTOR CF = block dimension

overstressed competent
100
jointed

10
poor good
5
massive

1
overstressed competent

0.01 1 100
GROUND QUALITY

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 continuous 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.

6.4.1 Stability and rock support in continuous materials

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

COMPRESSIVE STRENGTH
TYPE OF ROCK
OF ROCK
ROCK MASS INDEX
JOINT C ONDITION FACTOR
JOINTING PARAMETER RMi
BLOCK VOLUME
(brittle or ductile)
ROCK STIFFNESS

GROUND WATER
TANGENTIAL
DIAMETER

STRESSES
ROCK STRESSES
BLOCK

AROUND
SHA PE OF THE OPENING
TH E OPEN ING

DIAMETER OF
THE OPENING

CONTINUITY OF COMPETENCY
THE GROUND OF THE GROUND
CF Cg

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 competency
factor
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.
 6 - 17

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,
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 rock bursting; if the
deformations caused by overstressing occur more slowly, squeezing occurs, as further described in
the following.
s

ta tre
se

s
ng ss
support lining m as
ro k m

en (
t f c

tia σ θ)
en ro

l
em g
if n ndin
n
co rrou
su

m
ro
t f ng
en lini
m rt
ine o
nf pp
co su

TUNNEL

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
SIZE RATIO
massive

overstressed

GROUND QUALITY

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 competent 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 incompetent massive ground, i.e. Cg < 1, the overstressing of the rock mass will cause some
form of stress induced instability:
- In brittle 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 deformable, flexible or ductile 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
SIZE RATIO

overstressed
GROUND QUALITY

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 low to moderate state of stress instability is dominated by structurally controlled

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.

In overstressed ground 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 running ground and flowing ground 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 spalling, 4 popping or rock burst, 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.

A. Instability assessments

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

ty
vi
ti
ac
t
rs
bu
ck

w
ro

8 Lo e
at
o

er
N

od
M y
ivit
ac t
rs t
k bu
4
h roc
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

Competency factor FAILURE MODES
Cg = fσ⋅σc /σθ = RMi /σθ in massive brittle rocks

> 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
Stress problem Characteristic behaviour Rock support measures

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
Ground characteristics Applied rock support
Project Continuity factor Rock mass index Competency factor B( ) = rock bolt (spacing)
CF RMi RMi/σθ 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

SHOTCRETE 50 - 100 mm SCALING SCALING

+ + +
ROCK BOLTS ROCK BOLTS SPOT BOLTING
spaced 0.5 - 2 m spaced 1.5 - 3 m

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

B. Possible measures to reduce rock stress problems

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

A B

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.

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
 6 - 26

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)

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.
 6 - 27

TABLE 6-8 CLASSIFICATION OF SQUEEZING (based on Aydan et al., 1993)

Squeezing class
The tunnel behaviour according to Aydan et al. (1993)
competency range
No squeezing The rock behaves elastically and the tunnel will be stable as the face effect ceases.
RMi/σθ > 1

Light squeezing The rock exhibits a strain-hardening behaviour. As a result, the tunnel will be stable
RMi/σθ = 0.7 - 1 and the displacement cease.

Fair squeezing 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.

Heavy squeezing 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.

Very heavy squeezing 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)
Approx. convergence and rock support according to Seeber et al. (1978)
NATM for tunnel with diameter 12 m
English term 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

SHOTCRETE SHOTCRETE For highly jointed rock masses:

100 - 250 mm 50 - 150 mm
+ + Use support chart
ROCK BOLTS ROCK BOLTS for discontinuous
spaced 0.5 - 1.5 m spaced 1 - 2.5 m 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

A. The use of analytical methods to determine rock support in squeezing ground

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 m 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 arch action 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
arch action 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

6.4.2 Stability and rock support in discontinuous (jointed) materials

"Paradoxically, the excavation of an underground opening in a highly stressed environ-

ment is likely to be less hazardous when the rock is jointed than when it is intact."
Nick Barton (1990)

1000
crushed

overstressed
100
SIZE RATIO

jointed

10
poor good
massive

1
overstressed

0.01 1 1000
GROUND QUALITY

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 COMPRESSIVE

S T R E N G TH
R O CK M A S S I N D E X
JO I N T C O N D I TI O N FA C T OR
RMi
JOINTING PARAMETER
BLOCK VOLUME
D I A ME TER
B LOC K

G R O U N D W AT E R R OC K
S TR E S S
ROCK STRESS ES LE V EL

SIZE OF
THE OPENING

C ON T I N UI T Y OF
ADJUSTMENT FOR THE
TH E G R OU N D WALLS OF THE OPENING
CF
ORIENTATION OF
THE OPENING

SIZE OF THE
WEAKNESS ZONE

CONDITION OF ROCK
MASSES ADJACENT
TO THE ZONE

IMPACT FROM
THE EXCAVATION

U S E OF
SIZE RATIO TH E OP EN I N G
C OM P E T E NC Y
O F T H E G RO U N D
Sr Cg
fo r
disc
o n ti
nuo
us g
rou
n d
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:
 6 - 31

- 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.

6.4.2.1 The ground condition factor (Gc)

The ground conditions consist of the general inherent rock mass features of main influence on
stability and the external 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.

A. Effect of stress level in the ground

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 total stresses. The influence of pore pressure or joint water
pressure is generally difficult to incorporate in the stress level. Often, the joints
 6 - 32

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
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)

Maximum Approximate Stress level
Term stress overburden factor
σ1 (for k ≈ 1) (SL)*)
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)

B. Possible instability induced from high ground stresses.

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 "If the rock
masses around the tunnel is in a state of intense elastic deformation, the connections or interlocks
between blocks such as A and B in Fig. 6-21 and their neighbours, may suddenly snap, whereupon
 6 - 33

the block is violently thrown into the tunnel. If such an incident occurs, it is necessary to provide the
tunnel with the support prescribed for popping."

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
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 the equivalent block diameter 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.

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

(1974) the degree of freedom determined by the number of joint sets 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.)

Co is an orientation factor representing the influence from the orientation 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 Rating of
o o TERM orientation factor
for strike > 30 for strike <30 all strike values
Co
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 'common' hard rock mass conditions, 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
 6 - 35

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.
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.
Ground characteristics Applied rock support
Jointing parameter Ground
Project and location & condition Size ratio B( ) = rock bolt (spacing)
Rock mass index factor F( ) = fibrecrete (thickness)
JP & RMi Gc Sr 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
Co

2
100
Wt or Ht

80

3
Db e

(S50)

discontinuous (jointed) materials

60 (B1.5, F80)

(S50)
40
Size ratio Sr =

(B1.5, F70)
(B2, F50)
(B1.5, F70)

(B1.5, F70)
20
(B3)
(B2.5-3)
(B3) (B2.5)
(B spot)
x x (B spot)
10
8
x (B spot)
6

4 1.
5
(B spot)
x
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

Ground condition factor Gc = SL x RMi x C

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.

6.4.3 Stability and rock support of faults and weakness zones

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 geometry and dimensions of the zone.

The instability and problems in weakness zones will generally increase with the width of 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 reduced stresses 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 arching effect from 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 arch action 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 swelling, sloughing, or permeable 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.

CO MP RE S S I VE
TY P E O F RO CK S TR EN GT H

JO I NT CO NDI TI O N FACTO R
JOINTING PARAMETER
BLOCK VOLU ME
D I AM ET E R

G R O U N D W AT E R ROCK
BLOCK

S TRES S
ROCK STRESS ES LE VE L

S I ZE O F
TH E O PE NI N G

C O N T I N U I TY O F
ADJUSTMENT FOR THE
THE G ROUND WALLS OF THE OPENING
CF
ORIENTATION OF
THE OPENING

SIZE OF THE
WEAKNESS ZONE

CONDITION OF ROCK
MASSES ADJACENT
TO THE ZONE

IMPACT FROM
THE EXCAVATION

U SE O F
SIZE RATIO T H E O PE N IN G
C O M PE TE N C Y
O F T H E G R O UN D
Sr Cg

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.

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 z = × × (orientation of zone) = × eq. (6-19)
block size (in zone) diameter of tunnel Db zone 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).
Ground characteristics
Jointing parameter Ground Applied rock support
Project and location & condition Size ratio
Rock mass index factor B( ) = rock bolt (spacing)
JP & RMi Gc Sr 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
Size ratio Srz = Db Co

100
z
Tz

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

Ground condition factor Gc = SL x RMim x C

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 cosβ
(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

6.4.4 Comments to the RMi method for assessing rock support

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:
- continuous ground, which can be either massive rock or heavily jointed (particulate) rock
masses; and
- discontinuous ground, 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 continuous 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 discontinuous 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

rocks + ductile,
heavy fair light

massive rocks
highly jointed
REINFORCED REINFORCED FOR MASSIVE ROCKS: no support
SHOTCRETE SHOTCRETE
100-250 mm 50-150 mm
+ + FOR HIGHLY JOINTED ROCK MASSES:
ROCK BOLTS ROCK BOLTS Use support chart for discontinuous materials
spaced 0.5 - 1.5 m spaced 1.5-3 m

SHOTCRETE 50-100 mm
SCALING

massive
+ NO ROCK

brittle,
+

rocks
ROCK BOLTS ROCK BOLTS SUPPORT
SPOT BOLTING
spaced 0.5 - 2 m spaced 1.5-3 m

600
d G
an IN te
te LIN re

zones
fib rce
400 tc

fo
e

re me
cr E o

RO
0.
ot ET

re nt
h

5
G s

in-
sh CR

CK
IN ed

-1
N n
N LI ig

.5

BO
CO E es

LT
200 E T d
R ia l

1.
SP
NC ec

AC
p
C Or s

weakness
Size ratio Sr = Wt or Ht Co

IN
Nj

G
100

(m
2

)
80
Db

3
60

discontinuous ground
40

and
20
0.5

rocks
10
-
1.5

6
1.5

4
Jointed
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

Ground condition factor Gc = SL x RMi x C

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 'common' hard rock mass conditions, 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.

A. In low stability ground

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 timing is a factor in tunnelling, that can " hardly
be computed or even assessed by a rock mechanics specialist if he is not provided with deep
geological knowledge or if he does not intimately collaborate with an engineering geologist."
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.

B. In hard rock regimes

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.
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

Chapter 7

RMi PARAMETERS APPLIED IN PREDICTION OF TUNNEL BORING

PENETRATION

"Nothing has been more difficult than evaluating the rock mass characteristics and applying
the evaluation to a formula predicting penetration rate."
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

7.1 Factors influencing the TBM performance

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)
Rock mass factors Machine factors

- Rock mass jointing (ks) - Thrust per cutter (M)

° type and continuity
° frequency
° orientation
- Rock porosity
- Rock drillability (DRI)
- Rock hardness/abrasiveness
- Stress in rock
- Cutter edge bluntness (br)
- Cutter spacing (A)
- Cutter diameter (d)
- Torque capacity and RPM
- The machine's capacity for handling large chips or blocks
- General solidity against blows and vibrations
- 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.

7.2 Prediction models

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, "in the
interest of conservatism in most analyses, the improvement in boring rate due to jointing has been
neglected by testing unfractured specimens of solid rock and basing predictions on the strength
characteristics of intact rock" (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.

7.2.1 The NTH prognosis model

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.

7.3 The use of RMi parameters to characterize rock masses for TBM
 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

JOINT CONDITION
JOINTING APPLIED THRUST
FACTOR (jC)
PARAMETER PER DISC
BLOCK VOLUME (Vb)
JP MB

CORRECTION FOR
TBM JOINTING CUTTER SIZE (k d )
FACTOR
ks CORRECTION FOR
CUTTER SPACING (k a )
MATERIAL

TYPE OF ROCK (E-factor)

ROCK

EQUIVALENT TBM EQUIVALENT THRUST

JOINTING FACTOR PER CUTTER
keq Meq

input parameter

Fig. 7-1 Layout of a method to predict TBM penetration using RMi parameters based on the NTH model.

7.3.1 The rock material properties

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 30000
10000

50000
35000
20000

40000
15000

25000
5000

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:
Joints 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).
Fissures 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)
Massive rock 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

k DRI 1

0.75

10 30 50 70 90
DRI

IV

fissure class
4

joint class
TBM penetration factor ( ks )

NTH Joint/fissure
3 b class spacings
III-IV II-III

0 -
c
2
0-I 1.6 m
d I- 0.8 m
III II
e I 0.4 m
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
Angler between tunnel axis and discontinuity plane

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
NTH discontinuity class 0-I I I-II II II-III III III-IV IV

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 joints
 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

angle between tunnel
0-15o 15 - 30o 30- 45o 45 - 75o 75 - 90o
axis and joint set
1.5
average value of co = 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 Vb -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 fissures 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 jointed 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

For massive 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 massive 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)

7.3.3 Assessment of the net advance of boring

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
Penetration per pass ( io ) (mm/rev)

utter
00 kN/c
Meq = 3
10

250
8

200
6
150

4 100

0
0 1 2 3 4 5 6 7
Equivalent TBM jointing parameter ( keq)

Fig. 7-6 The influence of jointing on the TBM boring (from Norwegian Institute of Technology, 1994).

1.6

1.4
kd

1.2

1.0

0.9
280 300 340 380 420 460 500 540
Cutter diameter (mm)

Fig. 7-7 Correction factor kd for the size of the cutters (from Norwegian Institute of Technology, 1994)

1.1
ka

1.0

0.9

0.85
55 60 65 70 75 80 85 90
Average cutter spacing (mm)

Fig. 7-8 Correction factor ka for the mean spacing of the cutters. (from Norwegian Institute of Technology,
1994).
 7 - 11

7.3.4 Example

The following data have been given on rock mass conditions and TBM factors:
--------------------------------------------------------------------------------------------------------------------------------------
Rock mass factors mapped: Evaluations made:

- 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)

TBM machine factors:

- 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
-------------------------------------------------------------------------------------------------------------------------------------
Calculations:

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

ESTIMATING TBM PENETRATION RATE (valid for disc cutters)

Tunnel:
INPUT TBM PARAMETERS TBM type: CALCULATIONS
TBM diameter = 4,5 m Spacing between cutters Sc = 75 mm Reference
Number of cutters = RPM = 11,1 Cutter diameter Dc = 483 mm E= 750 eq. (7-2)
fs = 0,760 eq. (4-7)
INPUT PARAMETERS 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)
Orientation 1) 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)
CALCULATIONS 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]

7.3.5 Discussion of the RMi method for TBM penetration assessment

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
Calculated boring advance (m/h)

0
0 1 2 3 4 5 6 7 8 9 10
Real boring advance (m/h)

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)

Communicati on

Hoek-B rown
failure
i nput in rock mechanics
models and cal culations
used in rock engi neeri ng

Fragmentation criterion
classifi cation systems

and blasting RMR system

Numeri cal
input to exi sti ng

modelling
Stabil ity and
rock support Deformation
Q - system modulus of
assessments
rock masses

TBM progress Ground

NATM response
evaluat ions curves

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 eq. (8-1)

where σ1' is the major principal effective stress at failure.

σ3' is the minor principal effective stress.
σc 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:
σcm = σc × s½ 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
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
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
 8-3

RQD, which in Appendix 4, Section 6 has been shown to often poorly represent the variation in
jointing.

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
σ'3 = minor principal effective stress
σc = uniaxial compressive strength of
intact 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 s 0.000003 0.000003 0.000003 0.000003 0.000003
waste 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
Appendix 3). The constant mb is the same as m in the original criterion. It varies with the jointing
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
(1991).

As shown in Fig. 8-1, the variation of mb can be mathematically expressed as:

a) for undisturbed rock masses
mb = mi × JP 0.64 eq. (8-4)
b) for disturbed rock masses
mb = mi × JP 0.857 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 eq. (8-6)
where s and m are replaced by JP and mi

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:
σ ’3 a
σ = σ + σ c ( mb )
’
1
’
3 eq. (8-7)
σc

where mb and a are constants which depend on the composition, structure and surface of the
jointed rock mass.
 8-5

mb is found from the ratio mb /mi in Table 8-3. mb /mi varies between 0.001 in crushed rock
masses with highly weathered, very smooth or filled joints to 0.7 in blocky rock masses with rough
joints. In massive rock mb /mi = 1. 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.

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
mb and a are constants which depend on the

VERY POOR
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
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) eq.(8-8)
 8-6

where τ is the shear stress at failure, and

Φ i' is the instantaneous friction angle.

The value of the instantaneous friction angle is given by

Φi' = Arctan [4hCos 2 (π/6 + Arcsin h - 3/2) - 1] - 1/2 eq. (8-9)

where σ' = the effective stress

h = 1 + 16(m × σ' + s × σc) /3m2σc eq. (8-10)
m = mi × JP 0.64 (for undisturbed rock masses) eq. (8-11)
s = JP 2 eq. (8-12)

The instantaneous cohesive strength is found as

ci' = τ - σ' TanΦi' 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
STRENGTH (ci') FROM INPUT OF RMi PARAMETERS.
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.
 8-7

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.

8.3 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

ε1
p
Support C
reaction
ε1e ε1 line
+
D

0 p
-
δio
v
- Gradient = F
p
v
ε1
+ ε1p
 8-8

Fig. 8-4 Left: The material behaviour used by Brown et. al.(1983) in the closed-form solution.
Right: The ground response curve.

The input data used in the closed-form solution are:

ri the internal tunnel radius
σc the compressive strength of intact rock
po the in situ hydrostatic rock stress
f the gradient of line in the -ε3p , ε1p diagram (Fig. 8-4)
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.

The following calculation sequence is given by Brown et al. (1983):

1. M = ½ [(m/4)2 + (mpo/σc) + s] ½ - m/8.
2. G = E/[2(1 + ν)].
3. For pi ≥ po – Mσc, deformation around the tunnel is elastic: δi/ri = (po – pi /2G.
4. For pi < po – Mσc, plastic deformation occurs around the tunnel: ue /re = Mσc /2G.
5. N = 2{[(po – Mσc )/mr σc] + sr /mr2)}½.
6. re /ri = exp{N – 2[pi/mr σc) + (sr /mr2 )] ½}.
7. δi/ri = Mσc /[G(f + 1)]{[(f – 1)/2] + (re /ri )f+1}.

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 eq. (8-15)

s = JP2 can be found from eq. 4-4 or from Fig. 4-4 based on field characterization of the block 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
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
sr = JP2 = 0.06 Vb2/3 eq. (8-16)

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.
 8-9

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.

8.4 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,
Austria, has been one of the chief de-
velopers of the New Austrian Tun-
nelling Method. Goal of NATM: to
provide safe and economic support in
tunnels excavated in materials inca-
pable of supporting themselves – e.g.
crushed rock, debris, even soil. Sup-
port is achieved by mobilizing what-
ever humble strength the rock or earth
possesses.
The New Austrian Tunnelling
Method has several features:
• It relies on strength of surrounding
followed by final collapse – unless NATM aims at a temporary semi-rigid
resisted in time by a lining. lining stressed by a moderate rock load pr
The tunnel lining must be neither that will be just above the theoretical
too stiff nor too flexible. If stiff, pr minimum value.
will remain unnecessarily large – the (Our thanks to Mr. Herbert Nussbaum,
lining will be uneconomic. With an engineering consultant and expert on
increased ∆r, however, the pressure tunneling from Weat Vancouver, B.C.,
pr decreases. With a lining allowing Canada. Based on phone interviews with
too much yield, pr will be big and the him, we were able to write this box and
lining uneconomic and unsafe. captions for all figures – GD).
Before excavation After excavation

rock to provide tunnel support. This is

done by inhibiting rock deterioration, When
joint opening, and loosening due to pr 0 pr surrounding rock
moves into
excessive rock movements. ∆r pt tunnel cavity,
• It uses protective measures like stresses pr at
r
-∆

pt 0 cavity-rock
lining tunnel walls with shotcrete and
r0

interface
r=

driving anchors into unstable rock. In decrease

dramatically,
many cases, a second, inner linings not r0 Cavity Plastic zone Elastic zone making possible
needed – e.g. for water conduits, short the use of tunnel
linings that are
highway tunnels.
R

pr much less thick,

R
-r much less
• It involves installation of sophis- costly. But since
ticated instrumentation at the time total weight of
mountain is
initial shotcrete lining is placed, to constant,
provide info for designing a second stresses in rock
must increase
inner lining. elsewhere to
• It completely eliminates costly g,
pr 0 carry this
100% in re
en ailu constant load.
Too

os n f
Pressure on tunnel lining, (pr )

o
interior supports for tunnel walls, such L de Note stresses pt
stif

d
80% su
f,

as heavy steel arches. sti

ff increase to max
saf

oo at a short
e

,t
fe
Tunneling into hard or inferior 60%
Unsa distance from
cavity.
rock disturbs the existing equilibrium Au
40% str
of forces. A rearrangement of stresses ian
me
thod afe
within the rock surrounding the cavity 20%
, safe Uns
Incompetent rock
Competent rock
follows (See Fig.). As time progresses,
0%
the freshly excavated tunnel radius ro 1 2 3
Rock movement, ∆r
decreases to (ro – ∆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
 8 - 10

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.

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
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.
-----------------------------------------------------------------------------------------------------------------------------------------------
*)
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.
 8 - 11

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:
- 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 )
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
 8 - 12

5 Squeezing Continuous ground *) 0.35 < Cg < 0.7

6 Strongly squeezing Continuous ground *) Cg < 0.35
*)
Continuous ground is where CF < approx. 5 or CF > approx. 100 (CF = tunnel diam./block diam.)

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
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.
 8 - 13

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.

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=
φ=2

φ=

φ=
15

15
20
o

o
0
o

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
R o 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 gt ng

h th
100 24 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

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
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 &
in g

00
Character of Dis- D int Cl a Shear Strength / Stress Level
RQ 200 Jo ock Continuous y Clay Content
continuities 0 ide ive R 2
W 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 re n t
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 cs
ani

t io
m ech
o il
Se ont.

100 S

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 l os 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

M aj
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

)
mmm

m3
3
be = 0 so
0 - 46 m

BO
= 1 .0

UL
DE
( V Db

R
SILT

Uniaxial 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

RO CK
1.0 10 25 50 100 250 kPa
Uniaxial compressive strength of soil mass

1,000 3
σc =

m
12 m)
SIVE
1

inted
MAS

(Db =
y jo

Vb =
e
σc
=5
akl
we
E qu

σc =
i v

15
DE
ale

d
te
UL
n
BO
joi
n

σc =
ts

50
ly

1. m 3
o

)
te

m
il p

σc =
ra
ar

1
2
e

100
tic

od

b =
le

(D Vb
m

e =

σc =
250 te d LE
oi n BB
)

ly j CO
b

SA
ong (
V

ND
str
V cru shed
3 e
(Db b = 1 dm ) m
m GRAVEL =1 2m lu OF A R
e =
12 m 3 Vb = 0.1 TU
RE OC
mm (Db 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

O F A SOIL
RE
U
CT
STR U

)
(Db 2 mm Db =
c le s Db = GRAVEL 60 mm
arti
lp CO
oi ND
BB
LE Db
s

SA =
60
0
m
m
mm

BO
06
0.

UL
=

DE
Db

R
SILT

Uniaxial 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
Uniaxial compressive strength of soil mass
σc =

0 m3
SIVE
1

1,00
MA S
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

m
50

σc =
ar

100
tic

=
le

Vb

σc =
15

250 LE
05 BB
)

0. CO
b
5

SA
(
V

ND
0.1 .25
Vb 0 1
dm
3
e
= 1m =1 m
m3
GRAVEL
Vb lu
v o
ck
3
Vb = 1 cm OF A ROCK
blo RE
M
TU

AS
STRUC

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
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.
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

Chapter 9

DISCUSSION AND CONCLUSIONS

"A complete understanding of the features and mechanisms involved in the strength of rock
masses presents formidable theoretical and experimental problems and, hence, simplifying
assumptions are required in order to provide a reasonable basis for estimating the strength of
jointed rock masses for engineering design purposes."
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.

9.1 ON THE LAYOUT OF THE RMi SYSTEM

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:
i. It should describe the 'material' in well-understood terms that convey an idea of its type and
behaviour.
ii. The system should furnish an indication of 'material' properties and performance.
iii. It should be applicable from visual examination, both in a simplified form and, with experience,
in a more refined form.
iv. It should employ a simple system of notation for graphic abstracts of boring logs (sections,
plan maps, contract specifications, reports) on drawings recording boring information.
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

JOINTING INTENSITY jointing

JOINT CHARACTERISTICS parameter
Rock Mass index (RMi)
uniaxial
compressive
strength

LOCATION PARAMETERS
Communication
CONSTRUCTION PARAMETERS

REQUIREMENTS
Input in rock mechanics
models and calculations
classification systems
Input to exixsting
Applied in rock
engineering

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

9.1.1 Comparisons with the principles in other rock engineering systems

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 A b 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.

9.2 ON THE STRUCTURE OF THE RMi

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 × s½ .

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 s for jointed rock masses, have been considered "very approximate" 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 relative 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 s were conservative. Therefore, they adjusted the s-values in
1988. A general comparison between s 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 s (= JP 2), are mainly lower
than JP2 for the 'original' s, while they are mostly higher for the 'revised' s . 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' s and JP for the higher values.

Fig. 9-5 Comparison between the jointing parameter (JP) and the parameter s in the Hoek-Brown failure
criterion. Values of s have been found via the Q-system.

As for the Hoek-Brown failure criterion, also RMi - when applied directly in calculations - is
restricted to continuous 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.

9.3 ON THE INPUT PARAMETERS TO RMi

Basically, the properties of the rock material can be found from laboratory tests. Joints, 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.

9.3.1 The uniaxial compressive strength of the rock

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.

9.3.2 Jointing

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.

9.4 ON THE VARIATIONS AND UNCERTAINTIES IN ROCK MASSES

"When the starting point is least sound, most surprising results may be found."
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 before construction, uncertainties are introduced
from the interpolations made between more or less known conditions at the surface and from
various forms of extrapolations 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

9.5 COMPARISON BETWEEN RMi AND OTHER METHODS USED IN ROCK

ENGINEERING

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.

9.5.1 The rock quality designation (RQD)

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.

ROCK QUALITY
0 30 60 90 100 DESIGNATION (RQD)

VOLUMETRIC
3
JOINT COUNT
100 50 20 10 5 2 1 0.5 0.2 joints/m

BLOCK
VOLUME
1 10 0.1 1 10 0.1 1 10 100 1000 10 000 100 000
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

9.5.2 Rock support design systems

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.

Fig. 9-7 Comparison between RMR and RMi

As shown a general correlation 3 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.

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).
 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
-----------------------------------------------------------------------------------------------------------------------------------------------
ROCK MASS PARAMETERS REPRESENTED IN
PARAMETER The RMR system The Q-system Support applic. of the RMi system
-----------------------------------------------------------------------------------------------------------------------------------------------
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.

9.6 THE NEED FOR A 'LANGUAGE' IN ROCK MECHANICS AND ROCK

ENGINEERING

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. "Never should rock mechanics be used without the background of engineering
geology. Such a use or misuse would be worse than the use of engineering geology without rock
mechanics."

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.

9.7 BENEFITS AND LIMITATIONS IN APPLICATION OF THE RMi SYSTEM

Some of the benefits 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 s (= 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 limitations 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.

9.8 SOME CONCLUDING REMARKS

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 understanding of 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

9.8.1 Future developments

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.
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

Chapter 10

REFERENCES

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.
 From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

APPENDIX 1

ON JOINTS AND JOINTING

"From an engineering point of view, a knowledge of the type and intensity of the rock
defects may be much more important than of the types of rock which will be
encountered."
Karl Terzaghi, 1946

Joints and related features are certain types of discontinuities in the rocks. Discontinuity is
the general term used in rock mechanics for any mechanical discontinuity in a rock mass
having zero or low tensile strength. It is the collective term for most types of joints, weak
bedding planes, weak schistocity planes and faults or weakness zones.

As there are not any distinct and generally accepted rules or nomenclature for a
terminology of discontinuities for engineering purposes Brekke and Howard (1972) suggest
to make use of
- scale, based on aperture, persistence and occurrence, and
- character, based on occurrence of filling material.

Based on this it has in this work been chosen to divide discontinuities into 3 main groups:
micro-fissures length < 10 mm
joints length 0.1 - 100 m
weakness zones length > 30 m

Micro-fissure is usually considered as a defect or flaw in the rock material (Brekke and
Howard, 1972). Being a rock material parameter rather than genuine discontinuity, it is
generally included in the properties of rock; therefore it is not further dealt with here.

Weakness zones including faults, are described in Appendix 2.

Joints. There is a difficulty in giving a hard and fast definition of what constitutes a joint.
Many geologists will reserve the term 'joint' for typical extension discontinuities, without
any perceptible movement along it. During the years there has been several discussions if
'joint', 'fracture' or other terms should be preferred in rock mechanics, engineering geology
and rock engineering for those features forming the network of discontinuities in the rock
masses. ISRM (1975) has chosen 'joint' as this term defined as: "Joint is a discontinuity
plane of natural origin along where it has been no visible displacement." This is the same
definition as used by many geologists.

In this work, the term 'joint' is used for a group of discontinuities in the range of
approximately 0.1 - 100 m, as indicated in Table 2-1. This is the same definition as
suggested by Brekke and Howard (1972). A similar definition is used by the Norwegian
Rock Mechanics Group (1985).

JOINTING\text-1.wpd
 A1 - 2

Joints may be the result of failure under natural forces (Fig. A1-1), including small seams
(filled joints) and some fractures such as minor and moderate shear ruptures. They also
include weakness planes caused by rock texture or anisotropy. Thus, the main types of joint
are:
− Tectonic joints; 1 i.e. breaks formed from the tensile stresses accompanying uplift or
lateral stretching, or from the effects of regional tectonic compression (ISRM, 1975).
They commonly occur as planar, rough-surfaced sets of intersecting joints, with one
or two of the sets usually dominating in persistence.
− Sheeting joints; a set of joints developed more or less parallel to the surface of the
ground, especially in plutonic igneous intrusions such as granite; probably as a result
of the unloading of the rock mass when the cover is eroded away.
− Exfoliation joints; breaks developed as a product of exfoliation; the breaking or
− splitting off from bare rock surfaces by the action of chemical or physical forces,
such as differential expansion and contracting during heating and cooling over the
daily temperature range.
− Cooling joints, breaks formed as a result from cooling of igneous rocks.
− Foliation joints and partings, discontinuities developed along the foliation planes in
metamorphic rocks.
− Bedding joints and partings, discontinuities developed along the bedding planes in
sedimentary rocks.

Table A1-1 THE VARIOUS TYPES OF JOINTS USED IN THIS WORK

NATURAL JOINTS MAN MADE JOINTS
term length aperture term length
JOINT (small, medium, large size) 0.1 - 100 m < 100 mm rupture < 10 m
- parting <1m < 1 mm crack < 0.1 m
- seam (minor and moderate)*) >3m 10 - 100 mm
*)
Large seams or shears are included in the term weakness zone

I II III
Fig. A1-1 The three main modes for development of tectonic joints (from Scholz, 1990)

The terms for the various types of joints in Table A1-1 are generally chosen from their size
and composition. Some supplementary definitions of these are given as follows:

1
ISRM (1975) advises against the use of the terms tension joint and shear joint, since there are
many possible ways that they can be developed. For example, tension joints can be developed from
cooling of igneous rock, from shrinkage of sediments, from folding, or from ice retreat.
 A1 - 3

Crack a small, partial or incomplete discontinuity. (ISRM, 1975)

Parting a plane or surface along which a rock is readily separated or is naturally divided
into layers, e.g. bedding-plane parting. (Glossary of geology, 1980). 2
Seam 1) A minor, often clay-filled zone with a thickness of some centimetres.
When occurring as weak clay zone in a sedimentary sequence, a seam can be
considerably thicker. Otherwise, seams may represent very minor faults or
altered zones along joints, dikes, beds or foliation. (Brekke and Howard, 1972).
2) A plane in a coal bed at which the different layers of coal are easily separated.
(Dictionary of geological terms, 1962)
Rupture a fracture or break, in this work limited to have been caused by excavation
works or by other activities of man.
Fracture a break in rock due to intense folding or faulting. (Dictionary of geological
terms, 1962)
Fracture is a general term used in geology for all kinds of discontinuities
caused by mechanical stresses in the bedrock. Fractures include joints and
cracks and faults.
Shear a seam of sheared and crushed rock usually spaced more widely than joints
and is marked by several millimetres to as much as a metre thickness of soft or
friable rock or soil.
Singularity is used in this work as a general term for seams, filled joints or other persistent
discontinuities which often in engineering are not considered as belonging to
the detailed jointing.

1 JOINT CHARACTERISTICS

A joint is a three-dimensional discontinuity composed of two matching surfaces called

joint walls. The main joint characteristics are condition of joint walls, possible joint
infilling material together with joint size and continuity or termination. Also joint aperture
or thickness may have great impact on the rock mass behaviour.

Joints commonly terminate at another joint. Joints that terminate in massive rock, are in
this work called discontinuous joints. Such joints can be foliation partings, en echelon
joints in addition to many of the smaller joints (less than 1 metre long). One joint set will
often be more persistent than the other sets. The joints of the other sets will therefore tend
to terminate against the main joints. Joint terminations are also described in Section 2.3 in
this appendix, see also Fig. A1-3.

Pollard and Aydin (1988) mention that the shape and size of joints are largely related to the
rocks they penetrate and to the size and geometry of the rock mass. For example, joint
dimensions perpendicular to layering is controlled by the thickness of the jointed unit.

2
Partings, which often occur as bedding plane and foliation partings, are separations parallel to a
mineralogically defined structural weakness in the rock. They are most often tight and rough
except where flaky minerals (mica, chlorite) occur. In agreement with Burton (1965) they have
been included in the term joints : "I think it would be advantageous to regard bedding planes in
sedimentary rocks and foliation planes in metamorphic rocks as planes of weakness (discontinuity
planes) which do not differ in any significant way from joints."
 A1 - 4

These joints are roughly rectangular in shape. Their size is rarely more than some tens of
metres, while the dimension of joints parallel to layering can be more than a hundred
metres.

Joints formed by cooling of a lava flow or by desiccation of a sediment layer are also
rectangular in shape and are usually perpendicular to the flow surface or layer (Pollard and
Aydin, 1988). The long (vertical) dimension of the rectangle, which is limited by the
thickness of the flow layer or unit, seldom exceeds several tens of meters. The short
(horizontal) dimension controlled by the fracture process, is generally less than the
thickness of the flow.
Analyses of dip lengths and strike lengths performed by Robertson (1970) indicate that
joints tend to be of approximately isotropic dimensions. When terminating in solid rock
they may therefore tend to be circular, and presumably rectilinear when terminating against
other discontinuities.
The size of a joint plane may vary within the same joint set. This is specially the case for
foliation joints where small joints or partings occur between long, persistent joints. Pollard
and Aydin (1988) further note that joints rarely exceed several hundred metres.

Persistence implies the size or areal extent within a joint plane. It can be crudely quantified
by observing the joint trace lengths on the surface exposures (ISRM, 1978). Frequently
rock exposures are small compared to the area or length of persistent joints, and the real
persistence can only be guessed.
Persistence is an important rock mass parameter, but one of the most difficult to quantify in
anything but crude terms. Less frequently, it may be possible to record dip length and the
strike length of exposed joints and thereby estimate their persistence. Joint continuity or
persistence can be distinguished by the terms persistent, sub-persistent and non-persistent
(ISRM, 1978), or more simply as continuous and discontinuous.

Aperture or thickness of joints, i.e. maximum distance between joint walls is generally
small; mainly less than a millimetre, except for filled joints, seams or shears. Nieto (1983)
and Barton (1990b) have observed that persistence generally is proportional to the
thickness of the discontinuity, especially for the larger types in igneous and metamorphic
terrains. Also Kikuchi et al. (1985) have arrived at similar results from studies in andesite,
as shown in Fig. A1-2.
(m)
(m)
100 100

10

10
lenght

0.1

0.01 0.1
0.1 0.3 1 2 10 (mm) 0.01 0.1 1.0 10 (mm)
aperture mean val ue of ap erture

Fig. A1-2 Relation between joint length and aperture in andesite (from Kikuchi et al., 1985).
 A1 - 5

Large aperture can be a result from shear displacement of discontinuities having

appreciable roughness and waviness, from tensile opening, from outwash, or from solution
activity (ISRM, 1978). Steep or vertical discontinuities that have opened in tension as a
result of unloading due to valley erosion or glacial retreat may have very large apertures.

The character of the joint surface may be fresh, weathered, coated, stained, etc. and may
strongly influence on the strength and roughness of the surface and hence the frictional
properties of the joint. Weathering and alteration generally affect the walls of the
discontinuities more than the interior of the rock (Piteau, 1970). These processes may
completely change the joint after it has been formed. The properties of the rock may be
important in determining this feature (Piteau, 1970) as described later in this appendix.

Surprisingly, little has been published on the course or the planarity of joints. A general
trend is that homogeneous rocks often exhibit relatively planar joint walls. Bedding and
foliation joints show often waviness or undulation, especially where the rocks are folded.

Filling includes materials derived from breakage of the country rock due to movements
(i.e. fault zone materials such as breccia), in situ weathered materials (i.e. alteration pro-
ducts), foreign infilling materials deposited between the structural planes (such as calcite),
and also intruded igneous materials, which are different from the host rock. The filling can
therefore consist of several different minerals and materials. The main groups of these are:
− Hard and resistant minerals (quartz, epidote and serpentine).
− Soft minerals (clay, chlorite, talc and graphite).
− Soluble minerals (calcite, gypsum).
− Swelling minerals (swelling clays (smectites), anhydrite).
− Loose materials (silt, sand and gravel).
These materials are further described in Appendix 2 and 3.

Joints, seams and sometimes even minor faults may be healed through precipitation from
hydrothermal solutions of quartz, epidote or calcite. This may be the case for layered igne-
ous and metamorphic rocks in which the layers are strongly welded together (Burton,
1965); therefore such planes are not planes of weakness in the way bedding planes in sedi-
mentary rocks are, but can be regarded more appropriately as planes of reduced strength.
The influence of joint characteristics has been discussed in several papers, among others
Goodman (1970, 1989), Brekke and Howard (1972), Barton et al. (1974), Barton (1973,
1976), Nieto (1983), ISRM (1978).

2 JOINTING CHARACTERISTICS

Jointing is the occurrence of joint sets forming the system or pattern of joints as well as the
amount or intensity of joints. The network of joints in the massifs between the weakness
zones can according to Selmer-Olsen (1964) be characterized as 'the detailed jointing'.

2.1 Joint sets

Field studies of several workers have shown that rocks are invariably jointed in preferential
directions and occur in joint sets. Two or three prominent sets and one or more minor sets
often occur; in addition random joints may be present. Pollard and Aydin (1988) propose
that each continuous joint set have been formed during a single deformation episode.
 A1 - 6

The conditions of the joints in the various sets can vary greatly depending on their mode of
origin and the type of rocks in which they occur. Not only can the size and average spacing
of joints vary, but also the other characteristics mentioned above. Variations in these
properties cause that one joint set can have very different effect on the shear strength
characteristics than another.

Although some characteristics are common for joints of different sets in a structural region,
it does not, however, seem to be any general connection between all joint conditions in the
different types of rock. Thus, for each of the joint sets, within a structural region with
similar jointing characteristics, the various properties of each set must be considered
individually.

In many cases one joints set is dominant, being both larger and/or more frequent than joints
of other sets in the same locality. This set is often referred to as the main joint set (or by
geologists as primary joints). Often, only one more joint set is developed (Price, 1969).

2.2 Joint spacing

Joint spacings varying from some millimetres to many metres may often seem arbitrary.
There are, however, sometimes certain trends in the density of joints caused by spacings.

Nieto (1983) has observed variations in average spacing between joints from centimetres in
highly tectonized rocks (folded, faulted, and intruded) of all types to more than 10 metres
in massive, horizontally layered rocks. The regularity of joint spacing decreases with the
amount of tectonic activity of the area.

Similarly, Pollard and Aydin (1988) mention that spacing of joints in some sets in intrusive
igneous rocks is not uniform and that distances between joints range from less than 20 cm
to more than 25 m, and that clusters of joints crop out sporadically.

Pollard and Aydin (1988) have further observed regular distribution of joints in
sedimentary rocks and that the spacing of joints can scale with the thickness of the layer.
Nieto (1983) mentions a general trend to a marked increase in the spacing or even the
virtual disappearance of joints in flat-lying sedimentary sequences at depths of as little as
100 m.

Pollard and Aydin (1988) suggest from field data that the following other factors also
influence joint spacing:
− Two joint sets in the same lithologic unit often have different spacings.
− Spacing of joints in different lithologic units of comparable thickness can be
different.
− Spacing can change as a joint set evolves. For example, columnar jointing - initiated
at a flow base show an increase in spacing towards the interior - and the number of
joints in a sedimentary unit decreases with distance from the initiation surface. The
spacing of cooling joints that grow from the top of a lava flow is smaller than the
spacing of those that grow up from the base. This has been attributed to a faster
cooling rate at the flow top.
In addition two other trends should be mentioned:
− Rock masses that have undergone tectonic disturbance often present clusters of joints
(joint zones). Often the joint spacing decreases near faults and shear zones.
 A1 - 7

− Spacing is also influenced by weathering, as there often is an increase in jointing

density within the zone of weathering, especially where mechanical disintegration
has taken place.

2.3 Jointing pattern and block types

Joint patterns comprising of more than one set are common in nature. Piteau (1970) has
observed that in instances where jointing is considered to have random distribution, it is
usually the case that several joint sets occur simultaneously or are superimposed on earlier
sets and the resulting complexity gives the appearance of randomness.

Although there are many varieties of joint patterns in nature, there are few types of joint
intersection geometries, which can be classified as orthogonal (+ intersections) and non-
orthogonal (X intersections), see Fig. A1-3. Both types can be divided into three groups
according to the persistence of the joints at intersections:
1. All joints are persistent (crossing other joints)
2. Some persistent, some non-persistent
3. All joints are non-persistent
This is schematically illustrated in Fig. A1-3.

A B

C D

E F A. Orthogonal pattern, with persistent sets

(+ intersection)

B. Non-orthogonal pattern, with persistent sets

(X intersections)

C. Orthogonal pattern, one set is persistent

(T intersections)

D. Non-orthogonal pattern, one set with

persistent joints
G H
E. Orthogonal pattern, both sets have mainly
discontinuous joints

F. Non-orthogonal pattern, both sets have mainly

discontinuous joints

G. Triple intersections with all joints

o
H. Tripple intersections with 120 angles

Fig. A1-3 Schematic illustration of main joint patterns (from Pollard and Aydin, 1988).
 A1 - 8

Pollard and Aydin (1988) have observed that orthogonal joints often terminate against
persistent joints. They mention, however, that there are many examples of joints that
apparently cut across bedding interfaces and other joints. The + or X types of such
intersections seem to contradict the notion that older discontinuities act as barriers to joint
propagation, as implied by T intersections.

The results from analyses carried out by Kikuchi et al. (1985) of joint connections in
granitic rocks showed that most of the joints belonged to the X type, but also the T type
and the + type were frequently observed. The joint termination type mainly belonged to the
T type. Dershowitz and Einstein (1988) mention that 60% of joints in Stripa, Sweden
terminate at T-type intersections; in other places 42% of this type has been recorded.

According to Price (1969) joints frequently occur in relatively narrow zones, in which one
joint is replaced en echelon by another joint, which is slightly off-set, see Fig. A1-4.

nt
non-p si s te
ersist
ent pe r

Fig. A1-4 Joints are sometimes arranged in zones with reduced spacing and replacing each other en
echelon (modified from Price, 1969).

2.3.1 Block types and sizes

The joint sets and possible random joints divide the rock volumes into characteristic
blocks. The jointing pattern and the difference in spacing between the joints within each
joint set determine the shape of the resulting blocks, which can take the form of cubes,
rhombohedrons, tetrahedrons, sheets etc. The unit block can be described by its volume,
type and shape as described in Appendix 3.

Müller et al. (1970) have made the division of block shapes as shown below in Fig. A1-5.

Shape
of
rock block d3 d3
d3 d3
d2 d2
d2 d1 d2 d1 d2 d1 d1 d1
Joint spacing d3
d (cm)
Ratio
d1 / d3 : d 2 / d3 <1:5 1 : 2 to 1 : 5 1:1 2 : 1 to 5 : 1 >5:1
big block
d max > 100 column paralleliped slab plate
medium block
100 > d max > 10 decimetric cube medium slab medium plate
parallelepiped
pencil small block centimetric cube
d max < 10 parallelepiped
small slab small plate

Fig. A1-5 The classification of block types applied by Müller et al. (1970).
 A1 - 9

Another characterization into block types has been presented by Dearman (1991), based on
a description by Matula and Holzer (1978) as shown in Table A1-2 and Fig. A1-6.

TABLE A1-2 BLOCK TYPES AND JOINTING CHARACTERISTICS (after Dearman, 1991).
Type of block Jointing characteristics

Polyhedral blocks Irregular jointing without arrangement into distinct sets, and of small joints.

Tabular blocks One dominant set of parallel joints, for example bedding planes, with other
non-persistent joints; thickness of blocks much less than length or width.

Prismatic blocks Two dominant sets of joints, approximately orthogonal and parallel, with a
third irregular set; thickness of blocks much less than length or width.

Equidimensional blocks Three dominant sets of joints, approximately orthogonal, with occasional
irregular joints, giving equidimensional blocks.

Rhomboidal blocks Three (or more) dominant mutually oblique sets of joints, giving oblique-sha-
ped, equidimensional blocks.

Columnar blocks Several, - usually more than three, - sets of continuous, parallel joints; length
much greater than other dimensions.

2
3
1
1
1
2 2
1
3
2
2

Polyhedral blocks Equidimensional blocks Prismatic blocks

3 1
1
4
3
5
2
1
1
2 3
2
1
3

Tabular blocks Rhombohedral blocks Columnar blocks

Fig. A1-6 Various types of jointing pattern expressed as block shape. The numbers refer to various joint
sets (from Dearman, 1991, based on data from Matula and Holzer, 1978)

Sen and Eissa (1991) has given the following, simpler characterization of block types:
Prismatic blocks: The three dimensions of these blocks are individually significant
in their definitions.
Platy blocks: These are similar to slabs where two of the three dimensions are
relatively larger than the third dimension.
Bar blocks: Only one dimension is significant.

This division has earlier also been applied by Burton (1965).

However, regular geometric shapes as given above are the exception rather than the rule
since the joints in any one set are seldom consistently parallel (ISRM, 1978). Jointing in
sedimentary and plutonic rocks usually produces the most regular block shapes.
 A1 - 10

Block size delineated by the joint planes is a volumetric expression for jointing density.
Block size is determined by the joint spacings and the number of joint sets - partly also by
the joint length. Individual or random discontinuities may further influence the block size.
The connection between soil diameter (d) and block volume (Vb) is shown in Table A1-3.
Here, a transition Vb = 0.58 d 3 has been applied.

Table A1-3 PARTICLE OR BLOCK DIAMETER ACCORDING TO THE

MODIFIED WENTWORTH SCALE RELATED TO VOLUME
Term Diameter (d) Volume (Vb)
Boulders 200 - 600 mm 4.6 dm3 - 0.125 m³
Cobbles 60 - 200 mm 0.13 dm3 - 4.6 dm3
Coarse gravel 20 - 60 mm 4.6 cm3 - 0.13 dm³
Medium gravel 6 - 20 mm 0.13 cm3 - 4.6 cm³
Fine gravel 2 - 6 mm 4.6 mm3 - 0.13 cm3
Coarse sand 0.6 - 2 mm 0.13 mm3 - 4.6 mm3
Medium sand 0.2 - 0.6 mm 0.0046 mm3 - 0.13 mm3
Fine sand 0.06 - 0.2 mm
Silt, clay < 0.06 mm

The features influencing on the block size are further outlined in Section 3 in Appendix 3.

3 ATTITUDE OF JOINTS

Nieto (1983) has observed that in the subsurface, high to very high dips (60 - 90o) and low
dips (0 - 30o) appear to predominate over intermediate dips, in a large number of geologic
settings. This seems to be particularly true in intrusive rocks (granite, diorite, and gabbro
families), high-grade metamorphics (migmatites and gneisses of various types) and
horizontally layered rocks. Selmer-Olsen (1964) also reports that a major part of the joints
in the Palaeozoic mountain range (Caledonides) in Scandinavia are steep-dipping.

4 DEVELOPMENT OF JOINTING IN VARIOUS ROCKS

Often, the joints have certain common characteristics within the type of rock they are
developed. There are also certain patterns connected to the type of rock they penetrate. The
size of joint planes is largely related to the lithology and the size of the rock units being
studied. Knowledge of the rock type and its properties may often help in evaluating the
characteristics of the joints.

Terzaghi (1946) mentions that in many homogeneous rocks other than extrusive igneous
rocks, the joints are either discontinuous or so irregular that the blocks located between
them are intimately interlocked.

In some layered igneous and metamorphic rocks in which the layers are strongly welded
together, joints, seams and sometimes even minor faults may be healed through precipi-
tation from solutions of quartz, epidote or calcite (Burton, 1965). Such planes separating
the layers are not planes of weakness in the way bedding planes in sedimentary rocks are,
but can be regarded more appropriately as planes of reduced strength.

4.1 Jointing in igneous rocks

In plutonic igneous rock such as granite, gabbro, diorite, etc. usually three sets of joints are
developed caused by tensional forces set up in a rock body as a result of cooling. Among
 A1 - 11

these three sets of joints, two are often vertical and perpendicular to each other, while one
will be more or less horizontal (sheet joints). They divide the rock into more ore less
prismatic blocks. The set of joints, which commonly are called sheet joints, are more or
less parallel to the surface of the ground. The existence of such joints enable the extraction
of rock slabs (Terzaghi, 1946; Eshwaraiah and Upadhyaya, 1990).

The joint spacing in igneous rocks may range between some centimetres and many metres.
Fresh joints are often medium sized rough and planar. In some areas the orientation and the
spacing of the joints in granite is almost constant over large areas, whereas in other areas it
changes from place to place in an erratic manner. Regular, large blocks developed in
decorative rocks often facilitates extraction of natural stones in plutonic rocks.

In basaltic rocks where uniform cooling and contraction in a homogeneous magma has
taken place, columnar jointing is common. It consists of hexagonal columns orientated at
right angles to the surface of cooling. The columns commonly measure from one to three
decimetres across. Since the joints between the columns are open, water can circulate
freely through them. Terzaghi (1946) mentions that in igneous rocks which cooled rapidly
the joints are generally closely spaced, and that in contrast to basalt, rhyolite has a tendency
to develop closely spaced and irregular joints.

4.2 Jointing in sedimentary rocks

Sedimentary rocks also commonly contain three sets of joints, one of which is invariably
parallel to the bedding planes. The other joints commonly intersect the planes at
approximately right angles (Piteau, 1970; Terzaghi, 1946; Deere et al., 1969).

Even when strong sedimentary rocks like limestone and well cemented sandstones predominate,
thin argillaceous intercalations ("shale partings") can introduce pervasive weakness planes.

In limestone and sandstone, the joint spacing of each set are commonly of metre size. In
shale, they are generally closer, and they may be so close that no intact specimen can be
secured with a width of more than a centimetre (Terzaghi, 1946). During excavation, such
shales can disintegrate into small angular fragments along very small weakness planes. The
surfaces of the fragments of some shales are shining and striated from slickensides. Nieto
(1983) has observed that flat-lying sedimentary rocks display the most regular spacing.

4.3 Jointing in metamorphic rocks

In metamorphic rocks one joint set is often parallel or sub-parallel to the foliation or
schistocity with two or more sets of joints oriented approximately at right angles to this
direction (Deere et al., 1969; Piteau, 1970; Terzaghi, 1946). Varying amount of random
joints are present in addition to the joint sets. In many cases the jointing has a character of
irregular as the amount of random joints exceeds the joints connected to joint sets.

