Rock Mass Characterization

Minerva Access is the Institutional Repository of The University of Melbourne
Author/s:
Baczynski, Norbert Richard Przemyslaw
Title:
Rock mass characterization and its application to assessment of unsupported underground
openings
Date:
1980
Citation:
Baczynski, N. R. P. (1980). Rock mass characterization and its application to assessment of
unsupported underground openings. PhD thesis, Engineering, Department of Mining &
Metallurgy, The University of Melbourne.
Publication Status:
Unpublished
Persistent Link:
http://hdl.handle.net/11343/35714
File Description:
Rock mass characterisation and its application to assessment of unsupported underground
openings
Terms and Conditions:

Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the
copyright owner. The work may not be altered without permission from the copyright owner.
Readers may only download, print and save electronic copies of whole works for their own
personal non-commercial use. Any use that exceeds these limits requires permission from
the copyright owner. Attribution is essential when quoting or paraphrasing from these works.
ROCK MASS CHARACTERIZATION
a.nd I.u Appt,i..c.cttion :to AMe6.6me.n:t
on Un.6u.ppoll.:te.d UndeJtgJtound Ope.ning.6.
Thu-i-6 .6ubm..ute.d by
NORBERT RICHARD PRZEMYSLAW BACZYNSKI

B.Sc.. (Hon.6) (ANU), M.Sc.. (JCUNQ), Vip. Vo.-tame.:t. (UNISA)
NovembeJt 1980
a..6pa.Jl.:U.al. nuln-Ume.n:t on :the. Jte.qu-Ur.e.me.nt-6 no.l!..

:the. Ve.g.~te.e. on Voc.:toJt on Phil.oMphy in
:the. Ve.paJt:tme.n:t on Mining & Me.:ta..Uwtg y a.:t :the.
Unive!t.6Uy o6 Me1bouJtne.
(i)
ABSTRACT
The research study centres on the stabiZity assessment of

unsupported, sub-leveZ open stapes Zocated within the
siZver-Zead-zinc orebodies at the Mount Isa Mine in Queens-
Zand, AustraZia.
A systematic approach is adopted for characterization of

rock masses and a statisticaZZy-based modeZ is formuZated
to account for ZocaZ variabiZity in geoZogicaZ structure.
Investigations suggest that the degree of fracture cluster-
_ing observed in some masses may not have resuZted from a
random process. A zonal model is proposed for spatial
distribution of these weakness planes. The general
validity of the assumed modeZ with respect to the dolomitic
shales at the mine is confirmed by computer simulation of
fracture distributions within three-dimensional blocks. The
computer program used in the simuZation process has been
deveZoped as part of the project.
MechanicaZ properties of the mass are derived on the basis

of the statistical modeZ for geoZogical structure and
various pubZished rock mass classification systems. A
computer program is used to expedite these anaZyses.
AZternate approaches for derivation of properties are
mentioned.
SeveraZ techniques are employed to assess the significance

of statisticaZly variabZe material properties on stabiZity
of the stapes. These include rock mass classifications,
two-dimensional finite eZements with elasto-plastic
behaviour and sequential excavation, and the probabiZity of
faiZure anaZysis. Limitations of each technique are
discussed.
(ii)
Some consideration is given to broadly reviewing past

mining experience within stapes at the mine.
Results of analyses demonstrate that a greater depth of

understanding and appreciation of rock mass behaviour may
be gained from models with locally variable material pro-
perties. A reasonable degree of correlation is achieved
between model and prototype response.
(iii)
DECLARATION
I declare that this thesis is my own work. It has not been submitted
in any form for another degree or diploma at any other university or
institute of tertiary education.
Information derived from the published or unpublished work of others

has been acknowledged in the text and a list of references is given.
/
{_*rhf /:;::J:::-- Aj.
Norbert R.P. Baczynski
-
25th November 1980.
(iv)
ACKNOWLEDGEMENTS
Dr. G.D. Base, Mr. W.E. Bamford, Dr. J.R. Barrett and
Mr. D.R. Willoughby - staff members of the University of Melbourne
and the CSIRO Division of Applied Geomechanics - were involved or
were nominated at various times to act in a supervisory and
coordinating capacity.
Special thanks are extended to Dr. J.R. Barrett and Mr. W.E. Bamford
for their active guidance of the research project.
Thanks are due to Mr. M.C. Bridges of Mount Isa Mines Limited for
the interest in statistical approaches to geomechanics that he
managed to spark in the writer during the early 1970's.
The writer thanks Mr. F. Leahy of Mount Isa Mines Limited for his
continued interest and support of the work. Mr. R. Cowling of the
same company, Dr. G. Beer of the University of Queensland and
Mr. B. Perkins of the CSIRO Division of Applied Geomechanics offered
their enthusiastic assistance at various times during the course of
the project.
The work was basically funded from a research grant awarded by

Mount Isa Mines Limited. The cost of all travel, field accommodation
and on-site computing was borne by the Company. The assistance
offered by computing and geological staff at the mine is gratefully
acknowledged. Without the Company's generous and kind assistance
this project would not have been possible.
The grant was administered by the CSIRO, who also freely provided
their office, library and computer facilities. This arrangement
ensured a most congenial working environment during the term of the
studies.
(v)
Funding of the work was supplemented via a single award of the

Henry DeWitt Smith Scholarship (American Institute of Mining,
Metallurgical and Petroleum Engineers) and a partial award of the
Gilbert Rigg Scholarship (University of Melbourne) . The former was
initially granted in 1975 to pursue the postgraduate programme
overseas. However, it was later extended to permit studies at the
University of Melbourne during 1977 - 1980. The writer thanks
Mr. R.J. Searls of Newmont Proprietary Limited for his interest in
the project and for his patience during finalization of study
arrangements. The award helped to ease the cost of establishment in
Melbourne.
I thank my wife Carol for her support during the years that she has
had to live with this project.
(vi)
INDEX
Paragraph Subject Page No.
CHAPTER 1 INTRODUCTION 1
1.1 Definition of Problem and Project

Objectives 1
1.2 Description of Field Site 2
1.3 Conceptual Basis Adopted for Model 6
1.3 .1 General Implications Inherent in
Modelling 6
1.3 .2 Characterization of Mine Shales 8
1.4 Structure of Thesis 10
1.5 Constraints on Complexity of Assumed
Model 11
1.5 .1 Areas Investigated 11
1.5.2 Derivation of Rock Mass Properties 12
1.5.3 Stability Analyses 13
CHAPTER 2 MODEL FOR INTACT ROCK COMPONENT OF ROCK MASS 18
2.1 Strength of the Intact Rock 18

2 .1.1 Unconfined Compressive Strength 18
2 .1.2 Tensile Strength 23
2.1.3 Shear Strength, S , and Angle of Internal
0
Friction i 23
2.2 Modulus of Elasticity 25
2.3 Poisson's Ratio 26
2.4 Strain at Failure 26
2.5 Sununary 26
(vii)
INDEX
CHAPTER 3 PHYSICAL CHARACTERISTICS OF GEOLOGICAL

DISCONTINUITIES 28
3.1 Source of Data 28

3.2 Scope of Model 29
3.3 Characteristics of Discontinuities 31
3.3.1 Orientation 31
3.3.2 Spacing 34
3.3.3 Shape and Continuity 38
3.3.4 Frictional Properties of Fractures 42
CHAPTER 4 MODEL FOR SPATIAL DISTRIBUTION OF FRACTURES 46
4.1 General Comments 47