Intercalated gneisses and schists, phyllites and slates usually display well developed
foliation planes which contain concentration of weak, platy or elongated minerals of mica,
chlorite, amphiboles, pyroxenes. These planes are often easily splitted to form foliation
joints (Nieto, 1983).

The most significant direction of weakness (cleavage) in metamorphic rocks can be

independent of the primary layering after the rock has undergone regional metamorphism.
Selmer-Olsen (1964) mentions that where tensile and shear stresses probably had other
directions than the cleavage in the rock, cleavage partings and joints cut each other at
 A1 - 12

oblique angles to form rhombohedral blocks, see Fig. A3-32 in Appendix 3. This type of
pattern is often found in regions with metamorphism in connection with mountain range
foldings, and in fault zones of crushed rocks developed by shear stresses.

5 STATISTICAL DISTRIBUTION OF JOINTS

The most commonly measured geometric properties of jointing are spacing (or density),
trace length, and orientation. Based on results obtained by many workers, the statistical
distribution of joint density can often, as shown in Table A1-4, be modeled by an
exponential function.

Reported distributions of joint trace length are less consistent than those for spacing,
perhaps caused in part by strong biases implicit in many common sampling plans and in
part by the way data are grouped into histograms prior to analysis. Log-normal distributions
are perhaps the most frequently reported, but given size biases in the way samples are
collected, many different in situ distributions would produce approximately log-normal
samples (Baecher and Lanney, 1978). Many workers have used exponential distributions in
analysis, primarily for computational convenience, but there is little empirical verification
of this assumption, see Table A1-4.

From studies made for a probabilistic slope stability analysis Herget (1982) found log-
normal distribution for dip, dip direction, hardness, and strength of fillings, and negative
exponential distribution for spacing, trace length, and waviness.

As noted by Hudson and Priest (1970), since the exponential density of joints is fully
defined by one parameter, the following simple relation exists between rock quality
designation (RQD) and average joint spacing (l) for hard, unweathered rocks:

RQD = 100 × e - 0.1 l (0.1 l + 1) eq. (A1-1)

TABLE A1-4 STATISTICAL DISTRIBUTION OF JOINTS (based mainly on Merritt and Baecher (1981)
---------------------------------------------------------------------------------------------------------------------------------
Source Spacing Trace length Shape
---------------------------------------------------------------------------------------------------------------------------------
Snow (1968) exponential - -
Robertson (1970) - exponential equidimensional
Louis and Perrot (1972) exponential - -
McMahon (1974) - log-normal -
Steffen et al. (1975) - exponential -
Bridges (1976) - log-normal oblong
Call, Savely, Nicholas (1976) exponential exponential -
Priest and Hudson (1975) exponential - -
Baecher, Lanney, Einstein (1977) exponential log-normal equidimensional
Barton (1977) - log-normal equidimensional
Cruden (1977) - censored exp. -
Baecher and Lanney (1978) exponential log-normal or exp. -
Herget (1982) exponential exponential -
---------------------------------------------------------------------------------------------------------------------------------
 A1 - 13

6 SUMMARY

Many igneous and metamorphic rocks without planar orientation of the minerals may have
regular jointing systems with three ore more sets at right angles forming prismatic blocks.
Often, they are characterized by medium-sized, rough and planar joints.

Bedded sedimentary rocks and foliated or schistose metamorphic rocks show separations
as partings parallel to the bedding or foliation in the rock. This direction is often along
layers or bands of mica and/or chlorite, both minerals with elastic, anisotropic mechanical
properties. They mainly cause that the partings or joints have a soft and weak joint walls
with smooth surfaces.

Persistence of joints seems to be greater in rocks with well developed foliation, layering,
or bedding than in the homogeneous rocks. The other joint sets are often developed
perpendicular to the weakness direction.

Joint characteristics are often similar within the same joint set, while it can be different
between joint sets in the same region.

Although the jointing often is composed mainly of joint sets, additional random joints
frequently disturb the pattern. Variations in spacings, number of joint sets and random
joints generally result in great differences in block size and shapes within an area or a
location.

Weathering occurs more frequently along the joints than in the adjacent rock block.

Joint spacing and trace length exhibit generally a logarithmic or exponential distribution.
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

APPENDIX 2

ON FAULTS AND WEAKNESS ZONES

"It remains just as necessary today as it ever has been for the engineer to simplify to under-
stand and to visualise the problem. It is only the retention of this ability that enables the
engineer to react appropriately when, as ever, he encounters in nature some features unex-
pected in his original concept."
A. M. Muir Wood, 1979

According to ISRM (1978) a fault is "A discontinuity zone along which there has been
recognisable displacement, from a few centimetres to a few kilometres in scale. The walls
are often striated and polished (slickensided) resulting from the shear displacement. Fre-
quently rock on both sides of a fault is shattered and altered or weathered, resulting in
fillings such as breccia and gouge. Faults may vary from millimetres to hundreds of
metres."

To be characterized as 'fault' it is thus required that there is a proof of movement. Simi-

larly, a gouge is (ISRM, 1978; Dictionary of geological terms, 1962): "A clay-like material
occurring between the walls of a fault as a result of the movements along the fault
surfaces."

Knowledge of the origin or formation of the zone can be helpful in working out the de-
scription and evaluation of its composition and structure. But used in rock engineering and
construction the features of main interest are connected to the actual properties and behav-
iour of the zone. Therefore, the more general term 'weakness zone', including also other
weaknesses, is used in this work. This term has also been recommended by the Norwegian
Rock Mechanics Group (1985) for large structural lineaments in the earth's crust, defined
as: "A weakness zone constitutes a part of the ground in which the mechanical properties
are significantly lower than those of the surrounding rock mass."

Similarly, together with gouge which is related to faults, the more general term filling is
often used for the finer, often clay-like material occurring between the walls of a seam
(filled joint) or a weakness zone.

Although weakness zones basically can be said to be composed of mainly rock(s) in

addition to joints and seams with or without filling, a great variety exist. Many of them are
formed as a result of tectonic events, while other are related to layers of weak rocks
surrounded by stronger rock masses. Common for them all is that they form zones, lenses,
veins or layers in almost all types of rocks. Basically, there are two main groups of weak-
ness zones: those which are formed from tectonic events, and those consisting of weak
materials formed by other processes. Weathering, hydrothermal activity and alteration are
features that may have had a significant impact on the composition and properties of a
zone.

Selmer-Olsen (1964, 1971) who has studied many of the weakness zones encountered in
more than 2000 km of Norwegian tunnels in crystalline Precambrian and Palaeozoic rocks,
has worked out the following division for use in engineering geology:
 A2 - 2

• Zones of weak materials.

• Fault and fracture zones:
- tension fracture zones;
- shear fault and fracture zones (crushed zones);
- altered faults and fracture zones.
• Weathered and recrystallized zones.

The remaining part of this section is mainly short descriptions of zones based on the
division above.

1 ZONES OF WEAK MATERIALS

These types of weakness zones consist mainly of:

- Layers, veins or dykes of soft or weak minerals.
- Zones of weak rocks or of rocks which are heavily jointed.
- Deposits and weathered rocks.

Many of these types of weak materials are only regarded as weakness zones if they are
surrounded by other, stronger rock masses. Some of the types here can also sometimes be
regarded as crushed zones (see Section 2.5.1 - 2.5.4), but may belong to such weak materi-
als because the brecciation is related solely to the rock in the zone which has a limited
extension as a band, layer, lens, vein or zone in the surrounding rock masses.

weak layer

schistocity parting

foliation joint

Fig. A2-1 Zone of weak rock, for example chlorite or talc schist in a phyllite.

Fig. A2-2 The action of weathering along joints, rock boundaries and crushed zones in and near the
surface (from Morfeldt, 1976).
 A2 - 3

The (weak) material in these zones may consist of clay, pegmatite, mica or chlorite, poorly
cemented sedimentary layers (for example tuff layers in basalts), or coal layers (seams).
The zone has often a sharp boundary to the adjacent stronger rocks, as shown in Fig. A2-1.

Also weathered surface and near surface occurrences belong to this group. The weathering
process has often acted along rock layers, dykes or rock contacts, or along joints, seams,
and crushed zones to form zones, layers or pockets of weathering products with low
mechanical properties as shown in Fig. A2-2.

2 FAULTS AND FRACTURE ZONES

Large faults are the major rupture zones encountered in the earth's crust. It is important to
realize that most of the larger faults and fault zones are the result of numerous ruptures
throughout geological time and that their composition and magnitude may vary largely.

Faults and fracture zones are a result from the effects of regional tectonic compression or
tensile stresses accompanying uplift or lateral stretching (ISRM, 1978). They can have
been formed through failure in extension/tension, in shear, or in more complex failure
through a combination of both, see Fig. A2-3. Rupture surfaces from extension are charac-
teristically rough and clean with little detritus, while simple surfaces from shearing are
characteristically smooth with considerable detritus. Because these two main modes, shear
and tension, generally result in different structure and composition it is convenient to
distinguish between them as has been done in Section 2.4 and 2.5.

The tunnelling problems associated with a fracture or fault will generally increase with its
width. However, this factor should always be assessed in relation to the attitude of the fault
and to:
- the frequency, orientation, and character of adjacent joint sets;
- the existence of adjacent seams or faults (if any); and
- the competence of the wall rock type.
Several severe slides in tunnels have occurred where each individual seam or fault has
been of a small width, but where the interplay between several seams and joints has lead to
the instability and failure (Brekke and Howard, 1972).
The main type of weakness zones formed from faulting are:
1 Tension fracture zones
- filled zones
2 Shear fracture and fault zones
- coarse-fragmented crushed zone
- small-fragmented crushed zones
- sand-rich crushed zones
- clay-rich crushed zones, such as:
> simple, clay-rich zones
> complex, clay-rich zones
> unilateral, clay-rich zones
- foliation shears
3 Altered faults and fractures
- altered clay-rich zones
- altered veins/dykes
- altered, leached (crushed) zones.
 A2 - 4
σ2 σ1
6 5 6 σ
5 3

σ3
σ2
σ
1
4

1: NormalfFault
2: Listric normal fault
3: Sole or floor fault
4: Fault zone
5: Horst
6: Graben
7: Half graben
8: Footwall
3 8 9 2 9: Hanging wall
4
8 9
1
7

2 2
3
1

2
3

1: Reverse fault
2: Fault set

1 1 1

2
1: Sinistral fault
3
2 2: Dextral fault
1 3: Fault system
4: Fault set
1
5: Fault surface

3
2
5

3 σ3: Axis of minimum stress

σ2: Axis of mean stress
σ1: Axis of maximum stress
1

Fig. A2-3 Various types of faults and features related to extension regimes (upper), and compression
regimes (middle and lower) (from Nystuen, 1989)
 A2 - 5

Lineamen t

Surface trace of: Surface trace of: Surface trace of: Map view of: Map view of:
fracture zone, fold rock boundary linear linearly-
joint zone, geophysical aligned,
fault, anomaly rock bodies,
etc. volcanic
islands,
etc.

Fig. A2-4 Faults and weakness zones are one group of lineaments, i.e. topographical features reflecting
surface traces of various crustal structures (from Nystuen, 1989).

E
L AK
RN
NE
VÄ

E
AK
NL
TER
VÄT
GOTHENBURG

GOTHENBURG

MALMÖ 50 km

Fig. A2-5 Three dimensions of regional lineament (upper) and fault pattern in southern Sweden
(compiled from Tirén and Beckholmen, 1992).
 A2 - 6

2.1 Occurrence of faults and fractures

The outcropping of large faults and fault zones often form depressions in the topography
forming part of the lineaments in the earth's surface, see Fig. A2-4. From surface observa-
tions it is, however, seldom possible to observe their composition and size because of
overburden and weathering.

Faults constitute characteristic regional patterns in the earth's crust (Fig. A2 -5) consisting of
several mutual independent sets or systems (Selmer-Olsen, 1964). 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.

Although the filling material or gouge in a fault may be only some centimetres wide, the
overall affected zone with open or altered joints may be some metres wide and the extent of
the length of the fault zone may be from hundred metres to more than a kilometre. Fault
zones can be found hundreds of meters below surface (Brekke and Selmer-Olsen, 1965).
Sometimes clay fillings with a very low degree of consolidation have been encountered.

Nieto (1983) and Gillespie et al. (1992) have observed that the length generally is propor-
tional to the thickness of the fault, Fig. A2-6. Thus thick crushed- and gouge-filled faults can
be traced for very long distances (kilometres) whereas thin, gouge-filled faults and shears
usually extend for only some tens of metres. These trends are most prevalent in igneous and
metamorphic terrains. Cowie and Scholz (1992) have shown that the length of a zone can be
proportional also to the displacements caused by the ruptures, Fig. A2-6.
6
10 6
0
5
10 10 5
γ=
4
10 4
maximum displacement (m)

-3
log(displacement) [m]

10
γ=
3
10 3

2
10 2

1
10 1

0
10 0
L = 10 km

-1
10 -1

-2
10 -2

-3
10 -3
-4 -3 -2 -1 0 1 2 3 4
10 10 10 10 10 10 10 10 10 -2 -1 0 1 2 3 4 5 6 7
fault length, L (km)
log(width) [m]

Fig. A2-6 The correlation between displacement and fault length (left: from Cowie and Scholz, 1992) and
between displacement and fault width (righ: from Gillespie et al., 1992).

Generally, faults will mostly be developed in the rocks which from a mechanical point of
view are the weakest in the area. Where two faults of different orientations meet or cross
each other a larger part of the rock masses suffer from increased crushing or jointing, see Fig.
A2-7.
 A2 - 7

Fig. A2-7 The crossing of two faults (from Selmer-Olsen, 1964)

Nieto (1983) has observed that the strike and dip of fault zones in igneous and high-grade
metamorphic terrains can vary sometimes drastically, over relatively short distances. Faults in
anisotropic rocks at an acute angle to the schistocity partly follows along the structure of the
rock, and partly across it. By this a zigzag course is developed. Selmer-Olsen (1950) reports
that the width of such faults are thinner where they follow the rock structure, refer to Fig. A2-8
(A and B). The thickness of a fault is also often reduced when it passes from brittle (granite)
into less brittle (dolerite), as seen in Fig. A2-8 (C).

DIRECTION OF ZONE
A

MICASCHIST

Fault developed at an accute angle to the schistocity in a micaschist

DIRECTION OF ZONE

B
GNEISS

Fault developed at an acute angle to the schistocity in a gneiss

C
DIRECTION OF ZONE

QUARTZITE QUARTZITE

Fault passing a dolerite vein with other properties than the surrounding rocks

Fig. A2-8 Principles in variation in course, thickness and structure of a fault dependent on the rock (revised
from Selmer-Olsen, 1964).
 A2 - 8

2.2 Composition and structure of faults

Faults and fracture zones can vary in composition from mostly brecciated or crushed
material with relatively small amounts of clay to highly weathered or hydrothermally
altered, highly plastic, swelling clay gouge. The composition of these pieces can be similar
to the adjacent rock, or hydrothermal solutions can have altered the original rock material
and/or brought in and deposited new minerals that are not associated with the petrography
of the adjacent rock or the wall rock. This is most important since it implies that the
mineralization in faults and seams is not necessarily a function of the composition of the
'host' rock (Brekke and Howard, 1972). Therefore, any mineralization may be found in a
fault where hydrothermal activity and/or alteration has taken place.

Faults in brittle rocks such as granites, quartzites and some sandstones are likely to be
developed as relatively wide "crushed zones" from crushing or brecciation into blocky
material, consisting of relatively large angular fragments of broken rock surrounded by
finer gouge material. Since this type of faulting constitute underground drains (Terzaghi,
1946), they may be invaded by surface waters or by hot solutions coming from below, or
by both. Chemical alteration in such zones is likely to be more intense and more varied
than in zones containing mainly (impervious) gouge produced by the crushing.

Faults are seldom developed as large crushed zones in 'ductile' rocks such as shales, many
schists, and some of the basic igneous rocks. In these rocks faulting mainly occurs as
plastic deformation like flexure and intense folding, resulting in more narrow zones richer
in fine-grained material. A large number of shear surfaces and slickensides may be present
in the zone.

As mentioned in Section 2.2 in Appendix 1, faults are quite often associated with other
parallel discontinuities that decrease in frequency and size in the direction away from the
central zone of the fault (Terzaghi, 1946). Thus, as a tunnel approaches a zone of intense
crushing, it passes through rock masses which are more and more heavily jointed.

Apart from the sand-like materials, fault gouge itself is normally impervious. Possible high
permeability connected to faults can be due to the high jointing density often found adja-
cent to the fault as in unilateral crushed zones (see Section 2.5.4). High water inflows in
tunnels and underground openings encountered after excavating through the impervious
gouge is one of the most adverse conditions associated with faults as reported by Brekke
and Howard (1972).

Another significant problem in tunnelling through faults may be caused by materials that
have been crushed to an almost cohesionless (sand-like) material which may run or flow
into the tunnel immediately after blasting (see Section 2.5.3 and Chapter 6, Section 4.3).

2.3 Gouge (filling materials) in faults

Brekke and Howard (1972) reports that the character of the gouge material in faults is very
seldom uniform. The gouge will sometimes have the character of unaltered crushed rock.
"Blocks, or even plates, of intact rock may 'float' in a basic matrix of soft material. In
addition bands or seams of hard material such as quartz or calcite may occur." Thus, fault
gouge constitutes normally a very complex material both in regard to mineralization and in
regard to physical properties.
 A2 - 9

Subsequent possible hydrothermal alteration of the 'original' rock and the gouge material
and/or the deposition of hydrothermal products will complicate the mineralogical identifi-
cation since products not associated with the petrography of the crushed rock or wall rock
may be present.

Brekke and Howard (1972) have given an overview of the major types of materials that
can be found in weakness zones and faults. This is shown in Table A2-1 together with the
potential behaviour of the material in excavations. The basic division is made according to
the mineral or material that dominates the properties of the filling. This is not necessarily
the most abundant material.

Table A2-1 GROUPING OF FILLING AND GOUGE MATERIALS AND THEIR POTENTIAL BEHAV-
IOUR (modified after Brekke Howard, 1972).
----------------------------------------------------------------------------------------------------------------------------------
Dominant material in filling (or gouge) Characteristic behaviour
-----------------------------------------------------------------------------------------------------------------------------------
1. SWELLING CLAY Swelling, sloughing and squeeze

2. INACTIVE CLAY Slaking and sloughing caused by squeeze

3. CHLORITE, TALC, GRAPHITE, SERPENTINE Ravelling

4. POROUS OR FLAKY CALCITE, GYPSUM May dissolve

5. QUARTZ, EPIDOTE Durable, high strength

6. CRUSHED ROCK FRAGMENTS (gravel size) Ravelling or running.
OR SAND-LIKE FILLING

7. FILLING OF OTHER ROCKS

----------------------------------------------------------------------------------------------------------------------------------

The often complex composition and structure of the gouge in a given fault may well
overlap several of these classes. Fault gouge and filling materials are normally impervious,
with the major exception for sand and gravel-like compositions (Brekke and Howard,
1972).

The third group listed in Table A2-1 is intended to cover blocky gouge material that is
heavily interwoven with slickensided seams and joints filled or coated with the minerals
listed (Brekke and Howard, 1972). The characteristic property of such gouge is low shear
strength, in particular when wet.

In zones with swelling clays (smectite) the initial and later change of water content can be
important for the mobilized swelling pressures. In addition, swelling clays have both a low
shearing strength. Several authors (Piteau, 1970; Brekke and Selmer-Olsen, 1965; Selmer-
Olsen and Palmström, 1989,1990) describe stability problems with rock falls, slides, and in
some cases, collapses caused by the swelling of clays in joint fillings and faults.

2.4 Tension fracture zones

These zones are often developed with a filling of soft minerals between parallel walls
having a low degree of jointing. The filling material can be chlorite, (swelling) clay,
porous calcite, silt etc., and the zone can be named according to the dominant filling; for
example clay-filled zone, or calcite-filled zone. Feather or pinnate zones ("fiederspalten")
belong to this group. There is generally a sharp boundary to the adjacent rocks in tension
fault zones. These types of zones may sometimes also be classified as zones of weak
materials described in Section 1; this depends whether the characterization is based on the
formation of the zone or from its material.
 A2 - 10

2.5 Shear fault and fracture zones

Shear faults and shear fault zones are formed as a result from shear strain which causes the
rock mass to be crushed and brecciated by many intersecting joints and/or seams, as in Fig.
A2-9. Their central part may sometimes be weathered or completely altered to clay. Faults
formed in shear are in general much more susceptible to alteration than are those formed in
extension. Normally, such alteration leads to a lowering of the strength and to other
disadvantageous conditions for construction. 1

Shear zones can vary in width from a few centimetres to several metres, and may be termed
crushed zone in the case of hard rock. In the case of metamorphic rocks shear zones may
occur parallel with the foliation, typically along weak mica-rich rocks. In such cases they
are often termed foliation shear zones.

Fiederspalten Fiederspalten and Crushed zone Crushed zone

shear joints

Fig. A2-9 Principles in the development of crushed zones (from Selmer-Olsen, 1964).

The main parts of a typical shear zone may consist of:

• The central part. This is where the main movements mainly have taken place,
which have resulted in intense jointing or crushing of the rock, with possible hydrother-
mal activity and deposition of minerals and/or alteration.
• The transition part, i.e. the area disturbed by the movements with a higher degree of
jointing than in the adjacent rocks where also alteration can have taken place.
• The surrounding rock masses which have been little influenced by the movements, but
sometimes penetrated by seams and minor faults that branch out from the zone.

The main elements (both in the central and the transition part) of the zone are:
- Blocks and/or fragments.
- Filling (or gouge) of clay-like or sandy material. Thicker fillings are generally
restricted to the central part. Also the transition part often contain joints and seams
but in a lesser amount than in the central part.
- Coating or slickensides on joint surfaces.
- Alteration of blocks/particles both in the central and the transition part.

1
In the case of bedding plane slip caused by folding or basin formation, shear zones may
develop in, for example, interbedded layers of over-consolidated clay, with the formation of lenses
and slickensided shear surfaces. In both cases the relative shear displacement of the discontinuity
walls is insufficient for the zone to be characterized as a fault.
 A2 - 11

There is seldom a uniform composition of the central part; sometimes there are two or
more central parts with transition areas between them as is the case in complex zones
described in later in this section, see Table A2-2. Brekke and Howard (1972) mention that
the blocky gouge material may be heavily interwoven with slickensided seams and joints
filled or coated with clay, mica, chlorite etc.

Crushed zones are most commonly found in hard rock provinces. They are often pervasive
and relatively planar seen along a distance of some hundred metres. Crushed zones have
most often a gradually reduced degree of jointing from their central part to the surrounding
rock masses. There are numerous types of crushed zones of which the main are listed
below. Consequently, there are many intermediate occurrences between these.

Caine and Foster (1993) have developed the scheme shown in Fig. A2-10 for permeability
structures in fault zones. The four end members in this classification scheme are based on
the content of subsidiary structures (i.e. joints, veins, seams) and gouge (i.e. filling materi-
als) are the variables used for the main groups.

II low % gouge high IV

high high
accretionary Stillwater
prisms fault

% subsidary Permeability Structures % subsidary

structures In Fault Zones structures

Shawangunk San Gabriel

low Mountains low
ultracatalclasites

I low % gouge high III

Fig. A2-10 The main groups of fault zones with respect to permeability based on the content of joints,
and gouge (from Caine and Foster (1993).

End member I includes faults free of both gouge and subsidiary structures, which means
single discontinuities which in the field often are termed large joints. End member II and
III are cases where either gouge or subsidiary structures are absent, while group IV con-
sists of a well developed central gouge zone in addition to replete subsidiary structures.
Most of the various types of shear faults described in the following can be correlated to
group II, III or IV.
 A2 - 12

2.5.1 Coarse-fragmented crushed zones

These zones, shown in Fig. A2-11 (left), have blocks over the entire width, often with
larger blocks towards the adjacent rock masses, and belong generally to end member II in
Fig. A2-10. The blocks are often slickensided with or without clay coatings, and the
individual blocks have often rhombohedral shape.

Fig. A2-11 Left: Coarse-fragmented crushed zone.

Right: Small-fragmented crushed zone (from Rokoengen, 1973).

2.5.2 Small-fragmented crushed zones

Such zones have a central zone with fragments of gravel size. There are generally few
seams or clay filled joints but lots of small slickensided fissures, Fig. A2-11 right. A
gradual transition to larger blocks in the surrounding rock masses is common. This type of
crushed zone can be classified as end member II in Fig. A2-10.

2.5.3 Sand-rich crushed zones

If brittle rocks are subjected to intense deformation, they may fracture to such an extent
that "the fragments may even be reduced to powder as if it had passed a crushing machine"
(Terzaghi, 1946). In such cases the rock is completely crushed or decomposed to contain
rock fragments of gravel or sand size, i.e. materials with a typical cohesionless behaviour.

Sand-rich crushed zones may cause major excavation problems having very short stand-up
time where typical earth tunnelling conditions may be encountered. As further described in
Chapter 6, Sections 2 and 4.3 running or flowing ground may occur in connection with
such zones (Brekke and Howard, 1972). Ward (1978) mentions examples from the Alps
and Himalayas where wide thrust zones contain a dense pulverized mass, or a highly
slickensided mylonite with typical earth tunnelling conditions.

2.5.4 Clay-rich crushed zones

Clay-rich zones show wide variations in composition and structure. They have often a
central clay-rich zone in addition to more scattered clay-filled joints in the less crushed
transition zone to the adjacent rock masses, and they can therefore be classified as end
member IV in Fig. A2-10. The main types are:
 A2 - 13

i Simple, clay-rich, crushed zones with blocky composition. The joints have gener-
ally a spacing of 0.05 - 0.5 m and cut through the foliation of the rock. Sometime
they occur as long, smooth and planar joints or seams extending several tens of
metres out from the zone, see Fig. A2-12 left.

CLAY MATERIAL

Fig. A2-12 Left: Simple clay-rich, crushed zone.

Right: Complex clay-rich, crushed zone. (from Selmer-Olsen, 1971).

ii Complex clay-rich crushed zones, which in the central part show several clay-rich
zones or seams, often with crushed rocks between them (Fig. A2-12 right). Two or
more close crushed zones should be regarded as one complex crushed zone in
tunnel stability evaluations, provided that the distance between their central crushed
part is less than about 1.5 times the span of the tunnel/cavern.
Also small-fragmented and coarse-fragmented crushed zones can occur as complex
zones.
iii Unilateral clay-rich crushed zones are zones where the seams, shears or filled
joints are concentrated in one side of the zone with a sharp boundary to the
surrounding rock masses.. Such zones may contain highly permeable brecciated
rocks adjacent to highly impermeable clay gouge with strongly anisotropic water
conducting flow parallel to the plane of the fault. Tunnelling from the impermeable
side in a large zone of this type may, as mentioned, cause considerable excavation
problems as low stability and water inflow may occur at the same time.

2.5.5 Foliation shear zones

Foliation shear zones or "foliation shears" occur in metamorphic rocks as thin sheared
zones along the foliation of metamorphic rocks, often in mica-rich schists (Patton and
Deere, 1970; Deere 1971). Such mica-rich rocks are reported in thick schist sequences but,
more importantly, also as thin interbeds in massive metamorphic gneisses and quartzites.
"The shears trend parallel, or in some instances subparallel, to the foliation. Locally the
shears may cut across the foliation where they flatten or roll. Typically, the shears will
thicken and thin somewhat". They belong to end member III or between II and III in Fig.
A2-10.

In addition to the clayey filling which may be only some centimetres wide, the overall
affected zone with partially crushed, sheared and slickensided rock may be some metres
wide, Fig. A2-13. The extent of the zone along its trend may be from hundred metres to
more than a kilometre (Deere, 1971). Differential slippage along weak micaceous interbeds
during folding or stress relief probably accounts for the origin of most of them.
 A2 - 14

The joints in the hard rock adjacent mica-rich schist containing the shear zone, may be
somewhat disturbed with some loosening, slickensiding, and chemical alteration (thin clay
or chlorite coating on the joint surface). Thinner foliation shears may occur parallel or sub-
parallel to the main one at distance of some meters.

Shear
zone

Gouge

Fig. A2-13 Typical foliation shear zone (from Cording and Mahar, 1974).

2.6 Altered faults

Alteration of faults may take place in most types of the weakness zones described above.
The alteration processes may occur during the formation of the zone and/or later.

2.6.1 Altered clay-rich zones

These zones are characterized by alteration to clay of feldspar in the zone and in the
adjacent rocks. The (hydrothermal) alteration is mainly related to crushed zones, some-
times also to seams and clay-filled joints and zones. Compared to many other crushed
zones this type of altered weakness zones is often highly consolidated and almost impervi-
ous. The clay has often high swelling properties (Brekke and Howard, 1972).

Fig. A2-14 Altered, clay-rich zone (from Selmer-Olsen, 1971)

 A2 - 15

2.6.2 Altered, leached (crushed) zones

These are normally smaller coarse-fragmented or small-fragmented crushed zones where

the rocks in the transition zone have been dissolved to form permeable materials. The
minerals removed are generally quartz and feldspar or carbonate.

As mentioned earlier, a real hazard exists 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 then be released.

3 RECRYSTALLIZED AND CEMENTED/WELDED ZONES

Recrystallization may cause significant changes to the composition, properties and behav-
iour of a weakness zone. These types, which probably earlier have been coarse-fragmented
or complex, crushed zones, are still geologically named faults or thrust zones. Still they
often have some slickensided and clay filled joints, with secondary formed minerals of
epidote, quartz, feldspar, calcite siderite and/or chlorite which have "welded" the blocks
and 'reinforced' the zone.

4 DESCRIPTION OF WEAKNESS ZONES

Larger faults and weakness zones should be described and mapped as structural regions of
their own in connection with rock construction (Bieniawski, 1984). Core drilling from the
surface or probing ahead from an advancing heading are the most effective means of
collecting information of a zone before it is penetrated. It can, however, be difficult to
obtain enough data to fully describe its structure, especially in case of core loss.

After the zone has been encountered in the excavation its composition and structure can be
studied. However, only a small part of it is 'opened'. It is therefore difficult to observe and
measure other features of a weakness zone than its orientation, and local thickness, com-
position and structure. An adequate description of these features is very important for the
decision of excavation procedure(s) and for the various analyses included in the design for
appropriate rock support. Where time is available the description can be backed up by
laboratory tests to measure the properties of important features.
A description of a weakness zone would consequently consist of:
• Size, measured as total thickness, formed by
- the thickness of the central part, and
- the thickness of the transition part.
• Composition and structure (arrangement) of the zone, in
- the central part, made up of either
mainly blocks, or
blocks and fillings (gouge), or
mainly fillings (gouge);
- the transition part to the adjacent rocks, being either
sharp or gradual.
• Possible alteration different from that of the surrounding rocks) in
- the central part, and
- the transition part.
 A2 - 16

In the case of important weakness zones it is helpful to make idealized sketches showing
the estimates of the principal dimensions (ISRM, 1978). A verbal description of these
features should always be given so that extent and character of the discontinuity is commu-
nicated (ISRM, 1975).

Hints on description of weakness zones are also given in Appendix 3, Section 5.

Many weakness zones do not have a well defined thickness, but show a gradual transition
from the central part to the surrounding rock masses. The description of a zone should pay
attention to this. Also, the conditions of the surrounding rock masses may be of importance
for the rock mass behaviour in connection with the weakness zone. Of special importance
is the occurrence of nearby weakness zones, as well as seams, shears or small faults
connected to the main zone.

The following is an example of a description of a 10 m wide weakness zone encountered in

the 25 m2 headrace tunnel at the Haukrei power plant in Norway:
The zone has an orientation strike/dip = N 45oE/90o related to the tunnel. It consists of
a partly chloritized diabase. It is formed by several parts having different composition
as described in Table A2-2.
The surrounding rocks consist of Precambrian gneiss and granitic gneiss with
strike/dip = N 20-30oE/70-80o with 1 - 5 m long rough and undulating foliation joints
spaced 0.5 - 2 m. Some random joints occur.

TABLE A2-2 COMPOSITION OF THE WEAKNESS ZONE AT HAUKREI POWER PLANT

thickness of individual parts of the central part
FEATURE adjacent (m) adjacent
rock rock
0.5 - 1 1-2 2 1-2 1 -2 0.4 - 0.5 0.5 -1

Joint spacing (m), set 1 0.5 - 2 0.01-0.05 0.1-0.5 0.05-0.2 0.02-0.1 0.01-0.05 a zone 0.2-0.3 0.5-2
set 2 0.3-1 mainly of
Joint length (m), set 1 1-5 0.1-0.5 0.5-3 0.3-3 0.3-2 0.1-1 chloritic 0.5-2 1-5
set 2 0.01-0.1 clay
Joint smoothness rough smooth smooth smooth smooth smooth rough rough
Joint waviness undul. undul. undul. undul. undul. undul. planar undul.
Joint alteration fresh
or coating chlorite chlorite chlorite chlorite chlorite

Random joints a few a few a few

3 3 3 3 3
Block volume, min. 0.3 m 2 cm 10 5 dm 1 dm 2 cm 0.02 0.3m3
max. 3 m3 50 cm3 100 dm3 50 dm3 10 dm3 100 cm3 m3 3 m3
Block shape flat long flat flat flat long 0.1 m3 flat
(rhomb.) flat

Rocks granitic slightly altered diabase with chlorite coating on most joint planes granitic
gneiss gneiss

This zone is also described in Appendix 7, Section 1.3 in connection with the description
of ground condition and applied rock support used in Chapter 6.
 A2 - 17

5 SUMMARY

Faults and fractures are often complex features where several factors have influenced upon
development and the final result found today. However, some few general trends are:
- Crushing and brecciation occur mainly in brittle rocks.
- Where faults exhibit zigzag course in schistose rocks: they are thinner along than across
schistocity.
- Where two or more faults intersect, larger volumes of rocks are involved (increased
thickness of zones).
- There is often increased jointing in rock masses adjacent to a major fault.

TABLE A2-3 SUMMARY OF WEAKNESS ZONE CHARACTERISTICS

Swelling Inactive Chlorite, Crushed Porous or Possible other
clay clay talc, rock flaky materials
TYPE OF WEAKNESS ZONE graphite, fragments calcite, or
serpentine, (gravel- or gypsum
or mica sand size)

Zones of weak materials coal

Layers or lenses of soft or weak minerals ............ x
Zones of rocks, sometimes fractured, such as:
- some dolerite dykes * ........................................ x x x altered rock
- some pegmatites (broken)................................. x
Weathered near surface occurrences .................... x x x x weathered
rock
Faults and fault zones
Tension fault zones
- filled zones ...................................................... x x x x
Shear fault zones
- coarse-fragmented crushed zones *** ................ c c c
- small-fragmented crushed zones *** ................... x x x x x
- sand-rich crushed zones..................................... x
- clay-rich crushed zones *, such as:
> simple, clay-rich zones .................................. x x x x
> complex, clay-rich zones ** ............................ x x x x
> unilateral, clay-rich zones .............................. x x x x
- foliation shears .................................................. x x
Altered faults
- altered clay-rich zones ....................................... x x x x
- altered veins/dykes ............................................ x x x
- altered, leached (crushed) zones ........................ x

Recrystallized and cemented zones ................... x x epidote, quartz

*
May also occur as altered and as weathered zones where the adjacent rock may be affected
**
May occur as recrystallized/cemented zones
***
May occur as 'leached' zones
c Occur mainly as coating or thin filling

Faults show very wide variations in dimensions, structure, composition, occurrence of gouge,
as well as in the character of the transition zone to the adjacent rock masses. A division of
faults for rock mechanics, rock engineering and rock construction purposes can consist of:
1. Tension fault zones, which sometimes are developed as feather joints containing filling
of soft minerals. They consist mainly of infilling or secondary materials with a sharp
boundary to the adjacent rocks.
2. Compression faults, mostly developed as crushed zones believed to be formed by shear
movements. They can vary from "dry" fragmented zones to zones mainly consisting of
soft, clayey materials. Many of these faults show a transition zone with a gradually
reduced jointing quantity from the central part to the surrounding bedrocks.
 A2 - 18

3. Altered faults, which are characterized by alteration to clay of feldspar in the zone and
in the adjacent rocks. Alteration may have caused that minerals have been dissolved to
form a permeable materials. A summing up main characteristics in weakness zones is
presented in Table A2-3, and assumed RMi or JP values in Table A2-4.

TABLE A2-4 ASSUMED RANGE OF JP AND/OR RMi VALUES FOR THE MAIN TYPES OF WEAKNESS
ZONES. THE VALUES DO NOT INCLUDE THE EFFECT OF SWELLING.

Jointing Rock mass Assumed values of input parameters

TYPE OF WEAKNESS ZONE parameter index --------------------------------------------
JP RMi σc (MPa) Vb (10-3 m3) jC

Zones of weak materials

• Layers of soft or weak minerals, such as:
- clay materials 1) .................................................... 0.01 - 0.05 0.025 - 0.1 ** -
- mica, talc, or chlorite layers and lenses 2) ............ 0.05 - 5 0.1 - 10 ** -
- coal seams ............................................................ 0.04 - 0.1 0.6 - 3 15 - 25 10 - 100 1-2
• Zones of weak rocks or brecciated rocks, such as:
- some dolerite dykes 3) .......................................... 0.005 - 0.05 * 0.1 - 10 1-2
- some pegmatites, often heavily jointed ................ 0.005 - 0.05 * 0.1 - 10 1-2
- some brecciated zones and layers which
have not been "healed" ........................................ 0.005 - 0.05 * 0.1 - 10 1-2
• Weathered surface or near surface deposits .......... 0.005 - 0.05 0.05 - 3 1 - 10 1 - 100 0.2 - 0.5

Faults and fault zones

• Tension fault zones
- feather joints and filled zones, such as:
> clay-filled zones 1) .............................................. 0.01 - 0.05 0.025 - 0.1 ** -
> calcite-filled zones 2) .......................................... 0.5 - 5 0.1 - 10 ** -
• Shear fault zones
- coarse-fragmented, crushed zones ..................... 0.01 - 0.1 * 1 - 100 0.5 - 1
- small-fragmented, crushed zones ....................... 0.001 - 0.02 * 0.01 - 1 0.4 - 0.8
- sand-rich crushed zones ..................................... 0.0005-0.005 0.0005-0.005 1 0.001 - 0.1 0.5 - 1
- clay-rich, crushed zones, such as:
> simple, clay-rich zones ......................................... 0.001 - 0.015 * 0.1 - 10 0.2 - 0.5
> complex, clay-rich zones ...................................... 0.0005 - 0.01 * 0.01 - 10 0.2 - 0.4
> unilateral, clay-rich zones .................................... 0.002 - 0.02 * 0.1 - 10 0.3 - 0.6
- foliation shears 4)
• Altered faults
- altered, clay-rich zones ........................................ 0.005 - 0.05 0.006 - 3.5 0.1 - 10 1 - 100 0.2 - 0.5
- altered, leached (crushed) zones .......................... 0.002 - 0.02 0.003 - 2 0.1 - 10 0.1 - 10 0.3 - 0.6
- altered veins/dykes ............................................... 0.01 - 0.1 0.0003 - 0.3 0.01 - 0.5 10 - 1000 0.2 - 0.5

Recrystallized and cemented/welded zones It is difficult to indicate 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 as very soft - firm
2)
No strength data have been found. The values given are, therefore, assumed
3)
It is 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

In addition to the orientation and thickness the following features may be applied in description
of faults:
- joint and seam characteristics;
- filling or gouge type and properties;
- block sizes and shapes; and
- the types of rocks or minerals and their possible alteration.
- the composition of the rock masses in the transition zone between the zone and the
adjacent rocks.
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

APPENDIX 3

METHODS TO QUANTIFY THE PARAMETERS APPLIED IN THE RMi

"If no collective system and method exists, so much detail is usually recorded as to obscure the
essential data needed for design."
Douglas A. Williamson and C. Rodney Kuhn (1988)

From Chapter 2 on data collection it has been shown that some of the uncertainties and errors in
rock engineering stem from the rock mass characterizations, such as:
- 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
and methods for data collection.
- During the measurements of joints and jointing, as the joints exposed only may be a portion of
the joints considered to be representative of the entire rock mass.

This appendix has been worked out to remedy some of these tasks giving supplementary
descriptions of the various methods in use to find the numerical values of the RMi parameters,
either directly from measurements or observations, or from other types of measurements of rock
strength properties or jointing features. Some of the methods described are time-consuming and
expensive. Therefore, the methods used during collection of the field and laboratory investigations
should be chosen to meet the requirements to accuracy and quality of the input data which may be
determined by:
− the purpose and use of the construction;
− the stage of planning;
− the method of excavation;
− the availability to observe or measure the actual properties of the rock mass;
− the complexity of the geology; and
− the quality of the ground with respect to the type and use of the actual opening or the method
of excavation.

For exchange of engineering geological data and for information to other people involved,
additional documentation of the rock mass conditions is important. Therefore, a verbal description
should be included in the characterization. It should describe the composition and structure of the
rock mass with special emphasis on the parameters of importance for engineering properties,
including how the rock masses are related to the geological setting in the area. Important during the
development of the RMi system has been to:
− group the rock masses in such a way that those parameters which are of most universal
concern are clearly dealt with, and at the same time
− keeping the number of such parameters to a useful minimum. Thus, the input parameters
selected, which are determined numerically in this appendix, are only:
− the uniaxial compressive strength of intact rock;
− the joint characteristics given as the joint condition factor; and
− the block volume.
 A3 - 2

1 METHODS TO DETERMINE THE UNIAXIAL COMPRESSIVE STRENGTH OF

ROCKS

For some engineering or rock mechanical purposes the numerical characterization of rock material
alone can be used, for example boreability, aggregates for concrete, asphalt etc. Also in assessment
for the use of fullface tunnel boring machines (TBM) rock properties such as compressive strength,
hardness, anisotropy are often among the most important parameters. For stability evaluations and
rock support engineering, however, the rock properties are mainly of importance only where they
are weak or overstressed. Though rock properties in many cases are overruled by the effects of
joints, it should be brought in mind that their properties highly determine how the joints have been
formed, which in turn may explain their characteristics.

Although the geological classification of rocks is mainly based on formation and composition of the
material, it is so well established that other methods for division of this material have not come into
major use. Since, in rock mechanics and rock engineering, the rock behaviour rather than its
composition is of main importance Goodman (1989) has presented the classification shown in Table
A3-1.

TABLE A3-1 BEHAVIOURAL CLASSIFICATION OF ROCKS (from Goodman, 1989)

GROUPS OF ROCKS Examples

I Crystalline texture
A. Soluble carbonates and salts ........................................... Limestone, dolomite, marble, rock salt, gypsum
B. Mica or other planar minerals in continuous bands ....... Mica schist, chlorite schist, graphite schist
C. Banded silicate minerals without continuous mica sheets Gneiss
........................................................................................
D. Randomly oriented and distributed Granite, diorite, gabbro, syenite
silicate minerals of uniform grain size ...........................
E. Randomly oriented and distributed silicate minerals in a Basalt, rhyolite, other volcanic rocks
background of very fine grain and with vugs ................. Serpentinite, mylonite
F. Highly sheared rocks ......................................................

II Clastic texture Silica-cemented sandstones and limonite sandstones

A. Stably cemented ............................................................. Calcite-cemented sandstone and conglomerate
B. With slightly soluble cement ........................................... Gypsum-cemented sandstones and conglomerates
C. With highly soluble cement ........................................... Friable sandstone, tuff
D. Incompletely or weakly cemented................................... Clay-bound sandstones
E. Uncemented ....................................................................

III Very fine-grained rocks Hornfels, some basalts

A. Isotropic, hard rocks .......................................................
B. Anisotropic on a macro scale, but Cemented shales, flagstones
microscopically isotropic hard rocks ............................. Slate, phyllite
C. Microscopically anisotropic hard rocks .......................... Compaction shale, chalk, marl
D. Soft, soil-like rocks .........................................................

IV Organic rocks
A. Soft coal Lignite and bituminous coal
B. Hard coal
C. 'Oil shale'
D. Bituminous shale
E. Tar sand

A verbal description of the rock should, in addition to the rock name and its strength, include
possible anisotropy, weathering, and reduced long-term resistance to environmental influence
(durability). The description of the rock could also contain information on texture, colour, lustre,
small scale folds, etc. which can give information to a better understanding of the rock conditions as
well as the jointing.
 A3 - 3

As rocks are composed of an aggregate of several types of minerals with different properties,
arrangement and "welding", there are many factors which determine their strength properties. In
addition, possible weathering and alteration can highly influence on the final strength properties of a
rock. The effect of this is outlined in Section 1.6.4

Some minerals have a stronger influence on the properties of a rock than other. In rock construction
the mica and similar minerals have an important contribution where they occur as parallel oriented
continuous layers (Selmer-Olsen, 1964). Mica schists and phyllites with a high amount of mica
show, therefore, strongly anisotropic properties which often influence in rock construction works as
shown in Section 1.6.3.

1.1 The uniaxial compressive strength (σc )

In rock mechanics and engineering geology the boundary between rock and soil is defined in terms
of the uniaxial compressive strength and not in terms of structure, texture or weathering. Several
classifications of the compressive strength of rocks have been presented, as seen in Fig. A3-1. In
this work a material with the strength ≤ 0,25 MPa is considered as soil, refer to ISRM (1978) and
Table A3-1.
0.5 0.7 1 2 3 4 5 6 7 8 10 20 30 40 50 70 100 200 300 400 700

Very weak Weak Strong Very strong Coates

1964

Low Medium High Very high Deere and Miller

Very low strength
strength strength strength strength 1966

Weak Moderately Moderately Strong Very Extremely Geological Society

Very weak
weak streng strong strong 1970

Soil Rock

Extremely Low Very high Extremely

Very low Medium High Broch and Franklin
low strength strength high
strength strength strength 1972
strength strength

Very soft Hard Very hard Jennings

Soil Soft rock Extremely hard rock
rock rock rock 1973

Low Medium High Very high Bieniawski

Soil Very low strength
strength strength strength strength 1973

Very low Medium High Very ISRM

Low strength Moderate
high 1979

0.5 0.7 1 2 3 4 5 6 7 8 10 20 30 40 50 70 100 200 300 400 700

Uniaxial compressive strength, MPa

Fig. A3-1 Various strength classifications for intact rock (from Bieniawski, 1984)

The uniaxial compressive strength can be determined directly by uniaxial compressive strength tests
in the laboratory, or indirectly from point-load strength test (see Section 1.4.2). The tests should be
carried out according to the methods recommended by the ISRM (1972).
The classification of the uniaxial compressive strength suggested by ISRM is shown in Tables A3-
1a and A3-7.
 A3 - 4

TABLE A3-1a CLASSIFICATION OF THE UNIAXIAL COMPRESSIVE

STRENGTH OF ROCKS (σc ) (from ISRM (1978)
---------------------------------------------------------------------------------------------
Soil σc < 0.25 MPa
Extremely low strength σc = 0.25 - 1 MPa
Very low strength σc = 1 - 5 MPa
Low strength σc = 5 - 25 MPa
Medium strength σc = 25 - 50 MPa
High strength σc = 50 - 100 MPa
Very high strength σc = 100 - 250 MPa
Extremely high strength σc > 250 MPa
---------------------------------------------------------------------------------------------

The uniaxial compressive strength of the rock constitutes the highest strength limit of the actual
rock mass. ISRM (1981) has defined the uniaxial strength of a rock to samples of 50 mm diameter.
A rock is, however, a fabric of minerals and grains bound or welded together. The rock therefore
includes microscopic cracks and fissures. Rather large samples are required to include all the
components that influence strength. When the size of the sample is so small that relatively few
cracks are present, the failure is forced to involve a larger part of new crack growth than in a larger
sample. Thus the strength is size dependent. This scale effect of rocks has been a subject of
investigations over the last 30 to 40 years. Fig. A3-2 shows the results from various tests compiled
by Hoek and Brown (1980) where the scale effect (for specimens between 10 and 200 mm) has been
found as
σc = σc50 (50/d)0.18 eq. (A3-1)

where σc50 is for specimens of 50 mm diameter, and

d is the diameter (in mm) of the actual sample.
σc 50

1.3
σc
Uniaxial compressive strength of 50 mm diameter specimen

1.2
Uniaxial compressive strength of specimen

1.1
Symbol Rock Tested by
0.18 Marble Mog 166
(σc/σc50) = (50/d) Limestone Koifman 169
Granite Burchartz et al.169
1.0 Basalt Koifman 169
Basalt-andesite lava Melekidze169
Gabbro Ilnickaya 169
Marble Ilnickaya 169
Norite Bieniawski 167
Granite Hoskins & Horino 170
0.9 Quartz diorite Pratt et al.168

0.8

0.7
0 50 100 150 200 250
Specimen diameter d - mm

Fig. A3-2 Influence of specimen size upon the uniaxial compressive strength of intact rock.
(from Hoek and Brown, 1980)
 A3 - 5

Wagner (1987) has later shown that 'samples' with diameter more than 2 metres follow the
following similar equation:
σcf = σc50 (50/d) 0.2 eq. (A3-1a)

ISRM (1980) recommends that the uniaxial compressive strength of the rock material in an area is
given as the mean strength of rock samples taken away from faults, joints and other discontinuities
where the rock may be more weathered. When the rock material is markedly anisotropic in its
strength, the value used in the RMi should correspond to the direction along which the smallest
mean strength was found. However, in such cases it is usually of importance to record the uniaxial
compressive strength also in other directions.

Many compressive strength tests are made on dry specimens. ISRM (1980) recommends that the
samples should be tested at a water content pertinent to the problem to be solved. Because rocks are
often much weaker for wet than for dry materials, it is important to inform about the moisture
conditions. This is further discussed in Section 1.3.

The compressive strength σc50 is used directly in calculation of RMi. The accuracy and quality of
this measurement has been discussed by several writers, as different modes of failure occur during
the tests (Lama and Vutukuri, 1978; Hoek and Brown, 1980; Farmer
and Kemeny, 1992; among others). Undoubtably, compressive strength is the mostly used test on
rock samples and the loading resembles often situations in the field.

The uniaxial compressive test is time-consuming and is also restricted to those relatively hard,
unbroken rocks that can be machined into regular specimens. Although the strength classification is
based on laboratory tests, it can be approximated by simple methods. An experienced person can
make a rough five-fold classification of rock strength with a hammer or pick. Deere and Miller
(1966) have shown that rock strength can be estimated with a Scmidt hammer and a specific gravity
test with enough reliability to make an adequate strength characterization. According to Patching
and Coates (1968) the rock strength can be quickly and cheaply estimated in the field, and more
precision can be attained, if required, by laboratory tests. Also from a fully description of a rock
including composition and possible anisotropy and weathering it may in many instances be possible
to assess the strength. These matters are described in the following. The methods, their applications,
test procedures etc. are not dealt with, merely how the results can be used.

1.2 Effect of saturation upon rock strength

It is known that the influence of water on the strength of rocks may be considerable. From the
published papers no general expressions other than the fact that moisture may reduce the strength of
rocks drastically, has been established. Salustowicz (1965) has found that the decrease by moisture
in sandstones and shales can be 40% and 60% of the dry strength respectively. Colback and Wiid
(1965) mention a decrease from dry to saturated in the order of 50% both for quartzitic shale and for
quartzitic sandstone. Table A3-2 shows compilation of some published result between compressive
strength for saturated versus dry specimens. Results from Ruiz (1966) in Table A3-2 shows that wet
strength may be higher than dry probably caused by heterogeneity and possible error from few tests.
This effect has also been observed by Broch (1979) for some very fine-grained rocks (basalt, black
shale).
 A3 - 6

TABLE A3-2 EFFECT OF SATURATION UPON STRENGTH (worked out from Lama and Vutukuri, 1978,
with data from Feda, 1966; Ruiz, 1966; Boretti-Onyszkiewicz, 1966)
--------------------------------------------------------------------------------------------------------
Number of Saturated in %
rocks tested Rock type of dry strength
--------------------------------------------------------------------------------------------------------
5 basalt 45 - 116
2 diabase 92 - 125
1 dolomite 83
3 limestone 56 - 118
1 schist 48
3 gneiss 36 - 59
5 gneiss 88 - 112
2 granitic gneiss 68 - 85
4 granite 68 - 116
1 granulite 54
1 quartzite 80
1 sandstone 90
5 sandstone 67 – 87 tested normal to bedding
" 54 – 92 tested parallel to bedding
--------------------------------------------------------------------------------------------------------

TABLE A3-3 EFFECT OF MOISTURE CONTENT ON COMPRESSIVE STRENGTH (worked out

from Lama and Vutukuri, 1978, based on data from Price, 1960, and Obert et al., 1946)
--------------------------------------------------------------------------------------------------------------------------
No. of Relative compressive strengths for various
rocks ROCK TYPE Moisture contents expressed as % of oven-dry
tested Oven-dry Air-dry Saturated
--------------------------------------------------------------------------------------------------------------------------
5 sandstone 100 51 - 99 45 – 90
1 marble 100 99 95
1 limestone 100 96 83
1 granite 100 93 86
1 slate 100 94 80
--------------------------------------------------------------------------------------------------------------------------

20 90
Kota sandstone

80 Jamrani sandstone

gab
Singrauli sandstone
bro
Uniaxial compressive strength, MPa

B 70 x Jhingurta sandstone
Point load strength, I S (32) (MN/m )

15
2

60

50
gabb
ro A

10 quartzd 40
iorite
gneiss
amphibolite
30
gneiss
amphiboli
te
20
5
10
marble xx
x x x x x
0
0 25 50 75 100 0.0 0.4 0.8 1.2 1.6
Water content (%) Equilibrium moisture content, ml/g, %

Fig. A3-3 Left: Point-load strength as a function of water content in rock cores (from Broch, 1979).
 A3 - 7

Right: Variation in uniaxial compressive strength with equilibrium moisture content for sandstones (from
Seshagiri Rao et al. 1987).
The results shown in Table A3-3 differ whether the samples have been oven-dried or air-dried. This
effect may strengthen the variation published in many papers on the effect of water upon strength.
Broch (1979) has shown in Fig. A3-3 that the main strength reduction generally takes place for
water contents less than 25%. Also Seshagiri Rao et al. (1987) have found that the main strength
reduction for sandstones takes place at low moisture content (Fig. A3-3)

Broch (1979) concludes that the strength reduction from water increases with increasing amount of
dark minerals (biotite, amphiboles, pyroxenes) and for increasing schistosity (anisotropy), as seen in
Fig. A3-4.

From this complex picture of the effect of moisture and saturation upon rock strength Lama and
Vutukuri (1978) conclude that: "Moisture in the rocks has a very significant effect on compressive
strength in many instances. Unless the values to be used for design purposes are corrected to in situ
conditions, catastrophic failures can occur. Many times, saturated specimens are recommended for tests so
that the values obtained are conservative."

POINT LOAD STRENGTH OF WATER SATURATED ROCKS IN PER

TYPES AND GROUPS OF CENT OF THE STRENGTH OF OVEN DRY ROCKS
ROCKS
50 60 70 80 90 100 110

FINEGRAINED NONPOROUS ROCKS

ACIDIC INTERMEDIATE PLUTONIC ROCKS

GNEISSES

BASIC PLUTONIC ROCKS

MICASCHISTS

Fig. A3-4 The effect of water on point load strength of various groups of rocks (from Broch, 1979)

It is, therefore, important that the conditions at which the rocks are tested, are reported. ISRM
(1972, 1981) suggests that rock samples are stored in 50% humidity for 5 - 6 days before testing.

1.3 Compressive strength determined from the point-load strength

The principle of the point load strength test is that a piece of rock is loaded between two hardened
steel points. Details on the measuring procedure are described by ISRM (1985), and the method is
further dealt with in several textbooks and papers (Lama and Vutukuri, 1978; Hoek and Brown,
1980, among others).

Both Franklin (1970) and Bieniawski (1984) recommend the use of point load strength index (Is) for
rock strength testing. The reason is that Is can be determined in the field on specimens without
preparation, using simple portable equipment. Also Broch (1983) points out the great advantage
using the point load strength test as it does not require machined specimen. As long as the influence
of specimen size and shape are considered in the calculation of the strength index, any piece of rock,
whether the surface is smooth or rough, can in principle be tested. Although tests on irregular
specimens appear to be crude, Wittke & Louis (1969) have shown that the results need be no less
reproducible than those obtained in uniaxial compression.
 A3 - 8

1.3.1 The point load strength index (Is)

Greminger (1982) relates the test results, Is(50), to standard 50 mm thick samples; this has also been
applied by ISRM (1985) in a revised edition of the 'Suggested Method for determining point load
strength'. The point load strength test is a form of "indirect tensile" test, but is largely irrelevant to
its primary role in rock classification and strength as a tensile characterization (ISRM, 1985). Is(50)
is approximately 0.80 times the uniaxial tensile or Brazilian tensile strength.

Also from the point load strength test a strength anisotropy index (Ia(50)) can be measured (Broch,
1983) from the maximum and minimum strengths obtained normal to, and parallel to the weakness
planes, respectively, of the rock, i.e. bedding, foliation, cleavage, etc.

TABLE A3-4 CLASSIFICATIONS OF THE POINT LOAD STRENGTH ( Is)

----------------------------------------------------------------------------------------------------
TERM Bieniawski (1984) Deere (1966)
----------------------------------------------------------------------------------------------------
very high strength Is > 8 MPa Is > 10 MPa
high Is = 4 - 8 MPa Is = 5 - 10 MPa
medium Is = 2 - 4 MPa Is = 2.5 - 5 MPa
low Is = 1 - 2 MPa Is = 1.25 - 2.5 MPa
very low Is < 1 MPa Is < 1.25 MPa
----------------------------------------------------------------------------------------------------

As shown in Fig. A3-3 the point load strength varies with the water content of the specimens. ISRM
(1985) mentions that the variations are particularly pronounced for water saturations below 25%. At
water saturation above 50% the strength is less influenced by small changes in water content, so that
tests in this water content range are recommended unless tests on dry rock are specifically required.

1.3.2 The correlation between Is and σc

Point load strength test may often replace the uniaxial compressive strength test as it is, when
properly conducted, as reliable and much quicker to measure. Hoek and Brown (1980) are of the
opinion that a reasonable estimate of the uniaxial compressive strength of the rock can be obtained
by means of the point load test. Other authors (Greminger, 1982; Seshagiri Rao et al., 1987) have,
however, found that correlations with uniaxial compressive strength using
σc = k × Is eq. (A3-2)
or σc = k50 × Is50 (related to 50 mm thick samples) eq. (A3-2a)
are only approximate. They show that, though the factor k or k50 generally varies between 15 and
25 it may sometimes vary between 10 and 50 especially for anisotropic rocks, so that errors of up to
100% are possible in using an arbitrary ratio value to predict compressive strength from point load
strength. The variation in k and k50 published in various papers on point load strength results is
shown in Table A3-5.

Greminger (1982) found that estimation of uniaxial compressive strength of anisotropic rocks may
lead to significant errors, which is documented in Fig. A3-5. Seshagiri Rao et al. (1987) found for
sandstones that k50 = σc / Is50 varied with the strength of the rock according to the following:
k50 = 14 for σc = 80 MPa, k50 = 12 for σc = 43 MPa
k50 = 14 for σc = 24 MPa, k50 = 11 for σc = 6 MPa
 A3 - 9

TABLE A3-5 THE VALUE OF k PRESENTED IN VARIOUS PAPERS

REFERENCE value of k

Franklin (1970) k = approx. 16

Broch and Franklin (1972) k = 24
Indian standard (1978) k = 22
Hoek and Brown (1980) k= 14 + 0.175 D *)
Greminger (1982) no general correlation, see Fig. A3-5
ISRM (1985) k50 = 20 - 25
Brook (1985) k50 = 22
Seshagiri Rao et al. (1987) no general correlation
Hansen (1988) no general correlation
Ghosh and Srivastava (1991) k50 = 16
*)
D is the diameter or thickness of the tested sample

They argued that no general relation can be established between σc and Is50 . Also Hansen (1988) is
of the same opinion from numerous tests carried out at the Technical University of Norway. The
trend, therefore, seems to be that the factor k or k50 is higher for strong rocks. Based on the results
above it is suggested, where no other information is available, to use the values of k50 (related to
standard thickness) presented in Table A3-6 in the correlation from point load strength to uniaxial
compressive strength.

Loaded Loaded
parallel perpendicular
Augen gneiss
Ruhr sandstone
10 Chiandone gneiss
Nuttlar slate -

5
σc(50) = 24 Is(50)

0 50 100 150 200

Fig. A3-5 Relationship between uniaxial compressive strength and point load strength index (from Greminger, 1982)

TABLE A3-6 SUGGESTED VALUE OF THE FACTOR k50 VARYING

WITH THE STRENGTH OF THE ROCK
σc (MPa) Is 50 (MPa) k50
*)
25 - 50 1.8 - 3.5 14
50 - 100 3.5 - 6 16
100 - 200 6 - 10 20
> 200 > 10 25
*)
Bieniawski (1973) suggests that point load strength test are not carried out on rocks
having compressive strength less than approximately 25 MPa.

1.4 Compressive strength estimated from Schmidt hammer rebound number

The Schmidt hammer is used for non-destructive testing of the quality of rocks and concretes. It
measures the 'rebound hardness' of the tested material. The mechanism of operation is simple; a
 A3 - 10

plugger, released by a spring, impacts against the tested rock surface. The rebound distance of the
plunger is read directly from a numerical scale.

Measurements of rock properties with the Schmidt hammer are based on an imperfectly elastic
impact of two bodies, one of which is represented by the impact of the test hammer and the other by
the surface of the tested rock (Ayday and Göktan, 1992). Schmidt hammer models are designed in
different levels of impact energy, and the two types L and N are commonly adapted for rock
property determinations. The type L has an impact energy of 0.735 Nm, 1/3 of that of Type N.

ISRM (1978) suggests that the L hammer can be used for the testing of rocks which have uniaxial
compressive strength in the range of approximately 20 - 150 MPa . ISRM (1978) has also given a
complete test procedure including a chart for correlating Scmidt rebound hardness to uniaxial
compressive strength.

1.5 Compressive strength assessed from simple field test

Sometimes, particularly at an early stage in the description of the rock mass, strength may be
assessed without testing (ISRM, 1980). Such a first estimate of the uniaxial compressive strength (
σc) may be made by visual and sensory description of hardness of rock or consistency of a soil
(Piteau, 1970; Herget, 1982). The strength can be judged from simple hardness tests in the field
with geological pick by observing the resistance to breaking under impact, as shown in Table A3-7.

TABLE A3-7 SIMPLE FIELD IDENTIFICATION COMPRESSIVE STRENGTH OF ROCK AND CLAY
(from ISRM, 1978)
Approx.
GRADE TERM FIELD IDENTIFICATION range of σc
(MPa)
S1 Very soft clay Easily penetrated several inches by fist. < 0.025
S2 Soft clay Easily penetrated several inches by thumb. 0.025-0.05
S3 Firm clay Can be penetrated several inches by thumb with moderate effort. 0.05 -0.10
S4 Stiff clay Readily intended by thumb, but penetrated only with great effort. 0.10 -0.25
S5 Very stiff clay Readily intended by thumbnail. 0.25 -0.50
S6 Hard clay Intended with difficulty by thumbnail. > 0.50
--------------------------------------------------------------
R0 Extremely weak rock Intended by thumbnail. 0.25 - 1
R1 Very weak rock Crumbles under firm blows with point of geological hammer; 1-5
can be peeled by a pocket knife.
R2 Weak rock Can be peeled by a pocket knife with difficulty, shallow identi- 5 - 25
fications made by firm blow with point of geological hammer.
R3 Medium strong rock Cannot be scraped or peeled with a pocket knife; specimen can 25 - 50
be fractured with single firm blow of geological hammer.
R4 Strong rock Specimen requires more than one blow of geological hammer 50 - 100
to fracture it.
R5 Very strong rock Specimen requires many blows of geological hammer to fracture it. 100 - 250
R6 Extremely strong rock Specimen can only be chipped with geological hammer. > 250
The clays in grade S1 - S6 can be silty clays and combinations of silts and clays with sands, generally slow draining.

The hammer test should be made with a geologist's hammer on pieces about 10 cm thick placed on a
hard surface, and tests with the hand should be made on pieces about 4 cm thick. The pieces must
not have incipient fractures, and therefore several should be tested.
For anisotropic rocks tests should be carried out in different directions to the structure. The lowest
representative values should be applied. It is also possible to use the foliation anisotropy factor
 A3 - 11

described in Section 1.7.3 to find the lowest value of compressive strength to be applied in the
RMi.

1.6 Compressive strength estimated from rock description

Probably the most generally used single describer of rock structure has been "rock type". In this
term a wide variety of geological factors is embraced, ranging from basic rock origin: igneous,
sedimentary and metamorphic, to special properties such as texture and structure, mineral size and
composition, anisotropy, degree of alterations, etc.

According to Franklin (1970) over 2000 names are 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 and greater challenges in rock constructions. From the rock engineering point of view this
vast amount of rock types can be grouped into a reduced number of about 40, regarding strength and
behaviour.

Geological nomenclature emphasizes solid constituents whereas from the engineers' point of view
pores, cracks and fissures are of greater mechanical significance. Petrological data can, however,
make an important contribution towards the prediction of mechanical performance, provided that
additional information on the anisotropy and weathering is provided.

In most instances, a simple hand-specimen name will likely be adequate to determine the rock type
and name. An added term to indicate grain size may often be useful (Patching and Coates, 1968).
For many rocks the name of the rock, its homogeneity and continuity can be established by visual
observation in the field.

An accurate classification of rock materials used by geologists requires, however, often a detailed
consideration of mineralogy and petrography for instance from thin section analysis, an
investigation, which may seldom be important to engineers only in few cases.

1.6.1 Main geological characteristics

The minerals of most igneous rocks are hard and may display cleavage but are of a dense
interfingering nature resulting in homogeneous materials with only slight, if any, directional
differences in mechanical properties of the rock.

The minerals of sedimentary rocks are usually softer and are of generally weaker assemblage 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 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 sandstone rocks are usually the
most strongly anisotropic.

The metamorphic rocks, more particularly the micaceous and chloritic schists, are probably the most
outstanding with respect to anisotropy. The metamorphism have resulted in harder minerals in most
cases; however, the preferred orientation of platy minerals due to shearing movements results in
considerable directional differences in mechanical properties. Rocks with gneissic texture are
generally not strongly anisotropic. Slate, due to well developed slaty cleavage, is highly anisotropic.
 A3 - 12

1.6.2 Strength assessment from rock name

The rock types often give relative indications of their inherent properties (Piteau, 1970; Patching
and Coates, 1968). For many rocks, however, the correlation between the petrographic names of
rocks and their mechanical properties may be poor, caused by difference in composition, grain size,
porosity, cementation, anisotropy within each type. Nevertheless, a great deal of associated
information about a rock can be inferred from its geological name, such as whether it may be
homogeneous, layered, schistose or irregular.

TABLE A3-8 NORMAL RANGE OF COMPRESSIVE STRENGTH FOR SOME COMMON ROCK TYPES (data
from Hansen, 1988 and Hoek and Brown, 1980) AND VALUES FOR THE m FACTOR IN HOEK-
BROWN FAILURE CRITERION (from Hoek et al., 1992).
Uniaxial compressive Rating of Uniaxial compressive Rating of
Rock name strength σc the factor Rock name strength σc the factor
low average high mi 1) low average high mi 1)
Sedimentary rocks Metamorphic rocks
Anhydrite 120'? 13.2 Amphibolite 75 125 250 31.2
Coal 16" 21" 26" Amphibolitic gneiss 95 160 230 31 ?
Claystone 2' 5' 10' 3.4 Augen gneiss 95 160 230 30 ?
Conglomerate 70 85 100 (20) Black shale 35 70 105
Coral chalk 3 10 18 7.2 Garnet mica schist 75 105 130
Dolomite 60' 100'300' 10.1 Granite gneiss 80 120 155 30 ?

Limestone 50* 100' 180* 8.4 Granulite 80' 150 280

Mudstone 45 95 145 Gneiss 80 130 185 29.2
Shale 36" 95" 172" Gneiss granite 65 105 140 30 ?
Sandstone 75 120 160 18.8 Greenschist 65 75 85
Siltstone 10' 80' 180' 9.6 Greenstone 120' 170* 280* 20 ?
Tuff 3' 25' 150' Greywacke 100 120 145
Igneous rocks
Andesite 75' 140' 300' 18.9 Marble 60' 130' 230' 9.3
Anorthosite 40 125 210 Mica gneiss 55 80 100 30 ?
Basalt 100 165 355" (17) Mica quartzite 45 85 125 25 ?
Diabase (dolerite) 227" 280" 319" 15.2 Mica schist 20 80* 170* 15 ?
Diorite 100 140 190 27 ? Mylonite 65 90 120
Gabbro 190 240 285 25.8 Phyllite 21 50 80 13 ?

Granite 95 160 230 32.7 Quartz sandstone 70 120 175

Granodiorite 75 105 135 20 ? Quartzite 75 145 245 23.7
Monzonite 85 145 230 30 ? Quartzitic phyllite 45 100 155
Nepheline syenite 125 165 200 Serpentinite 65 135 200
Norite 290" 298" 326" 21.7 Slate 120' 190' 300' 11.4

Pegmatite 39 50 62 Talc schist 45 65 90 10 ?

Rhyolite 85'? (20)
Syenite 75 150 230 30 ?
Ultra basic rock 80' 160 360
Soil materials2):
Very soft clay σc = 0.025 MPa Soft clay σc = 0.025 - 0.05 MPa Firm clay σc = 0.05 - 0.1 MPa
Stiff clay σc = 0.1 - 0.25 MPa Very stiff clay σc = 0.25 - 0.5 MPa Hard clay σc = > 0.5 MPa
Silt, sand: assume σc = 0.0001- 0.001 MPa
* Values found by the Technical University of Norway, (NTH) Inst. for rock mechanics.
' Values given in Lama and Vutukuri, 1978.
" Values given by Bieniawski, 1984.
 A3 - 13

1)
Refer to the factor, m, in the failure criterion for rock masses by Hoek et al. (1992), mi is the parameter for intact rock, see
Chapter 9, Section 1. Values in parenthesis have been estimated by Hoek et al (1992); some others with question mark have been
assumed in this work.
2)
For clays the values of the uniaxial compressive strength is based on ISRM (1978) as presented in Table A3-7.

A good description of the rock material is a prerequisite when such strength evaluations are made.
Adjustments for possible anisotropy (schistosity, foliation, bedding) and weathering/alteration in
rocks are further described in Section 1.7.3 and 1.7.4.

Where representative, fresh specimens of the various types occur, it is, however, often possible if
not accurate values are required to make estimates of the compressive strength from the rock name
as presented in Table A3-8. Additional results from compressive strength tests are given in many
textbooks. Refer to Lama and Vutukuri (1978), Hoek and Brown (1980) etc. In the following some
methods to estimate the uniaxial compressive strength for anisotropic and weathered or altered
rocks are outlined.

1.6.3 Reduction in strength from anisotropy

Anisotropy in rock material is mainly caused by schistosity, foliation or bedding. The difference in
properties is determined by the arrangement and amount of flaky and elongated minerals (mica,
chlorite, amphiboles). This intrinsic rock property tends to be significant even at the scale of a
laboratory test specimen.
Quartzitic phyllite
120
Carbonaceous phyllite
Micaceous phyllite

Experimental
100 Theoretical
Uniaxial Compressive Strength (MPa)

80

60

40

20

0 30 60 90
β (degrees)

σ1
β

Case - I
Single
plane of Case - II
σ2 = σ3
Compressive strength, σ1

weakness

Splitting

Case-I Shearing
β

Multiple
planes of
weakness

0 30 60 90
Case-II
(a) (b) β (degrees)
 A3 - 14

Fig. A3-6 Experimental and theoretical variation of uniaxial compressive strength with angle between schistosity
plane and direction of testing (from Ramamurthy et al., 1993)

Tsidzi (1986, 1987, 1990) has presented results from tests on strength anisotropy in foliated rocks,
together with measurements of their intrinsic anisotropic foliation fabric. The lowest strength value
occurred when the orientation of the anisotropic fabric element (bedding, foliation) to the specimen
loading axis was between 30o and 45o, and the highest value for orientation either 0o or 90o. This
effect has also been shown by Ramamurthy et al., (1993) in Fig. A3-6, Hoek and Brown (1980) and
many other authors. Hence, the estimation of compressive strength anisotropy of rocks in terms of
values obtained only for tests parallel (0o) and normal (90o) to the foliation plan give only limited
information on the rock strength.

From his tests carried out on slates, schists, and gneisses with compressive strength ranging between
20 and 285 MPa Tsidzi (1990) has worked out the classification which is shown in Table A3-9.
From regressions Tsidzi arrived at the following expression for the uniaxial compressive strength
anisotropy factor:
fA = 0.95 + 0.17 Fi eq. (A3-3)

where Fi is the foliation index. Its ratings are indicated in Table A3-9.

TABLE A3-9 CLASSIFICATION OF FOLIATION AND ANISOTROPY OF ROCKS

(from Tsidzi, 1986, 1987, 1990)
FOLIATION FOLIATION
CLASSIFICATION ANISOTROPY
- - - - - - - - - - - - - - -- DESCRIPTION FACTOR
ANISOTROPY fA
CLASSIFICATION

Very weakly foliated Platy and prismatic minerals < 10%, which may occur as discon-
(or non-foliated) tinuous streaks or may be randomly oriented. Rock fractures are
Fi < 1.5 curved or folded. Usually found in high-grade regional metamorphic 1 - 1.2
----------- regions or in contact metamorphic zones.
Isotropic Typical rocks: Quartzite, hornfels, granulite.

Weakly foliated Platy and prismatic minerals 10 - 20%. Compositional layering is

Fi = 1.5 - 3 evident, but mechanically insignificant. Usually found in high-grade 1.2 - 1.5
----------- regional metamorphic regions.
Fairly anisotropic Typical rocks: Quartzofeltspatic gneiss, mylonite, migmatite

Moderately foliated Platy and prismatic minerals 20 - 40%. Thin to thick folia,
Fi = 3 - 6 occasionally discontinuous. Foliation is usually mechanically
----------- passive. Found in rocks formed by medium to high-grade regional 1.5 - 2
Moderately anisotropic metamorphism.
Typical rocks: Schistose gneiss, quartzose schist.

Strongly foliated Platy and prismatic minerals 40 - 60%. Thin wavy continuous folia
Fi = 6 - 9 which may be mechanically significant. Usually formed under
------------ medium-grade regional metamorphic conditions. 2 - 2.5
Highly anisotropic Typical rocks: Mica schist, hornblende schist.

Very strongly foliated Platy and prismatic minerals > 60% occurring as very thin, > 2.5
Fi > 9 continuous folia. Foliation is perfect and mechanically significant.
------------ Found in rocks formed by dynamic or low-grade regional metamorp-
Very highly anisotropic hism.
 A3 - 15

Typical rocks: Slate, small folded phyllite.

Fi = foliation index

The foliation index can be found from thin section analysis by measuring the mineral composition
and the shape of the minerals as presented by Tsidzi (1986).
From the values of fA in Table A3-9 combined with the content of platy and prismatic minerals,
the following equation is found:
fA = 1 + 2.5 c/100 eq. (A3-4)

where c is the content of platy and prismatic minerals in % .

According to Tsidzi (1989), the strength anisotropy index fA is directly proportional to foliation
regardless of the physical condition of rock. Thus, the minimum compressive strength of the
foliated rock can roughly be assessed as
σc min = σc max /fA = σc max /(1 + 2.5 c/100) eq. (A3-5)

In their works on anisotropic rocks Sing et al. (1989) have introduced the anisotropy ratio is defined
as Rc = σc 90 /σc min , where σc 90 is the uniaxial compressive strength measured at right angle to the
schistosity or bedding. Their results shown in Table A3-10 indicate that the strength reduction
caused by the anisotropy (Rc) is considerably higher than fA in Table A3-9.

TABLE A3-10 CLASSIFICATION OF ANISOTROPY

(from Sing et al., 1989 and Ramamurthy et al., 1993)
Anisotropy ratio Classification Rock types
Rc
1 - 1.1 Isotropic
1.11 - 2.0 Low anisotropy Shales
2.01 - 4.0 Medium anisotropy _ _
4.01 - 6.0 High anisotropy  Slates
> 6.0 Very high anisotropy _ Phyllites

A possible reason for this is that the rocks tested by Tsidzi generally exhibit stronger small scale
foldings which will reduce the effect of the foliation. Also possible differences in moisture content
may influence on the results. None of the authors have specified under which conditions the test
were carried out.

The sonic velocity anisotropy coefficient for some rocks is presented in Table A3-11 , which shows
values lower than the strength anisotropy factor of Tsidzi in Table A3-9.

TABLE A3-11 SONIC ANISOTROPY COEFFICIENTS FOR SOME ROCKS

(from Lama and Vutukuri, 1978)
anisotropy | anisotropy
ROCK coefficient | ROCK coefficient
Vº/VÁ. | Vº/VÁ.
Austin chalk 1.17 | Anhydrites 1.12 - 1.16
Limestones 1.04 - 1.30 | Marl 1.10
Salt no anisotropy | Sandstones 1.0 - 1.19
Shales 1.07 - 1.40 | Gneisses 1.20 - 1.27
Mica schist 1.36 | Granodiorite 1.33
Serpentine 1.18 |
 A3 - 16

In an earlier work performed by Bergh-Christensen (1968) correlations between compressive

strength anisotropy and wave velocity anisotropy were investigated. As shown in Fig. A3-7 there is
no clear correlation between sound velocity anisotropy
(Vmax /Vmin) and strength anisotropy (σc max /σc min). Most of the data fall, however, within the to
lines indicated in Fig. A3-7, which can be expressed as:
Vmax /Vmin < σc max /σc min < 4 ⋅Vmax /Vmin - 3 eq. (A3-6)
9.0
35
8.0
7.0 σmax Vmax
6.0 σmin = 4 Vmin - 3

5.0
σmax /σmin

20
4.0

3.0
37
σmax Vmax
=
σmin Vmin
2.0

1.0
1.0 2.0 3.0 4.0
Vmax / Vmin
Fig. A3-7 Correlation between strength and sound anisotropy. The symbols indicate measured rock blasting indexes.
(from Berg-Christensen, 1968)

Vmax c Vmax c
=1+ =1+5
60% Vmin 100 Vmin 100
+ Non-directional structure, no
B
signs of parallel orientation of
content of mica + chlorite in vol% (c)

minerals
50% Rock with more or less ma-
ked structural planaraty,
where the platy minerals oc-
40% cur as single grains
Platy minerals occur partly
H H
as continuous, planar and
30% parallel layers
Platy minerals occur partly
as continuous, small-folded
20% + layers
Platy minerals as con-
tinuous, planar and parallel
layers
10%
Platy minerals occur as con-
tinuous, small-folded layers
+
0% ++
1.0 2.0 3.0 4.0

Vmax / Vmin

Fig. A3-8 Correlation between content of flaky minerals (mica and chlorite) and sound velocity ratio for various
textures of rock (from Berg-Christensen, 1968).
(In the groups with continuous layers of flaky minerals there are rocks that do not follow the average trend.
One explanation can be that the classification into the structural groups are made from simple observations
(personal communication with Bergh-Christensen, 1992))

In Fig. A3-8 showing the relation between the sonic velocity anisotropy and the content of flaky
minerals it is seen that most data lies within the two lines represented by the following expression:
 A3 - 17

1 + c/100 < Vmax /Vmin < 1 + 5c/100 eq. (A3-7)

where V is the seismic velocity, and c is the content of flaky minerals (mica and chlorite) given
in %.

By roughly combining the lowest with the highest values in eq. (A3-6) and eq. (A3-7) it is possible
to find the rock anisotropy factor, fA, from the content of flaky minerals given as:
1 + 4c/100 < fA < 1 + 5c/100 eq. (A3-8)

This equation gives somewhat higher values for fA than eq. (A3-5) found by Tsidzi; for example,
for 50% content of flaky minerals fA = 3 - 3.5 with eq. (A3-8), compared to Tsidzi's fA = 2.25.
Eq. (A3-8) gives, however, lower values than Rc (see Table A3-10).

From the foregoing it is apparent that it is not a well defined, simple method to estimate the effect of
anisotropy upon strength. The anisotropy factor varies with the method used, in addition the
definition of the anisotropy class seems to be very approximate as possible small-scale folding is not
included. The method presented by Tsidzi in eq. (A3-5) seems, however to be better documented
and may, therefore, be the best for rough estimates. More investigations and are required to arrive at
better expressions for the rock anisotropy.

1.6.4 Reduction in strength from weathering and alteration

Weathering of rocks is a result of the destructive processes from atmospheric agents at or near the
Earth's surface, while alteration is typically brought about by the action of hydrothermal processes.
Both processes produce changes of the mineralogical composition of a rock, affecting colour,
texture, composition, firmness or form; features that result in reduction of the mechanical properties
of a rock. Deterioration from weathering and alteration generally affects the walls of the
discontinuities more than the interior of the rock (Piteau, 1970).

TABLE A3-12 ENGINEERING CLASSIFICATION OF THE WEATHERING OF ROCKS

(from Lama and Vutukuri, 1978)
CLASSIFICATION DESCRIPTION

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

Slightly Penetrative weathering developed on open discontinuity surfaces but only slight
weathered weathering of rock material. Discontinuities are discoloured and discoloration can extend into rock up
to a few mm from discontinuity surface.

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

Highly Weathering extends throughout rock mass and the rock material is partly friable. Rock
weathered has no lustre. All material except quartz is discoloured. Rock can be excavated with geologist's pick.

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

Residual soil Soil material with complete disintegration of texture, structure and mineralogy of the parent rock.

Characterization of the state of weathering or alteration both for the rock material and for the
discontinuities is therefore an essential part of the rock parameters to be applied (ISRM, 1978). In
rock engineering and construction it is seldom of interest to describe whether the process of
 A3 - 18

weathering or alteration has been acting; the main topic is to characterize the result. Knowledge of
the processes, which have taken place, can however be important for the understanding and
interpretation of the geological conditions and of the condition of the rocks likely to be found.

In general, the degree of weathering is usually estimated from visual observations, where only the
qualitative information is required. Table A3-12 shows classification of weathering/alteration
similar to that presented by ISRM (1978). A more precise characterization of alteration and
weathering can be found from analysis of thin sections in a microscope.

Papadopoulos and Marinos (1992) have made tests on a wide variety of petrological rock types,
consisting of clayey schists, phyllites, mica schists, sandstones with secondary anisotropy due to
weathering from tectonic action. Some of their results are shown in Fig. A3-9 where the grade of
weathering is according to the ISRM (1978) or the similar classification by Lama and Vutukuri
(1978) in Table A3-12.

Is (50) Is (50)

(a) (b)
MPa MPa

6 6

5 5

4 4

3 3

2 2

1 1

W.G. W.G.

I II III IV I II III IV

Fig. A3-9 Correlation between weathering grade I - IV and point-load strength values Is50 (from Papadopoulos and
Marinos, 1992)

The reduction factor from weathering is found as the ratio Is50fresh /Is50 weathered in Fig. A3-9. Results
from tests parallel and normal to anisotropy are shown in Table A3-13, in which also the rating of a
rock weathering factor (fW) has been suggested. From this the point load strength of the weathered
or altered rock can roughly be found as
Is50 = Is50 fresh /fW eq. (A3-9)

Assuming that the strength reduction for the compressive strength is similar it is approximately
found from
σc = σc fresh /fW = k50 × Is50 fresh /fW eq. (A3-10)

The suggested rating of the weathering/alteration factor in Table A3-13 is, however, based on very
few data and it is therefore considered very rough, especially for high grades of weathering. It
should, therefore, mainly be applied for rough calculations before better strength data are available.
 A3 - 19

TABLE A3-13 SUGGESTED WEATHERING/ALTERATION FACTOR (fW) OF ROCKS, fW

(worked out partly from ISRM (1978) and Papadopoulos and Marinos (1992).
REDUCTION FACTOR
GRADE (from Papadopoulos SUGGESTED
and DESCRIPTION and Marinos, 1992) RATING OF
TERM (from ISRM, 1978) TEST DIRECTION fW
Normal Parallel
I Fresh No visible sign of rock material weathering. 1 1 1

II Slightly Discolouration indicates weathering of rock 1.7 1.8 1.75

material and discontinuity surfaces. All the
rock material may be discoloured by
weathering and may be somewhat weaker
externally than in its fresh conditions.
III Moderately
weathered Less than half the rock material is decom- 2.6 2 2.5
posed and/or disintegrated to a soil. Fresh or
discoloured rock is present either as a
discontinuous framework or as corestones.
IV Highly
More than half the rock material is decom- 14 10 10
posed and/or disintegrated to a soil. Fresh or
discoloured rock is present either as a dis-
continuous framework or as corestones.

1.7 Summary

In addition to compression tests in laboratory the uniaxial compressive strength of the rock may be
found from several other methods which have been shortly described. These are:
• Compressive strength estimated from the point-load strength (Is), given as
σc = k50 × Is50
where values of k50 varying with the rock strength related to 50 mm samples, have been
suggested.
• Compressive strength found from the Schmidt hammer rebound number.
• Compressive strength assessed from simple field test using a geological hammer.

Where test data of the rock is not available the uniaxial compressive strength may be estimated from
the geological rock name and additional information on its structure and weathering/alteration. For
anisotropic and weathered/altered rocks a rough estimate of the uniaxial compressive strength can be
found from
σc = σc50 /(fA × fW) eq. (A3-11)

where σc50 can be found for fresh rocks from published strength tables, and
fA, fW are the foliation anisotropy and weathering/alteration factors, respectively. Sug-
gested ratings have been given for both factors, but additional investigations are
required to improve these approximate values.

The lowest value (σc min) is applied in RMi. Therefore, it is important to check if the effect of
anisotropy is included in σc50 . It is also important that the rock description - on which the estimates
are based - clearly delineate the composition and structure of the rock, and that the applied terms and
 A3 - 20

characterizations of the rock are well defined. Further information on useful description of rocks is
outlined in Section 5.

The content of moisture tends to reduce the compressive strength of most rocks.
 A3 - 21

2 METHODS TO DETERMINE THE JOINT CONDITION FACTOR ( jC)

"The success of the of the field investigation will depend on the geologist's ability to recognise
and describe in a quantitative manner those factors which the engineer can include in his
analysis."
Douglas R. Piteau, 1970

The usually large number of joints with various conditions involved in a rock mass, cause that
simplifications have to be made and that rapid and inexpensive measurements are preferred. The
joint condition factor, jC, is meant to represent the main inherent variables of the friction properties
of joints in a rock mass. Basically, the condition of joints are made up of the following parameters
(see alt. A in Fig. A3-11):
• The roughness of the joint walls, given as the joint roughness factor (jR), similar to Jr in the
Q-system. jR consists of:
- smoothness (or unevenness) of the joint wall surface, and
- waviness (or planarity) of the joint wall plane.
• The character of the joint wall including possible filling and its thickness, expressed in the
joint alteration factor (jA). 1
• Length and continuity of the joints, expressed as the joint size factor, jL, which is considered a
scale and geometry factor to mainly include the different importance between large, pervasive
joints and small, irregular joints on rock mass behaviour.

Similar, as for the Jr/Ja in the Q system, the ratio jR/jA is roughly a function of tan φ, the peak
friction coefficient of the joint. Thus, also measurements of the joint friction angle can be used to
find the ratio jR/jA as shown in Fig. A3-12. This method (alt. B) is based on measurements of the
joint roughness combined with input of the actual type and size of joints. Alt. A is based wholly on
observations.
field data

JOINT WAVINESS
JOINT
ROUGHNESS
JOINT SMOOTHNESS FACTOR ( jR )
JOINT
JOINT CONDITION
JOINT CHARACTER ALTERATION FACTOR
FACTOR ( jA ) jR
jC = jL
jA
JOINT LENGTH AND JOINT SIZE
TERMINATION FACTOR ( jL )

field data
JOINT LENGTH AND JOINT SIZE
TERMINATION FACTOR ( jL ) JOINT
CONDITION
FACTOR
JOINT TYPE BASIC
& FRICTION jR
jC = jL
ROCK TYPE ANGLE ( Φb ) PEAK FRICTION jA
ANGLE ( φ )
JOINT ROUGHNESS DILATION ( similar to jR/jA )
COEFFICIENT (JRC) ANGLE ( i )

alternativ input

Fig. A3-11 The two main methods to find the joint condition factor, jC.

1
jA is similar to Ja in the Q system, but some changes have been made to adapt it as input to RMi.
 A3 - 22

Alt. A is described in Section 2.1 and 2.2, alt. B in Section 2.3. The joint size factor, jL, which is
used in both alternatives, is described in Section 2.4

2.1 Estimating the joint roughness factor (jR)

The roughness of joint walls is characterized by a large scale waviness and a small scale smoothness
or unevenness, ISRM (1978), see Fig. A3-12. During shear displacement the waviness undulations,
if locked and in contact, cause dilation since they are too large to be sheared off, while the
asperities of the smoothness tend to be damaged unless the joint walls are of high strength and/or
the stress levels are low. In practice, it is seldom possible to observe and measure both these
features along the entire joint. Some sort of simplification has therefore to be made as outlined in
this section.

Fig. A3-12 The joint wall features can be characterized by the large scale waviness and the small scale smoothness
(or unevenness).

2.1.1 Field measurements of large scale roughness

Accurate measurements of joint waviness in rock exposures is relatively time-consuming by any of

the currently available procedures (Stimpson, 1982). The three most practical methods are:
1. To estimate the overall roughness (undulation) angle by taking measurements of joint
orientation with a Clar Compass to which the base plates of different dimensions are attached,
see Fig. A3-13 left.
2. To measure the roughness along a limited part of the joint using a feeler or contour gauge to
draw a profile of the surface, Stimpson (1982). Also Barton and Choubey (1977) makes use of
this method especially in connection with core logging.
3. To reconstruct a profile of the joint surface from measurements of the distance to the surface
from a datum (typically a rod or rule laid or supported over the joint).

The first technique provides information on large scale roughness angles, but does not give a record
of the joint profile. It is applicable primarily to large exposures of joint planes. Fecker and Rengers
(1971) have from measurements using a profilograph and geological compass, shown how these
results can be applied.
 A3 - 23

The second is a rapid method, where a few decimetres profile along the joint wall surface is
obtained by a contour gauge and pattern maker. The method is well applicable for joint smoothness
measurements on drill cores. By comparing the profile obtained for the joint with standard
roughness profiles, for example of JRC or Jr in Fig. A3-14, the actual roughness value can be easily
determined.
Direction of dip
Joint surface
Amplitude
Straight edge 36”

Large plate Compass Small plate

Length = 28” α

(a)
Amplitude

Straight edge

Length = 36”
(b)

Fig. A3-13 Left: Measurement of different scales of joint waviness (from Goodman, 1987).
Right: Principle for the measurement of waviness by a straight edge (from Piteau, 1970).

JOINT ROUGHNESS COEFFICIENT (JRC)

Relation between Jr and JRCn
Subscripts refer to block size (cm) Jr JRC20 JRC100
400 20
(a) rough
16 4 20 11
300 a1 a2 I
AMPLITUDE 12
200 10 smooth
8 II 3 14 9
6 slickensided
LENGTH 5 2 11 8
100 III
(L) 4
3 Stepped
50
40 2 rough
IV 3 14 9
30
smooth
AMPLITUDE OF ASPERITIES (mm)

20 1.0 V 2 11 8

slickensided
10 VI 1.5 7 6
0.5
Undulating
5
rough
4 VII 1.5 2.5 2.3
3
smooth
2 VIII 1.0 1.5 0.9

slickensided
IX 0.5 0.5 0.4
1.0
Planar

0.5
0.4 inclination of line u = a/L
0.3
0.2

0.1
0.1 0.2 0.3 0.5 1.0 2 3 4 5 10

LENGTH OF PROFILE (m)

Fig. A3-14 Left: Diagram presented by Barton and Bandis (1992) to estimate JRC for various measuring lengths. The
inclined lines exhibit almost a constant undulation as indicated.
Right: Relationships between Jr in the Q-system and the 'joint roughness coefficient' (JRC) for 20 cm and
100 cm sample length (from Barton and Bandis, 1992).
 A3 - 24

The third technique is generally laborious and, unless the data points are closely spaced along the
joint, it does not provide a detailed profile of the small scale roughness. Robertson (1970) and
Piteau (1970) have introduced measurement of waviness by using a standard length of 0.9 m (36
inch) straight edge placed on the exposed joint surface in a direction normal to the strike (i.e., down
dip).

Amplitude (a) of the wave is a measure of the maximum offset under the straight edge as shown for
two different cases in Fig. A3-13 right. This dimension is measured in millimetre. It is related to the
length (L), which is recorded as either 1) the distance between adjacent points of contact on the
peaks of the wave, Fig. A3-13 right (a), or 2) the standard length adopted, Fig. A3-13 right (b).
Also ISRM (1978) has described a similar method to measure the waviness of the joint plane.

Kikuchi et al., (1985) has measured waviness of joint plane by selecting portions on a joint plane
exposed to the ground, using a 2 m measuring scale. The height of the exposed portions at 2 cm
intervals in the direction of dip has been measured to draw the roughness of the joint plane. Also
Barton (1982) and Barton and Bandis (1990) make use of a ruler to determine the joint roughness
coefficient JRC. Fig. A3-14 shows how different lengths of the ruler can be used to determine JRC.
As indicated on this diagram the inclined lines approximately follow the expression for the
undulation factor u = a/L presented later in eq. (A3-12) has been used to find the ratings of the joint
waviness factor, jw, where u is applied as shown in Table A3-14.

2.1.2 The joint waviness factor (jw)

As waviness does not change with displacements along the joint surface, no shearing takes place
through asperities (Piteau, 1970). The result is that waviness is considered to modify the apparent
angle of dip of the joint but not the joint frictional properties. Tsidzi (1986, 1987, 1991) has
observed that waviness is a more characteristic feature of the foliation surface in many metamorphic
rocks than the smoothness of the surface.

Waviness of the joint appears as undulations from planarity of the joint wall. Ideally, the joint
waviness should be measured as the ratio between max. amplitude over the length of the joint. As it
is seldom possible to observe the whole joint plane, a simplified measurement is to find the ratio
between max. amplitude and a reduced measured length along the joint plane called the undulation
factor

amplitude from planarity (a)

u = eq. (A3-12)
measured length along joint (L)

This expression has been used to characterize the waviness of joint plane. From combination of the
two diagrams in Fig. A3-14 the division and ratings in Table A3-14 have been worked out. The
division here of the undulation is based on the characterization presented by Milne et al. (1992)
related to 1 m profile length:
Wavy joints with undulation, u > 2%
Planar to wavy joints, with u = 1 - 2%
Planar joints, defined as u < 1%.

Their measurements are also based on the 'joint roughness coefficient' (JRC) which is described in
Section 2.3.
 A3 - 25

TABLE A3-14 THE JOINT WAVINESS FACTOR ( jw). THE RATINGS ARE BASED ON Jr IN
THE Q-SYSTEM.
TERM FOR WAVINESS waviness factor
undulation jw

Interlocking (large scale) 3

Stepped 2.5
Large undulation u>3% 2
Small - moderate undulation u = 0.3 - 3 % 1.5
Planar u < 0.3 % 1

The longest possible ruler should be applied in measurement of waviness. In many cases, however,
the determination of (jw) is done from visual observations alone. Practice from ruler measurements
may reduce the possible errors in such cases.

5
500

=1
R
JC
300 a
200
AMPLITUDE (a) OF ASPERITIES (mm)

u = a/L
ti n ly
la g
g

100
du r on
un st

70

5
1.
50 R
=
JC

30
g ley o
t
la e ra tl y

20
du od ig h
tin t
u n m sl

10
7
3%
=

5
u

3
ar
an

2
pl
h
ug
ro

1
ry

3%
ve

0. 7
0.
=

0. 5
u

0. 3
el o
at t
h er t ly

0. 2
u g o d i gh
ro m s l

0. 1
0. 07
0. 05
th

0. 03
oo
sm

0. 02

0. 01
0. 01 0.02 0.05 0. 1 0. 2 0.5 1 2 5 10
LENGTH (L) OF PROFILE (m)

SMOOTHNESS WAVI NESS

Fig. A3-15 Waviness and smoothness (large and small scale roughness) based on the JRC chart in Fig. A3-14.
 A3 - 26

2.1.3 The joint smoothness factor (js)

Small asperities or second order projections are designated smoothness or unevenness. If the joint
surfaces are clean and closed these small asperities interlock to strongly contribute to the shear
resistance especially at low stresses (Barton and Chubey, 1977; Barton, 1976, 1987, 1990b, 1993).
Smoothness asperities usually have a base length of some centimetres and amplitude measured in
hundreds of millimetres and are readily apparent on a core-sized exposure of a discontinuity (see
Fig. A3-12 and A3-15).

As indicated in Fig. A3-12 and A3-15 the 'sample length' for smoothness is in the range of a few
centimetres. There is a general problem to arrive at a numerical estimate of joint smoothness from
measurements or visual observations of the joint wall surface. A possible solution is to simply touch
the surface with the finger and compare it with a reference surface of known roughness, for example
sand papers of various abrasivity (mesh) as indicated in Table A3-16.

The terms and ratings defined in Table A3-16 are based on Jr in the Q-system (Barton et al., 1974),
while the description is also partly based on Bieniawski (1984).

TABLE A3-16 THE JOINT SMOOTHNESS FACTOR (js). THE RATINGS ARE THE SAME AS FOR Jr IN THE
Q-SYSTEM. (The description is partly based on Bieniawski, 1984).
TERM FOR The smoothness
SMOOTHNESS DESCRIPTION factor js

Very rough Near vertical steps and ridges occur with interlocking effect on the joint 3
surface.

Rough Some ridge and side-angle steps are evident; asperities are clearly 2
visible; discontinuity surface feels very abrasive (rougher than
sandpaper grade 30)

Slightly rough Asperities on the discontinuity surfaces are distinguishable and can be 1.5
felt (like sandpaper grade 30 - 300).

Smooth Surface appear smooth and feels so to the touch (smoother than 1
sandpaper grade 300).

Polished Visual evidence of polishing exists. This is often seen in coatings of 0.75
chlorite and specially talc

Slickensided Polished and striated surface that results from friction along a fault 0.6 - 1.5 1)
surface or other movement surface.
1)
Rating depends on the actual shear in relation to the striations.

2.1.4 The joint roughness factor (jR) found from jw and js

As described in the foregoing the smoothness and waviness can both be characterized by the ratio
between the amplitude of asperities and the measuring length as shown in Fig. A3-15. A similar
principle has been presented by Milne et al. (1992) who have used the joint roughness coefficient
(JRC) to also assess small scale roughness (i.e. smoothness) using 10 cm profile lengths with the
following division:
- smooth surfaces have JRC < 10 (undulation u < 2%)
 A3 - 27

- rough surfaces JRC > 10 (undulation u > 2%)

The joint roughness factor, jR, is the product of the smoothness and waviness factors:
jR = js × jw eq. (A3-13)

which is shown in Table A3-17.

TABLE A3-17 JOINT ROUGHNESS FACTOR (jR) FROM SMOOTHNESS AND WAVINESS. THE
STRUCTURE OF THE FACTOR AND ITS RATING ARE SIMILAR TO Jr IN THE
Q-SYSTEM.
waviness
slightly to strongly interlocking
planar moderately undulating stepped (large scale)
smoothness undulating
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 filled joints without contact between joint walls: jR = 1

Joint roughness includes the condition of the joint wall surface both for filled and unfilled (clean)
joints. For joints with filling which is thick enough to avoid contact of the two joint walls, any shear
movement will be restricted to the filling, and as described later, the joint roughness will then have
minor or no importance. 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.

The joint roughness coefficient, jR (or Jr) can also be found from measured values of JRC using
Fig. A3-14 or A3-15.

The classification systems which often make use of many a large amount of input data, the ratings
for roughness are mostly found from observations. It is, therefore, not common to make detailed
measurements of joint roughness profiles.

2.2 Estimating the joint alteration factor (jA)

"It is often more important to characterize discontinuities according to surface character as it is

to note their scale parameters."
Tor L. Brekke and Terry R. Howard (1972)

The joint alteration factor is a collective parameter including the strength of the joint wall and the
possible joint filling, whether it is a clean or a filled joint. The following features are included in
this complex factor:
− The condition of the surface in clean joints, i. e. joints without coating or filling; with
indication if alteration of the joint wall may be other than the rock.
− The type of coating on the joint surface.
− The type of filling in joints and its thickness.
 A3 - 28

2.2.1 Clean joints

Clean joints are without fillings or coatings. For these joints the compressive strength of the rock
wall is a very important component of shear strength and deformability where the walls are in direct
rock to rock contact (ISRM, 1978).

Close to the surface in outcrops it is imperative not to confuse clean discontinuities with "empty"
discontinuities where filling material has been leached and washed away due to surface weathering.
Clean joints can be:
a. Healed or welded joints. Joints, seams and sometimes even minor faults may be healed
through precipitation from solutions of quartz, epidote or calcite. In such cases the joint plane
can be regarded more appropriately as a plane of reduced strength.
b. Healed joints may, however, have broken up again, forming new surfaces. Also, it should be
emphasized that quartz and calcite may well be present in a discontinuity without healing it.
c. Fresh rock walls. These are joint walls of unweathered or unaltered rock. They may, however,
show staining (rust) on the surfaces.
d. Altered or weathered rock walls. Some clean joints may show alteration of the rock material
on the joint surface (Piteau, 1970). The rock surface in these joints can be in the same
condition as the rock elsewhere. Often, however, when weathering or alteration has taken
place it is more pronounced along the joint surface than in the rock. This results in a wall
strength often considerably lower than that of the fresher rock found in the interior of the rock
blocks. The degree of weathering is usually estimated from visual observations. It can partly
be quantified applying the ISRM classification of alteration and weathering (see Table A3-
12), and should be applied in jA where it differs from the alteration of the rock material, as is
shown in Table A3-21.

2.2.2 Coated joints

Coating means that the joint surfaces have a thin layer or 'paint' with some kind of mineral. The
coating, which is not thicker than a few millimetres, can consist of various kinds of mineral matter,
such as chlorite, calcite, epidote, clay, graphite, zeolite. Mineral coatings will affect the shear
strength of joints to a marked degree if the surfaces are planar. The properties of the coating
material may dominate the shear strength of the joint surface, especially weak and slippery coatings
of chlorite, talc and graphite when wet.

2.2.3 Filled joints

Filling or gouge when used in general terms, is meant to include any material different from the
rock thicker than coating which occurs between two discontinuity planes. Thickness of the filling or
gouge is taken as the width of that material between sound intact rock. Table A3-18 shows the
classification of joint or seam thickness presented by ISRM (1978).

Unless discontinuities are exceptionally smooth and planar, it will not be of great significance to the
shear strength that a 'closed' feature is 0.1 mm wide or 1.0 mm wide, ISRM (1978). (However,
indirectly as a result of hydraulic conductivity, even the finest joints may be significant in changing
the normal stress and therefore also the shear strength.)
 A3 - 29

TABLE A3-18 SEPARATION OF DISCONTINUITY WALLS

(from Bieniawski 1984)

Very tight < 0.1 mm

Tight 0.1 - 0.5 mm
Moderately open 0.5 - 2.5 mm
Open 2.5 - 10 mm
Very open 10 - 25 mm

Aperture is the perpendicular distance separating the adjacent rock walls of an open discontinuity, in
which the intervening space is air or water filled. Aperture is thereby distinguished from the width
of a filled discontinuity (ISRM, 1978).)