4.2 "Even" and "Variable" Models 47
4.3 "Random" Mode 1 48
4.4 Testing of "Random" Model 50
4.5 Basis of "Zonal" Model 53
4.6 Derivation of "Zonal" Model Parameters 54
4.6.1 Field Investigations 54
4.6.2 Evaluation of Data 55
4.7 Some Statistical Considerations 58
4.8 Spatial Relationship Observed at the Mine 62
4.9 Model Testing 69
4.10 Spatial Model for Bedding Plane Partings 77
(viii)
INDEX
CHAPTER 5 MODEL FOR STRUCTURAL VARIABILITY OF

LARGE BLOCKS 81
5.1 Model for Each Fracture Set 81

5.2 Combined Model for Rock Mass 88
CHAPTER 6 ROCK MASS PROPERTIES 96
6.1 Method of Analysis 96

6.2 Computer Simulation 97
6.3 Model for RQD-Index 98
6.4 Basis of Model for Rock Mass Strength 100
6.4 .1 Intact Rock Mass 100
6.4.2 Discontinuous Rock Mass 101
6.5 Basis of Model for Rock Mass Modulus 103
6.5.1 Modulus of Intact Rock 104
6.5.2 Theoretical Approach to Rock Mass Modulus 105
6.5.3 Rock Mass Modulus 106
6.6 Poisson's Ratio Assumed for Rock Mass 108
6.7 Statistical Model for Material Properties 109
6.7.1 Definition of Domains 109
6.7.2 Material Types 110
6.7.3 Frequency of Occurrence 110
-
(ix)
INDEX
CHAPTER 7 STABILITY ANALYSIS 115
7.1 Rock Mass Classifications 116

7 .1.1 Classifications Considered 116
7 .1.2 Estimates of Stable Hanging-wall Spans 117 a
7.2 Finite Element Modelling 125
7.2.1 Choice of Computer Program 125
7.2.2 Complementary Programs 127
7.2.3 Geometry of Model 127
7.2.4 Mesh Formulation 129
7.2.5 Details of Base Mesh 130
7.2.6 Mesh Types 133
7.2.7 Accuracy of Mesh 135
7.2.8 Stress Field Assumed in Analysis 137
7.2.9 Extent of Finite Element Modelling
Programme 137
7.3 Results of Finite Element Modelling 140
7.3.1 Extent of Plastic Yield 140
7.3.2 Displacements 153.
7.3.3 Stresses 157
7.4 Probability of Failure Analysis 164
7.4.1 Description of Technique 164
7.4.2 Assessment of Probability 166
7.4.3 Advantages and Limitations of Method 183
(x)
INDEX
CHAPTER 8 COMPARISON BETWEEN MODEL AND PROTOTYPE

RESPONSE 186
8.1 Practical Difficulties 186

8.2 Scope of Analysis 188
8.3 Observed Response 189
8.4 Rock Mass Classification Predictions 194
8.5 Finite Element Predictions 196
8.6 Probability of Failure Analysis
Predictions 198
CHAPTER 9 SUMMARY AND DISCUSSION 199
9.1 Background Considerations 199

9.2 Basis of Adopted Model 201
9.3 Derivation of Model Parameters 201
9.4 Stability Analysis 203
9.5 Conclusions 205
9.6 Recommendations 208
REFERENCES 210
ADDITIONAL REFERENCES 222
APPENDIX I ROCK MASS CLASSIFICATION INPUT PARAMETERS Al-AS
Al.l Bieniawski's RMR Method Al

Al.2 Laubscher and Taylor's Modified RMR Method A2
Al.3 Barton's Q-Method A4
--
(xi)
LIST OF TABLES
Table No. De scription Page No.
1 Characteristics of Fracture Sets. 33
2 Mean Intensity of Fractures per Unit

Volume. 72
3 Relationship Between C-Factor and Intensity

of Fracturing per Metre. 103
4 Rock Mass Material Properties. 111
5 Average Material Types Used in Modelling. 113
6 Hanging-wall Spans (Metres) on Basis of

Rock Mass Classifications. 118
7 Magnitude of Pre-Mining Stress Field

Assumed in Modelling. 138
8 Sensitivity Testing: Different Initial

Stresses. 138
9 Finite Element Modelling Programme Using

VISPLAS. 139
10 Stability of 5 Orebody Stope with Uniform

(Homogeneous) Material Properties. 142
11 Stability of 5 Orebody Stope with

Heterogeneous (Statistically Variable)
Material Properties. 143
(xii)
LIST OF TABLES
Table No. Description Page No.
12 Stability of 11 Orebody Stope with Uniform

(Homogeneous) Material Properties. 144
13 Stability of 11 Orebody Stope with

Heterogeneous (Statistically Variable)
14 Case Study: Variations in the Width of

Stapes (in Metres). 193
15 Case Study: Stability of Hanging-walls &
Abutments. 193
(xiii)
LIST OF FIGURES
Figure No. Description Page No.
1 Typical Cross-Section Through Northern

Mining Area. 4
2 Diagrammatic Layout of a Sublevel Open

Stope in Silver-Lead-Zinc Orebodies. 5
3 Formulation of a Geotechnical Model. 7
4 Conceptual Basis for Rock Mass Characterization. 9
5 Relationship Between Core-Bedding Angle and

the Unconfined Compressive Strength of Cores. 20
6 Strength of 61 Specimens of Mineralized Shale

with Angle of 31-45 Between Loading Axis and
Bedding Plane. 22
7 Principal Fracture Sets. 32
8 Cumulative Frequency Plot on Logarithmic

Probability Paper for the Spacing of Fracture
Set 2 and 3 (Diamond Drill Core Data) . 35
9 Relationship Between Median Frequency of Extension

Fractures and Frequency of Bedding Plane
Partings. 37
10 Cumulative Frequency Plot on Logarithmic Probability

Paper for Continuity of Fracture Set No. 2 in Directions
Parallel and Normal to Bedding of Mine Shales
(Line SAmple Data) . 40
(xiv)
LIST OF FIGURES

Probability Paper for Continuity of Bedding
Plane Partings (Line Sample Data) . 41
12 Cumulative Frequency Distribution on

Normal Probability Paper for Proportion of
Graphitic Bedding Plane Partings in Samples
of 10 Partings (Total Sample Size = 1000) . 44
13 Pattern of Fracture Traces Generated on

Basis of "Random" Model for Spatial
Distribution of Fracture Plane Centres. 51

Logarithmic Probability Paper for Spacing
of Adjacent Fractures Derived on Basis of
"Random" Spatial Model. 52
15 Unit Areas with respect to a Particular

Set. 56
16 Conceptual Relationship: Influence of Immediately

Adjacent Areas on Observed Fracture Intensity
Within a Designated Unit Area. 57
17 Joint and Marginal Probability Density

Functions of Continuous Random Variables
X and Y. 61
-
(xv)
LIST OF FIGURES

Probability Paper for Local Variability in
the Intensity of Fracture Set No. 1. 64
19 Probability Density Functions for Intensities

of Fractures Observed Adjacent to Unit Areas
with a Known Intensity "i" in Direction Parallel
to Average Orientation of Set No. 1. 65
20 Probability Density Functions for Intensities

of Fractures Observed Adjacent to Unit Areas
with Known Intensity "i" in Direction Normal
to Average Orientation of Set No. 1. 66
21 Conceptual Relationship Between Test-Block

and Master-Block in Computer Program
CRACKS-2. 70
22 Typical Fracture Pattern Generated for

Fracture Set No. 1 on Basis of "Zonal}'
Model. 73

Logarithmic Probability Paper for Spacing
Between Adjacent Fractures of Set No. 1
Derived on Basis of "Zonal" Spatial Model. 74
-
(xvi)
LIST OF FIGURES
24 Fracture Pattern for Set Nos.l-6 Generated

on Several Planes through a 10m x Bm x 6m
Test Block on Basis of "Zonal" Spatial
Model. 76
25 Cumulative Frequency Plot on Normal

the Intensity (per metre) of Fracture Set
No.ll (Bedding Plane Partings) in the
Immediate Hanging-wall of Orebody No.ll. 79
26 Cumulative Frequency Plot on Normal

Probability Paper for Variability in the
Mean Intensity (per metre) of Fracture
Set No.ll (Bedding Plane Partings)
Between Stope Hanging-walls in Orebody
No.ll. 80
27 Approximate Relationship Between Volume

of Rock Mass & the Mean and Standard
Deviation of Normal Probability Density
Functions for Intensity of Fracture
2 3
Set Nos.l-11 (m /m ). 83-87

the Intensity of Fracture Set Nos.l-10
Combined. 91
29 Cumulative Frequency Plot on Probability

Paper for Local Variability in the Intensity
of Fracture Set Nos.l-11 Combined within the
Hanging~all of Orebody No.5. 92
(xvii)
LIST OF FIGURES
30 Cumulative Frequency Plot on Normal Probability

Paper for Local variability in the Intensity
of Fracture Set Nos.l-11 Combined within the
Hanging-wall of Orebody No.ll 93
31 Approximate Relationship Between Volume of Rock

Mass & the Mean and Standard Deviation of Normal
Probability Density Functions for the Intensity
of Fracture Set Nos.l-10 and 1-11 Combined within
the Hanging-walls of Orebody Nos. 5, 7 and 11,
respectively (m 2;m3 ). 94-95
32 Relationship Between Fracture Frequency and

RQD-Index: Mount Isa Mine. 99
32A Derivation of Rock Mass Cohesion. 102c
33 Cumulative Frequency Distributions Plotted on

Probability Paper for the Occurrence of Material
3
Types within Unit Volumes (1.0 m ) of the Rock
Mass. 114
34 Orebody No.5: Cumulative Frequency

Distribution for Stable Hanging-wall Spans
plotted on Probability Paper: Estimates Based
on Bieniawski (1973). 119
35 Orebody No. 11: Cumulative Frequency

Distribution for Stable Hanging-wall Spans
plotted on Probability Paper: Estimates Based
on Bieniawski (1973) 120
(xviii)
LIST OF FIGURES
36 Orebody No.5: Cumulative Frequency Distribu-

tion for Stable Hanging-wall Spans plotted
on Probability Paper: Estimates Based on
Laubscher and Taylor (1976). 121
37 Orebody No.ll: Cumulative Frequency Distribu-

tion for Stable Hanging-wall Spans plotted
Laubscher and Taylor (1976). 122

tions at Various "Stress Reduction Factor"
Levels for Stable Hanging-wall Spans plotted
Barton et al. (1974). 123

tions at Various "Stress Reduction Factor"
Levels for Stable Hanging-wall Spans plotted
Barton et al. (1974). 124
40 Cross-Section of Underground Sub-level Open

Stope Represented by Finite Element Model. 128
41 Base Mesh. 131-132
42 Dimensions of Various Mesh-Types Used in

Analysis. 134
43 Mid-span Displacement of Hanging-wall at

Various Excavation Stages: Comparison between
Boundary Integral and Finite Element Programs. 136
-
(xix)
LIST OF FIGURES
44 Key Diagram Indicating Elements that are

Considered to be Located within Hanging-
wall(x) and Abutment(@) Domains. 141

tions (plotted on Probability Paper) for
Number of Yielding Elements in Hanging-wall
of Stope with Uniform and Heterogeneous

tions (plotted on Probability Paper) for
Nwnber of Yielding Elements in Hanging-wall
of Stope with Uniform and Heterogeneous
47 Orebody No.5: Extent of Plastic Yielding

Around Stope with Statistically Variable
48 Orebody No.ll: Extent of Plastic Yielding

Around Stope with Statistically Variable
49 Extent of Plastic Yield in Model with Mean

Uniform (Homogeneous) Material Properties:
Orebody No.ll. 151
50 Typical Exarrple of the Extent of Plastic

Yield in Model with Statistically Hetero-
geneous Material Properties: Orebody No.ll. 152
51 Orebody No.5: Mid-span Hanging-wall

Displacements. 154
(xx)
LIST OF FIGURES
52 Orebody No.ll: Mid-span Hanging-wall

Displacements. 155
53 Orebody No.5: Mean Material Properties:

Mid-abutment Displacements at Various
Depths. 156
54 Orebody No.5: Model with Uniform Material

Properties: Principal Stress and
Displacement Vectors. 158-159
55 Orebody No.ll: Model with Uniform Material

Properties: Plot of Principal Stress and
Displacement Vectors. 160-161
56 Example of Typical Stresses Observed in

Finite Element Models with Uniform and
Statistically Heterogeneous Material
Properties. 163
57 A Hypothetical Example of Probability of

Failure Analysis. 167
58 Orebody No.5: Contours of Probability for

Occurrence of Local Ground Failure:
Y-stress = 36 MPa, X-stress = 16 MPa. 168-172
59 Orebody No.ll: Contours of Probability for

Occurrence of Local Ground Failure:
Y-stress = 36 MPa, X-stress = 16 MPa. 173-177
60 Orebody No.5: Excavation Stage No.5:

Contours of Probability for Occurrence of
Local Ground Failure: Y-stress = 26 MPa,
X-stress = 16 MPa. 178
--
(xxi)
LIST OF FIGURES



64 Schemrnatic Representation of Probability

of Failure Analysis for Extensive Zones. 182
65 Orebody Nos. 5, 7 and ll: Relationship

Between Strike Span of Hanging-walls and
Width of Stapes. 190
66 Orebody Nos. 5 and 7 Combined: Cumulative

Frequency Distr~bution plotted on
Probability Paper for the Strike Span of
Hanging-walls. 191
1.
CHAPTER 1 INTRODUCTION
1.1 Definition of Problem and Project Objectives
Most rock masses display several common attributes:
(1) They are physical systems whose principal components are

intact rock and geological discontinuities.
(2)- They are rarely, if ever, homogeneous (uniform)

throughout. Variability is apparent at all scales and
levels of detail.
(3) Their response under load may be potentially complex.
Logically, the above attributes tend to militate against the

characterization of real rock masses in terms'of simple, homogeneous
systems with single sets of material properties.
Accordingly, the objectives of the present study are to examine the

geological variability within a typical rock mass and then to assess
the importance of this variability on the stability of openings
located within such heterogeneous systems. The work is based on field
investigations undertaken in the silver-lead-zinc orebodies at the
Mount Isa Mine in north-western Queensland, Australia.
Three basic phases are involved in the study:
(1) Formulation of a statistically-based model to adequately

account for the observed structural variability within the
mine shales.
2.
(2) Equation of the structural model with a corresponding

model for variability in mechanical properties of the
rock mass.
(3) Examination of the effect and the significance, if any,

of a heterogeneous model in stability analyses; with
special reference to the hangingwalls of unsupported,
sublevel open stapes in the silver-lead-zinc orebodies
at the mine.
1.2 Description of Field Site
The sedimentary rocks which contain the Mount Isa deposits are part
of the Lower Proterozoic strata of the Australian Precambrian Shield.
The economic mineralization is confined to the Urquhart Shale Formation
within the Mount Isa Group.
Structurally, the Group is located on the western limb of a north

plunging anticline, the axis of which trends towards north-south
about 19 kilometres to the east of Mount Isa. A predominant north-
south strike and a 60-65 westerly dip persists for some 28 kilometres
northwards and 32 kilometres southwards from the mine.
On the local scale, the Urquhart Shales are composed of bedded, pale
and dark grey, fine grained dolomitic and volcanic shales, minor
tuffaceous bands, and silica-rich breccias.
The silver-lead-zinc orebodies are restricted to unaltered shale areas.