FILLED
DISCONTINUITIES

RECENTLY DISPLACED UNDISPLACED

CLOSE TO RESIDUAL STRENGTH CLOSE TO PEAK STRENGTH

THEREFORE WHETHER NORMALLY- WHETHER NORMALLY- OR OVER-
OR OVER-CONSOLIDATED IS NOT CONSOLIDATED OF CONSIDERABLE
OF GREAT IMPORTANCE IMPORTANCE

NEAR-SURFACE
FAULTS SHEAR CLAY BEDDING DISCONTINUITIES MAINLY
INTERBEDDED
ZONES MYLONITE PLANE CONTAINING HYDROTHERMALLY
CLAY
SLIPS WEATHERING ALTERED
OFTEN OFTEN BANDS
HYDRO- HYDRO- PRODUCTS FILLINGS
THERMAL THERMAL MOSTLY
ALTERATION MOSTLY
ALTERATION O-C. CLAY MOSTLY N-C. CLAY O-C. CLAY

Fig. A3-16 Simplified division of filled discontinuities into displaced and undisplaced, normally- and over-
consolidated categories (from Barton, 1974).

TABLE A3-19 MAIN TYPES OF COATING AND FILLING MATERIALS AND THEIR PROPERTIES, (mainly
based on Brekke and Howard, 1972)
TYPE OF MINERAL PROPERTIES
FILLING
Chlorite, talc, graphite Very low friction materials, in particular when wet.
Inactive clay materials Weak, cohesion materials with low friction.
Swelling clay Exhibits both a very low friction and swell with loss of strength because of swelling together
with considerable swelling pressure when confined.
Calcite May, particularly when being porous or flaky, dissolve during the lifetime of a construction in
rock, which reduces its contribution to the strength of the rock mass.
Gypsum May behave in the same way as calcite.
Sandy or silty materials Cohesionless, friction materials. A special occurrence of these are the thicker fillings of
altered or crushed materials being cohesionless (sand-like) materials which may run or flow
immediately after exposure by excavation.
Epidote, quartz and May cause healing or welding of the joint, resulting in an increased shear strength other hard
materials of the joint.

TABLE A3-20 PARTICLE SIZE ACCORDING TO THE MODIFIED

WENTWORTH SCALE (partly from ISRM, 1978)
TERM DIAMETER VOLUME

coarse sand 0.6 - 2 mm 0.13 mm3 - 4.6 mm3

medium sand 0.2 - 0.6 mm 0.0046 mm3 - 0.13 mm3
fine sand 0.06 - 0.2 mm
silt, clay < 0.06 mm
 A3 - 30

The effect of joint filling on the strength properties of a joint is of outstanding importance. If the
gouge is sufficiently thick, the filling (gouge) controls entirely the shear strength of the
discontinuity. With decreasing thickness, the asperities of the rock wall tend to become more
interlocked, and both the filling and the rock material contribute to the discontinuity shear strength.
Thus, the main cases with respect to gouge thickness which are worth examining are (Piteau, 1970):
a. No contact between the joint walls. The sliding plane passes entirely through gouge; the shear
strength is dependant only on the gouge material and no modification is considered for
roughness.
b. Partly contact between the joint walls. The sliding plane passes partly through gouge and
partly along the joint wall rock. The shear strength will be more complex being made up of
contributions of both gouge and wall rock.
c. Gouge is present but very thin as a coating; gouge is considered only as modification of the
friction angle.

In addition to the thickness of the filling the type of filling materials with different behaviour and/or
properties must be distinguish. The division in Table A3-19 covers most types of gouge and coating
materials. The size of the particles, fragments or blocks in the filling can be characterized according
to Table A3-20 or A3-26.

2.2.4 Characterization and rating of the joint alteration factor (jA)

The characterization of this complex parameter and its numerical values are mainly based on the Ja
in the Q-system and the features described above. The same main grouping and ratings are applied,
but some changes have been done to fit jA into the RMi system. Also the layout in Table A3-21 is
changed compared to the Q-system to possibly make field observations easier and quicker. The
main changes to Ja in the Q-system are:
1. The weathering/alteration of the rock in the joint wall.
2. As the RMi system has included the rock material (with its possible alteration/weath-ering), it
is only where the weathering of the clean joint wall is different from the rock, that jA
influences.
3. Zones or bands of disintegrated, crushed rock or clay are not included as such weakness zones
generally require special characterization as outlined in Appendix 2 and Chapter 6.

The joint alteration factor depends on the thickness, strength and basic friction angle of any material
on the joint surfaces. There are often difficult to determine if sufficient material exists to reduce
joint strength enough to warrant changing the jA rating from 1.0 to 4.0. Some preliminary
guidelines have been published by Milne and Potvin (1992) to augment the standard joint alteration
descriptions to better quantify the joint surface condition in this range:
Can be scratched with a knife Ja = 1.0 - 1.5
Can be scratched with a fingernail and feels slippery Ja = 2.0
Can be dented with a fingernail and feels slippery Ja = 4.0

These guidelines which do not remove the subjectivity from this parameter, only assist in finding a
small range of the jA ratings.

Major individual discontinuities (singularities) should as mentioned earlier be recorded on an

individual basis.
 A3 - 31

TABLE A3-21 CHARACTERIZATION AND RATING OF JOINT CHARACTER FACTOR ( jA)

(Partly based on Ja in the Q-system)
CONTACT BETWEEN THE TWO JOINT WALLS
WALL CHARACTER DESCRIPTION jA
CLEAN JOINTS
-Healed or "welded" joints Non-softening, impermeable filling (quartz, epidote etc.) 0.75
-Fresh rock walls No coating or filling on joint surface, except of staining (rust) 1
-Alteration of joint wall:
1 grade more altered The joint walls shows 1 grade stronger alteration than the rock 2
2 grades more altered The joint walls shows 2 grades stronger alteration than the rock 4

COATING OR THIN FILLING

-Sand, silt, calcite etc. Coating of friction materials without clay 3
-Clay, chlorite, talc etc. Coating of softening and cohesive minerals 4
FILLED JOINTS WITH PARTLY OR NO WALL CONTACT
Partly wall No wall
contact contact
TYPE OF FILLING thin fillings*) thick filling
MATERIAL DESCRIPTION (< approx. 5 mm) or gouge
jA jA
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)

2.3 Estimating the ratio jR/jA from friction angle recordings

Patton (1966) differentiates between first and second order projections on the joint wall surfaces,
corresponding to large scale undulations, given as waviness, and the small scale irregularities or
unevenness (smoothness). Surface roughness was determined by Patton as an angular measurement
of asperities from the general dip of the joint. Effective friction angle of a rock surface is the sum of
the basic friction angle (Φb) and the roughness angle
( i ). The effect of ( i ) is reflected as dilation under low normal loads, whereas at very high normal
loads there is very little dilation as the asperities are sheared through. At very low normal stresses
the small scale roughness (smoothness) is mainly contributing to the shear strength of joints,
whereas at higher stress the waviness is of main importance.

ISRM (1980) has classified the angle of friction as shown in Table A3-22.

TABLE A3-22 CLASSIFICATION OF THE ANGLE OF FRICTION FOR JOINTS (from ISRM, 1980).
Interval Descriptive terms
> 45o very high
35 - 45o high
25 - 35o moderate
15 - 25o low
< 15o very low

As the peak friction angle of joints can be expressed as

φ = φb + i

the ratio jR/jA can be found from

 A3 - 32

jR/jA ≈ tan-1 φ = tan-1 (φb + i ) eq. (A3-14)

where φb is the basic friction angle of the joint, and

i is the (peak) dilation angle.

The value of φb is mostly in the range 21o - 40o (Goodman, 1989). For most smooth unweathered
rock surfaces it varies between 25o and 35o, see Table A3-23. The values of φb can normally be
estimated with help of the data listed here, unless the joint walls are strongly weathered, coated or
the joint contain filling. Frequently, φb can be much lower when mica, talc, chlorite or other sheet
silicate minerals occur on the sliding surface or when filling is present. Value as low as 6o have been
reported in saturated fillings of montmorillonite clay (see Fig. 10-63 in Lama and Vutukuri, 1978).

TABLE A3-23 BASIC FRICTION ANGLES OF VARIOUS UNWEATHERED ROCKS OBTAINED FROM FLAT
AND RESIDUAL SURFACES (from Barton and Choubey, 1977)
Sedimentary basic friction Metamorphic basic friction Igneous basic friction
rocks angle φ b rocks angle φ b rocks angle φ b
Sandstone dry 26 - 35 (32)*) Amphibolite dry 32 Basalt wet 31 - 36
wet 25 - 34 (31) Gneiss wet 23 - 26 dry 35 - 38
Siltstone wet 27 - 31 dry 26 - 29 Granite, - fine-grained wet 29 - 31
dry 31 - 33 Slate wet 21 dry 31 - 35
Shale wet 27 dry 25 - 30 - coarse-grained wet 31 - 33
Conglomerate dry 35 dry 31 - 35
Chalk wet 30 Porphyry wet 31
Limestone wet 27 - 35 dry 31
dry 31 - 37 Dolerite wet 32
dry 36
*)
numbers in parenthesis are average values

Both waviness and smoothness contribute to the dilation or roughness angle, i, which can have any
value between 0o and 40o or more at low pressures. Two different methods to estimate ( i ) has been
roughly described in the following:
1. Using the joint roughness coefficient (JRC) as introduced by Barton (1973).
2. Combining the trace length and joint length found from photographs of natural cross sections
of joints, Turk and Dearman (1985).

The JRC method has been developed by Barton (1973), Barton and Choubey (1977) and Barton and
Bandis (1980, 1990) to estimate the shear strength of discontinuities. JRC ranges from 5 for smooth
planar joints to 20 for rough undulating joints (Fig. A3-13). It is subjectively estimated from
comparison with standard roughness profiles in Fig. A3-17. Barton and Bandis (1980) have
proposed a modified way of separating the components of sliding resistance where Patton's dilation
angle, i, is replaced by a bivariate measure
i = JRC × log10 (JCS/σn ) eq. (A3-15)

Here JCS = the joint wall compressive strength (for fresh rocks JCS = σc) , and
σn = the normal stress across the joint. For the peak dilation angle the normal stress is
very low (approximately 0.001 MPa, as in the tilt test).

This measurement of ( i ) in degrees includes both the small and large scale asperities. The factor
log10 (JCS/σn ) corrects JRC for asperity shearing. With low compressive strength they shear, and
with high compressive strength they are overridden, refer to Barton and Bandis (1990).
 A3 - 33

TYPICAL ROUGHNESS PROFILES for JRC range:

1 0-2

2 2-4

3 4-6

6-8
4

5 8 - 10

6 10 - 12

7 12 - 14

8 14 - 16

9 16 - 18

10 18 - 20

0 5 10 cm
SCALE

Fig. A3-17 Typical roughness profiles for various JRC ranges (from Barton and Choubey, 1977)

Barton, who recommends measuring JCS by correlation to Schmidt Hammer rebounds, had
originally suggested JRC by inspection. The procedure for estimating JRC for specific sites may,
according to Merritt and Baecher (1981), be prone to large measurements errors as it may be
dependent on how practised the field engineer or geologist is. To circumvent this problem, Barton
and Bandis (1980) recommends that tilt or push tests be used on typical block sizes in situ to obtain
a more reliable measure of JRC.

The second method for estimating the dilation angle ( i ) is proposed by Turk and Dearman (1985).
The joint trace seen on photographs of natural cross sections can be measured by some device as a
'map measure'. The roughness (dilation) angle ( i ) is determined by measurement of the direct
distance and the trace length along the surface between the starting and finishing points for the
profile measurement as
i = arc cos (L1 /L2) eq. (A3-16)

where L1 = the direct length, and L2 = the trace length on the measured joint surface.
The roughness angle ( i ) in eq. (A3-16 corresponds, according to Turk and Dearman, to the dilation
angle measured under very low normal load (σn = 0.001 MPa, as used in the first method). This
value can be determined on field and laboratory specimens, by simple linear measurement. The
method is simple to apply and gives the peak ( i ) value for a particular joint surface. Also profiles
found by feeler or contour gage in the second method mentioned in Section 2.2.1 may be used.
 A3 - 34

2.4 The joint size and continuity factor (jL)

The joint size and continuity factor consists of two contributions, the joint length and the termina-
tion (continuity) of the joint. The size of joints is, according to Piteau (1970), a difficult property to
determine, but is essential in evaluations, since the strength reduction on a failure surface which
contains a discontinuity is a function of the joint size. Also ISRM (1978) and Merritt and Baecher
(1981) mention the great importance of joint size and the problem in mapping its length, both for
small and large joints. A main reason for this is that the whole joint plane seldom can be seen in
rock exposures, most often only the joint is seen as a trace (line), and this trace does seldom
represent the largest dimension (Pollard and Aydin, 1988). In drill cores only a very small part of
the joint can be studied as shown in Fig. 3-5 in Chapter 3. Hudson and Priest (1983) recommend
that a measure of the joint length should only indicate the length/size interval of the joint.

TABLE A3-23 CLASSIFICATION OF JOINT PERSISTENCE (from Bieniawski, 1984)

very low persistence <1m

low 1-3m
medium 3 - 10 m
high 10 - 20 m
very high > 20 m

Less frequently, in large cuttings, for example in quarries and open pits, it may be possible to record
the dip length and the strike length of exposed joints and thereby estimate their persistence along a
given plane through the rock mass using probability theory (Jennings, 1970; Robertson, 1970;
ISRM, 1978).

The size of the joint is often proportional to the thickness or separation of the joint (Nieto, 1983;
Kikuchi et. al., 1985), see Fig. A3-18. From the visible parts of the joint traces in unweathered
exposures and excavated surfaces in tunnels or cuttings for example, the overall length interval of
the joint set can be roughly estimated (Piteau (1970).

Discontinuous joints, i.e. joints that terminate in massive rock, will result in a rock mass with great
inherent strength, since rupture must occur through intact rock before failure develops (Robertson,
1970). The ratings of joint size have been assessed from scale effect of joint length presented by
Barton (1992)
JCSn = JCS0(Ln /L0) - 0.03JRC eq. (A3-17)

where JCSn and JCS0 are the joint compressive strength (= σc) where subscripts refer to in situ
joint size and lab. strengths respectively,
Ln and L0 are the in situ and lab. size respectively
JRC refer to the lab. scale joint roughness coefficient.

Assuming an average value of JRC = 10 and that a 'normal' joint has a length in the range 1 - 10 m
(average L0 = 3 m), eq. (A3-17) can be expressed as
jL = JCSn /JCS0 ≈ 1.5 × L - 0..3 eq. (A3-18)

where jL = the joint size factor, and L = the trace length of the joint, given in metre.
 A3 - 35

(m)
100

10 (m)
100
length

1
10

length
0.1 1.0

0.01 0.1
0.1 0.3 1 2 10 (mm) 0.01 0.1 1.0 10 (mm)
mean value of aperture
aperture

Fig. A3-18 Relation between length and aperture. Mean values are shown in the right figure
(from Kikuchi et al., 1985).

Table A3-24 shows the characterization of joint size and the values of jL. From the foregoing the
characterization of the joint size and continuity has been based on the length intervals of partings,
short joints, medium and long joints as shown in Fig. 2-4 in Chapter 2. In a joint mapping or survey
it may often be possible to differ between these types of joints. Also the classification in Table A3-
23 has partly been applied.

TABLE A3-24 CHARACTERIZATION OF THE JOINT SIZE AND CONTINUITY FACTOR (jL).
LENGTH Joint size factor jL*)
TYPE
INTERVAL range (mean)
<1m bedding/foliation partings 2-4 (3)
0.1 - 1.0 m short/small joint 1.5 - 3 (2)
1 - 10 m medium joint 0.75 - 1.5 (1)
10 - 30 m long/large joint 0.55 - 0.75 (0.7)
> 30 m very long/large joint/seam**) approx. 0.3 - 0.55
*)
FOR DISCONTINUOUS JOINTS: multiply jL by 2
**)
Often a singularity, and should in these cases be characterized separately.

The influence of discontinuous joints, i.e. joints that terminate in massive rock, has been assumed as
a doubling of the jL value.

2.5 Summary

Joint surface characteristics are broken into the joint roughness factor, jR, and the joint alteration
factor, jA, which can be compared to Jr and Ja in the Q-system. Joint roughness is assessed at a
large and small scale and joint alteration is determined from thickness and character of the filling
material or from the coating on the joint surface. There is a significant degree of subjectivity in
determining both these terms; some adjustments have been developed to improve their
measurement.
 A3 - 36

in situ data calcu lations

JOINT LENGTH AND JOINT SIZE

TERMINATION FACTOR ( jL )
JOINT
FILLING
CONDITION
observations

MATERIAL FACTOR
JOINT
WEATHERING
OF JOINT WALLS
ALTERATION jR
FACTOR ( jA ) jC = jA jL
JOINT
SEPARATION

JOINT WAVINESS
or PLANARITY JOINT
ROUGHNESS
JOINT SMOOTHNESS FACTOR ( jR )
or UNEVENNESS

ut
alt. input

t. inp
al
jR/jA
measurement

JOINT ROUGHNESS DILATION

COEFFICIENT ( JRC ) ANGLE ( i )

alt. input FRICTION ANGLE alt.

OF JOINT input
JOINT
BASIC
CONDITION
FRICTION FACTOR
observation

TYPE OF JOINT ESTIMATION

ANGLE (Φb )
jR
jC = jA jL
JOINT LENGTH AND JOINT SIZE
TERMINATION FACTOR ( jL )

Fig. A3-19 The parameters involved and the methods applied to determine the joint condition factor (jC.)

Another method to determine the value of the joint condition factor, is to use the joint roughness
coefficient, JRC, to determine the friction angle of the joint.

Fig. A3-19 shows the two main methods and the parameters and principle calculations involved in
the joint condition factor jC. The factor for joint size and continuity is determined and applied in the
same way in both methods.
 A3 - 37

3 METHODS TO DETERMINE BLOCK SIZE

"Since joints are among the most important causes of excessive overbreak and of trouble with
water, they always deserve careful consideration."
Karl Terzaghi, 1946

The block size is a result of the detailed jointing in a rock mass mainly formed by the small and
moderate joints (Selmer-Olsen, 1964). The block dimensions are determined by joint spacings and
the number of joint sets. Individual or random joints and possible other planes of weakness may
further influence on the size and shape of rock blocks. Impact from excavation works may also
contribute to the splitting up of a rock mass into blocks.

ISRM (1978), Barton (1990) and several other authors mention that the block size is an extremely
important parameter in rock mass behaviour. A wide range of scale effects in rock engineering can
be explained by this feature, including compression strength, deformation modulus, shear strength,
dilation, conductivity, shear stiffness, failure mode, stress-strain behaviour etc.

In addition to briefly outline how various methods can be used to measure block size and/or the
quantity of joints, the following sections also show correlations between the measurements.

3.1 Types of block volume and joint density measurements

Different methods have been developed to measure the quantity or density of joints in the rock
mass. The selection of the method(s) to be applied at an actual site is often a result of the
availability to observe the rock and its jointing in an exposure, the requirement to the quality of the
collected data, the type and cost of the investigation or survey, and the experience of the engineering
geologist.

If all the blocks in a rock mass could be measured or "sieved" a block size distribution can be found,
much the same used to describe particle sizes of a soil. As the joint spacings generally vary greatly,
the difference in size between the smaller and the larger blocks can be large, Fig. A3-21. Therefore
the characterization of block volume should be given as an interval rather than a single value.

100

80
% SMALLER

60

40 ID = VOLUMES
D25 = 2m No. JOINTS = 43
D50 = 2.52m
20 D75 = 3.32m No. BLOCKS = 82
Cs = 1.66 Ju = 5.6
0
0.01 0.1 1 10m
B LO CK SI ZE (cu bic ed ge le ngt h)

Fig. A3-21 Example of a block size distribution curve for a rock mass (from Milne et al., 1992).

Where less than 3 joint sets occur it is often expected that defined blocks will not be found.
However, in most cases random joints or other weakness planes may contribute so that definite
blocks occur. Also where the jointing is irregular, or many of the joints are discontinuous, it can be
difficult to recognize the actual size and shape of individual blocks. Sometimes the block size and
 A3 - 38

shape therefore have to be determined from reasonable simplifications where an equivalent block
volume is assumed (see Section 3.2.3). Simplifications may also be necessary to possibly collect the
data within a reasonable amount of work.

TABLE A3-26 CLASSIFICATION OF BLOCK VOLUME RELATED TO PARTICLE SIZE (VOLUME) FOR
SOILS (Vb = 0.58 Db3 has been applied for the correlation between particle diameter (Db) and
block volume Vb).
TERM FOR TERM FOR VOLUME TERM FOR APPROX. SOIL
DEGREE OF JOINTING BLOCK SIZE (Vb) PARTICLE VOLUME
(or DENSITY OF JOINTS)

coarse sand 0.1 - 5 mm3

fine gravel 5 - 100 mm3
Extremely high Extremely small < 10 cm3 .. . . medium gravel 0.1 - 5 cm3
Very high Very small 10 - 200 cm3 .. . . coarse gravel 5 - 100 cm3
High Small 0.2 - 10 dm3 .. . . cobbles 0.1 - 5 dm3
Moderate Moderate 10 - 200 dm3 .. . . boulders 5 - 100 dm3
Low Large 0.2 - 10 m3 .. . . blocks > 0.1 m3
Very low Very large 10 - 200 m3
Extremely low (massive) Extremely large > 200 m3

TABLE A3-27 THE MAIN METHODS DESCRIBED IN THIS SECTION WHICH CAN BE USED TO MEASURE
THE QUANTITY OF JOINTS AND THE BLOCK SIZE.
DRILL CORE or
PARAMETER SURFACE OBSERVATIONS SCANLINE
MEASURED OBSERVATIONS
3-D registration 2-D registration 1-D registration
BLOCK SIZE Block volume estimated from
defined joint spacings (and
angles between joint sets).

- Block volume Block volume estimated from Jv

(see eq. (A3-27).

Block volume measured in the Block volume of drill

field. cores fragments 1).

- Equivalent Estimated block diameter Indirect block diameter

block diameter (Ib) according to ISRM measure (given as RQD).
(1978).
DEGREE OF
JOINTING

- Joint frequency Registration of the volumetric Measured number of joints Measured number of joints
joint count (Jv). intersecting an area. intersecting a line.

*Weighted jointing density *Weighted jointing density

measurement. measurement.

- Joint spacing Measured spacings for each joint Measured mean joint Measured length of cores
set. (Normally used to express spacings related to a plane. bits or spacings along a
the block size (Vb) or the line
volumetric joint count (Jv)). (fracture intercept, (ISRM,
1978)).
* Measurement introduced in this contribution.
1)
In drill cores the block volumes refer to have the size of core diameter or less (gravel or pebbles size)
 A3 - 39

The density of joints in a rock mass is mainly found from various types of observations made on
surfaces or on drill cores. The most common types are:
A. Surface observations, made as:
- Field registration of block volume,
- Joint spacing or frequency measurements
- 3-D jointing density (as for the volumetric joint count, Jv)
- 2-D jointing density (as for the number of joints in a surface)
- 1-D jointing density (as for the number of joints along a scanline)
B. Drill core logging, recorded as:
- Rock Quality Designation (RQD),
- 1-D jointing density (the number or length of core pieces)
C. Geophysical measurements; the jointing density is mainly estimated from
- Sonic velocities recorded by refraction seismic measurements.

Some of the measurements that can be made on rock outcrops, excavated surfaces and drill cores are
shown in Table A3-27. The correlations between them, which is developed in this appendix, enable
block volume to be determined from different sources.

As the blocks generally have varying sizes and shapes the measurements of characteristic
dimensions can be very time-consuming and laborious. To remedy this, easy recognizable
dimensions of the blocks and simple correlations between the different types of jointing
measurements are worked out.

3.2 Block volume measurements

The block volume is intimately related to the intensity or degree of jointing. Each one of such
blocks is more or less completely separated from others by various types of discontinuities. The
greater the block size the smaller will be the number of joints penetrating the rock masses. Hence,
there is an inverse relationship between the block volume and the number of joints.

Especially where irregular jointing occurs it is time-consuming to measure all (random) joints in a
joint survey. In such cases, as well as for other jointing patterns, it is often much quicker - and also
more accurate - to measure the block volume directly in the field. Where three or more regular joint
sets occur, the block volume can easily be found from the joint spacings.

For each of the joint sets the spacings vary within certain ranges. The block volume in a rock mass
should be characterized by a modal size together with the range i.e. typical largest and smallest
block indices. (ISRM, 1978; Burton, 1965). Ideally, the range should be between about 25% and
75% of the block sizes similar to what is often practised to characterize particle distribution of soils,
Fig. A3-21.

3.2.1 Block volume found from joint spacings

For blocks with less than 3 joint sets, the volume is often determined by the random joints in
addition to the joints occurring in sets. For more than 3 joint sets the volume is determined by the
jointing pattern and spacings as it is for 3 joint sets. Individual or random joints may further
influence on the type and shape of blocks.

The volume of a block determined by 3 joint sets is given as

 A3 - 40

S1 × S 2 × S 3 Vb0
Vb = = eq. (A3-19)
sin γ 1 × sin γ 2 × sin γ 3 sin γ 1 × sin γ 2 × sin γ 3

where γ1, γ2, γ3 are the angles between the joint sets, and
S1, S2, S3 are the spacings between the individual joints in each set.
Vbo is the volume for joints intersecting at right angles.

For a rhombohedral block with two angles between 45 - 60o, two between 135 - 150o and the last
two 90o, the volume will be between Vb = 1.3 Vbo and Vb = 2 Vbo as is further shown in Section
4. Compared to the variations in spacing the effect from the intersection angle between joint sets is
relatively small.

As earlier mentioned, no blocks will theoretically be formed where only one or two joint sets occur
and random joints are very few or absent. Section 3.2.3 outlines methods to establish an equivalent
block volume in such cases.

3.2.2 Block volume measured directly in situ or in drill cores

Where the individual blocks can be observed in a surface, their volume can be directly measured
from relevant dimensions by selecting several representative blocks and measuring their average
dimensions (ISRM, 1978). From this, the range in the block volumes can be determined.

For small blocks or fragments having volumes of dm3 or less this method of block volume
registration is often beneficial as it is easy to estimate volume compared to all the measurements of
the many joints. Block volume can also be found in drill cores where the presence of many joints
delineate small fragments. The laborious and time-consuming measurements of the many joints in
the core in such cases is often a main reason to apply a simpler method for core logging like the
RQD. By applying block volume in the rock mass characterization a significantly more accurate
registration of the joint density or frequency is achieved. The quality of such measurements has
been further outlined in Appendix 4 where various methods of jointing measurements have been
investigated.

Also where irregular jointing occurs it may be much more convenient to directly measure the block
size by eye during field inspection than to record all the joints and their locations.

3.2.3 Methods to find the equivalent block volume where joints do not delimit blocks

A minimum of three joint sets in different directions are theoretically necessary to delimit blocks in
a rock mass. As mentioned in Appendix 1 there are cases with irregular jointing where blocks are
formed mainly from random joints, and other cases where the blocks are delimited by one or two
joint sets and additional random joints.

Sometimes, however, where the jointing is composed of one or two joint sets with no or few
random joints, no defined blocks are formed. In such cases an equivalent block volume is applied in
the calculations. Such block volume may be found from the following methods:

• The equivalent block volume may be estimated directly in the field from joint observations.
• Where only one joint set occurs the equivalent block volume may be considered similar to the
area of the joint plane (i.e. L12 ) multiplied by the joint spacing (S1).
 A3 - 41

(Example: For foliation partings with lengths L1 = 0.5 - 2 m and spacing S1= 0.2 m the eq.
block volume is between Vb = 0.2 × 0.52 = 0.05 m3 and Vb = 0.2 × 22 = 0.8 m3)

• For two joint sets the spacing for the two sets (S1 ans S2) and the length (L) of the joints can be
applied: Vb = S1 × S2 × L.

• The equivalent block volume can be found from eq. (A3-27): Vb = β × Jv - 3 which requires
input from the block shape factor β. 2
β can be estimated from eq. (A4-1): β = 20 + 7 Smax /Smin
where Smax and Smin are the shortest and longest dimension of the block.
Another method to estimate the value of β from the length and spacing of the joints is outlined in
the following:
Eq. (A4-1) is developed for three joint sets. Where less than three sets occur, it can be
adjusted by a factor nj representing the rating for joint sets to characterize an equivalent
block shape factor:
β = 20 + 7 (Smax /Smin )(3/nj) = 20 + 21 × Smax /Smin × nj eq. (A3-20a)

or if the more accurate expression eq. (A4-2) is applied:

β = 20 + (21/nj) (Smax /Smin ) (1 + 0.1 log (Smax / Smin)) eq. (A3-20b)

The ratings of nj are given as:

3 or more joint sets nj = 3
2 joint sets + random joints nj = 2.5
2 joint sets nj = 2
1 joint set + random joints nj = 1.5
1 joint set only nj = 1

For fissures, partings and small joints where their length often can be found or estimated, the length
and spacing of the joints correspond to the longest and shortest block dimension, hence the ratio
length/spacing = L1/S1 can be applied in eq. (A3-20a):
β = 20 + 21 × L1/(S1 × nj) eq. (A3-20c)

For long joints it is generally sufficiently accurate to use L = 4 m.

Table A3-27A shows a comparison between the values of β found from various equations. The
rows for defined blocks typed in bold letters show the correct values for such blocks. With one joint
set the blocks are flat, while two joint sets have been assumed to form long&flat blocks. From the
table it is clear that the simplified eqs. (A3-20a) and (A3-20b) give higher values for β for ratio (=
longest/smallest block dimension) smaller than 20, while for the very flat or long&flat blocks the
values are lower.

Example
For 1 joint set (nj = 1) spaced S1 = 0.2 m having joint length L1 = 2 m the block shape factor
according to eq. (A3-20a) is β = 20 + 21 L1/(S1× nj) = 230. The volumetric joint count for this set,
Jv = 1/S1 = 5 , gives the equivalent Vb =β × Jv - 3 = 1.84 m3.
(For a defined block limited by 3 joints sets with spacings S1, L1, L1 the volume is
Vb = 0.2 × 2 × 2 = 0.8 m3 )

2
As the volumetric joint count can be measured also where joints do not delimit defined blocks, this approach can
be applied where few joints sets are found.
 A3 - 42

TABLE A3-27A COMPARISONS OF THE BLOCK SHAPE FACTOR β FOR VARIOUS TYPES OF JOINT
PATTERN AND EQUATIONS
Values of β for ratio*) =
JOINT PATTERN
1 2 3 5 7 10 20 30 50
FLAT BLOCKS:
3 joint sets (defined blocks) eq. (A3-28) 27 32 42 69 104 172 532 1092 2812
(dimensions: S1, S2 and S3 = S2)

1 set only (nj = 1) eq. (A3-20a) 41 62 83 125 167 230 440 650 1070
(dimensions: S1 and L1(= S3)) eq. (A3-20b) 41 63 86 138 193 284 640 1061 2061

1 set + random joints (nj = 1.5) eq. (A3-20a) 34 48 62 90 118 160 300 440 720
(dimensions: S1 and L1 (= S3)) eq. (A3-20b) 34 48.5 64 98 136 196 433 714 1380

LONG&FLAT BLOCKS:
3 joint sets (defined blocks) eq. (A3-28) 27 30.4 36 50 65 90 185 295 554
(dimensions: S1, S2 and S3 = S22 )

2 sets only (nj = 2) eq. (A3-20a) 30.5 41 51.5 72.5 93.5 125 230 335 545
(dimensions: S1 and L1 (= S3) eq. (A3-20b) 30.5 41 53 79 107 152 330 540 1040
or L2 (= S3))

2 sets + random (nj = 2.5) eq. (A3-20a) 28.4 37 45 62 79 104 188 272 440
(dimensions: S1 and L1 (= S3) eq. (A3-20b) 28.4 37 46.5 67 89 126 268 436 836
or L2 (= S3))

eq.(A3 − 28) : β =
[(S2/S1) + (S2/S1)(S3/S1) + (S3/S1)] 3
(the definition of β for 3 joint sets)
[(S2/S1)(S3/S1)] 2

eq. (A3-20a): β = 20 + 21⋅ Smax /Smin× nj

eq. (A3-21b): β = 20 + (21/nj) (Smax /Smin ) (1 + 0.1 log (Smax/Smin))
*)
ratio = longest/shortest block dimension
S1, S2, S3 = spacings for the joint sets; L1, L2, L3 = length of the joints; Nl = number of joints per m

3.3 Block diameter registrations

The block size index (Ib) introduced by ISRM (1987), is a measure of the block diameter. According
to ISRM this measure can be compared with the particle diameter for soils. The block size index can
be estimated by selecting by eye several typical block sizes and taking their average dimensions.
Since the index may range from millimetres to several metres, ISRM indicates a measuring accuracy
of ±10% as sufficient. "Each domain should be characterized by a modal (Ib) together with the
range i.e. typical largest and smallest block indices." The block size index Ib seems from published
papers to be seldom used and is therefore not further used in this work.

The equivalent block diameter, Db, used in the rock support charts in Chapter 6, Section 4.2 and 4.3
is another expression for the block diameter which uses the block volume and
its block shape factor β given in eq. (6-9): 3
Db = (27/β)Vb1/3
3)
(β0 /β) has been chosen in eq. (6-6) as a simple expression to roughly find the smallest block diameter. For most
cases it can be used with satisfactory accuracy for β < 150. For higher values of β a dominating joint set will normally
be present for which the average joint spacing should be applied. β is described in Section 4 in this appendix.)
 A3 - 43

Where a dominating joint set occurs the joint spacing (Sa) may preferably be used as the dimension
for the block diameter (Db = Sa).

3.4 Rock quality designation (RQD)

Rock quality designation is the method perhaps most commonly used for characterizing the degree
of jointing in bore hole cores, although this parameter also may implicitly include other rock mass
features as weathering and 'core loss' (Bieniawski, 1984).

RQD can be regarded as an indirect block size measure, as it is an expression of intact core lengths
greater than a threshold value of 0.1 m along any scan line (bore hole). Increase in the number of
joints in a rock mass causes decrease in RQD.

This parameter was originally restricted to characterize the amount of discontinuities in a drill core.
RQD is a main input both in the Q and the RMR system; in cases where no core drilling has been
carried out, the RQD value is roughly estimated from surface observations. An approximate
transition between surface observations of jointing density and RQD has been presented by
Palmström (1974, 1982) (see Fig. A3-22). Hudson and Priest (1983) and Sen and Eissa (1991,
1992) have later developed theoretical analytical approaches for determination of RQD from joint
spacings.

From its definition 'the sum of core pieces longer than 0.1 m for a certain length, taken as a percent'
and the fact that it is often assumed only from observation of rock surfaces, the use of RQD is a
crude and generally inaccurate measure of the degree of jointing. This is further explained in
Appendix 4.

The classification of RQD has been given by Deere (1966) as presented in Table A3-28.

TABLE A3-28 CLASSIFICATION OF ROCK QUALITY

DESIGNATION (RQD) (from Deere, 1966).
TERM RQD
very poor < 25
poor 25-50
fair 50-75
good 75-90
excellent 90-100

3.4.1 Correlation between the RQD and the volumetric joint count (Jv)

Also from its definition - being independent of the length of individual core pieces being longer
than 0.1 m - the RQD is a crude measure of the degree of jointing and hence the block size. It is,
therefore as mentioned in Appendix 4, not possible to obtain good correlations between RQD and
Jv or between RQD and other measurements of jointing. This is illustrated in Fig. A3-22 where the
following expression is given
RQD = 115 - 3.3 Jv eq. (A3-21)

As a consequence of this, especially when many of the core pieces have lengths around 0.1 m, the
correlation above must be regarded as a very rough. It is, however, difficult to recommend any
better transition from RQD via Jv to block volume than this correlation where RQD is the only
jointing data available.
 A3 - 44

From the volumetric joint count the block volume can be found provided input of the block shape
factor (β), see eqs. (A3-26) and (A3-27). Where β is not known it is recommended to use an
assumed 'common' value of β = 40.
GOOD VERY GOOD

100

90
RQD

75
FAIR

RQD = 115 - 3.3 Jv

50
POOR

25
VERY POOR

0
1 2 3 6 10 20 30 60
VOLUMETRIC JOINT COUNT (Jv)
VERY LOW LOW MODERATE HIGH VERY HIGH

Fig. A3-22 Correlation between RQD and Jv (from Palmström, 1982).

3.5 The volumetric joint count (Jv)

The volumetric joint count, Jv, has been described by Palmström (1982, 1985, 1986) and Sen and
Eissa (1991, 1992). It is a measure for the number of joints within a unit volume of rock mass,
defined by
Jv = Σ (1/Si) eq. (A3-22 )

where S is the joint spacing in metres for the actual joint set.

Also random joint can be included by assuming a 'random spacing' for these. Experience indicate
that this should be set to Sr = 5 m; thus, the volumetric joint count can be generally expressed as
Jv = Σ (1/Si) + Nr/5 eq. (A3-23 )

Here Nr is the number of random joints in the observation area adjusted for their length, see
Section 3.6.1.

TABLE A3-29 CLASSIFICATION OF THE VOLUMETRIC JOINT COUNT (Jv)

(revised after Palmström, 1982)

TERM FOR JOINTING TERM FOR Jv Value of Jv

massive extremely low < 0.3

very weakly jointed very low 0.3 - 1
weakly jointed low 1-3
moderately jointed moderately high 3 - 10
strongly jointed high 10 - 30
very strongly jointed very high 30 - 100
crushed extremely high > 100
 A3 - 45

The volumetric joint count can easily be calculated since it is based on common observations of
joint spacings or frequencies. In the cases where mostly random or irregular jointing occur the Jv
can be found by counting all the joints observed in an area of known size. The accuracy of this
measurement is described in Appendix 4. Table A3-29 shows the classification of Jv.

3.5.1 Block volume (Vb) estimated from the volumetric joint count (Jv)

Because both the volumetric joint count (Jv) and the size of blocks in a rock mass vary with the
degree of jointing, there exist a correlation between them (Palmström 1982). Jv varies, however,
with the joint spacings, while the block size also depends on the type of block as shown in Fig. A3-
24. A correlation between the two parameters has therefore, to be adjusted or corrected for the block
shape and the angle between the joint sets, as shown in the following.

The volumetric joint count determined from three joint sets intersecting at right angles, is expressed
as
1 1 1 S 2 × S 3 + S1 × S 3 + S1 × S1 S 2 × S 3 + S1 × S 3 + S1 × S 2
Jv = + + = = eq. (A3-24)
S S S S 2 × S 2 × S3 Vbo

where S1, S2, S3 are the joint spacings.

Using Vbo = Vb × sinγ1 × sinγ2 × sinγ3 for intersections at other angles eq. (A3-24) can be
expressed as
S 2 × S 3 + S1 × S 3 + S1 × S 2
Jv = eq. (A3-25)
Vb × sin γ 1 × sin γ 2 × sin γ 3

(α 2 + α 2 × α 3 + α 3)
3
1 β
3
Jv = × =
(α 2 × α 3)
2
Vb× sin γ 1 × sin γ 2 × sin γ 3 Vb× sin γ 1 × sin γ 2 × sin γ 3

By applying the ratio α2 = S2/S1, and α3 = S3/S1, provided S3 > S2 > S1, and
S13 = Vbo /(α2 × α3) eq. (A3-25) can be expressed as
From this the block volume is
1
Vb = β × Jv-3 × eq. (A3-26)
sin γ 1 × sin γ 2 × sin γ 3

In the cases where all angles between the block faces are 90o, the block volume is given as

Vbo = β × Jv-3 eq. (A3-27)

( α 2 + α 2 × α 3 + α 3)
3

The factor β= eq. (A3-28)

( α 2 × α 3)
2

depends mainly on the differences between joint spacings; i.e. the block shape, and has therefore
been named the block shape factor. A graph to determine its value from
eq. (A3-28) and its use is further described in Section 4.
 A3 - 46

1:1:1
1:1:5
1:3.5:3.5
1:1:8
1:5:5
1:8:8
1:12:12 1:12:24
1000
500
CUBIC

1:1:1
100 1
(Vo) BLOCK SIZE (m3)

1
50 1
ELONGATED
10 1:1:X
5
1 X
1
1 PLATY
0.5
Y
1:X:Y
soil classification

0.1
X
0.05
1

0.01
0.005

0.001
0.0005

0.0001
0.5 1 2 3 5 10 20 30 60 100 200
VOLUMETRIC JOINT COUNT (Jv)
VERY LOW LOW MODERATE HIGH VERY HIGH

Fig. A3-24 The relation between block size and volumetric joint count, Jv for various block shapes (from Palmström,
1982).

If three joint sets occur and both Jv and β are exactly measured, the exact block volume can be
calculated from eq. (A3-26) or eq. (A3-27) provided the joints intersect at right angles. Quite often,
however, not all faces of a block are known; in these cases the value of β can not be exactly found.
Therefore, a simplified method to determine the block shape factor given as
β = 27 + 7(α3 - 1) = 20 + 7(S3/S1) eq. (A3-29)
has been developed as outlined in Appendix 4, where S3/S1 = α3 is the ratio between the longest
and shortest block face. By applying this in eq. (A3-26) the block volume can be estimated from
20 + 7(S 3 /S1) Jv-3
Vb = eq. (A3-30)
sin γ 1 × sin γ 2 × sin γ 3

or Vbo = β × Jv -3 ≈ {20 + 7(S3/S1)} Jv - 3 eq. (A3-31)

These equations are valid for prismatic blocks where the joints or block faces intersect at right
angles.

As the volumetric joint count (Jv) by definition takes into account in an unambiguous way all the
occurring joints in a rock mass, it is often appropriate to use, Jv, in the correlation between jointing
frequency registrations and block volume estimates (Palmström, 1982). Important in these is the
block shape factor β which is included in all equations to estimate the block volume.

3.6 Joint frequency measurements

 A3 - 47

When the frequency is given for each joint set, it is possible to establish a correlation between joint
frequencies and block volume. In other cases, when an 'average frequency' is given, it is uncertain
whether the value refers to one-, two- or three-dimensional measurements, hence no accurate
correlation factor can be presented.

Some of the methods described in the following to estimate the block volume from joint frequency
measurements, are also described in Appendix 4.

3.6.1 2-D joint frequency in an area or surface

The 2-D joint frequency is the number of joints measured in an area. The length of the joints
compared to the size of the area will, however, influence on the frequency and some sort of
adjustments have to be made to estimate the block volume from this type of jointing measurement.
Fig. A3-25a shows three different observation areas for which the joints are larger than the
dimension of the area.

2
Area 1 (4 m ) with 4 joints
Area 2 (20 m2 ) with 10 joints
Area 3 (63 m2 ) with 17 joints

Fig. A3-25a Various sizes of the observation area and the number of joints observed. All joints are longer than the
dimension of the area.

In the table below the density of joints per square metre and per metre are given. As seen the latter
method gives a constant number of the joint frequency (Na), given as the density of joints per metre.

size of area number of joints joints/m2 joints/m

A na Na = na/A Na = na/ A
4 m2 4 1 2
20 m2 10 0.5 2.2
63 m2 17 0.27 2.1

If the joints are smaller, a higher amount of joints may be observed within the area, as shown in the
lower diagram of Fig. A3-25b.
 A3 - 48

2 observation area 2
Block area: 3 m 8 m x 18 m = 144 m

Observed: long joint bedding joint

9 bedding joints
6 other long joints 0 5m
0.5
Na = (9 + 6)/144 = 1.2 joints/m

Block area: 3 m2 observation area

8 m x 18 m = 144 m2

small joint bedding joint

Observed:
9 bedding joints
48 small (0.8 - 1.1 m long) joints
0 5m
Na = (9 + 48 .1.0/1440.5 )/1440.5 = 1.1 joints/m

Fig. A3-25b Two 'exposures' of joints with different jointing pattern. The blocks in both have the same size. A higher
amount of joints is recorded in the lower figure where the cross joints are short. Applying Na = na / A
the same value is found for both types of jointing.

The joint frequency Na should, therefore, be adjusted for the lengths of the joints if they are shorter
than the length of the observation plane, expressed as
Na = Σ(nai × Li / A ) eq. (A3-32a)

where na = the number of joints with length L and

A = the area of the observation plane.

Na varies with the orientation of the observation plane with respect to the attitude of the joints.
Recording of Na in several surfaces of various orientation gives a more accurate measure of the
jointing. Being an average measurement, Na should be measured in selected areas with the same
type of jointing. A larger area should be divided into smaller, representative areas containing similar
jointing, and the variation in jointing for the whole area calculated from these registrations.

The correlation between 2-D registrations of the joint density in a rock surface and 3-D frequency
values can, as shown in Appendix 4, be done using the empirical expression
Jv = Na × ka eq. (A3-32b)

where Na = number of joints per metre measured in a surface, and

ka = correlation factor shown in Fig. A3-25c.
 A3 - 49

As seen in Fig. A3-25c, ka varies mainly between 1 and 2.5 with an average value
ka = 1.5. It has its highest value where the observation plane is parallel to the main joint set.

35
Jv = 1.5Na
Jv = 2.5 Na

30
Volumetric joint count (Jv)

25

20

15
Jv = Na
10

0
0 5 10 15 20 25 30
2-D joint frequency (Na)

Fig. A3-25c Variation of the correlation Jv = ka × Na for various types of blocks and orientations of the observation
surface.

3.6.2 1-D jointing frequency measurements along a scanline or drill core

This is a record of joint frequency in a drill core or scan line given as the number of joints
intersecting a certain length (Ll ).

The 1-D joint frequency is an average measure along the selected length of the core. As in other
core logging methods and in surface observations it is important to select a section of the line or
core length, which shows similar jointing frequency. At the start of the logging it is rational to
divide the length into such sections of uniform or similar frequency.
35
Jv = 6Nl Jv = 2Nl
30
Volumetric joint count (Jv)

25

20

15
Jv = 1.25Nl
10

0
0 5 10 15 20 25
1-D joint frequency (Nl)

Fig. A3-26 Variation of the correlation Jv = kl × Nl for various types of blocks and orientations of the observation
surface.
 A3 - 50

The 'joint frequency' can, as mentioned, be very inaccurate if it is not strictly defined what is
included in the registration, and it should, therefore, be accompanied by additional information what
it covers.

The correlation between 1-D joint frequency registrations in drill cores and volumetric 3-D
frequency (Jv) can be done using a similar expression as eq. (A3-32)
Jv = Nl × kl eq. (A3-33)

where Nl = 1-D joint frequency, i.e. the number of joints per metre along a core or line, and
kl = correlation factor shown in Fig. A3-26 with an average kl = 2. As expected there is a
rather poor correlation between kl and Jv .

The joint spacing registrations presented in the following are similar to the joint frequency
measurements.

3.7 Joint spacing registrations

The terms joint spacing and average joint spacing are often used in the description of rock masses.
Joint spacing is the distance between individual joints within a joint set. Where more than one set
occurs this measurement for surface observations is often the average of the spacings for these sets.

However, when the recordings are made on drill cores the spacing is often the average length of
core bits. 4 Thus, the spacings or frequencies are not true recordings as joints of different sets are
included in the measurement. In addition, random joints, which do not necessarily belong to any
joint set, influence. As the term 'joint spacing' does not indicate what it includes, it is frequently
difficult to determine whether a 'joint spacing' referred to in the literature represents the true joint
spacing. Thus, there is often much confusion related to the use of joint spacing recordings.

TABLE A3-30 CLASSIFICATION OF JOINT SPACING

(from ISRM (1978)
TERM SPACING (S)
extremely close spacing < 20 mm
very close 20 - 60 mm
close 60 - 200 mm
moderate 0.2 - 0.6 m
wide 0.6 - 2 m
very wide 2-6m
extremely wide >6m

As joint spacing (S) is the inverse of joint frequency, the correlation factor between them for finding
Jv is
ca = 1/ka for 2-D observations on rock surfaces (average ca = approx. 0.67), and
cl = 1/kl for 1-D observations of scanlines or drill cores (average cl = approx. 0.5).

Deere et al.(1969) have experienced that where several joint sets are present, the resulting average
'volumetric spacing' is generally 2/3 to 1/3 of the average joint spacing of any of the joint sets. For
correlation purposes they consider it sufficiently accurate to use a ratio of 1/2. This is the same as

4
Joint or fracture intercept is the appropriate term for measurement of the distance between joints along
a line or bore hole.
 A3 - 51

has been found for average value of (cl) above. It must be realised, however, that this ratio may be
erroneous. If, for example, there is only one dominant joint system, the ratio would be closer to
unity. The ratio also depends on the orientation of the bore hole or observation surface relative to
the direction of the joints.

Franklin et al. (1971) have suggested to record a direct measure of joint spacing by using the
fracture spacing index (I f ), which refers to the average size of cored material within a recognisable
geological unit. When few joints are present in the core, the I f is the unit length divided by the
number of fractures within the unit. If the core is very broken, the I f is the average diameter of a
number of separate rock fragments. The latter can be compared with direct estimates of block
volume made on small fragments in drill cores which has been described in Section 3.2.2. Franklin
et al. (1971) and Hudson and Priest (1983) recommend two or more inclined bore holes in different
directions to obtain an accurate estimation of the fracture spacing index.

The use of weighted joint density measurements will generally improve the characterization of
block size, also where only results from a single bore hole is available. This method may positively
reduce the amount of drill holes in a site where measurement of joint density or block size is a main
reason in the investigation.

3.8 Weighted joint density measurements (wJd)

R. Terzaghi (1965) points out that the accuracy of jointing measurement can be increased by
replacing the number of joints measured in a surface or borehole, Nα, intersected at an angle α, by a
value N90. N90 represents the number of joints with the same orientation which would have been
observed at an intersection angle of 90o. This is expressed as
N90 = Nα/sinα eq. (A3-34)

Terzaghi stresses the problem of correcting for small values of α, because, in these cases, the
number of intersections will be significantly affected by local variations in spacing and continuity.
"Further, no correction whatsoever can be applied if α is zero. Hence N90 would fail to correctly
indicate the abundance of horizontal and gently dipping joints in a horizontal observation surface."
The method for weighted joint density measurement presented in the following is based on
measuring the angle between the joint and the observation surface or borehole. To solve the
problem of small intersection angles and to simplify the observations, the angles have been divided
into intervals as shown in Table A3-31. For 2-D measurements (surface observations) the weighted
joint density is defined as
wJd = Σ(1/sinδi) / A = Σ(fi )/ A eq. (A3-35)
and, similarly, for 1-D registrations along a scan line or drill core
wJd = Σ(1/sinδi) /L = Σ(fi ) / L eq. (A3-36)

Here δi is the angle between the observation plane (surface) and the individual joint.
A is the size of the area in m2, see Fig. A3-27
L is the length of the measured section along core or line, see Fig. A3-27
fi is the interval factor (1/sinδi) given in Table A3-31; its ratings have been
determined in Appendix 4.
 A3 - 52

1 - D borehole
measurement
2 - D surface measurement
drill core

Σ
1 1
wJd =
A sinδ i

Σ
1 1
wJd =
L sinδ i

Fig. A3-27 The intersection between joints and a drill core hole (left) and a surface (right).

TABLE A3-31 SELECTED INTERVALS OF THE ANGLE δi AND THE CORRESPONDING

FACTOR fi (= 1/sinδi ).
angle δi factor fi
> 60o 1
31 - 60o 1.5
16 - 30o 3.5
< 16o 6

In practice, each joint is multiplied by the (fi ) value for its actual angle interval.
It should be possible to quickly determine the intervals in Table A3-31 for the angle δi by the eye
after some training. The intervals chosen discards as mentioned the strong influence of the smallest
angles, i.e. angles parallel or nearly parallel to the observation plane or bore hole.