The copper orebodies occur in zones of recrystallized shale and
deformed Urquhart Shale locally known as silica-dolomite.
3.
The silver-lead-zinc orebodies occur in discontinuous, concordant

bands which may extend over a strike length of 1000 metres and over
a down-dip distance of 600 metres. Width of individual orebodies
ranges between 1.5 and 45 metres. Concentrations of economic mineral-
ization are known in 15 distinct groups of stratigraphic horizons.
Spatially, the orebodies are arranged in an en-echelon pattern. A
typical vertical cross-section normal to stratigraphy and looking
northwards is presented in Figure 1.
Minor folding is common in the mine, especially in mineralized areas.

Major folding is confined to distinct zones. Strike faulting and
shearing, transverse faulting and jointing are locally well developed
in the mine shales.
various mining techniques have been used at the mine over the last
fifty years. They include glory-hole mining, sublevel open stoping,
a combination of sublevel and shrink stoping, and a mechanised cut-
and-fill method (locally known as MICAF). The sublevel open stoping
and cut-and-fill methods are currently being employed to extract the
silver-lead-zinc orebodies.
Individual sublevel open stopes may extend over a strike length of

80 metres, over a down-dip distance of 250 metres and over a width
(in the direction normal to bedding) of 45 metres. The dimensions
of the average stope are much smaller, with the corresponding
distances of about 30 metres, 150 metres and 20 metres, respectively.
Commonly, the walls and crown of sublevel open stopes are not
supported or reinforced during mining. However, fill is now generally
placed in the stopes at the end of ore extraction.
The geometry and layout of a typical open stope is illustrated in

Figure 2.
4.
..
e "'e
8 ~
:::
D COPPER ORE
tm LEAD ORE
E3 "SILICA- DOLOMITE"
D GREENSTONE
CJ CARBONACEOUS MYLONITE
~- SlOPED AREA
llim FILLED STOPE
CROSS SECT IONS

r'--/
LOOKING NORTH
/
I /1
vJ/
. ?
L_..../ .
Sc4le of Me..tres
j .-,
~ .. (._.,./ i
TYPICAL CROSS SECTION THROUGH NORHRN MINING AREA
FIGURE l. Typical Cross-Section Through Northern Mining

Area.
--
5.
Dllll DR!H
a.ccos "r
lll.tU
UTUU ill
lONGITUDINAL OPEN SlOPING

MEDIUM WIDTH OREBODIES
FIGURE 2. Diagrammatic Layout of a Sub-Level Open

Stope in Silver-Lead-Zinc Orebodies
6.
1.3 Conceptual Basis Adopted for Model
l . 3.1 General Implications Inherent in Modelling
Any stability analysis of a rock mass system is a conceptual process.

This arises because the analysis invariably involves the idealization
of a real rock mass by means of a simplified model.
Taha (1976) has schematically depicted the level of abstraction that

is commonly involved in the formulation o a model from a real-life
situation. The appropriate diagram is reproduced in Figure 3.
The "assumed real world" is abstracted from the real situation by

concentrating on those variables that one considers or presumes to be
dominant in controlling the behaviour of the real system. The model,
being an abstraction of the "assumed real world", simplifies the
relationship among these variables in a form that is amenable to
analysis and testing under different sets of external conditions.
There are limitations associated with any model formulation and the
extrapolation of model results to in-situ or field situations.
Although the response of the model may be in accord with the "assumed
real world", it does not necessarily follow that the behaviour of the
simplified model will adequately represent the true response of the
"real world". This arises because the "assumed real world" may not
be representative of the "real world".
7.
FORMULATION OF A GEOTECHNICAL MODEL

After Taha (1976).
FIGURE 3.
8.
1.3 .2 Characterization of Mine Shales
The basis of the concept adopted in the present study for the
characterization of rock masses rests on the following idealization:
ANY LARGE BLOCK OF ROCK MASS MAY BE

REPRESENTED BY AN EQUIVALENT SYSTEM
OF SMALLER SUB-BLOCKS OR MODULES.
The concept is illustrated schematically in Figure 4.
The above concept is similar to the "systems approach" first formulated

in the 1930's by von Bertalanffy. This approach has been more recently
cited in literature by Couger and Knapp (1974), Ackoff (1974), Davis
(1974) and others with respect to the analysis of information systems.
Where a system is too complex to permit an evaluation of its response

as a whole, the problem can be simplified by dividing or factoring
the system into subsystems. The use of subsystems may also be termed
as a modular or building block concept. The process of factoring is
continued to a stage where individual subsystems are of a manageable
size.
This reduces the problem in the sense that it is now comprised of

modules or blocks, and the response of each of the units may now be
more readily assessed.
To a large extent, the response of the entire system is the total sum
of the responses of the individual subsystems that are located along
"critical paths" within the system. This view is supported by
von Bertalanffy (1969) when he states that" .. the properties
and modes of action of higher levels are not explicable by the
.summation of the properties and modes of action of their components
1
LARGE BLOCK OF ROCK MASS

EQUIVALENT SYSTEM OF SMALLER SUB-BLOCKS OR
MODULES.
FIGURE 4. Conceptual Basis for Rock Mass Characterization

1.0
10.
taken in isolation. If, however, we know the ensemble of the components

and the relations existing between them, then the higher levels are
derivable from the components .... ". In terms of rock masses, the
"critical paths" are likely to correspond with areas of high stress
gradient, or with zones of high compressive or tensile stresses.
Because str,ctural variability within a mass is statistically defined,
the interfaces or interrelations between the subsystems along the
"critical paths" may also be defined in a similar probabilistic
manner.
Thus, the systems approach to problems focuses on the system taken

as a whole, not their parts taken separately.
The writer is of the opinion that the systems concept offers a sound
approach for characterization of rock masses and for assessing rock
mechanics problems. It constitutes a framework for visualizing and
tackling the difficult task of defining structural and mechanical
variability within rock masses and then extending the process to
modelling of the response of the entire mass during mining.
1.4 Structure of Thesis
The results of the present study are presented in five main parts:
(1) A review of intact rock properties as derived from

laboratory testing.
(2) Summary of field investigations and the formulation of

a model for structural variability within the mine shales.
(3) Derivation of an equivalent model for variability in the

continuum material properties of the rock mass.
ll.
(4) Assessment of rock mass stability and a broad correlation

of the results with past mining experience in the silver-
lead-zinc orebodies.
(5) Overview of the approaches employed in the study and

formalization of the basic principles which may be
extended to the assessment of other rock masses.
1.5 Constraints on Complexity of Assumed Model
1.5.1 Areas Investigated
A dominant feature of the dolomitic shales at the mine is the

relatively high frequency of "minor" geological structures.
Accordingly, one of the prime considerations in this study has been

to examine the characteristics and spatial distribution of various
fracture sets within uniformly bedded shales and then to devise a
method of assessment whereby it would be possible to include their
effect in the analysis.
The scope of the investigation is limited to Orebody Nos. 5, 7 and