The weighted joint density method reduces the inaccuracy caused by the orientation of the
observation surface or bore hole with respect to the individual joints. Hence it leads to a better
characterization of the density of joints and may lead to reduced amount of bore holes.

3.8.1 Correlation between wJd and Jv

The weighted joint density is approximately equal to the volumetric joint count
(wJd = Jv) as is seen in Fig. A4-6 to A4-8 in Appendix 4.
 A3 - 53

3.9 Use of refraction seismic measurements to assess block volume

The use and possibilities to apply refraction seismic measurements are described in Appendix 5.
Refraction seismic is probably the geophysical method most closely related to rock mass properties
because the seismic velocity varies with many of the parameters in the rock mass. The results from
such measurements may therefore assist in the acquisition of geo-data.

3.9.1 Influence from the intact rock and the in situ conditions

Seismic refraction measurement in the field utilizes the propagation of compression or primary (P)
waves as these are the easiest to detect. Velocities of longitudinal waves vary considerably with the
type of rock materials involved, as shown in Table A3-32. The presence of pores, cracks and flaws
as in less compact and unconsolidated rocks, highly reduces the velocity, contrary to moisture,
which increases the velocity.

TABLE A3-32 AVERAGE VELOCITIES (km/s) OF PROPAGATION OF LONGITUDINAL WAVES FOR

SOME TYPICAL ROCKS AND SOILS (partly after Lama and Vutukuri, 1978)

Compact rocks Less compact rocks Unconsolidated materials

Dunite 7 Limestone 4 Alluvium 1

Diabase 6.5 Slate and shale 4 Loam 1
Gabbro 6.5 Sandstone 3 Sand 1
Dolomite 5.5 Loess 0.5
Granite 5

In anisotropic rocks the seismic velocity parallel to the layers V║ is always greater than the velocity
perpendicular to the layers V┴. The coefficients of anisotropy (defined as the ratio of the velocity
along and across the layers) for various rocks vary between 1.0 and more than 4, see Table A3-11.
Increase of pressure on rock reduces the effect of anisotropy.

The seismic velocities of rocks generally increase with increasing pressure. A generally rapid
increase in velocity at low pressures is due to a decrease in porosity from closing of cracks and
defects, and an increase in the mechanical contact between the grains.

In addition to the influence from the inherent rock properties the in situ seismic velocities are
mainly influenced by:
- the stresses acting;
- the block size (degree of jointing); and
- the opening (aperture) and possible filling of joints.

As for the rocks there is often a rapid velocity increase in rock masses with pressure increase at low
pressures due to a closing of the joints. In and near the earth's surface with low stress level the joints
are generally more or less open. Here, the jointing often has a strong impact on the seismic velocity.
With an increase in seismic velocities by depth from increased stress level, a direct comparisons of
seismic velocities in the surface and in a tunnel below can generally not be made.

Although there is generally a clear correlation between jointing and seismic velocities the latter also
include the averaged effect of the factors mentioned above as further dealt with later in this section.
 A3 - 54

3.9.2 Methods to asses the degree of jointing from in situ seismic velocities

Correlations between longitudinal seismic velocity and block size may be applied before or after
information on jointing from core drilling or surface mapping is available. In this way it is possible
to obtain information of the actual jointing at an early stage during investigations. It should be
noticed, however, that in these calculations local differences such as the composition of rock types,
mineral content, etc. are averaged.
Two different methods have been shown in Appendix 5 to evaluate the degree of jointing (or block
size) from seismic velocities measured in the field:
A. When no information is available on jointing at the site.
B. A minimum of two connections between jointing (jointing density and joint condition) and
seismic velocities are known.

A. The connection between jointing and seismic velocity is not known.

As the distribution of joints generally is exponential (see Table A1-4 in Appendix 1) the following
expression has been found to cover the 1-D joint frequency using data given by Sjögren et al.
(1979) and Sjögren (1984, 1993):
Nl = 3(v/Vo ) - Vo/2 eq. (A3-37)

Here Vo is the basic seismic velocity (km/s) for intact rock under the same stress level as in the
field, and
v is the measured in situ seismic velocity (km/s).

Important for the result is the magnitude and accuracy of Vo . Where Vo is not known, it is
recommended to use the velocity for intact rock under the same conditions as in the field (wet/dry,
same direction relative to the stresses, possible anisotropy, etc.).

From eq. (A3-37) the volumetric joint count (Jv) and the block volume may be calculated applying
eq. (A3-33) and the block volume from Vbo ≈ β × Jv-3 . If the block shape factor is not known, β =
approx. 40 may be applied. Joint openness and possible fillings may, however, highly disturb the
accuracy of this using input of V0 estimated from laboratory measurements or from standard
tables. Therefore, the method described in the following gives more accurate results as it includes
the site-dependent conditions.

B. Two or more correlations exist between jointing and seismic velocities

Sjögren et al. (1979) have presented a method to calculate the degree of jointing from longitudinal
sonic velocities. The method is based on known data on jointing and seismic velocity for two
different locations on a seismic profile. The degree of jointing given as joints/m is found from the
following expression:
Nl = (Vn - v)/(Vn × v × ks) eq. (A3-38)

where Vn is the maximum or 'natural' velocity in crack- and joint-free rock under the same stress
level as in the field. The velocities for some fresh rocks measured in the laboratory are
shown in Table A3-33. 5
v is the in situ seismic velocity recorded, and
ks is a constant representing the actual in situ conditions.

5
The difference between Vn and Vo is shown in Fig. A5-7 in Appendix 5.
 A3 - 55

TABLE A3-33 TYPICAL SEISMIC VELOCITY VALUES OF FRESH ROCKS, FREE

FROM CRACKS AND PORES (after Goodman, 1989).

Rock Vn (km/s) Rock Vn (km/s)

Gabbro 7 Basalt 6.5 - 7

Limestone 6 - 6.5 Dolomite 6.5 - 7
Sandstone and quartzite 6 Granitic rocks 5.5 - 6

It is, however, generally seldom possible to measure Vn at the surface as the (often weathered)
rocks in the surface seldom are free from joints, cracks and pores. Therefore, Vn is better found
indirectly using two data sets of measured values of jointing (Nl) and the corresponding seismic
velocity (v). From these data, it is possible to calculate the natural or maximum velocity
v1 × v2 ( Nl2 - Nl1)
Vn = eq. (A3-39)
Nl2 × v2 - Nl1 × v1

and the factor representing in situ conditions

1 1 1
ks = ( - ) eq. (A3-40)
Nl1 v1 Vn

Here Nl1, v1 and Nl 2, v2 are corresponding values of joints/m and in situ seismic velocity
respectively for the two pairs of measurements.

After Vn and ks are determined they are applied in eq. (A3-38) which can be used to work out a
curve representing the actual connection between the measured jointing density and the sonic
velocities can be established to quicker 'translate' velocities into block size (or jointing density).

From eq. (A3-38) the volumetric joint count (Jv) may roughly be calculated as described for alt. A.
Thus at the stage, when the joint frequency has been measured in drill cores or from observations in
rock exposures, the accuracy of the seismic velocity versus jointing can be significantly improved.
This can be used to improve the interpretation of jointing as the information collected in the very
limited volume of the rock mass covered by the borehole, can be extended to cover the whole
seismic profile.

3.9.3 Possible errors and limitations applying seismic velocities for jointing assessments

There are several limitations when using of seismic refraction velocities in rock mass quality
assessments. These mainly stem from the fact that there are several factors in a rock mass that
influence the seismic velocity. It is impossible to avoid errors and uncertainties when the velocity is
used to assess only one or a few of these.

Seismic velocities cannot be used to assess the joint condition (roughness and alteration of the joint
surface; filling and size of the joint). The probable effect of these features should be assessed in the
geological interpretation. A good knowledge of the geological conditions linked with practical
experience may possibly reduce these limitations.

The increased pressure by depth reduces the possibilities of the refraction seismic measurements to
effectively express variations in the degree of jointing. Therefore, refraction seismic measurements
give the best results near the surface where the stress level is low.
 A3 - 56

3.10 Summary of the correlations to determine the block size

The vast variations in jointing densities and block sizes as well as in jointing patterns cause that it is
very difficult to outline one or two methods for characterization which cover the whole range of
variations. Sometimes the type of characterization chosen must be accompanied by a second one to
fully include the conditions in the actual rock mass.

The variation range for three different methods for jointing density characterization is shown in Fig.
A3-28. RQD covers only a limited area compared to the volumetric joint count (Jv) and block
volume (Vb). As mentioned in Chapter 9 RQD is not a single measure for jointing density or block
size, but is also including core loss and possible weathering (Bieniawski, 1984, 1988). The often
poor quality of this method to characterize the degree of jointing is outlined in Appendix 4.

R O C K Q U A L I TY
0 30 60 90 100 D E SI G N AT IO N ( R Q D )

VO L U M ET R I C
100 50 20 10 5 2 1 0.5 0.2 joints/m
3 JOINT COUN T

BLOCK
1 10 0.1 1 10 0.1 1 10 100 1000 10 000 10 0 000 VO L U M E
3 3 3
cm dm m

Fig. A3-28 A rough correlation between three methods for joint density measurements

T YP E OF RE GI ST RAT IO N
PARAMETERS WHICH ARE
POSSIBLE TO MEASURE
S UR FA C E OBS E RVATI ON S

C OM MO N JO IN T
jR jA ( jL) β S 7
D ESCR IPT IO N 1
VOLUMETRIC JOINT
COUNT (Jv) 8
2
2- D JO IN TI NG D ENS IT Y wJd jR jA ( jL ) β Na 9
O BS ERVATIO N

D IREC T BL O C K VO L U ME
wJd jA ( jR ) ( jL ) β Vb
BLOCK VOLUME Vb

10
O BS ERVATIO N

R OC K Q UA LI TY
wJd RQD ( jR ) jA ma 11
D ESIG NAT IO N (R Q D)
DRILL CORE LOGGI NG

3 VOLUMETRIC JOINT
COUNT (Jv) 8
4
JO IN T F REQ U ENC Y ( jR ) jA ma Nl 12
R EG IST RATI O N

VO L UM E O F SMA L L
( jA ) ( jR ) ( β) Vb 13
C OR E F R AG ME NT S

JO IN T SPA CI NG 14
Fi ( jR ) jA ma
R EG IST RATI O N
5
VOLUMETRIC JOINT
8
R EGIS TRATION
GEOPH YS ICA L

COUNT (Jv)
6
R EFR AC T IO N SEISM IC
Nl 15
MEA SU REM ENT

A B B R E V I AT I O N S
jA = joint alt er ation f act or S = joint spacing appropriate equat ion number
jR = joint r oughness f act or Vb = block volume
jL = joint s ize fact or Na = number of joints per m2 ( ) c annot alway s be obs er ved
w Jd = w eight ed joint ing densit y Nl = number of joints per m
β = block shape factor Fi = fracture interscept (joint spacing)

Fig. A3-29 Correlations between various jointing density characterizations. The numbers refer to the equations shown
in Table A3-34.
 A3 - 57

Joint spacing and joint frequency or fracture frequency should not be applied as a single measures
for jointing density characterization unless it is clearly defined what they cover and how they have
been applied.

The connections between the main types of jointing density, degree of jointing and block volume
measurements are indicated in Fig. A3-29.

TABLE A3-34 TRANSITIONS BETWEEN VARIOUS TYPES OF JOINTING DENSITY REGISTRATIONS

MEASUREMENTS or OBSERVATIONS MADE ON REGISTRATIONS MADE ON
SURFACES DRILL CORES or SCANLINES
3-Dimensional 2-Dimensional 1-Dimensional
From measured spacings for From the average joint spacing (Sm): From the average length of core pieces
each joint set (S): Jv = Sm / ka [2] (fracture intercept, Fi):
Jv = Σ (1/Si ) [1] Jv = Fi / kl [5]
or
where also random joints (Nr) From the average joint frequency From the average joint frequency
occur: (Na): registration (Nl):
Jv = Σ (1(Si ) +Σ (Nr/5) [1] Jv = Na × ka [2] Jv = Nl × kl [4]

From weighted 2-D joint density From weighted 1-D joint density
measurement (wJd): measurement (wJd):
Jv = wJd [2] Jv = wJd [4]

From refraction seismic velocity (v): From rock quality designation (RQD):
Jv = 3 kl (v/V0) - 0.5 Vo [6] Jv = 35 - RQD/ 3.3 [3]

From the volumetric joint From weighted 2-D joint density From weighted 1-D joint density
count (Jv): measurement (wJd): measurement (wJd):
Vbo = β × Jv -3 [8] Vbo = β × wJd - 3 [9] Vbo = β × wJd - 3 [12]

From joint spacings (S1, S2, From average joint frequency (Na): From average joint frequency (Nl) or
S3) in 3 joint sets: Vb ≈ β (Na × ka) – 3 [9] spacing (Nl = 1/S) registration:
Vbo = S1 × S2 × S3 [7] Vb ≈ β (Nl × kl) 3 [12] [14]

From registration on site: From registration on site: From registration in drill cores (only
Vb = volume of block Vb = volume of block estimated pieces of core diam. or less size):
measured in situ [10] from jointing pattern [9] Vb = volume of small
core fragments [13]

From refraction seismic velocity (v): From rock quality designation (RQD):
Vbo≈ 0.04β × kl - 3 (v/Vo ) 1.5 Vo [15] Vbo ≈ β (35 - RQD/3.3) - 3 [11]

Comments:
Vbo = block volume where joints or block faces intersect at right angles. For intersections at other angles the
volume can be found from: Vb = Vbo / (sinγ1 × sinγ2 × sinγ3)
(γ1, γ2, and γ3 = angles between the joint sets or between the block faces)

Nr = Σ(nri ⋅ Lri ) / A where nr = the number of joints with length Lr, and
A = the area of the observation surface

β = block shape factor; it may be estimated from β = 20 + 7 S3/S1

(S3 and S1 are longest and shortest block dimension)
β = 27 - 35 for equidimensional (compact) blocks, β = 35 - 50 for slightly long or slightly flat blocks,
β = 50 - 150 for most long and flat blocks, β = 150 - 500 for very long or very flat blocks.
Values of the correlation coefficients in the joint frequency equations:
ka = 1 - 2.5 (average ka = 1.5) kl = 1 - 7 (average kl = 2)

Approximate basic velocity (km/s) of intact rock:

Vo = 7 for gabbo, Vo = 6 - 6.5 for limestone, Vo = 5.5 - 6 for granite, Vo = 6 for basalt and dolomite,
Vo = 6 for sandstone and quartzite
 A3 - 58

4 METHODS TO CHARACTERIZE THE TYPE AND SHAPE OF ROCK BLOCKS

"Incomplete geological information on time is worth far more than complete information
after decisions have been made and carried out."
Burwell and Roberts (1950)

The pattern of joints in the volumes of rocks in the earth's crust occurs as lines in a surface plane
where number of joint sets, the size (lengths), the relative differences in spacings, and the angles
between them present the main characteristics, refer to Appendix 1. In this work, the jointing pattern
in a rock volume is expressed as the type and shape of the rock block delineated by the joint planes
shown in Fig. A3-30.

S3

S1
S3
UNIT BLOCK

JOINT SET NO. 1

JOINT SET NO. 2
S1 JOINT SET NO. 3

Fig. A3-30 Rock blocks delimited by joint sets (from Palmström, 1982).

As described in Appendix 1, Section 2.3 and 2.4 the block type and shape is determined by
- the number of joint sets;
- the difference in joint spacings; and
- the angles between the joints or joint sets;
while the block volume depends mainly on:
- the spacings;
- the number of joint sets; and
- the angles between joints or joint sets.

The block volume Vb can be found from the joints spacings or, as shown in Section 3.5.1, from the
volumetric joint count and the block shape factor β. Methods to determine the block shape factor
are described in this section.

It is mentioned in Appendix 1, Sections 2.3 and 2.4 that the types of blocks delineated by joints
have been characterized in different ways and by different terms. Where relatively regular jointing
exists and extensive joint surveys have been carried out, it may be possible to give adequate
characterization of the jointing pattern according to the system presented by Dearman (1991), see
Section 2.3 and 2.4 in Appendix 1. In most cases, however, there is not a regular jointing pattern,
therefore a rougher type of block shape characterization is generally more practical, similar to that
presented by Sen and Eissa (1991), who divided the types of blocks into the three main groups
shown in Table A3-36 in which also the terms applied in this work are indicated.

The expression for the block shape factor given in eq. (A3-28) is in Section 3.5 is
 A3 - 59

( α 2 + α 2 × α 3 + α 3)
3

β=
( α 2 × α 3)
2

where α2 = S2/S1 and α3 = S3/S1 are the ratios between the two longest sides and the shortest
block face (S3 and S2 > S1, being the length of the faces). The connection between α2, α3 and
β is shown in Fig. A3-31 where also the division into various block types is shown.

TABLE A3-36 TERMS USED TO CHARACTERIZE THE MAIN TYPES OF BLOCKS

Common terms used for block type Terms used in this work
Equidimensional, cubical or blocky blocks Compact blocks
Elongated, long, columnar or bar blocks Long blocks
Tabular, platy or flat blocks Flat blocks
'Long & flat' blocks (a combination of platy and long blocks)

50

20
40

00
10
30

00
75
0
20
50
0

15
30
smid / smin

0
20
0

10
15
0

8
12
5

7
10
0

6
75

5
60

4
50
45

3
40
38 6
3
34
32

2
30
29

1.5
28
27
.5
27

1
1 1.5 2 3 4 5 6 7 8 10 15 20 30 40 50 60 70 80 100
slightly moderately very extremely

COMPACT BLOCKS
smax / smin
Fig. A3-31 Block types characterized by the block shape factor, β, found from the ratio between the longest and
shortest side or joint spacing. The data are based on block shapes at right angles. Refer to Section 2 in
Appendix 4.

TABLE A3-37 DEFINITION OF BLOCK TYPES, REFER TO FIG. A3-31

RATIO OR DIFFERENCE TYPE OF BLOCK
IN BLOCK LENGTHS
α2 ≤ 2 and α3 < 2 Compact or blocky block
α2 ≤ 2 and α3 > 2 Long block
α2 > {( α3 - 1)½ + 1} Flat block
2 < α2 ≤ {( α3 - 1)½ + 1} Long & flat block
 A3 - 60

The type of block is mainly determined by the difference in dimensions between the block faces.
For β = 27 - 32 the block term 'compact' is introduced; this term has been chosen to include cubical,
equidimensional, blocky and other existing terms for blocks not being elongated or flat. The
division chosen for block types is presented in Table A3-37.

The block shape factor β, described in Section 3.5 is used to further characterize the shape of the
different block types according to Table A3-38.

TABLE A3-38 CLASSIFICATION OF THE BLOCK SHAPE FACTOR, β.

VALUE OF β TERM (AND BLOCK TYPE)

32 - 50 slightly (long or flat block)

50 - 100 moderately (long or flat block)
100 - 500 very (long or flat block)
> 500 extremely (long or flat block)

The division into the block types above does not, however, include the impact of the angles between
joints or joint sets. As shown in Section 3.5 they also influence on the block volume, refer to eq.
(A3-26). As shown in Fig. A3-32 the angles between the block faces (or joint sets) delineate:
- right-angled or prismatic blocks;
- rhombohedral blocks; or
- obtuse-angled blocks (where more than 3 joint sets occur).

The value of β can be found using eq. (A3-28 ) or from Fig. A3-31 provided that the block is
limited by 3 parallel pairs of planes for example 3 joint sets. This requires that all the (three)
spacings or the dimensions of the (six) block faces are known. As blocks often have more than six
faces, it can be difficult to find β from eq. (A3-28). Therefore, a more practical method to estimate
β has been developed, as earlier mentioned in eq. (A3-29), from measurement of the longest (S3)
and shortest (S1) dimension of the block
β = 20 + 7 S3/S1 = 20 + 7 α3 This is also shown in Appendix 4.

Figs. A4-7 and A4-8 in Appendix 4 show that the shape factor of most types of blocks with β <
approx. 1000 can be found from this expression within reasonable accuracy (± 25%). For very -
extremely flat blocks eq. (A3-42) should be limited to β < approx. 100. See also Section 3.2.3.

TABLE A3-39 COMMON CONNECTIONS BETWEEN JOINT SETS, TYPES OF BLOCKS AND VALUES OF β
Assumed common
Number of joint sets Block shape Type of blocks range of β

One joint set only very - extremely flat blocks 100 - 5000
One set plus random moderately - very flat blocks 75 - 300

Two joint sets very - extremely long or flat blocks 75 -500

Two sets plus random moderately - very long or flat blocks 50 - 200

Three joint sets*)

Three sets plus random*) compact blocks to moderately long / flat blocks 27 - 75

Four or more sets*)

*)
Where there is a significant difference in spacing between the joint sets, very flat and very long blocks can occur also
for three or more sets. In these cases the values can be β = 100 - 1000.
 A3 - 61

Input of a factor for the number of joint sets (Jn) is applied in the Q-system to partly represent block
size. In Table A3-39 approximate values of β is correlated to the number of joint sets.

The influence on block volume from the angles between three joint sets can be roughly indicated as:
all angles 90o, volume = 100 %
two angles 90o, one 60o 116 %
one angle 90o, two 60o 130 %
all angles 60o 150 %
all angles 45o 280 %

Thus, the effect of joint intersection is relatively limited compared with the variations in
spacings.

Fig. A3-32 Examples of prismatic, rhombohedral and obtuse-angled blocks (Selmer-Olsen, 1964)

5 "TRANSLATION" OF QUALITATIVE DESCRIPTIONS INTO

NUMERICAL VALUES

The previous sections have shown various methods to find numerical values which can be applied
to find the rock mass index, RMi. In addition to improved communication a verbal description of a
rock mass is generally useful in rock engineering and construction when it contains information on
the material and its possible defects. Sometimes, the description of rock masses is qualitative
without numerical values. In such cases it is important that numerical values can be estimated. This
section briefly outlines some possibilities and problems connected to such 'translations'.

The more detailed and complete the description is, the better quality of the numerical values may be
found, especially when defined terms are used. To find the parameters in RMi the description
should contain information on:
- the rock material;
- the block volume, degree of jointing, or joint density; and
- the joint characteristics (roughness, alteration, length).

Weakness zones and faults, should as earlier mentioned, be described and characterized individually
as they often form separate structural regions. Also singularities, i.e. seams and filled joints may
often be described and evaluated individually as separate features when their characteristics are
recorded (ISRM, 1978).
 A3 - 62

Tables and figures covering most parameters in a complete rock mass description are
presented in this work. They give definitions, terms and numerical values useful in the 'translations'.

5.1 Rock material characteristics

The geological name generally ought to be fairly simple, such as could be obtained by examination
of a hand specimen (Burton, 1965). The strength of the material can be described according to the
classification given in Section 1 using terms as 'strong', 'weak', 'hard' etc. Any anisotropic structure
of the rock material such as degree of foliation or schistosity can be characterized using the
classification in Table A3-9.

As weathering and alteration of rocks are likely to have great influence on their properties and
behaviour the description should pay particular attention when such features occur. Terms like
'fresh', 'slightly weathered', etc. defined by to the ISRM (1978) indicate the effect of this feature as
described in Section 1.6.

Rock description terms such as outlook of the intact rock, its colour, possible folding details,
layering, minerals size etc., can not be directly used in the assessment of numerical values. Such
additional information can, however, be of value for better understanding the geological setting and
to improve communication. A complete rock description may contain the following features:

Rock -> Geological name, (type of rock)

material: * Orientation of foliation/bedding/layering
-> Anisotropy, (schistosity, foliation, bedding)
-> Weathering or alteration
-> Strength
additional information of interest:
Folding, colour, appearance, mineral size and texture
Porosity, density

-> used as input in RMi

* included in stability and boring index

5.2 Joint characteristics

According to Price (1966) joints may be classified and described with reference to one or more of a
number of their characteristics, such as shape, size, the way they occur (bedding joints, foliation
joints, etc.), or how they have been formed (cooling joints, tectonic joints, etc.) This tells the reader
a lot of the nature of the joints and the jointing.

The type of joint is often indicated by the size and outlook of the joint (termed parting, joint or
seam), how it occurs (foliation joint, bedding joint, sheet joint etc.) or how it is made/formed
(tension joint, shear joint etc). The type of joint may often indicate several of the other
characteristics, such as size, thickness, partly roughness. Therefore, describing this feature may
improve the understanding of the site conditions.
The various terms for joint condition described in Section 2.1, such as joint roughness, size and
alteration can easily be linked in a description, for example: smooth & planar, clay-coated foliation
joint.
 A3 - 63

Sometimes, when the character or condition of the joints is not given, it can be assumed from the
rock material in which the joints occur, because the development of the joint often is determined by
the properties of the material, refer to Appendix 1. The main elements to be included in a
description are:

Joint > Type of joint (foliation parting, bedding joint, etc.),

condition: -> Roughness (combination of smoothness and waviness)
-> Joint size and continuity
-> Joint alteration, included coating and filling
> Aperture, thickness
* Orientation

-> used as input in RMi

> partly used as input in RMi
* included in stability and boring index

5.3 Block size or density of joints

The blocks formed by the detailed jointing described in terms of blocks/fragments is meant to
express the density of joints and the jointing pattern. It is very important that this description is
carried out in a form that can be 'quantified' into block volume and block shape because this feature
cannot be estimated from other information in a description. By applying defined terms like small,
moderate, large etc. for the block volume or joint density the description can be readily translated
into numerical assessments.

Block type and shape, is a result mainly of the spacings or difference in spacings between the joint
sets, but often also influenced by the angle between the joints or joint sets. These features are not
directly used in RMi, but block shape is an important feature when the block volume is calculated
from joint density measurements.

Much work is required to measure angles between block faces, especially for irregular jointing. This
is the reason why this parameter is seldom used in descriptions, except where more detailed
characterizations are required. Rough terms like prismatic, rhombohedral and obtuse (see Section 4)
requires, however, an understanding of the distribution of joints in three-dimensions. In
consequence, they cannot be found from drill cores alone.

The main elements to be included in describing the jointing density (block size) and pattern are:

Joint density -> Block size (volume or any measure for density of joints)
and pattern: > Block shape
* Block diameter
additional information of interest:
Angles between joints or block faces

-> used as input in RMi

> partly used as input in RMi
* included in stability analysis

Example: Small prismatic blocks => The volume of the blocks are in the range
Vb = 0.2 - 10 dm3. and the block shape factor is probably β = 30 - 40.
 A3 - 64

5.4 Faults and weakness zones

Moderate and large weakness zones should, as suggested by Bieniawski (1984), be described
individually. A complete description useful in assessment of numerical values should, in addition to
location and orientation, contain such features as:

Weakness zone -> Type of zone/seam and structure/composition

or fault, and
singularity (seam):
> Size (thickness)
> Block size / fragment size
> Condition of joints
> Type of gouge or filling material
* Orientation
* Condition of adjacent rock masses

-> used as input in RMi

> partly used as input in RMi
* included in stability analysis

Description of weakness zones is also outlined in Appendix 2, Section 4.

5.5 Examples of numerical values found from qualitative descriptions

Complete descriptions may easily be comprehensive if all parameters and features of importance are
included. In many cases only some of the features applied in RMi are described; in such cases rough
estimates have to be made for those parameters missing. No assessments of the numerical values of
a rock mass can, however, be made unless the description contains
- the block volume, jointing density, or degree of jointing, and
- the type or the strength of the rock.
The joint condition can, however, be roughly estimated based on an assumption of 'normal'
characteristics, i.e. the joint condition factor varies between jC = 1 - 2.

5.5.1 Example 1 (from Gjövik Olympic (underground) Stadium)

Precambrian red and grey gneiss with a composition varying from granitic to quartzdioritic. The
jointing is sometimes irregular, but often two to three joint sets are found, resulting in moderate
degree of jointing. The main joint set occurs along the foliation of the gneiss, with mainly 2 - 5 m
long joints spaced 0.2 - 0.5 m. The joints are smooth to rough and undulating with no filling, but in
a few joints filling of clay, chlorite, silt/sand occurs; also calcite, epidote and quartz is found.
Tectonism has resulted in an additional network of micro-joints sometimes with clay coatings.

'Translation' of the description:

- In Table A3-8 the average uniaxial compressive strength of fresh gneiss is given as
σc = 130 MPa.
- The joint condition factor jC = jL × jR/jA = 3 (mainly slightly rough and undulation joints (in
Table A3-17) gives jR = 3.0; fresh joint character, i.e. jA = 1 (Table A3-20); 2 - 5 m long joints
gives jL = 1 (Table A3-24)).
- Average block volume calculated from spacing S1 = 0.2 - 0.5 m of the main joint set and
assumed spacing 0.5 m and 1 m for the two other sets give Vb = 0. 1 - 0.25 m3. (Moderate degree
of jointing indicates a block volume in the range Vb = 0.01 - 0.2 m3)
 A3 - 65

By applying eq. (4-4) and the values above the jointing parameter is
- JP = 0.2 jC × Vb D = 0.345 Vb0.297 = 0.175 - 0.23
(here D = 0.297 for jC = 3) (JP can also be found from Fig. 4-4.)
Thus RMi = σc × JP = 22.75 - 30.

5.5.2 Example 2

Fresh, amphibolitic gneiss. The main joint set which occurs along the foliation, has a joint spacing
of 0.3 - 1 m with 5 - 20 m long undulating, smooth - rough joint surfaces, often with staining.
Frequently, small joints with strongly undulating and rough, fresh surfaces occur across foliation.
They are generally spaced 0.5 -1 m. The blocks formed are long&flat blocks. They have often
calcite coatings.

'Translation' into numerical values:

- Assumed compressive strength of amphibolitic gneiss is:
σc = σc max /fA = 160/1.5 = 106 MPa (σc and fA have been found Table A3-8 and A3-9)
- The resulting joint condition factor: jC = 2 found from:
Main joint set: jR = 1.5 - 3 (smooth - rough & undulating, see Table A3-17)
jA = 1 (staining only (Table A3-20))
jL = 0.8 (5 - 20 m long, probably continuous, see Table A3-24)
jC = 1.2 - 2.4
Other joints: jR = 4 (rough & strongly undulating)
jA = 3 (joint wall contact, calcite coating)
jL = 2 (small joints)
jC = 2.7
- Block volume Vb = 0.1 - 3 m3 (found from spacing S1 = 0.3 - 1 for the main joint set; S2 =
S3 = 0.5 - 2 m for other joints)
JP = 0.15 - 0.4 is found from Fig. 4-4 and RMi = σc × JP = 30 - 80

5.5.3 Example 3

Palaeozoic mica schist with the main joint set along foliation spaced S = 0.2 - 0.5 m. The 1 - 5 m
long foliation joints are smooth, strongly undulating with thin clay coatings. Random, irregular,
short joints occur with 'random spacing' roughly 3 - 5 m. The jointing result in large, generally flat
blocks.

'Translation':
− Rocks: mica schist, assumed compressive strength σcmax = 100 MPa (assuming that this is
the maximum value). Applying an anisotropy reduction factor of fA = 2 (from Table A3-9) the
compressive strength is σc = 50 MPa.
− The resulting joint condition factor: jC = 1 has been estimated from the following:
- main joint set: jC = 0.75 (smooth & strongly undulating: jR = 2; wall contact, clay
coating: jA = 4; 1- 5 m length: jL = 1.5)
- random joints: jC = 5 (irregular joints: jR = 5, jA = 1; assume jL = 1)
− (The main joint set is considered to have the strongest influence on jC)
− Average block volume Vb = 1 - 10 m3 (estimated from joint spacings).
− With JP = 0.2 - 0.5 (from Fig. 4-4), RMi = 10 - 25
 A3 - 66

5.5.4 Example 4

Medium sized, long blocks of unweathered basalt with mainly rough and slightly undulating faces.

'Translation':
- Basalt may have a compressive strength σc average = 165 MPa.
- Joint condition factor, jC = 3 (rough & slightly undulating surface: jR = 3; assumed fresh joint
walls: jA = 1; assumed medium joint size: jL = 1)
- Medium sized blocks: Vb = 10 - 200 dm3
From this the jointing parameter JP = 0.09 - 0.2 and RMi = 15 - 33

5.5.5 Example 5

Strongly jointed, fresh granite.

'Translation':
- For fresh granite (from Table A3-8) average σc = 160 MPa.
- Strongly jointed means block volume Vb = 0.2 - 10 dm3
- As no information is given on joint condition, an it is assumed rough, planar joints with medium
length as is common in igneous rocks (see Appendix 1). The joint condition factor is then 1.5 - 2
and eq. (4-5) may be applied:
JP = 0.25 3 Vb = 0.015 - 0.054, and RMi = σc × JP = 2.4 - 8.6

5.5.6 Example 6, description of a weakness zone

The weakness zone (with strike/dip = 70o/20o in upstream direction) is approximately 20 m thick
with tectonized, folded, moderately weathered phyllite. It splits along very smooth and planar, 1 - 5
mm thick clay-coated foliation partings (length approx. 0.2 - 2 m) spaced 1 - 5 cm. Several short,
smooth joints cut across the foliation. Some few thin clay seams at an acute angle to schistosity
occur occasionally in the 10 m thick central part of the zone. The volume of the loosened fragments
are in general 50 cm3. There is 3 - 5 m of transition to the surrounding rock masses consisting of
fresh phyllite containing foliation joints spaced 1 - 3 m and some random joints. In the transition
zone the blocks are roughly 1 - 100 dm3. Most joints in the transition part and in the surrounding
rocks are fresh having smooth - slightly rough surfaces with large undulations. The zone may be
termed 'clay containing crushed zone'.

'Translation':
a. The central part:
− The moderately weathered phyllite has an estimated uniaxial compressive strength:
− σc = σc fresh /fA/fW = 50/2/2.5 = 10 MPa (from Table A3-8, A3-9, A3-13)
- The joint condition factor: jC = 0.15 (very smooth & planar, i.e. jR = 0.75;
partly wall contact + clay coatings, i.e. jA = 10; joint
length 0.2 - 2 m gives jL = 2)
Using an average block volume Vb = 50 cm3 the average jointing parameter is:
JP = 0.0003 and RMi = 0.003
(The presence of occasional clay seams may cause an even lower RMi than found above.)
 A3 - 67

b. The transition part:

− Here, the phyllite is fresh, therefore σc = σc fresh /fA = 50/2 = 25 MPa.
- The joint condition factor: jC = 2.5 (smooth - slightly rough & strongly undulating
surfaces, i.e. jR = 2.5; fresh joint walls, i.e. jA = 1;
joint length factor assumed 'normal', i.e. jL = 1)
− 3
With block volume Vb = 1 - 100 dm the jointing parameter is
JP = 0.04 - 0.15 and RMi = 1 - 3.75.

5.6 Summary

A complete rock mass description consists of information on the rock material and the jointing
expressed in defined terms. From such a description numerical values of rocks strength, joint
condition and block volume may be assessed.

TABLE A3-40 TERMS AND FEATURES WHICH MAY BE USEFUL IN A ROCK MASS DESCRIPTION
ROCK MATERIAL
Rock type Outlook Strength Weathering Anisotropy or Orientation
or alteration structure
---------------------- -------------------------- ---------------------- -------------------- ---------------------- ----------------------
Geological name (Colour) Extremely weak Fresh (Bedded) (Strike/Dip)
(Lustre) Very weak Slightly -Schistose
(Texture) Weak Moderately -Foliated
Medium strong Strongly Homogeneous
Strong Completely (Striped)
Very strong (Banded)
Extremely strong (Layered)
(Folded)
JOINT CONDITIONS
Smoothness Waviness Alteration Separation Joint size*) Joint type
---------------------- -------------------------- ---------------------- -------------------- ---------------------- ----------------------
Slickensided Planar Welded or Closed Very small Crack
Polished Slightly undulating Healed Small Small medium Parting
Smooth Strongly undulating Fresh separation Large Joint
Slightly rough Stepped Weathering grade Moderate Very large Seam* *)
Rough Interlocked Type of coating No wall contact Filled joint* *)
Very rough Type of filling *Fracture
Interlocked *Fissure
*)
In addition continuity of the joint (i.e. continuous or discontinuous)
* *)
In addition information of the type of filling material
BLOCK SIZE, JOINTING DENSITY OR DEGREE OF JOINTING
Block volume Term Block type Alternative description instead of block volume:
Joint spacing or frequency described as
--------------------- -------------------------- ---------------------- --------------------------------------------------------------------
Extremely small Slightly Compact or Very small
Very small Moderately Blocky Small
Small Very Long Moderate
Moderate Extremely Flat Large
Large Long&flat Very large
Very large
Extremely large
( ) not included in the definitions and numerical values given in this work
* must be defined
- additional characteristic term (small, moderate, large, etc.) should preferably be given
 A3 - 68

Tables A3-40 and A3-41 show features applied in descriptions, which may be useful in assessing
numerical values of the parameters applied in RMi. Most of the terms, except some terms for rocks,
have been connected to ratings in tables presented in this work.

TABLE A3-41 TERMS AND FEATURES WHICH MAY BE INCLUDED IN DESCRIPTION OF A WEAKNESS
ZONE
WEAKNESS ZONES
Type of zone*) Size (width) Composition
Central part & Transitional part Central part & Transitional part
------------------------------------------------------ ----------------------------------------- --------------------------------------------
- Zone of weak material(s)
- Tension (fault) zone
- Coarse-fragmented crushed zone Includes the dimension Includes the same terms as used for
- Small-fragmented crushed zone (thickness) of the different parts rock material, joint condition and
- Sand-rich crushed zone of the zone. block size in Table A3-40
- Clay-rich crushed zone
- Foliation shear zone
- Altered clay-rich zone
- Altered leached zone
- Recrystallized/cemented/welded zone
*)
Additional information: - orientation of the zone
- condition and composition of surrounding rock masses
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

APPENDIX 4

AN INVESTIGATION OF THE QUALITY OF VARIOUS JOINTING

MEASUREMENTS

"Present estimates of the intact block geometry and discontinuity properties can be
better quantified."
D. Milne, P. Germain and Y. Potvin (1992)

A problem which arises when methods for measuring jointing density is investigated, is the
variations in rock mass jointing and the associated difficulties in obtaining reliable
information on the jointing distribution. This problem can be omitted by applying analysis
where a known distribution of the joints is simulated in a computer. Lines, which represent
boreholes, cut the simulated model of joints, and planes constitute outcrop surfaces in which
the simulated number of joints can be found. In this way, results from different types of 'core
logs' and 'surface observations' can be easily compared. The benefit in this method is that the
block size, block type and shape are known; therefore, reasonable comparisons can be
carried out between the accuracy of the various measurements of the degree of jointing.

1 LAYOUT OF THE INVESTIGATIONS PERFORMED

The simulation was carried out using a computer spreadsheet where the jointing was
represented by three joint sets at right angles to each other. The spacing of the joint sets
could be varied so that different types of blocks, such as compact (or cubical) blocks, long
(or columnar) blocks, flat (or tabular) blocks, and long&flat (prismatic) blocks with varying
shapes were included.

Lines representing boreholes with varying angles to the joint sets were used, and the length
of the (core) pieces delineated by the joints calculated. From this a little more than 300
values of different "core logging" registrations were found. Similarly, for "observation
planes" (outcrops) more than 500 values were used in the comparisons. The following
variables were used:
• Angle between joint sets and borehole: 58o, 30o, 10o; 24o, 40o, 42o; 50o, 40o, 8o;
10o, 18o, 70o; 68o, 11o, 19o; 5o, 80o, 8o .
• Angle between joint sets and observation plane: 60o, 32o, 80o; 22o, 79o, 71o;
50o, 40o, 82o; 80o, 72o, 20o; 66o, 48o, 50o; 85o, 10o, 82o.
• Variation in the block shape factor: β = 27 - 1000 .
• Variation in the joint spacings which resulted in block volumes in the range of:
Vb = 0.0001 - 10 m3 .

The results shown in diagrams include the connection or relations between:

- the volumetric joint count (Jv), and other joint frequency measurements;
- the estimated and the known block shape factor (β); and
- the block volume (Vb) found from other jointing density measurements.

QUALITY\text-1.wpd
 A4 - 2

As the details of jointing are known, the correlations found can be compared directly.

2 BLOCK SHAPE MEASUREMENT

The block shape factor β is, according to eq. (A3-28) in Appendix 3, defined as
( α 2 + α 2 × α 3 + α 3)
3
β=
( α 2 × α 3)
2

β can also be found from Fig. A3-31, provided that the block is delimited by 3 parallel pairs
of faces or 3 joint sets intersecting at right angles. The error introduced by other intersecting
angles between the joint sets is probably small as their influence on the volume, as shown in
Section 3.2 in Appendix 3, is limited.

Eq. (A3-28) requires that all the (three) joint set spacings or the dimensions of the (six)
block faces are known. As blocks may have more or less than six faces, it can be difficult to
find β from this expression. Thus, a simplified measurement of β would be advantageous;
therefore the following expression to estimate the block shape factor has been found by trial
and error:
β = βo + 7(α3 - 1) = 20 + 7 α3 = 20 + 7 a3/a1 eq. (A4-1)

where βo = 27 is the lowest value of β, i.e. for a cubical block.

By this simplification β can roughly be estimated from the ratio between the longest and
shortest side of the block, given as a3/a1 = α3. The quality of eq. (A4-1) has been
investigated, and the results are given in Fig. A4-1.

Fig. A4-1 The correlation between the real and estimated block shape factor (β) found from eq. (A4-1)
for observation made on surfaces.

The figure shows that this simplified block shape factor β can be found with a reasonable
accuracy (± 25%) for most types of blocks with β < 1000, except for very - extremely flat
blocks. For the latter acceptable results are found for β < approx. 100. To also include very
flat and very long blocks eq. (A4-1) can be adjusted by an exponent
β = 20 + 7 a3/a1 = 20 + 7 (a3/a1)(1 + 0.1 log(a3/a1)) eq. (A4-2)
 A4 - 3

The effect of this exponent is significant only where a3/a1 > 10, so for most cases the
simpler expression eq. (A4-1) is sufficient as shown in Fig. A4-2.

Fig. A4-2 The correlation between the real and estimated block shape factor (β) found from eq. (A4-2)
for observation made on surfaces.

The application of this simplified expression for the block shape factor is further dealt with
in Appendix 3, Section 3.

3 2-D AND 1-D JOINT FREQUENCY REGISTRATIONS

The joint frequency can be found from 1-D or 2-D registrations measured in boreholes or
surfaces respectively. (From surface measurements it is also often possible to apply 3-D
registrations in the form of volumetric joint count.

3.1 2-D frequency measurements

As mentioned in Appendix 3 it is seldom defined whether the frequency (or spacing) given
in a report or paper is meant to represent the main joint set or if it represents the mean
frequency for the joint sets in an area. In this thesis, the 2-D joint frequency is defined as the
number of joints per unit length recorded in a surface. It can be found from the number of
joints recorded within a known area, given as Na = n / A where n is the number of joints
observed in the outcrop with an area, A.

The results from a simulation with

- variation in the density of joints (Jv) [or block volumes (Vb)],
- variation in block types and shapes (from variation in the joint spacings), and
- various angles between joints and the observation plane (i.e. outcrop)
is shown in Figs. A4-3 and A4-4.

As expected there is a rather poor correlation between the 'measured' 2-D joint frequency
and the real 3-D density of jointing (given as the volumetric joint count (Jv)) or the real
block volume. The Jv varies between 0.4Na and 1Na with an average value of
Jv ≈ 1.5 Na eq. (A4-3)
 A4 - 4

Fig. A4-3 The connection between measurement of 2-D frequency (Na) and volumetric joint count (Jv).

By applying eq. (A3-27) Vb = β × Jv-3, the block volume is found as

Vb ≈ 0.3β(Na)-3 eq. (A4-4)

(To obtain the best possible registration of the 2-D joint frequency the length of the joints
should be included in the measurement as described in Appendix 3, Section 3.6.1.)

Fig. A4-4 The relation between 2-D joint frequency measurement (Na) and block volume (Vb).

Another very rough correlation is found from Fig. A4-3 given as

Vb ≈ (Na/3)3 eq. (A4-5)
 A4 - 5

The error in these registrations can vary between + 500% and -50%, especially in cases
where very to extremely flat blocks occur.

3.2 1-D frequency measurements

Where the jointing frequency (or joint spacing) is recorded along a borehole or a scanline a
sort of average frequency is found. From a simulation with the same procedure as for 2-D
(surface) measurements the results in Fig. A4-5 and A4-6 show that the error in the jointing
measured by this method is, as expected, larger than for surface observations.

Fig. A4-5 The connection between 1-D joint frequency (Nl) and volumetric joint count (Jv).

Fig. A4-6 The connection between 1-D joint frequency (Nl) and block volume (Vb).

A transition between 1-D frequencies and real block volume or volumetric joint count which
depends on the block type and shape, will mainly vary between
Jv = 2 Nl and Jv = (Nl/2)3 eq. (A4-6)
Consequently, the block volume will vary between
Vb = β × Jv-3 and Vb = 0.13β(Nl)-3 eq. (A4-7)
 A4 - 6

As seen in Figs. A4-5 and A4-6 the error is significant, between +500% and -40%, when the
block volume (Vb) or the volumetric joint count (Jv) is estimated from 1-D measurements.

4 WEIGHTED JOINT DENSITY MEASUREMENTS

The orientation of a joint compared to an observation surface or a borehole influences the

number of joints observed (Franklin et al., 1971; Terzaghi, 1965). Joints perpendicular to
the surface plane will be more frequently intersected and may bias the observations. This
can to a large extent be corrected for by including the angle between the individual joint and
the observation surface or borehole in the measurement, thus obtaining a registration of the
'weighted joint density' (or frequency, given as number of joints per volume unit). This is
defined as
wJd = Σ(1/sinδi)/A for 2-D (surface) observations eq. (A4-8)
and
wJd = Σ(1/sinδi)/L for 1-D (borehole) observations eq. (A4-9)

Here δi is the angle between the observation plane (surface) and the individual joint,
A is the size of the actual area, and
L is the actual length of borehole.

The weighted joint frequency method for joint density registration has earlier been suggested
by R. Terzaghi (1965) who suggested to take into account the orientation of the joints and
the probability for them to be cut by the observation plane or the scan line (or drill hole).

Where accurate registrations have been performed both for the number of joints and the
angles between joint and observation surface or borehole the wJd should be equal to the
volumetric joint count (Jv = wJd).

For joints parallel or nearly parallel to the observation plane the factor (1/sinδ) will give
high contribution to the joint count. A single joint may easily disturb the measurement. For
this reason a common factor connected to an interval for the smallest angles should be
applied.

Where many joints occur with different angles to the observation plane the weighted
recording can be time-consuming. By grouping the angles into intervals with adjacent factor
will make the registration considerably simpler and easier. A factor fi representing 1/sinδ
is introduced for the intervals in Table A4-1.

TABLE A4-1 INTERVALS OF THE ANGLE δ AND THE

FACTOR fi REPRESENTING 1/sinδ
Angle between joint plane Factor
and surface/borehole f i
δ
> 60o 1
31 - 60o 1.5
16 - 30o 3.5
< 16o 6

In this interval method the wJd is found as the sum the actual value of (fi ) for each joint
for its angle interval. The results from a simulation including various orientation of joints
and with respect to observations planes and boreholes is given in Figs. A4-7 and A4-8 where
wJd is compared to the volumetric joint count (Jv).
 A4 - 7

Variations in block size and shape as well as angles between the joints and the borehole or
the observation plane were used in the simulation. A total of 640 registrations were made
including:
- Block shape factor between 27 and 2000
- Block sizes between 0,001 m3 and 1000 m3
- Angles between plane and joints:
10o and 85o for surface planes;
5o and 80o for boreholes.
The same values of fi were applied both for 1-D and 2-D registrations.

4.1 Surface observations

The "observations" were made on surface planes corresponding to 75 m2, where the number
of joints were recorded together with the angle between the plane and the joint. As seen in
Fig. A4-7 there is a good correlation within approximately ± 30%, except for the lower
values of Jv (i.e. large blocks). This is caused by the large spacing of joints and small angles
between joint and observation plane, resulting that the joints in the actual set is not recorded.
Therefore, the Jv will be lower and hence the block volume larger than the real.

Fig. A4-7 Weighted 2-D joint density measurement (wJd) made in 'surfaces' compared to the (known
value of the) volumetric joint count (Jv).

4.2 Borehole logging

In these simulations an observation length of 10 m was chosen. As expected, there is a

poorer correlation for 1-D jointing measurements, - having an inaccuracy between
approximately +35% and - 50% for Jv > 2 , - than for 2-D measurements. (See Fig. A4-8)
Also here, the poorest correlation is for the lowest values of Jv (Jv < 2), and this can
probably be explained from the fact that some of the joints were not encountered in the 10
m long measuring section in the 'boreholes'.

The interval method offers a relatively quick and simple way to measure jointing density as
the intervals chosen for the angle between joint and borehole should be familiar to most
people.
 A4 - 8

Fig. A4-8 Weighted 1-D jointing density measurement (wJd) made in 'boreholes' compared to the (known)
volumetric joint count (Jv).

5 CALCULATIONS OF BLOCK VOLUME FROM SIMPLIFIED JOINTING

MEASUREMENTS

Sections 3 and 5 show methods to measure the block shape factor (β) and the weighted
jointing density. These measurements can be combined to estimate the block volume
Vb = β × wJd - 3 eq. (A4-10)

From the simulations made in Section 4 the connection between the (known) block volume,
and the estimated volume is shown in Figs. A4-9 and A4-10.

Fig. A4-9 Estimated block volume found from weighted 2-D measurements (in surfaces) compared to the
known block volume.
 A4 - 9

Fig. A4-10 Estimated block volume found from weighted 1-D measurements (in boreholes) compared to
the known block volume.

For 2-D registrations the accuracy of the estimated block volume < 3 m3 lies within +100%
and -50%, except for very to extremely flat blocks (β > 100); for 1-D measurements the
inaccuracy is much higher as shown in Tables A4-2 and A4-3.

TABLE A4-2 ACCURACY IN 2-D MEASUREMENTS, OBSERVATION AREA 75 m2

Vb most blocks very to extremely flat blocks
< 3 m3 +100% to -60% +50% to -8%
> 3 m3 +300% to -50% +50% to -15%

TABLE A4-3 ACCURACY IN 1-D MEASUREMENTS, OBSERVATION LENGTH 10 m

Vb most blocks very to extremely flat blocks
< 0,5 m3 +500% to -50% +100% to -8%
> 0,5 m3 +2500% to -50% +1000% to -8%

The relatively poor connections for the larger block sizes stem for a main part from the fact
that some of the joints were not recorded along the observation length (or area) chosen. Also
small angles between the observation plane or borehole and joints will highly bias the
measurements. This is especially true for the 1-D registrations.

6 ROCK QUALITY DESIGNATION (RQD)

In the simulations made to compare RQD with other 1-D jointing density observations a
total of 310 registrations were made for
- various block shape factors between 27o and 58o,
- various block volumes, and
- 9 different angles between borehole and joints varying between 8o and 58o.
All simulations used 3 joint sets at right angles to each other.

In these simulations, also the volumetric joint count (Jv), the weighted jointing density
(wJd), and the estimated block shape factor (β) have been found in order to compare the
results obtained by the measurements.
 A4 - 10

6.1 Connection between the RQD and the volumetric joint count (Jv)

Fig. A4-11 shows that for the same type of jointing (same block type and shape) the RQD
can be both 0 and 100 for jointing within an interval of Jv = 13 - 22. A wide range of RQD
values is also found for Jv = 22 - 40. Therefore, a general and reliable connection between
RQD and Jv can hardly be found. According to Fig. A3-22 and eq. (A3-21) in Appendix 3
the general connection between RQD and Jv is RQD = 115 - 3.3 Jv (Palmström, 1974).