11. These orebodies are purposely selected because they are considered
to embrace the full range of ground conditions that have been experienced
in the past within Sublevel open stapes at the mine.
The scope of the analysis is also limited to exclude those areas of

12.
the mine where the rock mass is transected by major, gouge-infilled

faults.
1.5.2 Derivation of Rock Mass Properties
It is apparent that field testing offers the only reliable method

for evaluating various rock mass parameters (i.e., cohesion, effective
peak and residual friction angles, unconfined compressive and tensile
strengths, elastic modulus or stiffness, and Poisson's ratio).
Several aspects characterize such tests:
(1) The dimensions of the test samples are commonly quite

large to ensure that they are representative of the mass.
(2) Where the rock mass is locally variable, a large number

of tests are necessary to permit meaningful trends to
be deduced from the results.
(3) Testing programmes are costly and very time consuming.
Because of the above characteristics and difficulties in locating

suitable test-sites, most field testing programmes tend to be rather
limited in scope. Very few results are reported in technical literature.
Similarly, there was little opportunity to extend the scope of the

present study to include such tests.
Estimates of rock mass parameters used in the present analysis are

derived on the basis of various published rock mass classifications.
Details of the procedure employed to determine these values are
discussed in Chapter 6 of this thesis.
13.
The parameters are determined for the range of structurally possible

situations within subsystems or unit modules of the dolomitic shales.
A computer program GEOCLAS was developed to expedite the assessment.
The analysis centres on "Monte Carlo" sampling of probability density
functions for the intensity of each fracture set as well as other
input parameters and the derivation of appropriate ratings for
several classification systems. The process is iterated as many times
as considered necessary to derive a model for variability between
sub-blocks (of any specified dimensions}.
There is no doubt that the above assessment can only hope to yield
a crude estimate. Thus, various criticisms may be levelled at this
approach. Most of these relate to the simplicity of the ensuing
material model and the validity of extrapolating relationships
established at other sites to the mass under investigation.
In spite of these limitations, the values may be considered as first-

order approximations of the in-situ parameters. They suffice to
demonstrate the potential extent of variability that may exist locally
within the mine shales. Validation of the assumed values can only be
achieved by appropriate field testing.
1.5 .3 Stability Analyses
The formulation or choice of model(s} should be mainly governed by

two considerations:
(1) The model(s) should approximate as closely as possible

the real situation.
(2) The model(s) should be as simple as possible.

14.
The accuracy of rJrcdiction and therefore, the usefulness of the

results obtained on the basis of modelling, depends to a large
extent on whether the selected formulation reflects the realities
of the in-situ problem.
The degree of difficulty associated with the incorporation of "minor"

structural features in geotechnical models is directly proportional
to the scale factor between the continuity and intensity of the
structures and the physical dimensions of the excavation being
modelled. Low continuity and high frequency of structures confound
the task of realistically modelling the mass.
Several potential modelling approaches were examined.
Initial considerations were given to physical modelling. An outline

of a tentative testing progrillfu~e was drawn up and is presented in
Baczynski (1977b). A survey of relevant technical literature was
undertaken (Baczynski, 1980d). The Australian Coal Industry Research
Laboratories at Wollongong were visited in late-1977 to review
physical modelling techniques and to inspect testing facilities.
Full access was also available to the results of modelling work
undertaken at the CSIRO Division of Applied Geomechanics in
Helbourne.
Several difficulties became apparent during the review. These centre

on:
(l) Development of representative equivalent materials to

simulate intact rock.
(2) Quality control to ensure that uniformity of equivalent

rna terials is maintained throughout the testing programme.
(3) Adequate means for simulating gravitational loading.
(4) The time element involved in the construction of large,

three-dimensional models with high frequencies of non-
penetrating discontinuities to simulate geological structure.
1 J.
C"
( 5) Ensuring that r:rodel construction extending over several

months did not superimpose significant variability on
the properties of equivalent materials due to local
differences in drying and curing time.
(6) Development of a suitable technique for L~e incorporation

of relatively closed (tight) geological structures.
(7) Adequate simulation of the frictional characteristics of

joints and bedding plane partings.
The latter two factors are considered to be especially important.

This arises because the strength and response of a model is largely
a function of:
(1) Frictional properties of the weakness planes, and
(2) Degree of interlock that exists between the constituent

sub-blocks of the model.
It is suspected by the writer that physical models with fairly open

discontinuities would have resulted in unrealistic local stress
concentrations at the extremities of the weakness planes and the
response of the system under load may not have reflected the true
behaviour of mine shales.
Several numerical techniques for the assessment of continua and

discontinua were also examined.
The two most applicable techniques available for nili~erical modelling

of discontinua are:
(1) Finite element method with joint elements to simulate

geological discontinuities, as described by Goodman et
al. (1968, 1977) and others, and
(2) Blocky model method developed by Cundall (1971).

16.
The two methods permit a more realistic characterization of rock

masses. However, an enormous computational effort is involved in
the analysis of even relatively simple systems. This places a
severe limitation on the size and complexity of the problem that
can be successfully coded and analysed.
Three different approaches were finally selected to assess the

stability of stopes at the mine. These are:
(1) Rock Mass Classifications.
(2) Two-dimensional, Finite Element Method.
(3) Probability of Failure Analysis.
Rock mass classifications that are employed include Bieniawski (1973)

RMR-index, Laubscher and Taylor (1976) Modified ~ffi-index, and the
Barton et al. (1974) Q-index.
Considerable difficulty was experienced in securing a finite element

program that would model elasto-plastic behaviour and permit sequential
excavation of openings. All programs that were examined appeared to
perform well in modelling of a single-stage opening in an elastic
continuum. However, problems arose in most of them when the analysis
was extended to multi-stage models. Errors were mainly of the type
that is discussed in Christian and Wong (1973). They relate to the
dependency between observed displacements and excavation sequence.
Some simple, but potentially significant, logic errors were also
detected.
The finite element analysis was undertaken using VISPLAS, a two-

dimensional program develo:ped by Dr. J. Heck at the University of
Queensland in Brisbane. Several modifications were incorporated by
the writer in the program to reduce the data into a more meaningful
format and to permit computer-plotting of principal stress and
displacement vectors. Numerical modelling was conducted on the
UNIVAC 1100 system at the Mount Isa Mine.
17.
Stress input for the "probability of failure analysis" are determined

by the boundary element program BITEM developed by Crotty and Wardle
(1978) at the CSIRO Division of Applied Geomechanics. A simple computer
program was written to expedite the analysis. This aspect of the work
was completed on the CYBER 7600 system at the CSIRO.
18.
CHAPTER 2 MODEL FOR INTACT ROCK COMPONENT OF ROCK MASS
The mechanical properties of the intact rock component of the

dolomitic shales are reviewed in this chapter. The work discussed
below is based entirely on laboratory testing undertaken by Mount
Isa Mine personnel and other researchers over the last 15 years.
The writer's involvement is restricted to collation and assessment
of the available data.
2.1 Strength of the Intact Rock
An extensive laboratory testing programme, involving several

thousand core specimens, was undertaken between 1965 and 1978. The
results of some of the tests have been reported in Rosengren (1968) ,
MacLeod-Carey (1969), Mathews and Edwards (1969), Baczynski (1973,
1977c), Leask (1977) and Miller (1978).
2.1.1 Unconfined Compressive Strength
Both MacLeod-Carey (1969) and Mathews and Edwards (1969) note that
the unconfined compressive strength is anisotropic and is dependent
on the angular relationship between core axis and sedimentary
layering (i.e., bedding planes).
19.
Although there is some disagreement in regard to the a~solute
magnitude of the values, the indicated strength in directions

parallel, normal and at 25-30 to the bedding planes is about
200-210 MPa, 300-360 r.:Pa and 125-200 J.:Pa, respectively. The standard
deviation in each case appears to be approximately 35 MPa.
Brady (1977) suggested that the average strength of mineralized

shale in an area of trial stoping v;ithin Orebody No.7 ..,as 170 MPa,
whereas the assessment of rock strength undertaken by Leask (1977)
indicated a mean value of 188 MPa for the dolomitic shales. The
latter work was restricted to evaluation of cores from the 1100
Orebody (copper) area of the mine, but had included samples from
rock types which are typical of those in the silver-lead-zinc areas
of the mine.
Figure 5 summarizes the results of core tests reported in Baczynski