This expression is shown in Fig. A4-12 where it is seen that it gives lower RQD values than
in the simulation. The expression
RQD = 140 - 4 Jv eq. (A4-11)
seems to give a better connection. However, with values of RQD = 0 and 100 for a large
span of jointing, it is uncertain to compare these two correlations. In practice, the jointing is
less uniform and the threshold of t = 0.1 m often measured inaccurately because joints
frequently intersect the core at acute angles. Thus, the 'original' expression eq. (A3-21)
between RQD and Jv may give acceptable results.

Fig. A4-12 Comparison between RQD and Jv

Sen and Eissa (1991) have made theoretical studies of the connection between Jv and RQD.
Fig. A4-13 also shows that there theoretically is a non-linear connection between RQD and
Jv contrary to the 'original', linear eq. (A3-21) of Palmström (1982). According to Sen and
Eissa this expression is acceptable for long (bar) blocks up to Jv = 25 (RQD = 25). For the
high jointing density (Jv > 25) the threshold value has a significant influence on RQD and
accurate measurements of RQD is therefore important. Another important fact is that in such
highly jointed rock there are often difficult to obtain accurate core length measurements as
the dense jointing often affects the drilling process and hence the size of the core bits. This
may lead to inaccurate RQD values for highly jointed rock.

Another fact is that, in order to record RQD values in the range of 1 - 10, long measuring
sections are required. (For instance, only one 10 cm long core bit in a 10 m long measuring
section will give RQD = 1.)
 A4 - 11

100

80

Palm
ström
60
E ( RQD )

t=
40

0.
05
20
0.1
0.20 0.25
0.1

0
5

0
0 20 40 60 80 100
Volumetric joint count, Jv

Fig. A4-13 Connection between RQD and Jv for various thresholds (t) for long (bar) blocks. The eq. (A3-21)
of Palmström (1982) for a threshold of t = 0.1 m is indicated (from Sen and Eissa, 1991).

The main outcome from the investigations in this section is, as mentioned earlier, that it is
important to realize that RQD can vary largely. The way that it is defined and the fact that it
is a one-dimensional measurement and cause that there is generally inaccuracy connected to
RQD.

6.2 Connection between block volume and "block size" expressed as RQD/Jn
in the Q system

Barton et al. (1974) have, in their Q system, modified the RQD by dividing it by a factor
representing the number of joint sets occurring at each location. This quotient (RQD/Jn) is
meant to represents the overall structure of the rock mass as a crude measure of the relative
block size within the two extreme values 200 and 0.5, representing crude but recognisable
approximations of "particle size" (diameter) varying between 2 m and 0.005 m respectively.

Fig. A4-14 Connection between RQD/Jn in the Q-system and block volume, Vb.
 A4 - 12

As is shown in Fig. A4-14 there is a very poor connection between the block size (Vb) and
the ratio RQD/Jn. This is not unexpected, remembering the inability of RQD to give
accurate measurement of the density of joints (and hence the block size).

The factor Jn in the Q system can be considered as a factor representing the block shape as
it is varying with the number of joint sets. Jn can, therefore, crudely be compared with the
block shape factor β. As pointed out by Cecil (1970) the number of joint sets in a rock mass
is an important indication of the degree of freedom of blocks in a rock mass.

8 SUMMARY

The correlation between block volume (Vb) and volumetric joint count (Jv) found in
Appendix 3 have been applied to develop transitions between block volume and the
following joint measurements:
- 1-D frequency observation in boreholes or scanlines;
- 2-D frequency observations in surface areas;
- RQD measures in boreholes; and
- Weighted joint density measurements in boreholes or in surface areas.

Important in the correlation Vb = β × Jv-3 is β, the block shape factor. Where regular
jointing with 3 joint sets occurs, β may easily be found from the spacings. In other cases an
estimate of it can be made from
β = 20 + 7 a3/a1

where a3 is the length of the block, and a1 the thickness of the block.

2-D frequency registrations (in surfaces) characterizes, as expected, generally the degree of
jointing (i.e. block size) better than 1-D measurements (in boreholes and scanlines). The
errors can be as large as (+500%, -50%) especially for very to extremely flat blocks.

The weighted joint density measurements introduced in this work give better accuracy than
the existing 1-D and 2-D registrations. In surface observations the error is in the order
±30%; in boreholes and scanlines +35, -50%.

Another result from the investigation is that it is almost impossible to estimate the block
volume from RQD registrations within reasonable accuracy.

The introduction of the ratio RQD/Jn (Jn is the joint set number in the Q-system) improves
the RQD's possibility to characterize jointing, but still the use of RQD for volume estimates
is highly uncertain. The conclusion drawn is that RQD often expresses unreliable values of
the degree of jointing. RQD measurements should generally be regarded as very crude
measurements, and it is therefore difficult to recommend any better transition from RQD to
block volume than eq. (A3-27) (RQD = 115 - 3.3 Jv) where RQD is the only registration
available on the jointing density.
From Palmstrom A.: RMi - a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

APPENDIX 5

USING REFRACTION SEISMIC VELOCITIES TO CHARACTERIZE

JOINTING

"Judgement is thus the intelligent use of experience or, more cautiously expressed, it is the
recognition of one's limitations of the methods one uses, and of the limitations and uncertainties
of the materials one works with; and this brings us back to geology."
Herbert H. Einstein, 1991

Refraction seismic measurement is the geophysical method most closely related to rock
mass properties because the longitudinal sonic velocity recorded varies with rock
properties as well as jointing, stresses etc. in the rock mass. The results from such seismic
measurements may therefore assist in site selections and in rock engineering.

Seismic refraction measurements have been used in Scandinavia for at least 40 years in
connection with planning of dams, tunnels and portals. The earliest applications were
primarily for the determination of the depths to bedrock beneath soil cover. Since 1959 the
method has also been used successfully for the location of weakness zones, such as shear
zones and faults (Sjögren et al., 1979). Such zones give considerably lower and therefore
easily recognisable velocities in fresh rocks. From the beginning of the 1960s refraction
seismic velocities have also been used as indicator of rock mass quality in fresh igneous
and metamorphic rocks, as shown in Fig. A5-1.

40

SURFA CE
GROUND WATE RIVER
R LEVEL 30.5 800m/s
800m/s (SAND)
30
1600m/s (SAND, saturated) 1600m/s
XX

XX 2400m/s (MORAINE) 2400m/s

XX XX
XX
XX XX XX
20 4000m/s (WEA THERED ROCK) XX
0m /s
XX
350

5500m/s (ROCK)
10 20m 4800m/s (ROCK)
10 3500m/s
(WEAKNESS ZONE)

Fig. A5-1 Typical refraction seismic profile in hard, unweathered rocks with interpretations shown in
brackets (from Broch, 1988).

The seismic refraction measurement utilizes the propagation of elastic waves in the ground.
The compression or primary (P) sonic waves are easiest to detect and have mainly been
used in the measurements. The ratio of the shear (or transverse) and longitudinal sonic
velocities can be used to determine the dynamic moduli of the rock as described by Sjögren
et al. (1979) and several other authors. In this appendix only the use of the longitudinal
sonic velocities are described.
 A5 - 2

The field measurements can be carried out

- on the ground,
- in boreholes,
- on the seabed, or
- just below the sea surface.

In each case the refracted head wave travels parallel to the surface. Velocities are
calculated from the slope in a 'travel time versus distance' graph worked out from the
registrations. The determination of the sonic velocities in the various layers is a complex
process, and a great deal of practical experience is required of the operator before the
results can be regarded reliable.

The usefulness of the seismic exploration technique can be extended through use of cross-
hole techniques between boreholes, as described by Nord et al. (1992).

1 FEATURES INFLUENCING THE MAGNITUDE OF LONGITUDINAL

SONIC VELOCITIES

In the field there are several factors that, in a complex way, may influence the propagation
of sonic velocities. The main contributions stem from:
• The inherent properties and condition of the rock material consisting mainly of:
- rock type (mineral content, texture, density, porosity, anisotropy);
- weathering or alteration of the rock material;
- saturation;
- pressure; and
- temperature.
• The in situ conditions, with the main contributions from:
- distribution of rock types;
- quantity of joints;
- joint openings;
- rock stresses; and
- ground water condition.
The most important of these factors are briefly discussed in the following.

1.1 Factors influencing the sonic velocities in intact rock

Velocities of longitudinal waves vary considerably with the type of rock. A representative
selection of typical longitudinal (compressional) sonic velocities is given in Tables A5-1 to
A5-3.

TABLE A5-1 LONGITUDINAL VELOCITIES OF SOME MINERALS.

(data from Fourmaintraux (1976), presented by Goodman, 1989).
-----------------------------------------------------------------------------------------------------------------------------------------------------------------
Mineral Vm (km/s) Mineral Vm (km/s) Mineral Vm (km/s)

Amphibole 7.20 Epidote 7.45 Orthoclase 5.80

Augite 7.20 Gypsum 5.20 Plagioclase 6.25
Calcite 6.60 Magnetite 7.40 Pyrite 8.00
Dolomite 7.50 Muscovite 5.80 Olivine 8.40
Quartz 6.05
------------------------------------------------------------------------------------------------------------------------------------------------------------------
 A5 - 3

TABLE A5-2 AVERAGE SONIC VELOCITIES FOR INTACT ROCKS (partly from Lama and
Vutukuri, 1978)
----------------------------------------------------------------------------------------------------------------------------------------------------------------------
Compact rocks (km/s) Less compact rocks (km/s) Unconsolidated rocks (km/s)

Dunite 7 Limestone 4 Alluvium 1

Diabase 6.5 Slate and shale 4 Loam 1
Gabbro 6.5 Sandstone 3 Sand 1
Dolomite 5.5 Loess 0.5
Granite 5
----------------------------------------------------------------------------------------------------------------------------------------------------------------------

TABLE A5-3 TYPICAL RANGES OF LONGITUDINAL SONIC VELOCITIES FOR INTACT ROCKS
(from Sjögren, 1984)

0 1 2 3 4 5 6 7
Overburden
Above groundwater
Below groundwater: coarse sand
clay
gravel
moraine
Bedrock
Cemented gravel (Chile)
Lias sandstone (Öresund, Sweden)
Cretaceous limestone (Öresund, Sweden)
Miocene limestone (Libya)
Late Precambrian/Cambrian sandstones
Eocene limestone (Libya)
Cambrian-Silurian limestones
Jotnian sandstone (Lima, Peru)
Caledonian quartzite
Gneiss
Granite
Meta-anorthosite, / gabbro
Diabase (Liberia)

0 1 2 3 4 5 6 7
Seismic velocity in km/s

The density of the rock is one factor, which affects the velocity of longitudinal sonic
waves. In the elastic wave theory the velocity decreases with increasing density. The effect
of density is, however, overruled by the dynamic stiffness of the rock, which generally
varies with the density. Therefore, the sonic velocity seems incorrectly to vary with the
density of the rock as shown in Fig. A5-2.

11 7

13
mean velocity, km/s

6 12 1: Pumic (glass),
2: Vesicular basalt,
9 3: Rhyolite,
2 4 4: Serpentine,
4 8 5: Dacite,
10 3 6: Semiwelded tuff,
5
7: Dunite,
8: Altered rhyolite,
2 1 6
V = 0 .35 + 1.88ρ 9: Granodiorite,
2 10: Vesicular basalt,
r = 0. 83
11: Gabbro,
12: Obsidian,
0 1.0 2 .0 3.0 13: Tholeitic basalt,
3 14: Vesicular basalt
density, g/cm

Fig. A5-2 Mean longitudinal pulse velocity versus density of the rock.
(Data from Bur and Hjelmstad, 1970, compiled by Lama and Vutukuri, 1978).
 A5 - 4

The wetting of porous rocks leads to change of elastic wave velocities in them. The higher
the sound velocity in the pore filling material, the greater is the total velocity in the rock
sample (Lama and Vutukuri, 1978). Since the sonic velocity in water (Vl ) is five times
greater than air, water saturation leads to a rise in the elastic wave velocity in hard rocks.
However, the wave velocity in more porous rocks completely saturated with water is lower
than in slightly porous rocks, because Vl is less than the sound velocity in the mineral
skeleton.

5.50

measurements (σ − ρ) = 420 bars

5. 00 measurements (σ − ρ) = 0

4 .50

4.00
velocity, km/s

500 bars
3.50

3.00

2.50
0

2.00
0 0 .10 0.20 0 .30

porosity, Φ

Fig. A5-3 Longitudinal wave velocity versus porosity for two different effective stresses (from Lama
and Vutukuri, 1978)

TABLE A5-4 ANISOTROPY COEFFICIENTS FOR SOME ROCKS

(from Lama and Vutukuri, 1978)
Anisotropy Anisotropy
ROCK coefficient ROCK coefficient
V¦/V-. V¦/V-.
-------------------------------------------------------------------------------------------------------------
Austin chalk 1.17 Anhydrites 1.12 - 1.16
Limestones 1.04 - 1.30 Marl 1.10
Salt no anisotropy Sandstones 1.0 - 1.19
Shales 1.07 - 1.40 Gneisses 1.20 - 1.27
Mica schist 1.36 Granodiorite 1.33
Serpentine 1.18

In anisotropic rocks the velocity parallel to the layers (V¦) is always greater than the
velocity perpendicular to the layers (V-.). The coefficient of anisotropy (defined as the ratio
of the velocity along and across the layers) in Table A5-4 vary between 1.0 and 1.40.
Measurements performed by Bergh-Christensen (1968) on schists and phyllites show that
this coefficient can have considerably higher values. Increasing the pressure on rock
reduces the effect of anisotropy.

Usually, the velocity of longitudinal waves falls with rise in temperature. Below 400oC and
at atmospheric pressure, the temperature coefficient of variation of velocity has the values
given in Table A5-5.

The presence of pores, cracks and flaws highly influences the sonic velocity in the rock. At
low pressure the pulses are transmitted through the water or air-filled gaps and the velocity
is significantly affected by the gaps and voids. However, at high pressure the content of
 A5 - 5

gaps becomes less important since, as the cracks are closed; in this situation the increase in
velocity is determined by the rock framework. A generally rapid increase in velocity at low
pressures is due to a decrease in porosity from closing of cracks and defects, leading to an
increase in the mechanical contact between the grains. See Fig. A5-4.

TABLE A5-5 COEFFICIENT OF TEMPERATURE INCREASE FOR SOME ROCKS GIVEN AS %

INCREASE IN THE SEISMIC VELOCITY (from Ide, 1937, Lama and Vutukuri, 1978).
---------------------------------------------------------------------------------------------------------------------------------------------------------------------
ROCK % per 100oC ROCK % per 100oC
Quartzitic sandstone -1 Danby marble -3
Solenhofen limestone -1.2 Sudbury norite -4
French Creek norite -1.6 Vinal Haven diabase -5
---------------------------------------------------------------------------------------------------------------------------------------------------------------------

The velocity increase at higher pressure results from changes in intrinsic properties of the
rock, such as finite compression of the minerals. In some rock specimens, the sonic
velocity is observed to decrease (as in salt) and in others it decreases when pressure
exceeds a certain value (as in some limestones).

.
. .
8.0 .
. . dunite
gabbro

7. 0 . dolomite
velocity, km/s

granite

6 .0 gneiss
. .
.
. sandstone
.
5.0 .

4.0 .
0 4000 8000
pressure, bars

Fig. A5-4 Longitudinal wave velocity versus pressure (from Lama and Vutukuri, 1978, based on data
from Birch, 1960).

1.2 The influence from in situ factors on measured sonic velocities

In addition to the influence from the inherent rock properties mentioned above, Sjögren et
al. (1979) conclude from their investigations that, the in situ longitudinal velocities in
unweathered rock masses are mainly determined by:
- the stresses acting;
- the degree of jointing; and
- the presence of open joints or joints with filling.

Both compression and shear sonic velocities of rocks generally increase with increasing
pressure. There is often a rapid velocity increase at low pressures due to a closing of the
joints. In and near the surface with low stress level the joints are generally more or less
open. Thus, the degree of joints will strongly influence velocity of the waves. This is an
important feature in interpretation of refraction seismic measurements to assess the degree
of jointing.
 A5 - 6

The increase in stress level by depth, therefore, causes increase in longitudinal sonic veloc-
ities. Cecil (1975) mentions that the difference between velocities measured on the ground
surface and in the tunnel 50 - 60 m below is up to 17% for high quality rock and up to 38%
for low quality (highly jointed) rock. Sjögren et al. (1979) have from their investigations
found the same tendency with an increase of 5 - 15% at a depth 30 - 50 m from the rock
surface, and a usually greater increase in low velocity zones.

This trend can be explained by:

1. An overall loosening of rock at the bedrock surface that has occurred after
glacial retreat (as in Scandinavia). Weak zones at the bedrock surface probably are
considerably more relieved and loosened than sounder rock.
2. A tightening effect of joints with depth caused by the increase of ground
stresses.
3. The effect of weathering in upper the 10 - 30 m of the ground (in
Scandinavia). This may have resulted in development of additional joins as well as
opening up of joints in addition to increased weathering along joints.

Thus, it is obvious that direct comparisons of velocities in the surface and in the tunnel
cannot be made. As the ground pressure increases by depth the effect of jointing on sonic
velocities is reduced. This feature reduces the ability of the refraction seismic
measurements to effectively characterize the degree of jointing in deep tunnels.

2 EARLIER METHODS USED TO CHARACTERIZE ROCK MASSES FROM

SEISMIC VELOCITIES

Although there is a clear correlation between jointing and seismic refraction velocities, the
latter also includes the averaged effect of other factors such as rock properties and stress
conditions as further dealt with later in this section.

TABLE A5-6 APPROXIMATE CONNECTIONS BETWEEN REFRACTION SEISMIC VELOCITIES,

ROCK MASS CONDITIONS AND ROCK SUPPORT IN SCANDINAVIAN TUNNELS
(partly based on Sjögren et al., 1979)
In-situ
velocity Probable ground conditions Possible rock support
m/s
------------------ ---------------------------------------------------------------------------------------- --------------------------------------------------
< 3000 Cavities in the bedrock filled with soil, or completely Extensive rock support.
crushed and fragmented rock material in weakness zones.

< 4000 Ground related to faults, contact zones etc. with highly High amount of rock support.
fractured rock.

4000-4400 Strongly - moderately jointed rock masses. Moderate to high amount of rock
support.

4500-5000 Slightly - moderately jointed rock masses. Small to moderate amount of

rock support.

> 5000 Massive rock masses. Generally little need for rock
support.
 A5 - 7

2.1 Connections between jointing and seismic velocities

In Scandinavia, an approximate method to utilize seismic velocities measured in the field

to estimate rock mass quality and tunnel support requirements has been frequently used for
30 years. Cecil (1971) concludes that "It can be said almost without exception that on the
surface of the hard crystalline bedrocks of Sweden seismic velocities of 4000 m/s or less
are indicative of weak zones in the bedrock. Stretches on a profile with such a velocity thus
can be considered to be weak zones without further question, provided the surrounding
higher velocities are greater than 4900 m/s."

An example of a classification often used in Norway is shown in Table A5-6. It should be

noted that this classification is crude and that it is related to unweathered, hard, crystalline
rocks. The method may in many occasions be inaccurate and even wrong.

2.2 Rock quality estimated from the seismic velocity ratio

Some authors have compared in situ velocities of longitudinal waves in the rock mass with
the velocity of intact rock cores tested in the laboratory in order to characterize rock
quality. The 'Seismic Velocity Ratio' introduced by Merritt (1968) is defined as
SVR = Vf /Vr eq. (A5-1)

where Vf is the longitudinal seismic wave velocity measured in the field, and
Vr is the basic sonic wave velocity measured in the laboratory.

For a massive rock mass containing only a few joints, the velocity ratio (Vf /Vr ) should
approach unity; but as the degree of jointing becomes higher, (Vf /Vr ) will be reduced.

The squared seismic velocity ratio named 'Velocity Index' VI = (Vf /Vr )2 of Coon and
Merritt (1970) is comparable to SVR. The ratio has been squared to make the velocity
index (VI) equivalent to the ratio of the dynamic moduli. Table A5-7 illustrates the
relationship between the velocity index, velocity ratio and rock mass quality. Cecil (1971)
did not find that the squaring of the seismic velocity ratio had any advantages over the first
power of seismic velocity ratio (SVR) other than a wider numerical band for the
intermediate support category.

TABLE A5-7 RELATIONSHIP BETWEEN TUNNEL SUPPORT CATEGORY AND SEISMIC

VELOCITY RATIO (based on data from Cecil, 1970 and Bieniawski, 1984)
Seismic velocity ratio Velocity index Description Tunnel
of rock mass support
(SVR = Vf /Vr) VI = SVR2 quality category
----------------------------------------------------------------------------------------------------------------------------------------------------
> 0.9 > 0.8 very good minimum
0.8 - 0.9 0.6 - 0.8 good intermediate
< 0.8 0.4 - 0.6 fair )
0.2 - 0.4 poor } maximum
< 0.2 very poor )
 A5 - 8

2.3 Correlations between seismic velocities and rock mass characteristics

A vast amount of experience has been gained from more than 30 years with refraction
seismic measurements in Scandinavia. Sjögren et al. (1979) have from a comprehensive
investigation of field measurements shown correlations between seismic velocities and
joints measured in drill cores taken in seismic profiles. The investigation comprised 113
km of refraction seismic profiles and 2850 m of drill cores from 8 sites in unweathered,
igneous and metamorphic rocks such as amphibolite, granite, gneiss, meta-anorthosite,
pegmatite, porphyry, quartzite, and mylonite. From the results they have equated longi-
tudinal seismic velocity with 1-D joint frequency in boreholes as shown in Fig. A5-5.

The curve in Fig. A5-5 may be representative for jointed unweathered hard, crystalline
rocks near the surface. The high velocities are well represented in the curve, but for
velocities below about 3500 m/s the data are more scattered. The curve may, therefore, be
less representative in this part. Additional uncertainties may stem from the fact that logging
of drill cores is generally less accurate in highly jointed and crushed rock. Cecil (1971)
proposed a multiplication factor of 0.8 for the higher in situ velocities in order to overcome
this problem.

20

18

16

14

12

10

0
0 1 2 3 4 5 6 7

Fig. A5-5 Average regression curve of the correlation between longitudinal sonic velocity (v) and joint
density. The curve is derived from 5 sites comprising igneous and metamorphic rocks with
1670 m of cores (from Sjögren et al.,1979).

Sjögren et al. (1979) found from their studies that a factor of 0.75 for the rocks of higher
velocities and 0.85 for the relatively low velocities gave a better result. This idea has not
been further developed, probably because it may be difficult to define more accurately for
which velocity range the factor should be applied.

Another investigation has been presented by Cecil from various projects in Sweden.
Refraction seismic measurements in tunnels and at the ground surface above were
compared. The seismic velocities were correlated to the amount of rock support installed in
the tunnel. Unfortunately, the descriptions of the rock mass quality have not been included
in this investigation; therefore, the results are of limited value in this work.
 A5 - 9

3 METHODS FOR ASSESSING THE DEGREE OF JOINTING FROM IN SITU

SEISMIC VELOCITIES

Fig. A5-6 shows 4 different correlations between seismic velocity and joint density. The
curves represent various geological conditions including both fresh and weathered rocks. It
is interesting to notice that all the four curves are almost parallel for the higher degrees of
jointing (>10 joints/m). The basic longitudinal velocity (V0) for the various curves is dif-
ferent, probably caused by different properties and compositions of the intact rock. V0 is
considered to represent the velocity of intact rock under the same conditions (i.e. stress and
water conditions) as in the field, see Fig. A5-7. Thus, the jointing is considerably lower for
the same sonic velocity as the basic velocity (V0 ) decreases.

Curve 2 and 3, contrary to the other two curves, show that there is a stronger influence
from few joints on the sonic velocities. This may be due to a generally larger aperture
(opening) of the joints in these two sites.
1-D joint frequency (joints/m)

20

18

16

14

12

10

8
KEY
6 1: Average results of jointed, unweathered, igneous and meta-
morphic rocks of Palaeozoic age in Scandinavia
4 2: Jointed granite, granodiorite, and andesite from the Andes, Chile
(based on data from Helfrich, Hasselström and Sjögren, 1970).
3: Jointed and weathered metamorphic rocks from the Andes
2 (based on data from Sjögren, 1993). The rock are quartzite, and
various schists and shales.
4: Jointed Triassic and Permian sandstones from Tanzania
0
(Sjögren, 1984).
0 1 2 3 4 5 6 7
seismic velocity (km/s)

Fig. A5-6 Correlations between velocity and joint density for various types of rocks based on results from
Sjögren et al. (1979) and Sjögren (1984, 1993).

The theoretical correlation between seismic velocities and the degree of jointing can be
found from two different approaches:
1: No information is available on the jointing versus seismic velocities.
2: At least two correlations between jointing and seismic velocities are known.

The application of these two methods is described in the following.

 A5 - 10

3.1 Alt. 1. Correlations between jointing and sonic velocity are not known.

The generally exponential distribution of joints (see Table A1-4 in Appendix 1) has been
applied to express the curves in Fig. A5-6 as
Nl = b ⋅ v a eq. (A5-2)

where Nl is the 1-D joint frequency (joints/m) along the drill core or scanline,
v is the seismic velocity (km/s) measured in the field,
a and b are constants related to the local conditions (rock material, stress conditions,
jointing features etc.).

Two different methods have been developed to find the values of a and b:
i: The correlation 1 is expressed as Nl = b ⋅ v - 2.8. The factor b, which represents the local
conditions, is found from b = V03.4 which gives
Nl = V03.4 ⋅ v - 2.8 eq. (A5-3)

where V0 is the basic velocity (velocity of intact rock under the same condi-
tions as in the field).

As seen in Table A5-8 the deviation is generally less than 25% for the four curves, except
for very high and very low degree of jointing.

ii: The constants (a) and (b) vary with the basic velocity (V0 ). The following correlations
have been found based on the curves in Fig. A5-6:
a = -V0 /2 eq. (A5-4)
b = 3/V0a = 3V0Vo/2 eq. (A5-5)

From this eq. (A5-2) can be written

Nl = b ⋅ v a = (3/V0a) v - Vo/2 = 3(V0 /v)Vo/2 eq. (A5-6)

The comparison between eqs. (A5-3) and (A5-6) for the curves 1 - 4 in Fig. A5-6 is shown
in Table A5-8. There is a relatively acceptable accuracy for all curves except curve 3 for
which the calculated results are somewhat higher than the real. Method ii gives a better
correlation between jointing density and seismic velocity than method i.

Important in both methods is the assessed magnitude of the basic velocity (V0 ) which
represents the site-dependent (in situ) velocity for intact rock. Where V0 is not known, it
is recommended to use the velocity for intact rock under the same conditions as in the field
(wet/dry, orientation of anisotropy, stress conditions, etc.).

Joint openness and possible fillings may, however, disturb the accuracy of both correlations
described above where V0 is estimated from laboratory measurements, or from Table A5-2
or similar tables in textbooks. Therefore, alt. 2 described in the next section gives more
accurate results as it includes the site-dependent features.

1
The constant value of a = -2.8 and the eq. (A5-3) have been found from trial and error
fitting to the curves in Fig. A5-6.
 A5 - 11

TABLE A5-8 CORRELATIONS BETWEEN SEISMIC VELOCITY AND JOINTING FOR VARIOUS
VALUES OF THE FACTORS (a) AND (b) IN Eq (A5-2). THE CURVES REFERRED TO
ARE THOSE IN FIG. A5-6.

Correlation | FIELD VELOCITY in km/s

Nl = b ⋅ v a | 1.5 2 2.5 3 4 5 5.5
----------------------------------- --------- ------- ------- -------- ------- ------- --------
Joints/m in curve 1 19 9.5 4.5 3.5
(V0 = 5.8 km/s)
i: Nl = 394 v-2.8 = 18.2 8.1 4.3 3.3
ii: Nl = 490 v-2.9 =  20 8.8 4.6 3.5

Joints/m in curve 2 15 7.5 3 2

(V0 = 5.5 km/s)
i: Nl = 329 v-2.8 = 15.2 6.8 3.6 2.8
ii: Nl = 325 v-2.75 =  15.8 7.2 3.1 3.0

Joints/m in curve 3 15 10.5 4 1.5

(V0 = 5.3 km/s)
i: Nl = 290 v-2.8 = 22 3.4 6.0 3.2
ii: Nl = 229 v-2.65 =  20 12.5 5.8 3.2

Joints/m in curve 4 13 8 5.5 4

(V0 = 3.5 km/s)
i: Nl = 70 v-2.8 = 22 10 5.4 3.2
ii: Nl = 27 v-1.75 =  13.3 8 5.4 3.9

V0 is the basic velocity (for intact rock).

i: The factors a = constant = -2.8 and b = V03.4 (eq. (A5-3))
ii: The factors a = - V0/2 and b = 3/V0a (eq. (A5-6))
 best fit equation of of i: and ii:

3.2 Alt. 2. Two or more correlations exist between jointing and velocities

Sjögren et al. (1979) have presented a method to calculate the degree of jointing from
measured velocities in seismic profiles. The method is based on known data on jointing
and velocities in least two different locations in the profile. They developed the following
relation:
x/Vz + (1 - x)/Vn = 1/v eq. (A5-7)

where Vn is the maximum or 'natural' velocity of rock, see Table A5-9. Both V0 (the
basic velocity of intact rock) and Vn refer to seismic velocities in intact rock. As shown
in Fig. A5-7 Vn are for rocks without cracks and pores.
Vz is the velocity in the crushed or highly jointed rock material,
v is the in situ velocity recorded in the field,
x is the length with the velocity Vz along the measured profile.

This expression has been derived into

ks ⋅ Nl = 1/v - 1/Vn eq. (A5-8)

in which ks is a constant representing the actual in situ conditions,

Nl is the 1-D joint frequency (joints/m) along the core, and
v is the measured in situ velocity (km/s).
 A5 - 12

joint frequency

V0 pores and cracks

0
Vn
0 seismic velocity

Fig. A5-7 The principle difference of the basic seismic velocity (V0.), and the natural or maximum
velocity (Vn ).

TABLE A5-9 APPROXIMATE (NATURAL) VELOCITIES OF FRESH ROCKS, FREE FROM

CRACKS AND PORES. THE VELOCITIES ARE BASED ON THE CONTENT AND
VELOCITIES OF THE MINERALS SHOWN IN TABLE A5-1
(from Goodman, 1989, based on data from Fourmaintraux, 1976).
Rock Vn (km/s) Rock Vn (km/s)
------------------------------------------------------------------------------------------------------------------------------------------------------------
Gabbro 7 Basalt 6.5 - 7
Limestone 6 - 6.5 Dolomite 6.5 - 7
Sandstone and quartzite 6 Granitic rocks 5.5 - 6

It is generally seldom possible by seismic measurements to find (Vn ) at the surface as the
rocks in the surface seldom are free from joints, cracks and pores. Therefore, (Vn ) can best
be found from eq. (A5-8) using the calculation procedure described in the following.

The two unknown constants (ks) and (Vn ) can be found using two data sets of
measured values of (Nl) and the corresponding (v). From this, the natural or
maximum velocity is
v1 ⋅ v 2 ( Nl2 - Nl1)
Vn = eq. (A5-9)
Nl2 ⋅ v 2 - Nl1 ⋅ v1
When Vn has been determined the factor representing the in situ conditions is
1 1 1
ks = ( - ) eq. (A5-10)
Nl1 v1 Vn
Here Nl1 , v1 and Nl2 , v2 are corresponding values of joint frequency and in situ
seismic velocity, respectively, for the two pairs of
measurements.

After ks and Vn have been found from eq. (A5-9) and (A5-10), the degree of
jointing given as joints/m is found from
Nl = (Vn - v)/(Vn ⋅ v ⋅ ks) eq. (A5-11)

From eq. (A5-11) a curve similar to that in Fig. A5-7 which represents the actual
connection between the measured jointing density and the sonic velocities can be
established.
 A5 - 13

Sjögren et al. (1979) showed that these theoretical calculations of average jointing
frequencies have shown a satisfactory agreement with those empirically obtained. The
discrepancies between them have been less than 0.5 joints/m. Thus, seismic refraction
measurements provide a useful tool to characterize the degree of jointing which may be
found very attractive.

3.3 Worked examples

3.3.1 No field information is available on jointing versus sonic velocities

During the initial planning stage a geological survey was carried out which showed that the
bedrocks in the area consisted of fresh dolomite. No information was available on the joint-
ing. Refraction seismic measurements were performed in an area covered by loose deposits
as shown in Fig. A5-8. The rocks in this area were below the ground water table.

Based on the velocities of intact rock in Table A5-2 and A5-3 the basic velocity of
dolomite is estimated as V0 = 5.5 km/s (from Table A5-2). V0 is used to find the value of
the constants 'a' and 'b' in eqs. (A5-3) to (A5-6)

m.a.s.l
s urfa ce
140
1.1 - 1.3 km/s
(loose material)

4.5 km/s
4.2
4.9 km/s 4.5 km/s 3.9 km/s
120 km/s 3.3
km/s
velocity
zone
low

0 20 m 40 m

100

Fig. A5-8 The velocities measured in a refraction seismic profile located in dolomite.

method i: The constant a = const. = -2.8

The factor b varies with V0. Based on the general expression Nl = b ⋅ v - 2.8 the constant
b = V03.4 = 5.5 3.4 = 329
The correlation between the degree of jointing (given as joints/m) and sonic velocity is
Nl = b ⋅ v -2.8 = 329 v - 2.8

method ii: Both constants (a) and (b) vary with the basic velocity (V0).
From eqs. (A5-4) and (A5-5) the constants
a = - V0/2 = - 5.5/2 = - 2.75
b = 3/V0a = 3/5.5-2.75 = 326
The degree of jointing versus sonic velocity is then
Nl = b ⋅ v a = 326 v - 2.75

These two expressions for jointing versus velocity have been illustrated in Fig. A5-10 as 'a'
and 'b'.
 A5 - 14

3.3.2 Field data on jointing and sonic velocity are available.

Two core drillings were later carried out in the seismic profile in Fig. A5-8 as shown in
Fig. A5-9.

BH
BH

1
m.a.s.l

2
s urfa ce
140 1.1 - 1.3 km/s ts/m
/m 5 join 4 6
nts 1 (loose material) 2
joi 0
10
5 4.2 5
4.5 km/s
0 km/s
4.9 km/s 10
4.5 km/s 5
120 15
10
3.9 20
km/s 15

20
3.3
km/s 0 20 m 40 m

100

Fig. A5-9 Refraction seismic profile and core drilling results.

Three pairs of data from core drilling and seismic measurements are used to establish the
relation between the degree of jointing and the longitudinal velocities. These are shown in
Table A5-10.

Table A5-10 THE DATA USED FROM DRILL CORES AND SEISMIC MEASUREMENTS
sonic velocity joints/m borehole comment

1. v1 = 4.5 km/s Nl1 = 4.5 BH 1 (average along the whole borehole in rock)
2. v2 = 3.5 km/s Nl2 = 12 BH 2 (average for 0 - 10 m along borehole)
3. v3 = 3.9 km/s Nl3 = 8 BH 2 (average for 10 - 20 m along borehole)

Combining data set 1 and 2 in Table A3-10 the two unknown constants, ks and Vn, in
eqs. (A5-9) and (A5-10) are found as:
v1 ⋅ v 2 ( Nl2 - Nl1) 4.5 ⋅ 5.5(11 - 4.5)
Vn = = = 5.61 km/s
Nl2 ⋅ v 2 - Nl1 ⋅ v1 11 ⋅ 3.5 - 4.5 ⋅ 4.5
and
11 1 1 1 1
ks = ( - ) = ( - ) = 0.010
Nl1 v1 Vn 4.5 4.5 5.61
The correlation between the degree of jointing given as joints/m and velocity is then
Nl = (Vn - v)/(Vn ⋅ v ⋅ ks) = (5.61 - v)/(5.61 ⋅ 0.01 ⋅ v) = 17.8(5.61 - v)/v

This has been illustrated in Fig. A5-10 as curve 'c'. Similarly, combination of data set 2 and
3 gives curve 'd'. As is seen there is a good accordance between all curves for joint fre-
quencies higher than 6 joints/m. For the lower frequencies curve 'c' or 'd' are considered as
being the most representative.

From the known value of 1-D joint frequency, Nl, the block volume can be calculated
applying the expressions derived in Appendices 3 and 4.
 A5 - 15

16

1-D joint frequency (joints/m)

14

12

10

6
b
4 No field information on jointing exists
a: method i:
2 c a b: method ii:
d Correlation betrween jointing and sonic velocity
0 c: combining data from 1. and 2. in Table A5-10
0 1 2 3 4 5 6 7 d: combining data from 2. and 3. in Table A5-10
seismic velocity (km/s)

Fig. A5-10 Correlation between the results in the example.

4 Summary

Mathematical correlations between seismic velocities and the degree of jointing have been
developed so that the degree of jointing can be estimated at an early stage during investi-
gations where field data on jointing are lacking. It should, however, be noticed that in these
calculations local differences such as the composition of rock types, mineral content, etc.
are averaged, and that the calculations require input of an assumed 'basic velocity' (V0) of
the intact rock. The accuracy of V0 highly determines the quality of the assessments.

At a later stage, when the degree of jointing has been measured in drill cores or from obser-
vations in rock exposures, the accuracy of the assessments of jointing from seismic
velocities can be significantly improved. The jointing found in this way can be used to
characterize the block size along the entire seismic profile provided it is located in the
same type of rock; thus, the information collected in the very limited volume of the rock
mass covered by the borehole can be largely extended.

There are limitations in the use of seismic refraction interpretations in rock mass quality
assessments. These stem mainly from the fact that there are several properties and features
influencing on the velocity, and it is impossible to avoid uncertainties when variations in
the velocity is linked only to one or a few of these.

Refraction seismic measurements cannot be used to assess the joint condition (roughness
and alteration of the joint surface; filling and size of the joint). Cecil (1975) points out
"that clay and other weak or low friction joint fillings which may cause instability in a rock
mass with few joints, may not influence the seismic velocity. On the other hand, one or two
open joints that may not have any effect on the stability of an opening, can significantly
lower the seismic velocity and give the impression of low quality rock." The possibility that
such conditions may exist, must be considered in the geological interpretation of the
seismic refraction results. Therefore, a thorough knowledge of the geological conditions
linked with comprehensive experience in seismic measurements is important

As an increase in the stress level causes closing of the joints, the possibility to indicate
variations in jointing is reduced. Therefore, refraction seismic measurements are preferably
performed near the surface where the stress level is low or moderate.
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

APPENDIX 6

DESCRIPTION OF THE TESTS AND DATA USED IN THE

CALIBRATION OF THE RMi

"Engineers constantly need to simplify the problems that they encounter; this fact
was widely accepted until the days of the computer and its numerical techniques
introduced a dangerous proclivity for complexity."
A. M. Muir Wood, 1979

The problems and costs in performing in situ strength tests on rock masses cause that very
few such test results exist. Only results from 7 sets of such data have been found during this
work. This appendix briefly describes the composition of the material and the results from:
- large samples tested in laboratory;
- in situ tests on pillars in a mine (one 'sample'); and
- tests combined with back analysis from a slide in a mine.
The values found for the block volume (Vb), the joint condition factor (jC) and the jointing
parameter (JP) for the samples have been plotted in Fig. 4-3 in Chapter 4 to develop a
relation between the three parameters.

1 Sample 1. Results from triaxial laboratory tests on Panguna andesite

One of the most complete sets of triaxial test data available is that on the Panguna andesite
which comprises the host rock for a large copper deposit on the island of Bougainville in
Papua New Guinea. Jaeger (1969) has given a description of the samples tested. Essentially,
the rock mass consists of an aggregate of small interlocking blocks. Major joint systems and
probably many minor planes of weakness cause the blocks in the quarry to be of almost
random shape. The material of the test samples is divided up by a network of open joints
and veins with a rather weak filling, the spacings of these being so close "that a cross
section of the core would usually contain 50 - 100 individual areas separated by planes of
weakness." It was difficult to obtain sufficient small 25 x 50 mm cylinders of rock contain-
ing no planes of weakness for triaxial test on intact rock.

The test samples for triaxial tests were taken by very careful drilling, using 150 mm diame-
ter triple tube drilling equipment. A triaxial pot to take specimens 150 mm diameter by 300
mm long was developed and used to produce a number of undisturbed core samples of
jointed Panguna andesite. These samples were prepared and tested triaxially by Jaeger
(1969). The measured unconfined compressive strength of intact rock was 269 MPa. From
the triaxial tests at low confining stress carried out on 150 mm cores an unconfined com-
pressive strength of 3.7 MPa was evaluated by Jaeger, who noticed that an important effect
during testing was the interlocking of adjacent blocks.
As very little has been published on large triaxial tests on rock mass, Hoek and Brown
(1980) concluded from studies of test results of the Panguna andesite that these may be
taken as a reasonable model for the in situ strength of a heavily jointed hard rock mass.
 A6 - 2

Hoek and Brown (1980) used the following characteristics of the Panguna andesite in their
work on the original Hoek-Brown failure criterion, see Fig. A6-1:
- rock intact compressive strength σc = 265 MPa
1
- joint roughness number : Jr = 3
- slightly rough surfaces, with separation < 1 mm
1
- joint alteration number : Ja = 2
- 3 joint sets, with joint spacing < 60 mm

5 Undisturbed core specimens

m = 0.278 , s = 0.0002
PANGUNA ANDESITE
σc = 265 MPa
4
Recompacted graded specimens
Sh ear st ren g th τ - MP a

m = 0.016 , s = 0

3
Fresh to slightly weathered
m = 0.040 , s = 0
Moderately weathered specimens
m = 0.030 , s = 0
2
Highly weathered specimens
m = 0.012 , s = 0

1
Test range
Extrapolated

0
0 1 2 3 4 5 6
Normal stress σ - MPa

Fig. A6-1 Triaxial test results and Hoek and Brown's estimated values of s and m
(from Hoek and Brown, 1980)

The following values have been applied in the calibration of RMi:

• Jointing characteristics: Most joints are small, undulating/interlocking and partly
discontinuous with:
- rough joint surfaces, i.e. the joint smoothness factor is js = 2
2
- undulating/interlocking joint planes , i.e. jw = 2 - 3
- the joint alteration factor 2 jA = Ja = 3 - 4
- small joints (length ≈ 0.15 m), partly discontinuous
(a continuity factor 1.5 seems adequate here), gives a
joint size factor between jL = 2 × 1.5 and 2.5 × 1.5 , i.e. jL = 3 - 4
- The joint condition factor jC = js × jw × jL/jA is then
jCmin = 2 × 2 × 3/4 = 3, jCmax = 2 × 3 × 4/3 = 8.
• Based on the probabilistic method for calculation described in Chapter 5 the following
range has been chosen 3 jC = 4 - 6
• The rock compressive strength σc = 265 Mpa

1
The roughness and alteration factors have later been re-assessed by Wood (1991) to Jr = 2, and
Ja = 3 after reviewing the original publication by Jaeger.
2
Comment: As Jaeger has described the joints as interlocking and the joints on the photo shown
appear as wavy, I have used a higher value for the planarity or waviness than Wood and probably
also Hoek and Brown. With a "rather weak filling" and "open joints and veins" a rating of
Ja = 3 - 4 has been chosen above.
3
The probability calculation gave jC = 4.2 - 5.7 for 2 standard deviations (covering 68% of data).
 A6 - 3

• The average block size, assessed from description and

photo published by Jaeger (1969) is found as: Vb = 2 - 6 cm3
• The uniaxial compressive strength of the rock mass
calculated from triaxial tests: σcm = RMi = 3.7 MPa
From the data above the jointing parameter is found as JP = RMi/σc = 3.7/265 = 0.014

2 Sample 2. Large, compressive laboratory test made on granitic rock from Stripa

The data used have been found from laboratory tests on a 1 meter diameter by 2 meter long
sample of granitic (quartz monzonite) rock from the Stripa mine in Sweden. The sample was
cut by a slot drilling technique.

The following description has been reported by Thorpe et al. (1980): "The granitic rock in
the Stripa mine is pervasively fractured. The sample was intersected by two principal sets of
fractures and a large number of secondary discontinuities with lengths and spacings
ranging from the scale of the core to the microscopic. There were two dominant joint sets in
the sample:
- Set 1 (A,B,C) of essentially continuous joints approximately normal to the loading
- Set 2 (D,E,F) generally discontinuous and oriented 25 - 30o to the loading. They have a
wide distribution of trace lengths.
- In addition there are several small joints in the sample, generally discontinuous.
In general, joints appeared to be about a millimetre or less wide. The joint filling materials
were predominantly chlorite and sericite. There were not any marked difference between the
characteristics of the two sets. However, in set 2 (D,E,F) where a occasional thickening of
the filling material occurred, calcite was prominent."

The behaviour of the large rock mass sample in Fig. A6-2 under uniaxial compressive stress
was very complicated with stress-strain behaviour at low stresses was generally similar to
that of small cores containing single healed joints. Much of the induced fracturing during
compression followed the major pre-existing joints. It appeared that the observed overall
failure mode was a combination of shearing failure in the Mohr-Coulomb sense and brittle
fracturing typical of uniaxial failure of intact specimens. The test of the large sample was
made on wet material. The pervasively jointed nature of the large sample complicated its
mechanical behaviour. It is important to remember that the stress and displacement of such a
compressive test are not generally representative of the boundary conditions in an in situ
rock mass.

The large sample failed at a peak stress of 7.55 MPa, much lower than the typical strength
measured in small cores. 2 unconfined compressive tests performed on 52 mm cores of
intact rock from the big sample gave σc = 208 and 178 MPa. Earlier, a compressive strength
of σc = 214 MPa and 208 MPa have been found for the Stripa granite.

The following values have been applied in the calibration of RMi:

• Joint characteristics:
* Joint set 1: (joint A, B, C in Fig. A6-3) of long, continuous joints with:
- rough joint surface; i.e. the joint smoothness factor js = 2
- large undulations; i.e. the joint waviness factor jw = 2
- hard chlorite coatings; i.e. the joint alteration factor jA = 3 - 4
- continuous joints with length > 1 m; i.e. jL = 1
- The joint condition factor is then jC = 1 - 1.3
- the joint spacing S1 = 0.25 - 1.5 m
A6 - 4
 A6 - 5

* Joint set 2: (joint D, E, F) of short, partly discontinuous joints with

- rough joint surface; i.e. the joint smoothness factor js = 2
- small undulations; i.e. the joint waviness factor jw = 1.5
- hard chlorite coating; i.e. the joint alteration factor jA = 3 - 4
- party discontinuous, short joints (0.5-3 m); i.e. jL = 2 × 1.5 = 3
- The joint condition factor is then jC = 2.25 - 3
- the joint spacing S2 = 0.15 - 0.5 m
* Small, discontinuous joints (little information of these joints) with
- rough surfaces; i.e. the joint smoothness factor js = 2
- small undulations; i.e. the joint waviness factor jw = 1.5
- assumed fresh joints; i.e. the joint alteration factor jA = 1
- discontinuous, short joints (0.1-0.5 m); i.e. jL = 2 × 2 = 4
- The joint condition factor is then jC = 15

• Blocks are formed by the large and partly also by the small joints. Average block
volume is assumed to be in the range Vb = 5 - 15 dm3.
• As most of the displacement during failure followed joint set 1 and 2 the value for the
resulting joint condition factor is chosen as jC = 1.5 - 2.5
• The uniaxial compressive strength of intact rock:
- tests performed on the large sample σc = ½(208+178)MPa = 193 MPa
- earlier tests σc = ½(214+208)MPa = 211 MPa
From this a value of σc = 200 MPa has been applied.
• The uniaxial compressive strength of the rock mass calculated from triaxial tests was:
σcm = RMi = 7.55 MPa (with 0.06% strain at failure)
• From the data above the jointing parameter has been calculated as
JP = RMi/σc = 7.55/200 = 0.04

3 Sample 3. In situ tests on mine pillars of sandstone in the Laisvall mine

In the Laisvall Mine, Northern Sweden, a full scale test was conducted on 9 pillars in order
to obtain realistic design values for the load bearing capacity of the roof and to check the
available experience of pillar strength. The tests were performed in the Nadok orebody of
the mine. The rocks of Late Precambrian age consist of sandstone interlayered with thin
shale partings, overlain by Lower Cambrian clayey schists, see Fig. A6-4. Here, the roof
consists of quartzitic and shaly schist which is weaker than the sandstone in the pillar. The
mine roof in the test area was supported with cement grouted rebar rock bolts (dia. 25 mm,
length 2.3 m with an average density of 0.42 bolts/m2 roof.

The test procedure and the results from the test are described by Söder and Krauland (1990).
9 pillars were subjected to increasing stresses. This was accomplished by blasting off a slice
of approximately 0.8 m thickness from two sides of all the test pillars in each mining step.
The process was continued until pillar or roof/floor failure occurred. Cautious blasting was
applied with 22 mm reduced charges in 51 mm holes spaced 1 m. The damage from blasting
was reported minimal. Stresses, deformation and failures phenomena were monitored during
the test. Some of the results have been made available by the kind help from Norbert
Krauland of Boliden Mines, Sweden.

The average uniaxial compressive strength of intact rocks was found as:
- sandstone σc = 210 MPa
- schist σc = 130 MPa
 A6 - 6

At failure the cross section of the pillar was 23.4 m2 (4.8 x 4.8 m). The height of the pillar
was 5 m. The average uniaxial compressive strength of the rock mass in the pillars was
found as 19.8 ± 1.4 MPa.

Fig. A6-5 Photo of pillar no 8 showing the horizontal joints along the foliation. Pillar height is 5 m
 A6 - 7

Fig. A6-6 Detail from Fig. A6-5.

The jointing characteristics of two of the pillars (no 8 and no 9) have been further evaluated.
The characteristics of the joints have been assumed from descriptions and from several
photos (Figs. A6-5 and A6-6) and from communication with Krauland (1992). The
horizontal, thin shale layers approx. 2 cm thick and parallel joints form major weaknesses in
both pillars. These occur almost normal to loading. In addition steep dipping, mainly short
joints occur with varying intensity.

The following values have been applied in the calibration of RMi:

• The following characteristics of the joints have been assumed:
* Horizontal joints (normal to loading), mainly along the thin shale layers:
- the presence of shale cause that the joints
generally have smooth surfaces, i.e. js = 1
- small undulations of wall, i.e. the joint waviness factor jw = 1.5
- assumed clay on joint wall, i.e. the joint alteration factor jA = 2
- mainly continuous, long joints; i.e. jL = 1
- The joint condition factor is then jC = 0.75
- The joint spacing for this set is S1 = 0.2 - 1.2 m
* Vertical, mainly short joints (parallel to loading):
- slightly rough joints, i.e. the joint smoothness factor js = 1.5
- planar joints, the joint waviness factor jw = 1
- fresh joint walls, i.e. the joint alteration factor jA = 1
- mainly continuous joints with length
in the range 0.1-3 m ; i.e. the joint size factor jL = 1.5
- The joint condition factor is then jC = 2.25
- The joint spacing for this set is S2 = 0.3 - 1.5 m
• There is a wide variation in block size within the pillar.
Estimated block volume is Vb = 0.1 - 0.3 m3.
• The uniaxial compressive strength of the sandstone in the pillars is σc = 210 MPa
• The uniaxial compressive strength of the rock mass
found from the tests was σcm = RMi = 20 MPa
• From the data above the jointing parameter JP = RMi/σc = 20/210 = 0.095.
 A6 - 8

4 Sample 4. Strength data found from back analysis of a slide in the Långsele mine

A large slide occurred in the Långsele mine, Sweden in 1975. In connection with stability
analyses and recommendation for further mining, analysis of the rock mass compressive
strength was performed. The analysis was based on stress measurements and laboratory test
in the mine. The slide has been described and evaluated by Eriksson and Krauland (1975). It
had the shape of a large prism about 100 m long, 60 m thick, and 80 m deep and theinvolved
volume was approximately 300,000 m3. The rocks included in the slided material are:
− sericite-quartzite (1 - 5 m thick) σc = 80 - 280 MPa
− dacite-tuff (25 - 30 m thick) σc = 100 - 270 MPa
− grey schist and greenstone, in which
− most of the slide plane is located σc = 110 - 160 MPa

The foliation is steeply dipping (60 - 70o). The schistosity and foliation partings are signi-
ficant weaknesses. The failure plane seems in the upper part to partly pass through intact
rock. Where it crosses the dacite-tuff, the plane mainly followed existing joint planes. Also
in its the lower part the failure plane partly cut through intact rock. The result of the stability
calculation (Eriksson and Krauland, 1975) gave an uniaxial compressive strength of the rock
mass σcm = 1.85 ± 0.4 MPa.