(l977c). The majority of test samples were obtained from diamond
drill cores located within Orebody No.5 stratigraphic horizon at
14 Level of the mine.
The following trends may be deduced from Figure 5:
(l) The unconfined compressive strength is markedly

controlled by the angle between the core axis and
sedimentary layering (i.e., bedding planes) of the mine
shales. This angle is refered to as the "core-bedding
angle".
(2) The highest strengths are observed in specimens with

0
core-bedding angles near 90 and the lowest values are
recorded for s~~cimens with angles near 30.
(3) For specimens failing by shear through intact rock, the

strength of barren shale is greater than that of mineral-
ized shale, for all specimens with core-bedding angles
between 30 and 90. At lower angles, specimens of each
rock type appear to have similar strengU1s.
- 20.
400
300
z
H
200
100
o;---------.---------r--------,---------.---------r---------r
0 15 30 45 60 75 90
ANGLE BETWEEN WADING AXIS AND BEDDING PLANE
~------------~ Barren Shale: Failure through intact rock.

- - - ----- Barren Shale: Failure along potential weal<ne ss plane.
- - --- - . Mineralized Shale: Failure through intact rock.

X--------- X Mineralized Shale: Failure along potential weakness plane.
(number) Number of tests on which the mean strength value is based.
FIGURE 5. Relationship Between Core-Bedding Angle and

the Unconfined Compressive Strength of Cores.
21.
(4) For specimens failing along potential, pre-existing,

weakness planes such as bedding or calcite infilled
veins, the strength of both the barren and mineralized
shale specimens is similar in the range of core-bedding
0 0
angles between 0 and 60 At greater angles, the
strength of specimens appears to tend towards that dis-
played by specimens where failure is by shear through
intact rock (for each respective rock type).
(5) Where a large sample of data is available, the strength

values appear to be normally distributed about the mean
value. An example of a distribution is presented in
Figure 6. This observation is at variance with the trends
reported by Torrent (1978), where a lognormal model is
suggested for strength data.
(6) The average strength of barren and mineralized shale

failing by shear through intact rock is estimated as
253 MPa and 221 MPa, respectively. These estimates are
derived by partioning data into six sub-groups. Grouping
intervals correspond to ranges of core-bedding angles
equal to 0-15, 16-30, 31-45, 46-60, 61-75 and 76-90,
respectively. The average value is determined by equal
weighting of each subgroup mean.
(7) The average strength of barren and mineralized shale

failing along a potential weakness plane in the cores
is 139 MPa and 128 MPa, respectively. The same weighting
process as described above was also applied to data.
The above results suggest that the unconfined compressive strength

of the dolomitic shales is locally variable. A standard deviation
of about 40-50 per cent of the mean value is commonly observed for
each core-bedding angle. Strength is also dependent on the angle
that bedding makes with the core axis of the test specimen.
A weighted average deduced for all the data plotted in Figure 5,

22.
99.9
;r- +3
I
I
99.5 .. -----~ ----- --- -~---~~--- --
99.0 - /
7
98.0
95.0
---;+ +2
90.0 A-
/
~ t-- +1
80.0
~
;;
70.0 --------- --
~ "~"'
""0'
"><
60.0 / .....0.
0.
" 50.0 /
"'
.....,"> 40.0
; -- 0 0.
......<"'
...."'
30.0 ~ ",.._
0
:>
D>
u
20.0 I
10.0
/
I ----
-1
05.0 I
I
II
-2
02.0
01.0 --
00.5
----
-3
00.1 I I ------,--
0 50 100 150 200 250
Unconfined Compressive Strength in MPa.
FIGURE 6. - STRENGTH OF 61 SPECIMENS OF MINERALIZED SHALE WITH ANGLE
OF 31 - 45 BETWEEN LOADING AXIS AND BEDDING PLANE

23.
including specimens failing along potential weakness planes in the

cores, is 185 MPa with a standard deviation of 83 MPa.
2.1.2 Tensile Strength
Published information is restricted to that of MacLeod-Carey (1969) ,

Mathews and Edwards (1969) and Mathews (1970). The results are based
mainly on Brazilian tests. A diametral compression is applied between
the platens to a rock cylinder. Tensile cracks parallel loading axis.
Again, there are differences between the various sources.
The mean tensile strengths for directions of loading normal and

parallel to bedding are indicated as:
30 MPa and 6 MPa,

38 MPa and 9 MPa, and
27 MPa and 9 MPa,
for the three sources cited above, respectively.
On the basis of these results, the average tensile strength in the

two directions of loading is calculated as 32 MPa and 8 MPa,
respectively. The overall mean tensile strength is 20 MPa.
2.1.3 Shear Strength, S , and Angle of Internal Friction, .

l
The following parameters are determined with respect to shear failure

24.
through intact rock specimens.
A diversity of terminology appears in technical literature to des-

cribe the two parameters. The definition used in this thesis is
derived from Obert and Duvall (1967) .
With respect to the Mohr's circle for stress, the shear strength S
0
of the rock refers to the intercept of Mohr's failure envelope on
the shear stress axis of the plot. Angle of internal friction.~
refers to the angular relationship between Mohr's failure envelope
and the normal stress axis of the plot.
Assessment of the two parameters appears to be restricted to results

reported in MacLeod-Carey (1969) and Miller (1978). A related study
was undertaken by Rosengren (1968) where triaxial tests were conducted
on cores with a range of core-bedding angles favourable to failure
along bedding planes. This latter work yielded results on the shear
characteristics of bedding planes and not that for shear failure
through intact rock.
The results of MacLeod-Carey (1969) suggest a shear strength of

about 30-35 MPa and an angle of internal friction at levels of
normal stress within the range of stresses anticipated at the mine
of about 55-60. The corresponding values suggested by Miller (1978)
are 18 MPa and 54, respectively.
An alternative approach for the determination of the two parameters

has been suggested by Wuerker (1959) . The shear strength S is
0
deduced from a consideration of Mohr's representation of the
unconfined tensile and compressive strengths. On the basis of the
mean values of 20 MPa and 185 MPa deduced for the two strengths,
this technique yields estimates of 34 MPa and 55 for shear strength
and the angle of internal friction, respectively.
Horibe (1970) suggested that the shear strength intercept may be

determined by the following formula:
s0
25.
where,
s0 shear strength
cr unconfined compressive strength
c
unconfined tensile strength
~
Utilizing the same mean values for the compressive and tensile
strengths, the above equation yields a shear strength value of 38 MPa.
Overall, there is close agreement between various sets of estimates.

A shear strength of 34 MPa and an angle of internal friction of
55 are assumed to be representative of the dolomitic shales at the
mine.
The range of in-situ confining stresses is unlikely to exceed 60-80

MPa at the mine . Since these stresses are considerably lower than
the average unconfined compressive strength of the shales, a signi-
ficant variation in the operative angle of internal friction is not
anticipated.
2.2 Modulus of Elasticity
Results of modulus determinations have been reported in MacLeod-Carey

(1969), Mathews and Edwards (1969), Mathews (1970), MacKavanagh and
Lee (1977) and Leask (1977)
On the basis of these, the mean modulus parallel and normal to

bedding is estimated as 104 GPa and 91 GPa, respectively. The overall
mean value for the static modulus of the intact rock is 100 GPa.
26.
2.3 Poisson's Ratio
Results reported in the references cited in the previous paragraph indicate

that the average values of Poisson's ratio in directions parallel and
normal to bedding are 0.30 and 0.24. An overall mean value for the shales
is 0.27.
2.4 Strain at Failure
On the basis of the unconfined compressive strength and modulus values

for shale specimens with the bedding plane anisotropy normal and parallel
to the direction of loading, the corresponding strain at failure in the
two directions is 0.30 and 0.37 per cent, respectively. A value of 0.19
per cent is estimated when the overall mean compressive strength of all
shale test specimens is considered.
Thus, failure of the intact rock may be anticipated to occur at strains

of 0.2 - 0.4 per cent.
2.5 Sunnary
The mean material properties determined for the intact rock component of
the rock mass system are summarized below.
27.
Unconfined compressive strength 185 MPa.

Tensile strength 20 MPa.
Cohesion 34 .MPa.
Angle of Internal Friction 55
Static modulus of elasticity 100 GPa.
Poisson's Ratio 0.27
Strain at Failure 0. 3 per cent.
28.
CHAPTER 3 PHYSICAL CHARACTERISTICS OF GEOLOGICAL DISCONTINUITIES
Several types of geological discontinuities are present within the

dolomitic shales at the mine.
The aim of this chapter is to describe the physical characteristics

of some of the fracture sets. Orientation, spacing, continuity and
frictional properties are examined. Most of these parameters are
included in the model for structural variability within the shales.
Chapter 4 supplements this work. It develops a model to account for

the observed spatial distribution of fracture sets within the rock
mass.
3.1 Source of Data
The information described below is largely based on the results of

investigations that were undertaken at the mine during the early
1970's. A summary of this work has been reported in Baczynski (1974)
and Bridges (1975). The former was concerned with the silver-lead-
zinc orebodies. The latter concentrated on the development of a
conceptual model for structure within the copper orebodies at the
mine.
Some additional parameters have also been derived during field

investigation undertaken as part of the present study.
Information on major structural features was collected by systematic-

ally mapping all accessible underground openings on several successive
29.
levels at the mine.
Detailed investigations were undertaken at selected locations. Data

was collected by means of line and area sampling of openings, as
well as, by analysis of orientated diamond drill cores.
The line sampling method is a procedure whereby all fractures inter-

cepted by a continuous, straight sample line are included in the
data. The principal advantage of this technique is that it tends to
yield unbiased results. It also permits a direct comparison to be
made between in-situ mapping and orientated drill core data.
Area mapping entails the recording of all fractures that occur within
a designated area. A more detailed discussion of this technique is
presented in the next chapter. The underlying principles of the
method are also reported in Baczynski (1980f).
3.2 Scope of Model
The most frequently observed geological weakness planes within the

mine shales are "bedding plane partings" and "extension fractures"
(more commonly termed joints) The latter refers to that group of
discontinuities which are not bedding plane partings (i.e., not
parallel to the sedimentary layering within the shales) and along
which displacement of the two complementary rock surfaces has been
normal to the fracture plane.
Locally, the shales are transected by major "shear fractures". Shear

fractures are defined as those discontinuities with some relative
displacement between the two complementary rock surfaces of the
fracture plane. These geological structures are commonly referred
30.
to as "faults". Transverse faults strike across the general strike

of stratigraphy whereas strike faults trend parallel or semiparallel
to stratigraphy.
Several different sets of shear fracture orientations have been

recognised in the mine shales (Baczynski, 1974, 1977c). In areas
where such shear fractures constitute major zones of weakness within
the mass and where their continuity is proportional to or exceeds
the dimensions of underground excavations, ground behaviour is
largely governed by their presence.
In general, the occurrence of major shear fractures is restricted

to well defined zones. Thus, their presence only affects the
stability of relatively few stapes within the silver-lead-zinc
orebodies at the mine. In view of this,
MAJOR SHEAR FRACTURES ARE NOT CONSIDERED IN

THE PRESENT STUDY.
The scope of the model is limited to bedding plane partings,

extension fractures and two sets of shear fractures (called Set Nos.
3 and 4 in this thesis) with continuities similar to those of the
extension fractures.
31.
3.3 Characteristics of Discontinuities
3.3.1 Orientation
Orientation refers to the attitude of weakness planes with respect

to the x-, y- and z-coordinate axes.
Bedding plane partings constitute the most common plane of weakness

in the shales. Three sets of fractures are also evident throughout
the areas investigated. Other fracture orientations may attain
local prominence.
The mean orientation of the various sets is not persistent. It

appears to be locally variable and a considerable overlap between
the sets is evident.
For purposes of the present study, geological structure within the

dolomitic shales is defined in terms of:
TEN DIFFERENT SETS OF FRACTURES AND A SET

OF BEDDING PLANE PARTINGS,
The range of orientations for each set is represented on a lower

hemisphere, equal area, stereographic projection plot of poles to
weakness planes in Figure 7.
The mean orientation of each set is indicated in Table 1.
Rigorous statistical methods have been suggested by Kiraly (1969),

McMahon (1967), Markland (1974), Bridges (1975) and others for the
definition and evaluation of fracture set orientation in terms of
circular or elliptical normal probability density functions. However,
32.
w E
s
~Principal Fracture Sets
FIGURE 7.
33.
TABLE 1.
Characteristics of Fracture Sets
Fracture Set No. Mean Orientation* Spacing (rn) #
1 000/80 0.46
2 090/26 0. 46
3 075/60 0. 89
4 079/81 4.07
5 040/75 8.30
6 030/50 3.52
7 130/75 21.30
8 140/52 4.26
9 180/48 6. 72
10 000/48 8.30
11 (5 0/B Hanging-wall) 270/65 0.17

11 (5 0/B Orezone) 270/65 0.34
11 (7 0/B Hanging-wall) 270/65 0.19

11 (7 0/B Ore zone) 270/65 0. 34
11 (11 0/B Hangingw all) 270/65 0.06

11 (11 0/B Orezone) 270/65 0.22
* Dip Direction I Dip Angle

# Derived from Line Sample Data
34.
such an approach is not considered to be practical in terms of

the present study.
ALL FRACTURES OR MEMBERS BELONGING TO A

PARTICULAR SET ARE DESIGNATED BY A SINGLE
MEAN ORIENTATION,
This model is similar to that adopted by Robertson (1970) for a