There are no complete description of the jointing. The assessments made are based on
colour photos in the report by Eriksson and Krauland (1975) and on the rock mass descrip-
tions. Figs. A6-7 to A6-9 show some details from the mine and the slide.

The following values have been applied in the calibration of RMi:

• Joint set 1 occurs along foliation. The joint characteristics have been assessed as:
* In schist and quartzite:
- very smooth surfaces; i.e. the joint smoothness factor js = 0.75
- planar joint walls; i.e. the joint waviness factor jw = 1
- altered or clay coatings; i.e. the joint alteration factor jA = 4
- generally 1-10 m long, i.e. continuous joints jL = 1
- The joint condition factor is then jC ≈ 0.2
* In dacite and greenstone:
- smooth surfaces; i.e. the joint smoothness factor js = 1
- planar joint walls; i.e. the joint waviness factor jw = 1
- altered or clay coating; i.e. the joint alteration factor jA = 4
- 5-20 m long, continuous joints, i.e. jL = 0.75
- The joint condition factor is then jC ≈ 0.2
• Cross joints are short, continuous planar, smooth joints. They seem to occur in two
directions and occur:
* In schist and quartzite with:
- smooth surfaces; i.e. the joint smoothness factor js = 1
- planar joint walls; i.e. the joint waviness factor jw = 1
- altered and clay coating; i.e. the joint alteration factor jA = 4
- 1 -10 m long, continuous joints, i.e. jL = 1
- The joint condition factor is then jC = 0.25 .
A6 - 9
 A6 - 10

Fig. A6-9 The rock masses in the slide surface.

* In dacite and greenstone with:

- smooth surfaces; i.e. the joint smoothness factor: js = 1
- planar joint walls; i.e. the joint waviness factor: jw = 1
- clay coating; i.e. the joint alteration factor: jA = 4
- 5-20 m long, continuous joints, i.e. jL = 0.75
- The joint condition factor is then jC ≈ 0.2
• Assessed average block volume:
- in schist and quartzite Vb = 0.008 - 0.02 m3
- in dacite and greenstone Vb = 5 - 25 m3
• As the uniaxial compressive strength of the rock mass found from the back analyses
was
σcm = RMi = 1.85 MPa, the jointing parameter is JP = RMi/σc = 1.85/160 = 0.01
• It is probable that the block size of the schist and quartzite had the bearing effect on
the slide. The following data have therefore been used:
- the intact compressive strength: σc = 160 MPa
- the block volume: Vb = 0.008 - 0.02 m3
- the joint condition factor: jC = 0.2 - 0.3
- the jointing parameter JP = 0.01

5 Sample 5 - 7. Results from large-scale laboratory triaxial tests

The large scale laboratory tests were carried out at the University of Karlsruhe, Germany.
The test procedure and apparatus (in Fig. A6-10a) are described in ISRM (1989) and in
Natau et al. (1983) and Mutschler and Natau (1989). Many of the following data have been
kindly given by Mutschler (1993) during a visit to the University of Karlsruhe. The assess-
ments are also backed by descriptions and several photos taken of the sites as well as of the
samples (Fig. A6-10b) before and after the tests.
 A6 - 11

5.1 Sample 5. Caledonian clay-schist, Germany

This is a dark, fine-quartzitic, striped to banded clay-silt schist from Thüringer Wald. It
belongs to the 'Phycodenschiefer' of Ordovician - Silurian (Caledonian) age. Two test series
were performed, each of three tests, as described by Natau et al. (1995). The first series was
parallel to schistosity on cylinders dia. 0.6 m and length 1.2 m, the other normal to schistosity
on prismatic samples 0.6 x 0.6 x 1.2 m .

Fig. A6-10 a. Test equipment used in the large scale triaxial tests at the University of Karlsruhe
(from Mutschler and Natau, 1991).
b. Photo of a typical test sample with dimensions 0.6 m diameter and 1.2 m length.

The following values have been applied in the calibration of RMi:

• Three joint sets occur. (The indication of strike/dip is the orientation of the joint sets
measured in the field):
* The main joint set occurs along the steep-dipping schistosity with spacing less than
20 cm. The joint characteristics are:
- smooth surfaces; i.e. the joint smoothness factor: js = 1
- small undulations of joint wall; i.e. the joint waviness factor jw = 1.5
- fresh joint walls with some rust, no coating or filling; i.e. jA = 1
- short - medium, continuous joints; i.e. jL = 1 - 1.5
- The joint condition factor is then jC ≈ 1.5 - 2 .
* The two other sets consist of:
° Joints along the layering (strike/dip = 70/15) of mainly 'welded', dm-long joints.
° Other joints with strike/dip = 215/75 with cm - m size.
 A6 - 12

Both sets have similar joint characteristics with:

- slightly rough surfaces; i.e. the joint smoothness factor js = 1.5
- slightly undulating joint walls; i.e. the joint waviness factor jw = 1.5
- fresh joint walls with some rust, but no coating or filling;
i.e. the joint alteration factor jA = 1
- short - medium, continuous joints; i.e. jL = 1 - 1.5
- The joint condition factor is then jC ≈ 2 - 3
• 3 3
The block sizes are in the range of 100 cm to several dm . According to Mutschler
(1993) the average size has been estimated as Vb = 5 - 10 dm3.
(This means that there were only 35 - 70 blocks in the cylindrical sample tested, and 45 -
90 blocks in the prismatic sample.)

The results from the tests are:

• Uniaxial compressive strength of intact rock:
- parallel to schistosity σc = 55 MPa
- normal to schistosity σc = 100 MPa.
• Uniaxial compressive strength of jointed block mass:
- parallel to schistosity RMi = σcm = 3 MPa
- normal to schistosity RMi = σcm = 8 MPa
• From the test results above the jointing parameter is found as:
JP = RMi/σc = 3/55 = 0.055 parallel to schistosity; and
JP = RMi/σc = 8/100 = 0.08 normal to schistosity.
• The difference in JP for the two directions can probably be explained by the difference in
the joint condition for the schistosity joints and the other joints. The joint condition is,
therefore, estimated as:
jC = 1.5 - 2 for the test parallel to schistosity, and
jC = 2 - 2.5 for the test normal to schistosity
• The values used for calibration in Fig. 4-3 are:
* Tests parallel to schistosity:
- the joint condition factor: jC = 1.5 - 2
- the block volume: Vb = 5 - 10 dm3
- the jointing parameter: JP = 0.055
* Tests normal to schistosity:
- the joint condition factor: jC = 2 - 2.5
- the block volume: Vb = 5 - 10 dm3
- the jointing parameter: JP = 0.08

5.2 Sample 6. Mesozoic sandstone from Germany

This is a Lower Triassic sandstone with intercalations of silty claystone from Hessen. The
tests (named MTG) were performed normal to the layering on cylindrical samples of 0.6 m
diameter and 1.2 m length. They have been published by Natau et al. (1983).

The following values have been applied in the calibration of RMi:

• The following joints were present:
* Joints along the layering are most prominent, they had the following characteristics:
- slightly rough - rough surfaces; i.e. the joint smoothness factor js = 1.5 - 2
- planar joint walls; i. e. the joint waviness factor: jw = 1
- unaltered walls without coating/filling; i.e. the joint alteration factor jA = 1
- continuous, medium length (1 - 10 m) joints; i.e. jL = 1
- The joint condition factor is then jC = 1.5 - 2 .
 A6 - 13

* Cross joints, partly discontinuous with the following characteristics:

- rough surfaces; i.e. the joint smoothness factor: js = 2
- undulating joint walls; i.e. the joint waviness factor: jw = 1.5 - 2
- unaltered, often 'welded'; i.e. the joint alteration factor jA = 0.75
- short (< 1 m) joints, often discontinuous; i.e. jL = 2 × 1.5 = 3
- The joint condition factor is then jC = 12 - 16
• There is a great difference between the jC for the two types of joints.
Therefore it is difficult to decide what will be the resulting
jC during the test. It is assumed that jC = 5 -10 .
• Based on information from Mutschler (1993) the average block size is Vb = 1 - 5 dm3

The results of the triaxial tests are:

• Uniaxial compressive strength of intact rock - sandstone σc = 10.5 MPa
- claystone σc = 4.8 MPa
• Uniaxial compressive strength of jointed rock mass performed
on cylindrical samples σcm = RMi = 1.33

The sample consists of two types of rock with different strength and it is therefore uncertain
what compressive strength should be used in this case with such inhomogeneous rock mass.
An assumed value of σc = 8 MPa may be reasonable.This gives JP = RMi/σc = 1.33/8 ≈ 0.17

Based on this the values used for calibration are: - the joint condition factor: jC = 5 - 10 (?)
- the block volume: Vb = 1 - 5 dm3
- the jointing parameter: JP = 0.17

5.3 Sample 7. Palaeozoic siltstone from Germany

This is a Carboniferous, dark siltstone from Hagen in Germany. The tests were carried out on
cylindrical samples 0.6 m diameter and 1.2 m long. According to Mutschler (1993) no
prominent joint set occurred in this sample where mainly short joints were orientated in
various directions.

The following values have been applied in the calibration of RMi:

• The joints have the following characteristics:
- slightly rough joint surfaces; i.e. the joint smoothness factor js = 1.5
- small to large undulation of walls; i.e. the joint waviness factor jw = 1.5 - 2
- fresh joint surface without coating or fillings; i.e.
the joint alteration factor: jA = 1
- continuous joints with average length approx. 1 m:, i.e. jL = 1.5
- The joint condition factor is then jC ≈ 3.5 - 4.5
• The average block volume (Mutschler, 1993) is in the range Vb = 5 - 10 dm3

The laboratory tests on intact rock and on jointed rock mass gave the following results:
• Uniaxial compressive strength of intact rock σc = 65 MPa
• Uniaxial compressive strength on cylindrical
samples of jointed rock mass RMi = σcm = 6.8 MPa
• From this the jointing parameter is found as JP = RMi/σc = 6.8/65 = 0.10

The values used in the calibration are: - the joint condition factor: jC = 3.5 - 4.5
- the block volume: Vb = 5 - 10 dm3
- the jointing parameter: JP = 0.10
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

APPENDIX 7

COLLECTED DATA ON GROUND CONDITIONS AND ROCK

SUPPORT IN CONSTRUCTED UNDERGROUND OPENINGS

"A good engineering design is a balanced design in which all the factors which interact, even those which
cannot be quantified are taken into account. Therefore the responsibility of the design engineer is not to
compute accurately but to judge soundly."
Evert Hoek and Pierre Londe (1974)

Collection of rock mass characteristics, field data on rock mass conditions and rock support has
been carried out for the following underground excavations:
1. Gjövik Olympic mountain hall; Gjövik, Norway
2. Granfoss road tunnels; Oslo, Norway
3. Haukrei hydropower plant; Telemark, Norway
4. Vinstra hydropower plant; Gudbrandsdalen, Norway
5. Horga hydropower plant; Sigdal, Norway
6. Tromsö road tunnel; Tromsö, Norway
7. Nappstraumen road tunnel; Lofoten, Norway
8. Stetind road tunnel; Tysfjord, Norway
9. Njunis tunnel; Bardu, Norway
10. Sumbiar road tunnel; The Faroe Islands
11. Thingbæk chalk mines; Ålborg, Denmark

The information collected has been used to calculate the quality of the ground (given as the ground
condition factor for discontinuous materials, or the competency factor for continuous materials) and
the size ratio (tunnel size/block size). The calculations and data are presented in Section 2.

1 DESCRIPTION OF THE LOCATIONS

1.1 Gjövik Olympic mountain hall, Norway

This underground stadium has the following dimensions: span = 62 m, length = 91 m, total height =
24 m, wall height = 10 m. Orientation of the opening: N60oE

Description of the general ground conditions:

Precambrian red and grey gneiss. It has a composition varying from granitic to quartzdioritic with
about 30% quartz, 65% feldspar and a few percent of chlorite, mica and hornblende. The
unconfined compression strength is in the range 63 - 94 MPa. The cavern is not intersected by any
major weakness zone, except for a fault zone at the entrance of the main access tunnel.

The jointing is mostly irregular, usually three joint sets are found. The main joint set occurs along
the foliation of the gneiss (strike/dip = N-E/50 SW), with joint spacing usually 0.2 - 0.5 m with and
joint length mainly 2 - 5 m. The joints are smooth to rough and undulating. Most of the joints have

SUPPDATA\text-1.wpd
 A7 - 2

no filling, but in a few joints filling of clay, chlorite, silt/sand occurs; also calcite, epidote and
quartz is found. Tectonism has resulted in an additional network of microjoints sometimes with
clay coatings.
The average joint condition factor: jC = 3.0

The joint wall compressive strength obtained by means of Schmidt hammer readings, is between 50
and 100 MPa, i.e. the same as the strength of the rock.
Average block volume is Vb = 0.1 m3.
The overburden is 20 - 50 m. The stresses measured (at a depth 40 m below surface) are:
horizontal stress = 3.5 MPa; vertical stress = 1 MPa, i.e. k = horizontal/vertical stress = 3.5

In general, the initial support consisted of 27 mm diameter and 4 m long expansion shell bolts in
unstable blocks. For the permanent use, 6 m long and 25 mm diameter rebar rock bolts were
installed in pattern 2.5 × 2.5 m in addition to 12 m long twin-strand cable bolts in 5 × 5 m pattern.
All the permanent bolts are fully grouted. Initially, a 50 mm thick of fibre reinforced shotcrete was
sprayed, followed by additional 50 mm after the cable bolts had been placed.

1.2 Granfoss road tunnels, Oslo

The four tunnels have all horse-shoe shape. The following locations have been inspected:
Lysaker N, span Wt = 10.7 m,
Ullern N, span Wt = 12 m,
Ullern S, span Wt = 9 m.

Description of the general ground conditions:

Ordovician sedimentary rocks consisting of bedded clayschists, claystones and limestones
("knollekalk"). The rocks were folded during the Caledonian orogenesis and therefore the
orientation of the bedding varies. The layers consisting of schistose rocks have been most strongly
folded. Assumed compressive strength: clay schist σc = 40 MPa, claystone and limestone σc = 40
MPa.

The tectonic activity during Permian time resulted in several small and large faults, mainly striking
N-S. Several veins and dykes have been formed from volcanic activity in this period. Veins of
maenaite and syenite can be several 10 m thick. The 0.5 - 3 m thick dolerite dykes are steep-
dipping.

Two main joint sets occur in the tunnel. One is located along the bedding with strike N45oE and dip
of 70oN. The other is steep with striking between East and South.
In the locations described the joint are mainly rough and planar to undulating. Most walls are fresh;
they have seldom coating or filling. Generally there are mainly small, irregular joints with length
0.5 - 3 m. The degree of jointing (i.e. block size) vary in the various locations described. Some
damage from the drill and blast excavation has been made in the tunnel periphery.

The overburden for tunnel varies between 10 and 50 m. Assumed k-value: k = ph /pv = 2.

The following input values have been used:

 A7 - 3

Lysaker N (chainage 400):

-Rocks: clay schists
-Average block volume: Vb = 0.2 m3
-Average block shape factor: β = 40
-Average joint condition factor: jC = 2.0
-Main joint set -strike between tunnel and joint set: αj = 20o
-dip between tunnel and joint set: βj = 90o
Roof support used: fibrecrete 100 mm, bolts spaced 1.5 m (length 3 m)

Ullern N (chainage 1320):

-Rocks: clay schists
-Average block volume: Vb = 0.05 m3
-Average block shape factor: β = 30
-Average joint condition factor: jC = 2.0
Roof support used: fibrecrete 80 mm thick, rock bolts spaced 1.5 m (bolt length = 4 m)

Ullern N (chainage 1420):

-Rocks: limestone
-Average block volume: Vb = 0.2 m3
-Average block shape factor: β = 40
-Average joint condition factor: jC = 2.5
-Main joint set: -average spacing: Sa = 0.3 m
-strike between tunnel and joint set: αj = 70o
-dip between tunnel and joint set: βj = 80o
Roof support used: fibrecrete up to 100 mm thick, and rock bolts spaced 1.5 m (bolt length = 3 m)

Ullern N (chainage 1700):

-Rocks: limestone
-Average block volume: Vb = 0.5 m3
-Average block shape factor: β = 40
-Average joint condition factor: jC = 2.0
-Main joint set: -average spacing: Sa = 0.4 m
-strike between tunnel and joint set: αj = 70o
-dip between tunnel and joint set: βj = 80o
Roof support used: fibrecrete up to 100 mm thick, and rock bolts spaced 1.5 m (bolt length = 3 m)

Ullern S (chainage 1875):

-Rocks: limestone
-Average block volume: Vb = 0.1 m3
-Average block shape factor: β = 40
-Average joint condition factor: jC = 2.0
Roof support used: fibrecrete up to 100 mm thick, and rock bolts spaced 1.5 m (bolt length = 3 m)

1.3 Haukrei hydropower plant, Telemark, Norway

The observation have been made in the headrace tunnel which has a horse-shoe shape with a cross
section area: 12 m2, the tunnel span is Wt = 3 m.
 A7 - 4

Description of the ground conditions:

Precambrian gneiss and granitic gneiss (strike/dip related to tunnel = 20-30o/70-80o to the right)
with main joint set along foliations with spacing S = 0.5 - 2 m. Some random joints occur. The
foliation joints are rough, undulating with fresh walls and a length of 1 - 5 m
Overburden at the locations described below: z = 50 m,
There are probably high horizontal stresses in this area; assumed k = ph /pv = 4

Rock conditions at chainage 200:

-Rocks: grey gneiss, assumed compressive strength: σc = 130 MPa
-Average block volume: Vb = 3 m3
-Block shape factor: β = 40
-Average joint condition factor : jC = 3.5
-Main joint set: -average spacing: Sa = 0.75 m
-strike between tunnel and joint set: αj = 30o
-dip between tunnel and joint set: βj = 70o
Roof support: no rock support, occasionally some scaling work.

Rock conditions at chainage 110:

A 10 m wide weakness zone was encountered at chainage 110. Its orientation related to the tunnel
is (strike/dip = 45/90). It consists of a partly chloritized diabase (assumed σc = 100 MPa). The zone
consists of several parts having somewhat different composition as described in the table below.

Thickness of individual parts of the zone between the adjacent rock masses
Adjacent Adjacent
FEATURE rock on left (m) rock on
side right side
0.5 - 1 1-2 2 1-2 1 -2 0.4 - 0.5 0.5 -1
Joint spacing (m), set 1 0.5 - 2 0.01-0.05 0.1-0.5 0.05-0.2 0.02-0.1 0.01-0.05 0.2-0.3 0.5-2
set 2 0.3-1 a zone
Joint length (m), set 1 1-5 0.1-0.5 0.5-3 0.3-3 0.3-2 0.1-1 mainly of 0.5-2 1-5
set 2 0.01-0.1 chloritic
Joint smoothness rough smooth smooth smooth smooth smooth clay rough rough
waviness undul. undul. undul. undul. undul. undul. planar undul.
Joint alteration fresh - - - - -
or coating chlorite chlorite chlorite chlorite chlorite

Random joints a few a few a few

Block volume, min 0.3 m3 2 cm3 10 dm3 5 dm3 1 dm3 2 cm3 0.02 m3 0.3m3
max 3 m3 50 cm3 100 dm3 50 dm3 10 dm3 100 cm3 0.1 m3 3 m3
Block shape flat long flat flat flat long flat flat
(rhomb.)
Rocks granitic slightly altered diabase with chlorite coating on most joint planes granitic
gneiss gneiss

From the data presented in table above the following average values have been applied for the
various parameters:
-Width of weakness zone: Tz = 10 m
-Strike between tunnel and zone: αz = 45o
-Dip between tunnel and zone: βz = 90o
-Average joint condition factor: jC = 0.75
-Average block volume: Vb = 0.02 m3
-Average block shape factor: β = 40
Roof support:
-initial support: rock bolts and wire mesh;
-permanent support: additional rock bolts spaced 1.5 m and fibrecrete 100 mm thick.
 A7 - 5

1.4 Vinstra hydropower plant, Norway

The horse-shoe headrace tunnel has a gradient 1:11 and cross section area is 35 m2.
The span is Wt = 6.5 m.

Description of the ground conditions:

The actual location described at chainage 3250 is in a weakness zone in which a large slide occurred
after the tunnel had been in operation for 4 years. The slide involved approximately 15 000 m3.
Slided material from the progressive slide was transported by the water stream in the tunnel almost
down to the power house; a distance of 3 km. This is the reason why such large volume could be
involved in the slide.

The weakness zone (fault) had been supported by 15 cm thick fibrecrete and rock bolts 2 x 2 m. The
main reason for the slide was that clay seams were washed out during infilling of the tunnel system
(the system had been emptied 3 times). This outwash was possible because fibrecrete was not
applied in the lowest part of the tunnel walls.

To remedy the tunnel it was decided to excavate a 150 m long by-pass tunnel 40 m from the old
tunnel in the area where the weakness zone was located. Also this tunnel had to cross the weakness
zone.

The rocks in the weakness zone consist of tectonized, folded phyllite. The rocks are moderately
weathered with assumed compressive strength, σc = 10 MPa
The weakness zone (with strike/dip = 70o/200) is 10 - 15 m thick with a transition zone of
approximately 2.5 m on each side. The rock masses in the zone split up along thin clay-coated
partings (length approx. 0.2 - 2 m) spaced 5 - 50 mm along foliation. Some short (0.3 - 2 m long)
joints cut across the foliation. Laboratory tests showed that the clay-coatings contain swelling
minerals.

The overburden at the site is approximately z = 250 m

The following other input values have been used:

-Average volume of blocks: Vb = 25 × 150 × 250 mm = 0.001 m3
-Assumed k-value: k = ph /pv = 2.5
-Strike between tunnel and zone: αz = 70o
-Dip between tunnel and zone: βz = 20o
-Average joint condition factor: jC = 0.4
-Average block shape factor: β = 100

Rock support used:

- initial support: fibrecrete after each round, approx. 80 - 100 mm thick and rock bolts spaced
2.5 m
- permanent support: 1 additional 50 - 80 mm thick fibrecrete, plus 1 m wide reinforced ribs of
fibrecrete 200 mm thick spaced 2.5 m; the ribs are strengthened by rock
bolts at approx. 1 m distance.

1
This support may be considered equal to:
Initial fibrecrete 60 - 100 mm thick strengthened by cast in place concrete lining as permanent support. The reason
why this solution was not chosen was that it would have taken longer time.
 A7 - 6

1.5 Horga hydropower plant, Buskerud, Norway

The observation were made in the headrace tunnel, which has horse-shoe shape with a cross section
area 8 m2
The span is Wt = 3 m

Description of the general ground conditions

Precambrian red, striped gneiss (σc = 100 MPa) with some thin schistose layers or zones. The main
joint set occurs along the foliation, mainly consisting of smooth to rough and undulating joints.
Overburden: z = 50 - 200 m
Assumed k-value: k=2-3

The following input values have been used:

Chainage 470:
The main joint set occurs as flat-dipping, short foliation joints spaced 0.2 - 0.5 m. Some other short
(0.1 - 1 m long) joints (25/90) occur.
-Average joint condition factor: jC = 2
-Average block volume: Vb = 0.05 m3
-Average block shape factor: β = 40
-Main joint set: -average spacing: Sa = 0.4 m
-strike between tunnel and joint set: αj = 0o
-dip between tunnel and joint set: βj = 0o
Roof support: some spot bolting (average spacing estimated to 2.5 - 4 m)

Chainage 810:
Crushed weakness zone (strike/dip = 90/45), 4 m thick. The individual 0.3 - 2 m long joints in the
main joint set along the zone are spaced 0.05 - 0.3 m, average 0.08 m. It is thin clay coating on a
few joint walls.
-Strike between tunnel and zone: αz = 90o
-Dip between tunnel and zone: βz = 45o
-Average joint condition factor: jC = 0.5
-Average block volume: Vb = 0.05 m3
-Average block shape factor: β = 40
Roof support:
- initial support: fibrecrete 60 mm thick, rock bolts spaced 2 m,
- additional permanent support: fibrecrete 60 mm thick, rock bolts spaced 2 m

Chainage 1485:
Foliation joints are the main joint set; in addition to many steep, irregular, small, planar joints.
-Average joint condition factor: jC = 6
-Average block volume: Vb = 0.1 m3
-Average block shape factor: β = 40
Roof support: rock bolts spaced 2.5 m

1.6 Tromsö road tunnel, Tromsö, Norway

 A7 - 7

The actual location is in a roundabout located in rock with the following dimensions:
Roundabout shape: horse-shoe
Roundabout span is: Wt = 13 - 20 m, its wall height: Hw = 8 m
The horse-shoe tunnels leading into it have cross section area 50 m2
The tunnel span is: Wt = 9 m, wall height: Hw = 5 m

Description of the geological conditions:

Similar ground conditions occur in the tunnel and in the roundabout consisting of striped, folded
gneiss with its main joint set along foliation. These flat dipping joints are smooth and undulating
having length 1 - 3 m, occasionally up to 10 m.
Other joints occur mainly as random joints, which are smooth and planar; length 0.5- 5 m
Some 0.5 - 2 m wide joint zones occur with individual joints being smooth and planar; joint length
0.5- 1 m, sometimes up to 10 m.

The following other input values have been used

- Overburden: z = 20 m
- Assumed rock compressive strength: σc = 100 MPa
- Assumed k-value (ph /pv): k =3
- Joint condition factor: jC= 1.5 - 2 (jR = 1.5 - 2, jA = 1, jL = 1)
- Average block volume: Vb = 3 m3 (Vbmax = 10 m3, Vbmin = 1 m3)
- Block shape factor: β = 40

Rock support used:

- In tunnel (span 9 m): rock bolts spaced 3 m
- In roundabout (span 13 - 20 m): rock bolts spaced 1.5 m;
fibrecrete 60 - 80 mm thick

1.7 Nappstraumen road tunnel, Lofoten, Norway

The cross section area of this horse-shoe shaped tunnel is 50 m2, the span Wt = 10 m.

Description of the general ground conditions:

Precambrian gneisses varying between dark and light coloured types with assumed average
compressive strength, assumed σc = 150 MPa. The main joint set occurs along the foliation,
constituting rough - smooth, undulating walls, mainly with spacing 0.5 - 2 m. The joint walls are
often stained. Often, several small (5 - 20 cm long) cracks occur; they are partly healed, rough and
undulating.

Occasionally, up to 2 m wide zones occur consisting of more closely spaced joints spaced 5 - 20 cm.
The joints are sometimes long. The following input values have been applied for the general ground
conditions and rock support:
-Overburden: z = 30 - 50 m
-Assumed k-value (ph /pv): k =3
-Average block volume: Vb = 5 m3
-Average block shape factor: β = 35
-Average joint condition factor: jC = 2
-Average spacing of main joint set: Sa = 1 m
Roof support used: rock bolts spaced 2 m.
1.8 Stetind road tunnel, Nordland, Norway
 A7 - 8

Cross section area (horse-shoe): 50 m2, span: Wt = 10 m

Description of the ground conditions, chainage 15750:

Caledonian granitic gneiss, coarse-grained, exhibiting few very rough and undulating joints. Heavy
rock burst occurred during blasting. It is assumed that many of the joints which were observed
during the inspection may have been developed from rock burst. Originally, the block size was
probably in the range Vb = 5 - 30 m3. After excavation and redistribution of stresses the block
volumes are in the range: Vb = 0.5 - 3 m3.

The high peak Stetind amounts to 1300 m above the tunnel at an inclination >30o. The vertical
stress component is assumed to be equal to approximately 900 m overburden. The strongly
anisotropic stresses are assumed with a value: k = 5
The following other input values have been applied:
-Average joint condition factor: jC = 2.5
-Average block size (before): Vb = 15 m3
-Average block shape factor: β = 40
Roof support used:
- initial support: rock bolts spaced 1.5 - 2 m, fibrecrete 60 mm thick.
- permanent support: some additional bolting; (the resulting bolt spacing is 1 - 1.5 m)

1.9 Njunis tunnel; Bardu, Norway

This access tunnel (horse-shoe) has gradient 1:7, cross section area: 10 - 12 m2, and span: Wt = 3 m

Description of the general ground conditions:

Dark amphibolitic gneiss (meta-basalt), with foliation often parallel to the tunnel. Assumed
compressive strength is σc = 200 MPa.
The main joint set which occurs along the foliation, has a joint spacing of 0.3 - 1 m. The joints are
undulating, smooth - rough with length generally 5 - 20 m. Staining on joint wall is frequent.
Frequently, small (0.1 - 1 m) joints occur across foliation. They are strongly undulating, rough,
often with calcite coatings.
The overburden is 200 m
Assumed k-value (ph /pv): k = 5

The following input values have been used from this tunnel:

Chainage 6250:
-Average joint condition factor: jC = 3
-Average block volume: Vb = 0.5 m3
-Average block shape factor: β = 40
-Main joint set: -average spacing: Sa = 0.4 m
-strike between tunnel and joint set: αz = 45o
-dip between tunnel and joint set: βz = 10o
Roof support: rock bolts spaced 3 m

Chainage 6300:
 A7 - 9

Here, a steep-dipping, crushed weakness zone, 6 m thick (strike/dip = 70o/90o) was encountered. It
consists of 1 - 10 mm thick clay-filled joints spaced 0.2 - 0.3 m along the zone. Additional, parallel
small joints spaced 0.05 - 0.2 m and some short joints along the foliation (across the other joints)
with spacing 0.1 - 0.3 m.
-Average joint condition factor: jC = 0.25
-Average block size: Vb = 0.005 m3
-Average block shape factor: β = 40
-Strike between tunnel and zone: δz = 70o
-Dip between tunnel and zone: βz = 90o
Roof support:
- initial support: wire mesh and rock bolts spaced approx. 1.5 m.
- additional permanent support: fibrecrete 60 mm thick.

1.10 Sumbiar road tunnel, The Faroe Islands

Tunnel cross section (horse-shoe) area: 50 m2, tunnel span Wt = 10 m

Description of the ground conditions:

Tertiary basalt, flat-layered, characterized by relatively thick layers intersected by thin layers of tuff.
Assumed rock compressive strength: σc = 200 MPa. There are varying compositions of the basalt
layers, also within each layer the rocks vary. All basalts in the locations described have a dense
texture.

The degree of jointing is generally low to moderate. Most joints are smooth and undulating. There
are sometimes joints with calcite coating. Generally, the block volume is 0.5 - 2 m3 caused by a few
meter long joints. Smaller blocks are mainly caused by shorter (dm long) strongly undulating joints.

The stability in the tunnel (i.e. in the roof) is mainly influenced by the flat-dipping joints.
The weakness zones are generally steep, often at right angles to the tunnel. There are often parallel
joints with calcite coating in the adjacent rock masses.
Overburden: z = 100 - 300 m
Assumed k-value (ph/pv): k=2

The following other input values are:

Chainage 650:
-Average joint condition factor: jC = 1.5
-Average block volume: Vb = 1 m3
-Average block shape factor: β = 30
-Main joint set: -average spacing: Sa = 0.8 m
-strike between tunnel and joint set: αj = 75o
-dip between tunnel and joint set: βj = 90o
Roof support: spot bolting
 A7 - 10

Chainage 1315:
-The joints in this location are mostly rough, short and
discontinuous, therefore the joint condition factor is: jC = 7
-Average block volume: Vb = 0.025 cm3
-Average shape factor: β = 35
Roof support: 50 mm thick shotcrete

Chainage 2100:
This i a 3 m thick crushed weakness zone consisting of rough, undulating joints with silt/sand
coatings intersects the tunnel at strike/dip = 45o/90o.
-Average joint condition factor: jC = 0.75
-Average block volume: Vb = 0.05 m3
-Average block shape factor: β = 75
-Strike between tunnel and zone: αz = 45o
-Dip between tunnel and zone: βz = 90o
Roof support: 50 mm thick shotcrete

Chainage 600:
Another crushed weakness zone, 1.5 m thick (65/90), with many clay-filled joints. Some water
leakage occurs.
-Average joint condition factor: jC = 0.5
-Average block volume: Vb = 100 dm3
-Average block shape factor: β = 40
-Strike between tunnel and zone: αz = 65o
-Dip between tunnel and zone: βz = 90o
Roof support: rock bolts spaced 1.5 m, straps and wire mesh 2

1.11 Thingbæk chalk mines, Ålborg, Denmark

The Thingbæk mines, which are located near Ålborg in Denmark, have not been in operation since
1920. The observations are from the old mine in the branch tunnel leading to the newly excavated
(1992) connection tunnel between the old mine and the new mine where a museum has been
established. The tunnel with a rectangular shape, has a span of Wt = 5 m, and a height Ht = 7 - 8 m

Description of the ground conditions:

The Jurassic - Tertiary rocks are flat-layered chalk with thin bands rich in flint. The compressive
strength is measured as: σc = 1 - 2 MPa. The main joint set occurs along bedding planes generally
spaced 1 - 2 m. The joints are rough, partly planar with length
mainly in the range 3 - 15 m. In the rock blocks between the bedding joints there are short (0.1 - 0.3
m long) cross joints spaced 1 - 3 m. Average block volume is Vb = 2 - 50 m3. The flat (top of the)
roof at the actual location is shaped along one of the bedding weaknesses.
Some few large (> 10 m) steep-dipping joints cut the mine tunnel at 60 - 80o. Many of the tunnels in
the mine are located along these. The roof is often flat where it is developed along pervasive
bedding joints.

2
Rock bolt spaced 2 m and 60 mm shotcrete could also have been applied.
 A7 - 11

The following input values have been used:

-The overburden: z = 5 - 10 m
-Assumed k-value: k =1
-Joint condition factor: jC = 3.0
-Average block volume: Vb = 30 m3
-Block shape factor: β = 45
-Main joint set, average spacing: Sa = 1.5 m
-Strike between tunnel and joints: αj = 45o
-Dip between tunnel and joints: βj = 10o

Roof support used:

- Spot bolting (spacing approx. 3 m) to prevent loosening along bedding joints.
- Some time-dependent loosening of fragments was observed in the rock blocks are caused by
the small cross joints and a probably slight overstressing of the rock material.
- Long-term deterioration of the rock in the tunnel surface requires the use of shotcrete. This
has been applied in the mine museum which occupies a part of the mine. The museum is
open to the public.

2 CALCULATION OF GROUND CHARACTERISTICS APPLIED IN THE ROCK

SUPPORT TABLES

The rock mass characteristics described in the foregoing have been used to calculate the jointing
parameter (JP), the RMi value, and the ground characteristics which have been applied to develop
the support charts in Chapter 6.

The data from the various locations are inserted in the 'input part' of the tables presented in the
following. The various ground parameters are calculated using equations presented in Chapters 4
and 6.
 A7 - 12

ROCK SUPPORT EXPERIENCE - CASES

Project / tunnel: Gjövik OL stadium Granfoss road tunnels
INPUT DATA Location: general conditions Lysaker S, chain. 400 Ullern S, chain.1875
Rock(s): red and grey gneiss clay schist claystone "knollekalk"
Rock deformability (brittle = 1 "ductile" = 2) 1 2 1
Tunnel span (width) (m) Wt 62.00 10.70 9.00
Tunnel wall height (m) Hw 10.00 6.00 6.00
Tunnel shape (2=high horse-shoe, 4=horse-shoe, 6=circular, 8=square) 7 4 4
Overburden (m) z 40 20 30
Rock compressive strength (MPa) σc 80.0 40.0 60.0
Assumed stress ratio (Ph /Pv) k 3.5 2.0 2.0
Joint condition factor jC 3.0 2.0 2.0
Blocks: - Block volume (m3) Vb 0.2 0.2 0.1
- Block shape factor β 35 40 40
Main joint set: - Average spacing (m) Sa
- Strike (related to tunnel ) αj 20
- Dip (related to tunnel) βj 90
For weakness zone: - Indicate type of weakness zone
- Thickness of weakness zone (m) Tz
- Strike (related to tunnel ) αz
- Dip (related to tunnel) βz
- Rock Mass index in adjacent rocks RMia
Rock support used: - bolt spacing (m) in - roof 2 (l=4-6m) and 5(l=12m) 1.5 1.5
- wall
- shotcrete (mm) (F = fibrecrete) in: - roof (F) 100 (F) 70 (F) 70
- wall
- concrete lining
Remarks:

CALCULATIONS Scale effect compr. strength fσ 0.633013651 0.644383724 0.674856168

Jointing parameter JP 0.214776 0.168424 0.134723
Rock Mass index RMi 17.182109 6.736963 8.083371
Equivalent block diameter (m) Db 0.49 0.45 0.36
Intersection between main joint set and tunnel -roof favourable
-wall very unfavourable
Intersection between weakness zone and tunnel -roof
-wall
Assumed vertical stress (MPa) pz 1.08 0.54 0.81
Assumed horizontal stress (MPa) ph 3.78 1.08 1.62
Tunnel shape factors (according to Hoek & Brown, 1980): Factor A = 2.0 3.2 3.2
Factor B = 5.0 2.3 2.3
Tangential stress (MPa) in: - roof σθ r 6.48 2.92 4.37
- wall σθ w 1.62 0.16 0.24
Continuity of ground in: - roof CF continuous (particulate) discontinuous discontinuous
- wall CF discontinuous discontinuous discontinuous
Characteristics in jointed (discontinuous) rock masses
Approximate stress level (MPa) 3.78 1.08 1.62
Stress level coefficient SL 1 1 1
Ground condition factor (RMi x SL) in: -roof Gc 17.18 6.74 8.08
(RMi x SL x wall factor) in: -wall Gc 85.91 33.68 40.42
Orientation factor for joints in: -roof Co 1.5 1.0 1.5
-wall Co 1.5 3.0 1.5
Size ratio [(Wt/Db) Co] in: - roof Sr 4342494.9 23.8 868122.2
Size ratio [(Ht/Db) Co] in: - wall Sr 700402.4 40.0 578748.1
Characteristics in weakness zone or fault (No weakness zone) (No weakness zone) (No weakness zone)
Orientation factor (intersection zone - tunnel) in: -roof Coz
-wall Coz
Ground condition [SL x RMi(m)] in: -roof Gcz
-wall Gcz
Size ratio [(Tz/Db)Coz] or [(Wt/Db)Coz] in: - roof Srz
Size ratio [(Tz/Db)Coz] or [(Ht/Db)Coz] in: - wall Srz
Characteristics in continuous rock masses
Competency factor in: - roof Cg 2.65
- wall Cg
Possible stress behaviour of massive rock in: - roof high stress level
- wall
Possible behaviour of continuous, particulate rock masses: light squeezing
 A7 - 13

ROCK SUPPORT EXPERIENCE - CASES

Project / tunnel: Granfoss road tunnels, Ullern N
INPUT DATA Location: chainage 1320 chainage 1420 chainage 1700
Rock(s): clay schist claystone "knollekalk" claystone "knollekalk"
Rock deformability (brittle = 1 "ductile" = 2)
Tunnel span (width) (m) Wt 12.00 12.00 12.00
Tunnel wall height (m) Hw 6.00 6.00 6.00
Tunnel shape (2=high horse-shoe, 4=horse-shoe, 6=circular, 8=square) 4 4 4
Overburden (m) z 30 30 30
Rock compressive strength (MPa) σc 40.0 60.0 60.0
Assumed stress ratio (Ph /Pv) k 2.0 2.0 2.0
Joint condition factor jC 2.0 2.5 2.0
Blocks: - Block volume (m3) Vb 0.05 0.2 0.3
- Block shape factor β 30 40 45
Main joint set: - Average spacing (m) Sa
- Strike (related to tunnel ) αj 70 70
- Dip (related to tunnel) βj 80 80
For weakness zone: - Indicate type of weakness zone
- Thickness of weakness zone (m) Tz
- Strike (related to tunnel ) αz
- Dip (related to tunnel) βz
- Rock Mass index in adjacent rocks RMia
Rock support used: - bolt spacing (m) in - roof 1.5 1.5 1.5
- wall
- shotcrete (mm) (F = fibrecrete) in: - roof (F) 80 (F) 70 (F) 70
- wall
- concrete lining
Remarks:

CALCULATIONS Scale effect compr. strength fσ 0.680175601 0.644383724 0.637126946

Jointing parameter JP 0.107765 0.192613 0.191922
Rock Mass index RMi 4.310607 11.556791 11.515295
Equivalent block diameter (m) Db 0.34 0.45 0.48
Intersection between main joint set and tunnel -roof favourable favourable
-wall unfavourable unfavourable
Intersection between weakness zone and tunnel -roof
-wall
Assumed vertical stress (MPa) pz 0.81 0.81 0.81
Assumed horizontal stress (MPa) ph 1.62 1.62 1.62
Tunnel shape factors (according to Hoek & Brown, 1980): Factor A = 3.2 3.2 3.2
Factor B = 2.3 2.3 2.3
Tangential stress (MPa) in: - roof σθ r 4.37 4.37 4.37
- wall σθ w 0.24 0.24 0.24
Continuity of ground in: - roof CF discontinuous discontinuous discontinuous
- wall CF discontinuous discontinuous discontinuous
Characteristics in jointed (discontinuous) rock masses
Approximate stress level (MPa) 1.62 1.62 1.62
Stress level coefficient SL 1 1 1
Ground condition factor (RMi x SL) in: -roof Gc 4.31 11.56 11.52
(RMi x SL x wall factor) in: -wall Gc 21.55 57.78 57.58
Orientation factor for joints in: -roof Co 1.5 1.0 1.0
-wall Co 1.5 2.0 2.0
Size ratio [(Wt/Db) Co] in: - roof Sr 1203839.9 26.7 25.2
Size ratio [(Ht/Db) Co] in: - wall Sr 601920.0 26.7 25.2
Characteristics in weakness zone or fault (No weakness zone) (No weakness zone) (No weakness zone)
Orientation factor (intersection zone - tunnel) in: -roof Coz
-wall Coz
Ground condition [SL x RMi(m)] in: -roof Gcz
-wall Gcz
Size ratio [(Tz/Db)Coz] or [(Wt/Db)Coz] in: - roof Srz
Size ratio [(Tz/Db)Coz] or [(Ht/Db)Coz] in: - wall Srz
Characteristics in continuous rock masses
Competency factor in: - roof Cg
- wall Cg
Possible stress behaviour of massive rock in: - roof
- wall
Possible behaviour of continuous, particulate rock masses:
 A7 - 14

ROCK SUPPORT EXPERIENCE - CASES

Project / tunnel: Headrace tunnel, Haukrei Headrace, Vinstra
INPUT DATA Location: chainage 200 chainage 110 chainage 3250
Rock(s): gness dolerite, slightly altered phyllite, slightly altered
Rock deformability (brittle = 1 "ductile" = 2) 1 1
Tunnel span (width) (m) Wt 3.00 3.00 6.50
Tunnel wall height (m) Hw 4.00 4.00 5.00
Tunnel shape (2=high horse-shoe, 4=horse-shoe, 6=circular, 8=square) 4 4 4
Overburden (m) z 50 50.0 250.0
Rock compressive strength (MPa) σc 120.0 100.0 10.0
Assumed stress ratio (Ph /Pv) k 4.0 4.0 2.0
Joint condition factor jC 3.5 0.8 0.4
Blocks: - Block volume (m3) Vb 2 0.02 0.001
- Block shape factor β 40 40 100
Main joint set: - Average spacing (m) Sa 1.0
- Strike (related to tunnel ) αj 30
- Dip (related to tunnel) βj 70
For weakness zone: - Indicate type of weakness zone crushed, chlorite-containing crushed, clay containing
- Thickness of weakness zone (m) Tz 10.0 15.0
- Strike (related to tunnel ) αz 45 70
- Dip (related to tunnel) βz 90 20
- Rock Mass index in adjacent rocks RMia 45.0000 5.0000
Rock support used: - bolt spacing (m) in - roof 1.5 2.0
- wall 2.5
- shotcrete (mm) (F = fibrecrete) in: - roof (F) 80 (F) 200 + ribs
- wall (F) 100 + ribs
- concrete lining (could have been used)
Remarks: no rock support Initial support by fibrecrete
and rock bolts

CALCULATIONS Scale effect compr. strength fσ 0.549280272 0.751284731 1.036536462

Jointing parameter JP 0.456835 0.037386 0.005872
Rock Mass index RMi 54.820252 3.738630 0.058723
Equivalent block diameter (m) Db 1.00 0.21 0.04
Intersection between main joint set and tunnel -roof favourable
-wall unfavourable
Intersection between weakness zone and tunnel -roof favourable unfavourable
-wall unfavourable fair
Assumed vertical stress (MPa) pz 1.35 1.35 6.75
Assumed horizontal stress (MPa) ph 5.40 5.40 13.50
Tunnel shape factors (according to Hoek & Brown, 1980): Factor A = 3.2 3.2 3.2
Factor B = 2.3 2.3 2.3
Tangential stress (MPa) in: - roof σθ r 15.93 15.93 36.45
- wall σθ w -2.30 -2.30 2.03
Continuity of ground in: - roof CF continuous (massive) discontinuous continuous (particulate)
- wall CF continuous (massive) discontinuous continuous (particulate)
Characteristics in jointed (discontinuous) rock masses (The ground is continuous) (The ground is continuous)
Approximate stress level (MPa) 5.4 5.4 13.5
Stress level coefficient SL 1 1 1.5
Ground condition factor (RMi x SL) in: -roof Gc 54.82 3.74 0.09
(RMi x SL x wall factor) in: -wall Gc 274.10 18.69 0.44
Orientation factor for joints in: -roof Co 1.0 1.5 1.5
-wall Co 2.0 1.5 1.5
Size ratio [(Wt/Db) Co] in: - roof Sr 3.0 21.5 233.3
Size ratio [(Ht/Db) Co] in: - wall Sr 8.0 28.7 179.5
Characteristics in weakness zone or fault (No weakness zone)
Orientation factor (intersection zone - tunnel) in: -roof Coz 1.0 2.0
-wall Coz 2.0 1.5
Ground condition [SL x RMi(m)] in: -roof Gcz 4.7 0.1
-wall Gcz 23.4 0.6
Size ratio [(Tz/Db)Coz] or [(Wt/Db)Coz] in: - roof Srz 14.36 311.10
Size ratio [(Tz/Db)Coz] or [(Ht/Db)Coz] in: - wall Srz 38.30 179.48
Characteristics in continuous rock masses
Competency factor in: - roof Cg 3.44 0.00
- wall Cg tension 0.03
Possible stress behaviour of massive rock in: - roof high stress level
- wall lack of stresses
Possible behaviour of continuous, particulate rock masses: weakness zone
 A7 - 15

ROCK SUPPORT EXPERIENCE - CASES

Project / tunnel: Horga hydropwer plant, headrace tunnel
INPUT DATA Location: chainage 810 chainage 470 chainage 1485
Rock(s): striped gneiss striped gneiss striped gneiss
Rock deformability (brittle = 1 "ductile" = 2) 1 1 1
Tunnel span (width) (m) Wt 3.00 3.00 3.00
Tunnel wall height (m) Hw 4.00 4.00 4.00
Tunnel shape (2=high horse-shoe, 4=horse-shoe, 6=circular, 8=square) 4 4 4
Overburden (m) z 100 100.0 100.0
Rock compressive strength (MPa) σc 100.0 100.0 100.0
Assumed stress ratio (Ph /Pv) k 3.0 3.0 3.0
Joint condition factor jC 0.5 2.0 6.0
Blocks: - Block volume (m3) Vb 0.001 0.1 0.1
- Block shape factor β 50 40 40
Main joint set: - Average spacing (m) Sa 0.4
- Strike (related to tunnel ) αj 60 60
- Dip (related to tunnel) βj 10 10
For weakness zone: - Indicate type of weakness zone crushed, clay-coatings
- Thickness of weakness zone (m) Tz 4.0
- Strike (related to tunnel ) αz 90
- Dip (related to tunnel) βz 45
- Rock Mass index in adjacent rocks RMia 30
Rock support used: - bolt spacing (m) in - roof 3.0 3.0
- wall (F) 120
- shotcrete (mm) (F = fibrecrete) in: - roof (F) 50
- wall
- concrete lining
Remarks:

CALCULATIONS Scale effect compr. strength fσ 0.945042337 0.659753955 0.674856168

Jointing parameter JP 0.007507 0.134723 0.270109
Rock Mass index RMi 0.750689 13.472286 27.010933
Equivalent block diameter (m) Db 0.07 0.40 0.36
Intersection between main joint set and tunnel -roof unfavourable unfavourable
-wall favourable favourable
Intersection between weakness zone and tunnel -roof fair
-wall fair
Assumed vertical stress (MPa) pz 2.70 2.70 2.70
Assumed horizontal stress (MPa) ph 8.10 8.10 8.10
Tunnel shape factors (according to Hoek & Brown, 1980): Factor A = 3.2 3.2 3.2
Factor B = 2.3 2.3 2.3
Tangential stress (MPa) in: - roof σθ r 23.22 23.22 23.22
- wall σθ w -1.89 -1.89 -1.89
Continuity of ground in: - roof CF discontinuous discontinuous discontinuous
- wall CF discontinuous discontinuous discontinuous
Characteristics in jointed (discontinuous) rock masses
Approximate stress level (MPa) 8.1 8.1 8.1
Stress level coefficient SL 1 1 1
Ground condition factor (RMi x SL) in: -roof Gc 0.75 13.47 27.01
(RMi x SL x wall factor) in: -wall Gc 3.75 67.36 135.05
Orientation factor for joints in: -roof Co 1.5 2.0 2.0
-wall Co 1.5 1.0 1.0
Size ratio [(Wt/Db) Co] in: - roof Sr 1558331.4 15.0 16.8
Size ratio [(Ht/Db) Co] in: - wall Sr 2077775.2 10.0 11.2
Characteristics in weakness zone or fault (No weakness zone) (No weakness zone)
Orientation factor (intersection zone - tunnel) in: -roof Coz 1.5
-wall Coz 1.5
Ground condition [SL x RMi(m)] in: -roof Gcz 1.6
-wall Gcz 7.8
Size ratio [(Tz/Db)Coz] or [(Wt/Db)Coz] in: - roof Srz 67.84
Size ratio [(Tz/Db)Coz] or [(Ht/Db)Coz] in: - wall Srz 90.46
Characteristics in continuous rock masses
Competency factor in: - roof Cg
- wall Cg
Possible stress behaviour of massive rock in: - roof
- wall
Possible behaviour of continuous, particulate rock masses:
 A7 - 16

ROCK SUPPORT EXPERIENCE - CASES

Project / tunnel: Tromsö road tunnel to Breivika Nappstraumen road t.
INPUT DATA Location: tunnel (average cond.) roundabout in rock typical rock conditions
Rock(s): grey, striped gneiss grey, striped gneiss gneiss
Rock deformability (brittle = 1 "ductile" = 2)
Tunnel span (width) (m) Wt 10.00 20.00 10.00
Tunnel wall height (m) Hw 5.00 8.00 5.00
Tunnel shape (2=high horse-shoe, 4=horse-shoe, 6=circular, 8=square) 4 4 4
Overburden (m) z 30 30 40
Rock compressive strength (MPa) σc 100.0 100.0 120.0
Assumed stress ratio (Ph /Pv) k 3.0 3.0 3.0
Joint condition factor jC 1.3 1.3 2.0
Blocks: - Block volume (m3) Vb 3 3 2
- Block shape factor β 40 40 35
Main joint set: - Average spacing (m) Sa 1.0
- Strike (related to tunnel ) αj
- Dip (related to tunnel) βj
For weakness zone: - Indicate type of weakness zone
- Thickness of weakness zone (m) Tz
- Strike (related to tunnel ) αz
- Dip (related to tunnel) βz
- Rock Mass index in adjacent rocks RMia
Rock support used: - bolt spacing (m) in - roof 2.5 2.0 2,5 - 3
- wall
- shotcrete (mm) (F = fibrecrete) in: - roof (F) 50
- wall
- concrete lining
Remarks:

CALCULATIONS Scale effect compr. strength fσ 0.537955459 0.537955459 0.549280272

Jointing parameter JP 0.329849 0.329849 0.353596
Rock Mass index RMi 32.984854 32.984854 42.431572
Equivalent block diameter (m) Db 1.11 1.11 1.00
Intersection between main joint set and tunnel -roof
-wall
Intersection between weakness zone and tunnel -roof
-wall
Assumed vertical stress (MPa) pz 0.81 0.81 1.08
Assumed horizontal stress (MPa) ph 2.43 2.43 3.24
Tunnel shape factors (according to Hoek & Brown, 1980): Factor A = 3.2 3.2 3.2
Factor B = 2.3 2.3 2.3
Tangential stress (MPa) in: - roof σθ r 6.97 6.97 9.29
- wall σθ w -0.57 -0.57 -0.76
Continuity of ground in: - roof CF discontinuous discontinuous discontinuous
- wall CF continuous (massive) discontinuous discontinuous
Characteristics in jointed (discontinuous) rock masses
Approximate stress level (MPa) 2.43 2.43 3.24
Stress level coefficient SL 1 1 1
Ground condition factor (RMi x SL) in: -roof Gc 32.98 32.98 42.43
(RMi x SL x wall factor) in: -wall Gc 164.92 164.92 212.16
Orientation factor for joints in: -roof Co 1.5 1.5 1.5
-wall Co 1.5 1.5 1.5
Size ratio [(Wt/Db) Co] in: - roof Sr 310465.9 620931.7 344550.0
Size ratio [(Ht/Db) Co] in: - wall Sr 155232.9 248372.7 172275.0
Characteristics in weakness zone or fault (No weakness zone) (No weakness zone) (No weakness zone)
Orientation factor (intersection zone - tunnel) in: -roof Coz
-wall Coz
Ground condition [SL x RMi(m)] in: -roof Gcz
-wall Gcz
Size ratio [(Tz/Db)Coz] or [(Wt/Db)Coz] in: - roof Srz
Size ratio [(Tz/Db)Coz] or [(Ht/Db)Coz] in: - wall Srz
Characteristics in continuous rock masses
Competency factor in: - roof Cg
- wall Cg tension
Possible stress behaviour of massive rock in: - roof
- wall
Possible behaviour of continuous, particulate rock masses:
 A7 - 17

ROCK SUPPORT EXPERIENCE - CASES

Project / tunnel: Stetind road tunnel Njunis access tunnel
INPUT DATA Location: chainage 15750 chainage 6250 chainage 6300
Rock(s): granitic gneis, coarse-grained amphibolite amphibolite
Rock deformability (brittle = 1 "ductile" = 2) 1 1 1
Tunnel span (width) (m) Wt 10.00 3.00 3.00
Tunnel wall height (m) Hw 5.00 4.00 4.00
Tunnel shape (2=high horse-shoe, 4=horse-shoe, 6=circular, 8=square) 4 4 4
Overburden (m) z 900 200 200
Rock compressive strength (MPa) σc 90.0 200.0 200.0
Assumed stress ratio (Ph /Pv) k 3.0 5.0 5.0
Joint condition factor jC 3.0 3.0 0.3
Blocks: - Block volume (m3) Vb 30 0.3 0.005
- Block shape factor β 40 40 40
Main joint set: - Average spacing (m) Sa 0.5
- Strike (related to tunnel ) αj 45
- Dip (related to tunnel) βj 10
For weakness zone: - Indicate type of weakness zone crushed, clay-containing
- Thickness of weakness zone (m) Tz 6.0
- Strike (related to tunnel ) αz 70
- Dip (related to tunnel) βz 90
- Rock Mass index in adjacent rocks RMia 40.00
Rock support used: - bolt spacing (m) in - roof 1.5 spot bolting 1.5
- wall 2,5 ?
- shotcrete (mm) (F = fibrecrete) in: - roof (F) 50 - 80 (F) 60
- wall
- concrete lining
Remarks:

CALCULATIONS Scale effect compr. strength fσ 0.461409274 0.630957344 0.824020244

Jointing parameter JP 0.951299 0.242264 0.009045
Rock Mass index RMi 85.616877 48.452715 1.809018
Equivalent block diameter (m) Db 2.39 0.50 0.13
Intersection between main joint set and tunnel -roof unfavourable
-wall favourable
Intersection between weakness zone and tunnel -roof favourable
-wall unfavourable
Assumed vertical stress (MPa) pz 24.30 5.40 5.40
Assumed horizontal stress (MPa) ph 72.90 27.00 27.00
Tunnel shape factors (according to Hoek & Brown, 1980): Factor A = 3.2 3.2 3.2
Factor B = 2.3 2.3 2.3
Tangential stress (MPa) in: - roof σθ r 208.98 81.00 81.00
- wall σθ w -17.01 -14.58 -14.58
Continuity of ground in: - roof CF continuous (massive) discontinuous discontinuous
- wall CF continuous (massive) discontinuous discontinuous
Characteristics in jointed (discontinuous) rock masses (The ground is continuous)
Approximate stress level (MPa) 27 27
Stress level coefficient SL 1.5 1.5
Ground condition factor (RMi x SL) in: -roof Gc 72.68 2.71
(RMi x SL x wall factor) in: -wall Gc 363.40 13.57
Orientation factor for joints in: -roof Co 2.0 1.5
-wall Co 1.0 1.5
Size ratio [(Wt/Db) Co] in: - roof Sr 12.0 785403.6
Size ratio [(Ht/Db) Co] in: - wall Sr 8.0 1047204.9
Characteristics in weakness zone or fault (No weakness zone) (No weakness zone)
Orientation factor (intersection zone - tunnel) in: -roof Coz 1.0
-wall Coz 2.0
Ground condition [SL x RMi(m)] in: -roof Gcz 4.2
-wall Gcz 21.1
Size ratio [(Tz/Db)Coz] or [(Wt/Db)Coz] in: - roof Srz 22.80
Size ratio [(Tz/Db)Coz] or [(Ht/Db)Coz] in: - wall Srz 60.79
Characteristics in continuous rock masses
Competency factor in: - roof Cg 0.41
- wall Cg tension
Possible stress behaviour of massive rock in: - roof heavy rock burst
- wall lack of stresses
Possible behaviour of continuous, particulate rock masses:
 A7 - 18

ROCK SUPPORT EXPERIENCE - CASES

Project / tunnel: Sumbiar road tunnel, Suderey, the Faroe Islands
INPUT DATA Location: chainage 650 chainage 1315 chainage 2100
Rock(s): basalt basalt basalt
Rock deformability (brittle = 1 "ductile" = 2) 1 1 1
Tunnel span (width) (m) Wt 10.00 10.00 10.00
Tunnel wall height (m) Hw 5.00 5.00 5.00
Tunnel shape (2=high horse-shoe, 4=horse-shoe, 6=circular, 8=square) 4 4 4
Overburden (m) z 200 200 200
Rock compressive strength (MPa) σc 200.0 200.0 200.0
Assumed stress ratio (Ph /Pv) k 1.5 1.5 1.5
Joint condition factor jC 1.5 7.0 0.8
Blocks: - Block volume (m3) Vb 1 0.025 0.05
- Block shape factor β 32 35 35
Main joint set: - Average spacing (m) Sa 0.8
- Strike (related to tunnel ) αj 75
- Dip (related to tunnel) βj 90
For weakness zone: - Indicate type of weakness zone
- Thickness of weakness zone (m) Tz
- Strike (related to tunnel ) αz
- Dip (related to tunnel) βz
- Rock Mass index in adjacent rocks RMia
Rock support used: - bolt spacing (m) in - roof
- wall
- shotcrete (mm) (F = fibrecrete) in: - roof 50.0 50.0
- wall 50.0
- concrete lining
Remarks: spot bolting

CALCULATIONS Scale effect compr. strength fσ 0.574349177 0.72713166 0.694298769

Jointing parameter JP 0.244949 0.209854 0.053539
Rock Mass index RMi 48.989795 41.970764 10.707801
Equivalent block diameter (m) Db 0.80 0.25 0.31
Intersection between main joint set and tunnel -roof favourable
-wall unfavourable
Intersection between weakness zone and tunnel -roof
-wall
Assumed vertical stress (MPa) pz 5.40 5.40 5.40
Assumed horizontal stress (MPa) ph 8.10 8.10 8.10
Tunnel shape factors (according to Hoek & Brown, 1980): Factor A = 3.2 3.2 3.2
Factor B = 2.3 2.3 2.3
Tangential stress (MPa) in: - roof σθ r 20.52 20.52 20.52
- wall σθ w 4.32 4.32 4.32
Continuity of ground in: - roof CF discontinuous discontinuous discontinuous
- wall CF discontinuous discontinuous discontinuous
Characteristics in jointed (discontinuous) rock masses
Approximate stress level (MPa) 8.1 8.1 8.1
Stress level coefficient SL 1 1 1
Ground condition factor (RMi x SL) in: -roof Gc 48.99 41.97 10.71
(RMi x SL x wall factor) in: -wall Gc 244.95 209.85 53.54
Orientation factor for joints in: -roof Co 1.0 1.5 1.5
-wall Co 2.0 1.5 1.5
Size ratio [(Wt/Db) Co] in: - roof Sr 12.5 1400707.7 1111768.1
Size ratio [(Ht/Db) Co] in: - wall Sr 12.5 700353.9 555884.1
Characteristics in weakness zone or fault (No weakness zone) (No weakness zone) (No weakness zone)
Orientation factor (intersection zone - tunnel) in: -roof Coz
-wall Coz
Ground condition [SL x RMi(m)] in: -roof Gcz
-wall Gcz
Size ratio [(Tz/Db)Coz] or [(Wt/Db)Coz] in: - roof Srz
Size ratio [(Tz/Db)Coz] or [(Ht/Db)Coz] in: - wall Srz
Characteristics in continuous rock masses
Competency factor in: - roof Cg
- wall Cg
Possible stress behaviour of massive rock in: - roof
- wall
Possible behaviour of continuous, particulate rock masses:
 A7 - 19

ROCK SUPPORT EXPERIENCE - CASES

Project / tunnel: Sumbiar road tunnel Thingbæk chalk mine
INPUT DATA Location: chainage 600 old mine
Rock(s): basalt chalk
Rock deformability (brittle = 1 "ductile" = 2) 1
Tunnel span (width) (m) Wt 10.00 5.00
Tunnel wall height (m) Hw 5.00 7.00
Tunnel shape (2=high horse-shoe, 4=horse-shoe, 6=circular, 8=square) 4 4
Overburden (m) z 200 20
Rock compressive strength (MPa) σc 200.0 2.0
Assumed stress ratio (Ph /Pv) k 1.5 1.0
Joint condition factor jC 0.5 3.0
Blocks: - Block volume (m3) Vb 0.1 20
- Block shape factor β 40 45
Main joint set: - Average spacing (m) Sa 1.5
- Strike (related to tunnel ) αj 45
- Dip (related to tunnel) βj 10
For weakness zone: - Indicate type of weakness zone crushed, clay coatings
- Thickness of weakness zone (m) Tz 1.5
- Strike (related to tunnel ) αz 65
- Dip (related to tunnel) βz 90
- Rock Mass index in adjacent rocks RMia 40.00
Rock support used: - bolt spacing (m) in - roof 1.5 spot bolting
- wall
- shotcrete (mm) (F = fibrecrete) in: - roof
- wall
- concrete lining
Remarks: straps and wire mesh in roof some loosening of rock
fragments in surface
CALCULATIONS Scale effect compr. strength fσ 0.674856168 0.506495684
Jointing parameter JP 0.053149 0.843364
Rock Mass index RMi 10.629837 1.686729
Equivalent block diameter (m) Db 0.36 1.50
Intersection between main joint set and tunnel -roof unfavourable
-wall favourable
Intersection between weakness zone and tunnel -roof favourable
-wall unfavourable
Assumed vertical stress (MPa) pz 5.40 0.54
Assumed horizontal stress (MPa) ph 8.10 0.54
Tunnel shape factors (according to Hoek & Brown, 1980): Factor A = 3.2 3.2
Factor B = 2.3 2.3
Tangential stress (MPa) in: - roof σθ r 20.52 1.19
- wall σθ w 4.32 0.70
Continuity of ground in: - roof CF discontinuous continuous (massive)
- wall CF discontinuous continuous (massive)
Characteristics in jointed (discontinuous) rock masses (The ground is continuous)
Approximate stress level (MPa) 8.1 0.54
Stress level coefficient SL 1 0.5
Ground condition factor (RMi x SL) in: -roof Gc 10.63 0.84
(RMi x SL x wall factor) in: -wall Gc 53.15 4.22
Orientation factor for joints in: -roof Co 1.5 2.0
-wall Co 1.5 1.0
Size ratio [(Wt/Db) Co] in: - roof Sr 964580.2 6.7
Size ratio [(Ht/Db) Co] in: - wall Sr 482290.1 4.7
Characteristics in weakness zone or fault (No weakness zone)
Orientation factor (intersection zone - tunnel) in: -roof Coz 1.0
-wall Coz 2.0
Ground condition [SL x RMi(m)] in: -roof Gcz 18.1
-wall Gcz 90.3
Size ratio [(Tz/Db)Coz] or [(Wt/Db)Coz] in: - roof Srz 4.20
Size ratio [(Tz/Db)Coz] or [(Ht/Db)Coz] in: - wall Srz 8.40
Characteristics in continuous rock masses
Competency factor in: - roof Cg 1.42
- wall Cg 2.40
Possible stress behaviour of massive rock in: - roof no info on rock deform.
- wall info on rock def. required
Possible behaviour of continuous, particulate rock masses:
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

APPENDIX 8

COLLECTED DATA ON GROUND CONDITION AND TBM

BORING PERFORMANCE

"The geotechnical engineer should apply theory and experimentation but temper them by
putting them into the context of the uncertainty of nature. Judgement enters through engi-
neering geology."
Karl Terzaghi, 1961

The data applied to correlate the model for TBM boring performance in Chapter 7 with
practical experience have been collected from Svartisen hydropower plant and Meråker
hydropower plant, both in Norway. The conditions used are described in the following.

1. Boring experience and ground conditions at Svartisen Power plant

58 km of the 79 km water tunnels at Svartisen hydropower plant, Nordland county has been
excavated by tunnel boring machines during the years 1988 - 1993. The tunnels have
diameter between 3.5 m and 8.5 m.

Data from three of the tunnels have been used in this work to find how the predicted
performance developed in Chapter 7 fits with the boring advance obtained. The specifica-
tions of the TBM machines are as follows

TBM machine: Robbin high performance type 251-1 type 252 type 257

Diameter of TBM 5.0 m 4.3 m 3.5 m

Recommended max. thrust per cutter, MB 314 kN 314 kN 314 kN
Cutter spacing, SC 100 mm 100 mm 90 mm
Number of cutters 35 29 25
Cutter size 19" (483 mm) 19" 19"
Cutterhead RPM 11.94 11.94 12.6

NTH factor for cutter diameter, kd 1 1 1

NTH factor for cutter spacing, ka 0.85 0.85 0.9

1.1 Measured rock properties

The rocks in the three tunnels are of Precambrian - Cambrian age. During metamorphism
the sedimentary rocks were changed to mica schist, various gneisses, marble and quartzite.
Small sized core drilling, diameter 32 mm, was used to collect samples from three of the
tunnels. The results from the laboratory test are shown in Table A8-1. Uniaxial compres-
sive strength could not be measured due to the small sample size.
 A8 - 2

TABLE A8-1 RESULTS FROM LABORATORY TEST ON WET SAMPLES FROM SVARTISEN
LOCATION SONIC POINT LOAD STRENGTH Calculated
VELOCITY σc = k50 × Is50
TYPE OF
TBM chainage velocity angle test strength no. of thickness ROCK k50 σc
direc- Is50 tests of sample
(m/s) tion (Mpa) (mm) (Mpa)
1200 4680 60o 2.0 10 32 Marble 14 30
║ 1.1 8 32 14 15
252 2950 4120 45o Mica schist
⊥ 7.9 4 12 - 16 20 160
║ 6.9 6 32 Gneiss- 20 140
6480 4435 -
⊥ 8.5 8 13 - 25 granite 20 170
10180 5260 35o ║ 5.5 6 32 Mica gneiss 16 90
║ 0.6 5 32 Mica schist 14 10
11755 3160 45o
⊥ 4.7 5 12 - 18 (laminated) 16 75
║ 2.0 3 32 14 30
1650 5525 0o Mica gneiss
⊥ 6.5 6 14 - 16 16 100
257 ║ 1.1 4 32 14 15
3965 3910 35o Mica gneiss
⊥ 3.0 3 12 - 16 14 40
Mica gneiss
4180 3850 40o ║ 1.1 6 32 14 15
(mica layers)
║ 1.5 2 32 Mica schist 14 20
6335 5585 5-10o
⊥ 5.3 2 14 - 18 (laminated) 16 85
251-1 ║ 3.6 5 32 Marble 16 50
7460 6110 35o
⊥ 6.6 7 16 - 32 (layered) 20 130
8445 5150 30o ║ 2.8 4 32 Mica schist 14 40

In addition, laboratory tests carried out by the Robbins company gave the results in Table
A8-2.

TABLE A8-2 LABORATORY TEST PERFORMED BY THE ROBBINS COMPANY

LOCATION COMPRESSIVE POINT LOAD Calculated
STRENGTH STRENGTH TYPE OF σc = k50 Is50
σc ROCK σc
TBM chainage angle*) test direc- Is50 k50
(Mpa) (Mpa)
tion (Mpa)
║ 1.7 14 20
251-1 8142 10o 113 Mica schist
⊥ 6.1 18 110
║ 1.6 14 20
8400 10o 110 Mica schist
⊥ 4.7 16 75
║ 2.6 14 35
8456 10o 119 Mica schist
⊥ 4.8 16 75
║ 1.8 14 25
8494 10o 75 Mica schist
⊥ 6.1 18 110
*)
Angle between loading and schistosity

The drilling rate index for mica schist and mica gneiss is DRI = 55, for marble DRI = 80.
 A8 - 3

1.2 Rock mass description in tunnel locations

The detailed registration of the rock mass conditions in the tunnels, Table A8-3, has been worked
out from observations made by graduate students at the Technical University of Norway (NTH)
in connection with their final dissertation work in tunnel engineering; it is referred to Isaksen and
Solberg (1990) and Stang and Aadal (1991). The mica rich rocks encountered in the three tunnels
are strongly anisotropic with discontinuities occurring as foliation partings (fissures) spaced 0.1
m - 1 m. Some few random joints occur. Generally, the joints and fissures do not delimit defined
blocks.

TABLE A8-3 DESCRIPTION OF THE GROUND CONDITIONS AT VARIOUS LOCATIONS

LOCATION
DESCRIPTION
TBM chainage
Relatively massive marble with some actinolite. Joint spacing S = 1.6 m, angle with folia-
252 1200
tion α = 35o. Applied thrust per cutter MB = 236 kN.
Mica schist, occasionally containing garnets. The rock is dark often with veins or lenses of
2950 quartzpegmatite. Thin layers of mica often occur along foliation, and the rock is partly
folded. Spacing of foliation partings (fissures) S = 0.2 m, α = 10o. MB = 236 kN.
Granitic gneiss. The rock is light grey and folded. Some few thin layers of mica occur along
6480 the foliation. High content of feldspar; hard rock to bore in. Joint spacing S = 0.8 m,
α = 90o. MB = 265 kN.
Folded mica gneiss, often with quartzpegmatite. Some garnet and thin mica layers along
10180 foliation. Many of the joints have mica filling. Foliation partings (fissures) are spaced
S = 0.2 m, α = 44o. MB = 236 kN.
Folded mica gneiss containing some quartzpegmatite and some few enrichments of mica
11755 occurring as veins, thin zones. Dark diorite and granitic gneiss partly occur. Foliation
partings are spaced S = 0.3 m, α = 27o, MB = 228 kN..
Partly folded mica schist, frequently containing quartzpegmatite. Some veins or thin layers
257 1650
of mica occur. Spacing of foliation partings (fissures) S = 0.2 m, α = 0o. MB = 298 kN.
Mica gneiss with quartzpegmatite and veins of granitic gneiss and layers of mica. Foliation
3965
partings are spaced S = 0.4 m, α = 15o. MB = 289 kN.
Same as 3965, but the degree of jointing is higher. Foliation partings are spaced S = 0.3 m
4180
α = 15o. MB = 289 kN.
Mica schist with some quartzpegmatite veins. Foliation partings are spaced S = 0.075 m,
251-1 6335
α = 25o. MB = 160 kN.
Marble with a few veins containing amphibole, sometimes folded. Joint spacing S = 0.8 m,
7460
MB = 240 kN.
Mica gneiss with quartzpegmatite, sometimes folded, sometimes schistose, sometimes
8400 quartzrich. Veins of mica are present. Foliation partings spaced S = 0.4 m, α = 15o.
MB = 272 kN.
8445 Rock: same as 8400. Foliation partings spaced S = 0.3 m, α = 15o, MB = 240 kN.
8456 Rock: same as 8400. Foliation partings are spaced S = 0.3 m, α = 15o, MB = 280 kN.
8494 Rock: same as 8400. Foliation partings are spaced S = 0.4 m, α = 20o, MB = 240 kN.

The marble is almost massive, also the gneissgranite generally contains few joints. In these types
of rocks the discontinuities occur mainly as joints. Table 7-2 in Chapter 7 has been applied to
find the equivalent block volumes from the spacings. The results of the calculations made to find
the TBM jointing factor are shown in Table A8-4.
 A8 - 4

TABLE A8-4 ROCK MASS FEATURES AND CALCULATED JOINTING FACTORS

------------------------------------------------------------------------------------------------------------------------------------------------
TBM Compr. Rock OBSERVATIONS CALCULATIONS*)
q and Rock type strength factor joint jC angle eq. Vb co JP keq
chainage σc E spacing (m3)
------------------------------------------------------------------------------------------------------------------------------------------------
252
1200 marble 30 1000 massive 1.0 0.74 (0.9)
2950 mica schist 15 750 0.2 m 6 10o 1.5 1 0.54 1.35 (1)
6480 gneissgranite 140 1000 0.8 m 2 90o 50 1.5 0.98 0.66 (0.75)
10180 mica gneiss 90 750 0.2 m 6 40o 1.5 1.5 0.54 1.18 (1.25)
11755 mica gneiss 10 750 0.3 m 6 27o 4 1.25 0.70 1.47 (1.1)

257
1650 mica schist 30 750 0.2 m 6 0o 1.5 1 0.54 1.10 (1.1)
3965 mica gneiss 15 750 0.4 m 6 15o 10 1.25 0.89 1.0 (0.8)
4180 mica gneiss 15 750 0.3 m 6 15o 4 1.25 0.70 1.30 (1.0)

251-1
6335 mica schist 20 750 75 mm 6 25o 0.08 1.25 0.25 3.35 (2.27)
7460 marble 50 1000 0.8 m 2 - 50 1 0.98 0.60 (0.42)
8400 mica schist 20 750 0.4 m 6 15o 10 1.25 0.89 0.94 (0.68)
8445 mica schist 40 750 0.3 m 6 15o 4 1.25 0.70 0.97 (1.0)
8456 mica schist 35 750 0.3 m 6 15o 4 1.25 0.70 1.0 (1.0)
8494 mica schist 25 750 0.4 m 6 20o 10 1.25 0.89 0.88 (0.74)
------------------------------------------------------------------------------------------------------------------------------------------------
*)
The values in brackets have been found using the NHT model, version 5.

TABLE A8-5 MACHINE PARAMETERS AND DRILLING RATE

------------------------------------------------------------------------------------------------------------------------------------
TBM Applied TBM jointing CALCULATIONS*) Real
and thrust factor Eq. thrust ----- TBM penetration ----- boring
chainage MB keq Meq per rev. boring rate rate
( kN) (kN) io (mm) I (m/h) I (m/h)
------------------------------------------------------------------------------------------------------------------------------------
252
1200 236 0.74 (0.9) 200 3.0 (3.6) 2.1 (2.6) 2.89
2950 236 1.35 (1) 200 4.8 (4.0) 3.4 (2.9) 4.52
6480 265 0.66 (0.75) 225 3.0 (3.6) 2.1 (2.6) 1.47
10180 236 1.18 (1.25) 200 4.4 (4.6) 3.1 (3.3) 4.8
11755 228 1.47 (1.1) 195 5.0 (4.2) 3.6 (2.9) 4.3

257
1650 298 1.10 (1.1) 268 6.4 (6.4) 4.8 (4.8) 3.14
3965 289 1.0 (0.8) 260 6.1 (5.4) 4.6 (4.1) 3.86
4180 289 1.30 (1.0) 260 6.8 (6.4) 5.1 (4.8) 4.1

251-1
6335 160 3.35 (2.27) 135 4.5 (4.0) 3.2 (2.9) 3.85
7460 240 0.60 (0.42) 205 2.7 (1.6) 1.9 (1.2) 3.54
8400 272 0.94 (0.68) 230 4.8 (3.6) 3.4 (2.6) 4.0
8445 272 0.97 (1.0) 230 4.8 (4.8) 3.4 (3.4) 4.5
8456 280 1.0 (1.0) 240 5.3 (5.3) 3.8 (3.8) 5.06
8494 240 0.88 (0.74) 205 3.8 (3.2) 2.7 (2.3) 4.19
------------------------------------------------------------------------------------------------------------------------------------
*)
The values in brackets have been found using the NHT model, version 5.
 A8 - 5

2. Boring experience and ground conditions at Meråker hydropower plant

9634 metres of the diversion tunnel Dalåa - Torsbjørka at Meråker hydropower plant,
Tröndelag county has been excavated using a full face tunnel boring machine. The boring
performed 1991 - 1992, started in greywacke and meta-gabbro with unconfined compres-
sive strength 180 - 300 MPa. In fact, the first 400 m of this tunnel was excavated by drill &
blast, because the meta-gabbro in this part was assumed too strong for economic TBM
boring. Later, a section more than 50 m long was successfully bored in this type of rock.
The first of the two locations described in the following, is from the meta-gabbro.

The average net boring advance for the whole tunnel was 6.38 m/h. More than 1000
m/month has been achieved from the beginning, including boring through the strong meta-
gabbro mentioned above. The TBM machine specifications are:

Machine: Robbins high performance TBM, type 265

Diameter of TBM 3.5 m

Recommended max. thrust per cutter, MB 312 kN
Cutter spacing, SC 90 mm
Number of cutters 26
Cutter size 19" (483 mm)
Cutterhead RPM 13.4

NTH factor for cutter diameter, kd 1.0

NTH factor for cutter spacing, ka 0.9

The actual tunnel is located in the Caledonian mountain range. The rocks are mainly of
sedimentary or volcanic origin, but have later through metamorphism been changed into
phyllite/clayschist and greenstone, respectively. Some eruptive intrusions of gabbro occur
occasionally.

The rock mass condition in the tunnel have been mapped in two locations; one in meta-
gabbro (chainage 750) and the other in greenstone (chainage 10020). Point load strength
tests have been made on wet and dry rock samples as shown in Table A8-6.

TABLE A8-6 RESULTS FROM LABORATORY TEST ON SAMPLES FROM MERÅKER

LOCATION SOUND POINT LOAD STRENGTH Calculated
VELOCITY σc = k50⋅ Is50
Condition Test TYPE OF
TBM chainage of sample direc- v Is50 no. of thickness ROCK k50 σc
tion tests of sample
(m/s) (Mpa) (mm) (Mpa)

265 750 wet 8000 12.0 1) 17 20 - 32 Meta- 25 300

dry 4100 13.4 9 28 - 29 gabbro

10020 wet ║ 5160 1.9 6 25 - 26 Greenstone 14 30

⊥ 4700 7.1 2) 12 18 - 25 (laminated) 20 140
dry ⊥ 3500 14.8 10 15 - 21
1)
Tests carried out by SINTEF, Trondheim gave Is50 = 11.0 MPa
2)
Tests carried out by SINTEF, Trondheim gave Is50 = 7.1 MPa
 A8 - 6

2.1 Observations at chainage 750

The rock consists of meta-gabbro, slightly chloritized. It contains irregularly distributed

white stripes, partly as thin layers up to 2 m long, and lenses 0.1 - 0.2 m thick up to 0.5 m
long. The drilling rate index determined at SINTEF, Trondheim is DRI = 24 - 45 (average
DRI = 35).

The 1-3 m long joints are unevenly distributed. They are slightly undulating and rough with
spacing 0.3 -1 m. Some of the white stripes in the rock represent weakness planes. The
block size is mainly Vb = 0.25 - 1 m3. This could more easily be observed in the
drill&blast part of the tunnel excavated before TBM started (chainage 0 - 400).
The applied TBM thrust MB = 275 kN/cutter resulted in a net advance I = 6 m/h.

2.2 Observations at the brook intake, chainage 10020

The rock consists of a laminated greenstone containing many white and light coloured
stripes. It splits along thin foliation partings with chlorite/mica which are spaced 0.03 m -
0.2 m. The drilling rate index has been measured as DRI = 43.
The strike of the foliation is 20 - 30o related to the orientation of the tunnel with dip 30o
towards the intake.

The joints consist mainly of the 0.1 - 2 m long foliation partings which have slightly un-
dulating, smooth joint surfaces. Occasionally more massive (approx. 1 m thick) 'layers'
occur. Where locally some few 3 - 4 m long random joints occur the expereince was that
the same boring advance could be achieved with less cutter thrust.

The applied TBM thrust is MB = 265 kN/cutter which resulted in a net advance
I = 7.7 m/h.

TABLE A8-6 ROCK MASS FEATURES AND CALCULATED JOINTING FACTORS

----------------------------------------------------------------------------------------------------------------------------
Compr. Rock OBSERVATIONS CALCULATIONS s
Chainage Rock type strength factor Vb jC angle co JP keq
σc E (m3)
----------------------------------------------------------------------------------------------------------------------------
750 meta-gabbro 300 1000 0.5 2 0 - 5o 1 0.22 1.55
10020 greenstone 30 750 0.2 2 12o 1 0.17 3.5
----------------------------------------------------------------------------------------------------------------------------

TABLE A8-7 MACHINE PARAMETERS AND DRILLING RATE

------------------------------------------------------------------------------------------------------------------------------------
Applied TBM jointing CALCULATIONS Real
Chainage thrust factor Eq.thrust ---- TBM penetration ---- boring
MB keq Meq per rev. boring rate rate
( kN) (kN) io (mm) I (m/h) I (m/h)
------------------------------------------------------------------------------------------------------------------------------------
750 275 1.55 250 6.8 5.5 6.0 1)
10200 265 3.5 240 8.3 6.7 7.7 2)
------------------------------------------------------------------------------------------------------------------------------------
1)
Calculations carried out by NTH: average net advance I = 4.05 m/h with 275 kN/cutter and DRI = 40.
2)
NTH calculation: average net advance I = 5.19 m/h with thrust MB = 265 kN/cutter and DRI = 42.
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

APPENDIX 9

A METHOD TO ESTIMATE THE TANGENTIAL STRESSES AROUND

UNDERGROUND OPENINGS

"Soil mechanics is the study of the strength of soils while rock mechanics is the study of the
weaknesses of rock."
A. M. Muir Wood, 1979

The stresses developed in the ground surrounding an underground opening are mainly a result of the
original, in situ (virgin) stresses, the impact from the excavation works, and the dimensions and
shape of the opening. Their distribution may, however, be influenced by joints occurring around the
opening.

1 Estimating the magnitude of the in situ ground stresses

Several authors have contributed to the understanding and knowledge of ground stresses in the
earth's crust from in-situ measurements. Many of the results from these have been summarized and
linear regression analyses performed to find the distribution by depth. Fig. A9-1 shows a summary
of some results.
pz (MPa) ph (MPa)
0 50 100 0 50 100
z z
HAST (1973)
(m) (m)
σH = 9.31 + 0.05 H
HAST (1973)
σV = 0.026 H
1000 1000
HERGET (1974) HERGET (1974)
σV = 1.9 + 0.0266 H σH = 8.3 + 0.0407 H
WOROTNIKI WOROTNIKI (1976)
(1976) σ H = 7.7 + 0.021
2000 σV = 0.023 H 2000
VAN HEERDEN
VAN HEERDEN (1976) (1976)
σV = 0.027 H σ H = 6.7 + 0.012 H

3000 3000 ORR (1975)

σV = 0.025 H σH = 6.5 + 0.015 H σHav = 4 + 0.0215 H

HAIMSON (1976)
4000 σV = 0.0226 H 4000 σHmin = 2 + 0.017 H

HAIMSON
(1976)

5000 5000

Fig. A9-1 Vertical and horizontal stresses versus depth below surface according to various authors.
Left: Vertical stresses. Right: Horizontal stresses (from Bieniawski, 1984)
 A9 - 2

Hoek (1981) found that the approximate increase of the vertical stress in excess of 1000 m depth
can be reasonably well predicted by:
pz = 0.027 z eq. (A9-1)

where pz is the vertical stress (in MPa), and

z is the depth below surface (in m).

Ph / Pz
0 1 2 3 4

H (m)

500

HAST (1973)
SCANDINAVIA
AFRICA

1000
VAN HEERDEN (1976) SOUTH

HERGET (1974)
CANADA
1500
WOROTNICKI (1976)
AUSTRALIA

2000

2500 HAIMSON (1976)

USA

3000

Fig. A9-2 Variation of ratio of average horizontal stresses to vertical stresses versus depth below surface
according to various authors (from Bieniawski, 1984).

For the horizontal stresses there is not a similar general increase with depth (Fig. A9-1 right and Fig.
A9-2). Especially in the upper 500 meters the horizontal stresses are generally higher the vertical
stresses. No simple method exists for estimating the horizontal stresses which often vary in
magnitude and direction (Fig. A9-3). The following trends of the horizontal stresses were
formulated by Hoek (1981):
− With the exception of deep level South African gold mines, average horizontal stresses are
generally higher than vertical stresses for depths of less than 1,000 m below surface.
− At a depth of 500 m below surface, the average horizontal stress is approximately 1.5 times
the vertical stress with higher ratios being evident at shallower depths.
− For depths in excess of 1,000 m below surface, the horizontal and vertical stresses tend to
equalize, except in South African mines in quartzites where the ratio of average horizontal to
vertical stress is 0.75.
− In the Scandinavian Precambrian and Palaeozoic and in the Canadian crystalline rocks the
horizontal stresses are significantly higher than the vertical stress down to a few hundred
meters.
 A9 - 3

Where the stresses cannot be measured they may be evaluated from theory and the stress conditions
experienced at other nearby locations. There are, however, features that tend to complicate the stress
pattern underground. Bieniawski (1984) is of the opinion that in a competent rock the horizontal
stress probably varies considerably from its mean value on account of jointing and inhomogeneity.
In addition, tectonic features may cause k = ph /pz to vary.

VADSÖ

TROMSÖ ALTA

NARVIK

BODÖ
ARTIC CIRCLE

PRECAMBRIAN
PERMIAN
CAMBRIAN - SILURIAN
E
NG
RA

TRONDHEIM MEASURING SITE

N
A

30 MPa
NI
O
LED
CA

BERGEN

OSLO
STAVANGER

KRISTIANSAND

Fig. A9-3 Directions and magnitudes of measured horizontal stresses in Norway (from Nilsen and Thidemann,
1993, based on data from Myrvang)

Similarly, Selmer-Olsen (1964 and 1988) and Muir Wood (1979) mention that in the vicinity of
steep cliffs or valley sides a strongly anisotropic stress distribution is often found (Fig. A9-4). Thus,
the regional stresses may be influenced by local perturbations arising from topographic relief, or
from geological structures such as faults or folds, or adjacent excavations. Where the ground
stresses may strongly influence the stability of a planned underground excavation, stress
measurements should be carried out. Haimson (1993) suggests accurate information on the local
stress regime within the volume of rock containing the underground structure that is to be designed
and excavated (typically a block of 10 - 1000 m on each side). Also Bieniawski (1984)
recommends, in critical cases or in cases where the rock mass has been subjected to recent tectonic
activity, in-situ measurements of rock stress be carried out as early in the project as possible.
 A9 - 4

For the method of estimating rock support in discontinuous (jointed) rock masses, described in
Section 4.1 in Chapter 6, only a rough estimate of the stress level is required. For continuous rock
masses, however, the effect of rock stresses developed around the opening may turn out to be far
more important if they result in incompetent ground.

C B A

FAULT ZONE WITH CLAY

Fig. A9-4 A fault or weakness zone in a valley side may cause large local variations in the ground stress
distribution. The length of the arrows indicate the relative magnitude of the stress. At 'A' there is a
general stress situation, at 'B' highly anisotropic and at 'C' a destressed situation (from Selmer-Olsen,
1988).

2 The tangential stresses developed around an underground opening

During and after an underground opening is excavated in a rock mass, the in-situ (virgin) stresses
which acted upon the rock volume that is removed, are redistributed in the remaining rock mass.
This results in local stress concentrations in the immediate vicinity of the opening. This
phenomenon has been described by many authors. The calculations made are generally based on
idealized conditions: isotropic, homogeneous materials. The theoretical considerations show,
however, that the stress concentrations fall off quickly as shown in Fig. A9-5.

The stresses induced around the opening will change by time. Short time effect occurs during and
shortly after excavation and a secondary, long term effect takes place when deformations and cracks
develop in the surrounding rock mass. Sauer (1988) has made investigations of the stress
redistribution pattern in the vicinity of a tunnel in weak ground during excavation. He found that
stresses increase relatively rapidly at about 1 - 1.5 tunnel diameter in front of the working face to a
first (initial) maximum 0.5 diameter in front of the face. From this point there is a small decrease to
a stress minimum approximately 0.5 diameter behind the face. A second maximum occurs about
1.5 - 2 diameter behind the face, provided use of adequate support. Relaxation of the second
maximum takes several months, depending on the rheological properties of the rock. This is shown
in Fig. A9-6.

This development with the short-time increase and long-time reduction of the tangential stresses
surrounding a tunnel is utilized in the NATM, as mentioned in Chapter 8.
 A9 - 5

σθ / p2

3
K=0 K = p1 / p2
K = 1.0
2

1
K = 2.0
0 1 K = 3.0 r/a
2 3

-1

0 5
-1 1 1 2 3 4 6 7 8 σθ / p2

K=0 K = 1.0 K = 2.0 K = 3.0

p1 and p2 are principal stresses

K = p1 / p2
3
σθ = tangential stress around the opening

r/a

Fig. A9-5 Stresses around a circular hole in an isotropic, linearly elastic, homogeneous continuum (from
Goodman, 1989).

2
σ (kp/cm )
2.0

1.5

1.0

0.5

15 10 5 0 5 10 15 20 Excavation ( m )
5 4 1 0 1 3 4 6 99 - 186 (Days)

Direction of
excavation

Fig. A9-6 The variation of stress level before, during and after excavation (from Sauer, 1988).
 A9 - 6

3 A practical method to estimate the magnitude of the tangential stresses

A practical method to estimate the magnitude of the tangential stresses around various types of
underground openings has been developed by Hoek and Brown (1980). From a large number of
detailed stress analyses by means of the boundary element technique, they have presented the
following correlations:
- The tangential stress in roof σθr = (A × k - 1)pz eq. (A9-2)
- The tangential stress in wall σθw = (B - k)pz eq. (A9-3)

where A and B are roof and wall factors for various excavation shapes in Table 6-13
k is the ratio horizontal/vertical stress
pz is the vertical virgin stress

Eqs. (A9-2) and (A9-3) are shown graphically in Fig. A9-7.

TABLE A9-1 VALUES OF THE FACTORS 'A' AND 'B' FOR VARIOUS SHAPES OF
UNDERGROUND OPENINGS (from Hoek and Brown, 1980).

VALUES OF CONSTANTS A & B

A 5.0 4.0 3.9 3.2 3.1 3.0 2.0 1.9 1.8

B 2.0 1.5 1.8 2.3 2.7 3.0 5.0 1.9 3.9

Applying eq. (A9-2) and (A9-3) approximate estimates of the stresses acting in the rock masses
surrounding a tunnel can be found where the magnitudes of the vertical stresses and the ratio k = ph
/pv is known. As an example the tangential stresses in the ground around a horseshoe shaped tunnel
are considered. From eq. (A9-2) and Table A9-1 the tangential roof stress is
σθr = pz (3.2 k - 1)

For many situations where the vertical stress can be expressed by eq. (A9-1) the above expression
can be written as
σθr = 0.027 z (3.2 k - 1)

In Scandinavia, the value of (k) is often in the range 2 - 3. In such cases the tangential roof stresses
varies between
σθr = 5.5 pz to 8.6 pz
or if expressed by the depth below surface
σθr = 0.14 z to 0.224 z

Similarly, the tangential wall stress can be expressed as

σθw = pz (2.3 - k)

In the same ground conditions as listed above

σθw = 0.008 z to -0.018 z
which indicates that tension stresses will occur where high values of (k) occur.
 A9 - 7

12
σθr
11 = (Ak - 1)
pz

10

9
σθr
pz

8
VERTICAL IN SITU STRESS
MAXIMUM ROOF STRESS

7 SIDEWALL
6 6
σθw
5 5 = (B - k )
pz

σθr
4 4

VERTICAL IN SITU STRESS pz

MAXIMUM ROOF STRESS
3
RATIO

2 2

1 1

0 0

-1 -1

RATIO
0 1 2 3 4
HORIZONTAL IN SITU STRESS -2
RATIO =k
VERTICAL IN SITU STRESS

-3
ROOF 0 1 2 3 4
HORIZONTAL IN SITU STRESS
RATIO =k
VERTICAL IN SITU STRESS

Fig. A9-7 Influence of excavation shape and ratio horizontal/vertical stresses on mobilized tangential
stresses around underground openings (from Hoek and Brown, 1980).
From Palmström A.: RMi – a rock mass characterization system for rock engineering purposes.
PhD thesis, Oslo University, Norway, 1995, 400 p.

APPENDIX 10

SYMBOLS USED

1. General

γ weight per unit volume

n porosity
v Poisson's ratio
μ friction coefficient (= tan φ)
E Young's modulus
V deformation modulus
φ friction angle
c cohesion

2. Rock properties

w water content, dry weight basis

d the diameter (in mm) of the actual specimen
σc uniaxial compressive strength of intact rock material
σc 90 uniaxial compressive strength measured at right angle to the schistocity or
σc50 uniaxial compressive strength for 50 mm diameter sample size
Rc strength anisotropy (σc max /σc min)
Ia(50) strength anisotropy index
Is point load strength index
Is(50) point load strength measured on standard 50 mm thick sample
k correlation factor between compressive and point load strength (k = σc /Is)
k50 correlation factor related to 50 mm thick samples (k50 = σc50/Is50)
Fi rock foliation index, as given in Table A3-I.
fA rock anisotropy factor
fW rock weathering and alteration factor
c the content of platy and prismatic minerals in %

3. Jointing and block characteristics

i dilation angle for a joint plane

φj friction angle for a joint
Sj shear strength intercept ('cohesion') for a joint
JRC joint roughness coefficient
JCS the joint wall compressive strength (for fresh (unweathered) rocks JCS = σc)
u undulation of joint plane
L1 direct measured length along a joint surface (Turk and Dearman, 1982)
L2 the trace length measured on joint surface (Turk and Dearman, 1982)
 A10-2

γ angle between joint sets

S spacing of joints within a set
Sa average joint spacing
S1, S2, S3 spacing in various joint sets
α2 ratio between medium joint spacing and minimum spacing (S2/S1)
α3 ratio between maximum joint spacing and minimum spacing (S3/S1)
a3 length of the block
a1 thickness of the block.
β block shape factor
βe estimated block shape factor from βe = βo + 7(α3 - 1) = 27 + 7(a3/a1 - 1)
βo the lowest value of β, i.e. β = 27 for a cubical (equidimensional) blocks
Jv volumetric joint count (= the number of joints per m3 )
wJd weighted joint density
Ib block size index (eq. block diameter) introduced by ISRM (1978).
Db block diameter applied in rock support assessments (= 3√V _ b_)
Dbe eq. block diameter
Vb block volume
Vbo block volume delimited by 3 joint sets intersecting at right angles
A the size of the observation area (in m2, see Fig. A3-27)
na number of joints on an observation area with length Li
na* number of joints adjusted for the length and size of observation area (see
eq. (A3-32a))
Na 2-D joint frequency, i.e. the number of joints in a defined area, Na = na/A
Nl 1-D joint frequency, i.e. the number of joints intersecting a defined length along a
line or borehole
Nr the number of random joints in the observation area
Nα number of joints intersected at an angle α
N90 the number of joints with the same orientation which would have been observed at
an intersection angle of 90o
nj the rating for joint sets applied in eq. (A3-20 and (A3-21)
ka correlation factor from 2-D frequency measurement to 3-D (volume) (see
Fig. A3-25c and eq. (A3-32b))
kl correlation factor from 1-D frequency measurement to 3-D (volume) (see
Fig. A3-26 and eq. (A3-33))
ca 1/ka for 2-D observations on rock surfaces
cl 1/kl for 1-D observations of scanlines or drill cores
L length of the measured section along core or line, see Fig. A3-27
δ the angle between the observation plane (or drill core) and the individual joint,
which is used in the weighted joint density method
fi factor for the angle between joint and observation plane (or (1/sinδi) used in the
weighted joint density measurement, as given in Table A3-31

4. Stresses and related parameters

σo initial stress
σ1, σ2, σ3 principal stresses; σ1 > σ2 > σ3
σmin minimal principal stress
σmax maximum principal stress
σ1' the major principal effective stress at failure.
 A10 - 3

σ3' the minor principal effective stress

σn normal stress
pz or pv vertical stress
ph horizontal stress
p0 in situ hydrostatic rock stress
σθ tangential stress around underground openings
σr radial stress around underground openings
σθw tangential wall stress
σθr tangential roof stress
τ shear stress at failure
Φi' instantaneous friction angle
c cohesion
ci' instantaneous cohesive strength
k ratio of horizontal and vertical stresses (ph /pv)
f the gradient of line in the -ε3p , ε1p diagram (Fig. 8-4)

5. Refraction seismic properties and features

Vp longitudinal (compressional) wave velocity

Vs shear wave velocity
Vl sonic velocity in water
Vf longitudinal sonic velocitiy measured in the field
Vl longitudinal sonic velocity measured in the laboratory
V║, V┴ wave propagation parallel and across layers/schistocity
v seismic velocity measured in the field
V0 basic seismic velocity (km/s) for intact rock under the same stress level as in the
field (measured in the laboratory)
Vn maximum or 'natural' velocity in crack- and joint-free rock under the same stress
level as in the field. Natural velocities for some fresh rocks measured in the
laboratory are shown in Table A3-33
a, b constants related to the local ground conditions (rock material, stress condition,
jointing features etc.)for in-situ seismic velocities
ks factor representing in-situ conditions in seismic velocity assessments
Nl1, v1 and Nl2, v2 corresponding values of joints/m and in-situ longitudinal velocity, respectively, for
two pairs of measurements
SVR 'seismic velocity ratio' (SVR = Vf /Vl)
VI sonic velocity index (VI = SVR2 )

6. Rock mass properties and features

σcm the compressive strength the rock mass,

m undisturbed material constant in the original Hoek-Brown failure criterion
m disturbed material constant in the original Hoek-Brown failure criterion
mr material constants in the Hoek-Brown failure criterion for broken rock mass
mi material constants in the Hoek-Brown failure criterion for intact rock
mb constant in the modified Hoek-Brown failure criterion (1992)
s undisturbed material constant in the original Hoek-Brown failure criterion
s disturbed material constant in the original Hoek-Brown failure criterion

SYMBOLS\text-1.wpd
 A10 - 4

sr material constants in the Hoek-Brown failure criterion for broken rock mass
a constant in the modified Hoek-Brown failure criterio (1992)
Cg the reduction factor which Hansagi named 'gefüge-factor' (joint factor) being
"representative for the jointed effect of a rock mass".

6.1 Classification systems and parameters

RSR rock structure rating

RMR rock mass rating in the Geomechanics classification system
RQD rock quality designation
Q rock mass quality value in the Q classification system
Jn factor for joint set number in the Q-system
Jr factor for joint roughness in the Q-system
Ja factor for joint alteration and filling in the Q-system
Jw factor for joint water pressure or inflow in the Q-system
SRF stress reduction factor in the Q-system
ESR excavation support ratio in the Q-system

6.2 Parameters and features in the Rock Mass index (RMi)

jR joint roughness factor, representing the small and large scale unevenness of the joint
surface (jR = jw × js)
js joint smoothness factor (small scale evenness of joint surface)
jw joint waviness factor (large scale planarity of joint wall)
jA joint alteration factor, characterizing the strength of the joint surface
jL joint length and continuity (joint termination) factor
jC joint condition factor (combination of jR, jA and jL)
JP jointing parameter (i.e. combination of jC and Vb)
D factor in eq. (4-4) to calculate the jointing parameter [JP (D = 0.37 × jC - 0.2 )]

7. Parameters in the RMi rock support method

z the depth of the actual location below surface

Db equivalent block diameter
CF continuity factor for the rock mass (CF = tunnel size/block size)
Cg competency factor for continuous ground (Cg = RMi /σc)
Gc ground condition factor for discontinuous ground (Gc = JP × SL)
SL stress level factor used for discontinuous ground
Sr size ratio (Sr = CF × Co)
Co orientation factor for joints and zones
C gravity adjustment factor (of Gc) for tunnel walls (Milne and Potvin, 1992)
α the strike between tunnel surface and discontinuity
β the dip between tunnel roof (or floor) and discontinuity
Tz the width (thickness) of weakness zone
Ts the width (thickness) of singularity
σcz compressive strength of rock material in weakness zone
JPa the jointing parameter of the rock masses adjacent to the weakness zone
Gcz the ground condition factor for zones with Tz < JPa × σcz

SYMBOLS\text-1.wpd
 A10 - 5

Srz size ratio (Srz = Co × Tz /Db) for weakness zones for Tz < Wt or Tz < Wt
Gcs the ground condition for singularities
B rock bolt
S shotcrete
F fibrecrete
Wt width (span) of tunnel
Ht (or Hw) height of tunnel (or wall height)
ri internal tunnel radius
A roof factor for various excavation shapes (used by Hoek and Brown, 1980)
B wall factor for various excavation shapes (used by Hoek and Brown, 1980)

8. Parameters and features applied in the method for TBM penetration assessment

E factor for various groups of rocks

ks a TBM jointing factor (applied in the NTH method)
co factor representing orientation of the main joint set relative to the tunnel axis
keq 'equivalent TBM jointing factor' (applied in the NTH method)
kDRI adjustment factor of ks to arrive at keq = ks ⋅ kDRI (applied in the NTH method)
Meq equivalent thrust per cutter (also applied in the NTH method)
MB thrust capacity per disc (also applied in the NTH method)
kd correction factor for cutter diameter in Fig. 7-7 (also applied in the NTH method)
ka correction factor for cutter spacing given in Fig. 7-8 (also applied in the NTH
method)
I TBM advance rate (m/h)
io TBM penetration rate in mm per revolution io = F × keqG
F a factor in the expression for TBM penetration (F = 0.0015 Meq1.5 )
G an exponent in the expression for io (G = 30 keq- 0.5 × Meq- 0.8 for keq < 3.5)

SYMBOLS\text-1.wpd