"design joint"; a joint which has the properties which are average
properties for a particular joint set.
3.3.2 Spacing
Spacing between discontinuities of the same set is the normal

distance between successive fracture plane intersections along a
straight line sample.
Studies at the mine suggest that the spacing between fractures of

the same set can be best described by a lognormal probability density
function. Frequency distributions for Set Nos. 2 and 3 are plotted
on logarithmic probability paper in Figure 8. The illustrated
relationships are based on diamond drill core data. Accordingly,
this sample includes all fractures with continuities down to a lower
limit of about 0.02 metre.
A lognormal model for fracture spacing has been deduced for each of
the sets, including bedding plane partings.
The frequency of bedding plane partings per metre of stratigraphy

has been recorded at some 150 locations within the immediate hanging-
wall of the silver-lead-zinc orebodies. These measurements suggest
35.
c-.
c-.
0~
>-
LJ
z
w
::::> C)
d c-.
w
a:
u..
w C)
> r-
i=
<(
...J C)
::::> Ln
::L
::::>
w
C)
rn
Ln
015 05 20 3-0 60 100 200 300
SPACING (m)
Cumulative Frequency Plot on Logarithmic Probability Paper for the

Spacing of Fracture Sets 2 and 3 (Diamond Drill Core Data) .
FIGURE 8.
36.
a strong correlation between observed frequencies and the strati-

graphic position of the orebody within the Urquhart Shales. A
detailed discussion and summary plots are presented in Baczynski
(1974).
A number of researchers have noted a relationship between the

frequency of extension fractures and bed thickness. Harris et al.
(1960) observed that for a given lithological type, the concentration
of fractures was inversely related to the thickness of the bed.
Price (1966) cites other studies which have yielded similar trends.
An investigation undertaken on approximately 400 metres of suitably

orientated drill cores failed to establish a similar relationship
between the frequency of bedding plane partings and other fracture
sets within the mine shales. In contrast, this work indicates that
the mean and median fracture frequency remains fairly consistent
with increasing frequency of bedding plane partings. The median
trend is presented in Figure 9. The scatter of results observed at
bedding plane parting frequencies above 15 per metre arises from
the lack of sample data in that region. The scatter of values is
similar to that observed at lower bedding frequencies.
The main implications of this observation are that similar ranges

of extension fracture frequencies are likely to occur throughout
the mine shales, irrespective of the stratigraphic location of the
orebody hanging-wall. Interpreted statistically, it means that:
THE OCCURRENCE OF BEDDING PLANE PARTINGS AND

OTHER FRACTURE SETS IS INDEPENDENT.
Statistical testing confirms that there is no difference at the 5%

level of significance between mean fracture frequency per metre and
increasing frequency of bedding plane partings. As a result of an
inadequate sample size, the hypothesis could not be tested for bedding
plane parting frequencies above 10 per metre.
37.
Sample Size
-4-107-32~ ..__ 22-11~ +-- 8-4---------- 2-1 - - - - - - - - - - - - - - - - -

15 I
.,
Q)
""
""u
"'
"'"r:::
10

....0.,
r:::
1l
>< 5

. . .--. .. -----------
Pl
....
0
~---~-..- -----~-------4--- e---

0
0 5 10
15 20 25 30
I
35 40
Frequency of Bedding Plane Partings per Metre of Core
Relationship between Median Frequency of Extension

Fractures and Frequency of Bedding Plane Partings.
FIGURE 9.
38.
3.3.3 Shape and Continuity
Continuity defines the extent or size of geological discontinuities.
It is apparent that unless the fractures have a circular shape, such

as has been assumed by Secor (1968), Robertson (1970) and others,
the measured continuity will be dependent upon:
(1) the chord of intersection of the fracture plane with the

sample area, and
(2) the orientation of the sample area with respect to that

of the major and minor axes defining the fracture plane.
The determination of the two dimensional shape of fractures is a

difficult task. It requires that the extent of individual fractures
should be observed in the plane of the fracture. Such observations
are rare, especially for the more continuous fractures within a rock
mass. Thus, one is forced to adopt a compromise approach and to
estimate statistically the continuity distribution in several
directions. In practice, geometry of mine layouts further restricts
directions of observation. Commonly, it is only possible to observe
fracture traces in two, mutually perpendicular, directions.
Measurement of fracture continuities is also constricted by the

finite dimensions of the area available for observations. It is
impossible to observe the true continuity of vertically or near-
vertically dipping fractures whose continuity exceeds the height of
underground openings. One or both extremities of such planes may be
located outside the field of view. This problem is well recognised,
e.g., Piteau (1973), Cruden (1977) and others.
Thus, one is restricted to the study of horizontal or near-horizontal

fracture planes that may be observed in mutually perpendicular under-
ground openings such as drives and cross-cuts within the mine.
39.
On the basis of such a study undertaken for fracture Set No.2, it

has been found that:
(1) fracture continuity parallel and normal to bedding in

the mine can be best approximated by lognormal probability
density functions,
(2) the average ratio of continuities in the two directions

is approximately 1.0 : 0.44, respectively, and
(3) the minimum continuity of fractures in the direction

parallel to bedding is about 0.1 metre.
The above results are presented graphically on logarithmic probabi-

lity paper in Figure 10.
Although the results may be fortuitous, Skempton et al. (1969)

suggested a similar ratio of major and minor axes for continuity of
fractures.
On the basis of the rather limited study, a rectangular shape with

some possible minor rounding of corners is tentatively postulated
for the average fracture plane. The major axis of this average plane
trends parallel to bedding.
A similar shape and continuity model is assumed for fracture Set

Nos. 1 - 10.
The continuity of bedding plane partings (Set No. 11) also appears
to be in accord with a lognormal model. The observed frequency
distributions are presented on logarithmic probability paper in
Figure 11.
Although rigorous field evidence is not available, the writer is of

the opinion that the shape of the average bedding plane parting is
probably equidimensional.
40.
99.9 f +3
99.5
____/_
/1 I
.L
I
I I
I
99.0
98.0 I I
I
I
I t- +2
I
/ I
95.0
I
I
I
90.0 I
~
80.0
1_.
1
.
"
;f
:P
I i
I
i ------------
+1
i;'
70.0 ..."'
c
"'"C' 60.0 / _:{_ &
...."'... 50.0 ~ I i a"'
I 0 It
If
.........~
'L
40.0 ....<
"'0......
"'
.....
"!3 30.0 I i'" - ii:
.,:il.~r__
u
81 I
20.0
-1
_';,f
--
10.0
~ ?/ :;;~;+
:!!
05.0
4v 8,; -2
02.0 .? I
l I
01.0
" ,
00.5 I
I
I
-3
00.1 I I I I .
0.1 1.0 10.0
Continuity in Metres
Cumulative Frequency Plot on Logarithmic Probability

Paper for Continuity of Fracture Set No.2 in Directions
Parallel and Normal to Bedding of Mine Shales (Line
Sample Data)
FIGURE 10.
41.
99.9
;,-~/
+3
99.5
-/ ___ -
99.0
!I/
1//
98.0
I +2
// '
95.0 //'
;I/
/ I
90.0
80.0
/// +1
;;;
(;' 70.0
./// -- Ill
rt
.,
c:
60.0 //_/ ~
g.
.,"...
tJ'
,/J/ ...D.
r... 50.0 -- - 0
Q)
....>
.., 40.0 ,W_/ Do
...<"'
.....Ill It"/ ..."'
rt
0
" 30.0 .,"
u
~~~IL~ ~
20.0
:f'
. r7/ ~~ I ...., '
-1
f/
10.0 .-<>"> IC:
1T '-~~
ICJ
w
.(!} !.(!?"'
05.0
o""
~
;;?
0
--~
1.$"
-2
02.0
01.0
"'!If"+ ,,.
o"" '"'"'
~7 '::;
I~
00.5
I
I
I -3
00.1 II I I
0.1 1.0 10.0
Continuity in Metres

Paper for Continuity of Bedding Plane Partings (Line
Sample Data)
FIGURE 11.
42.
3.3.4 Frictional Properties of Fractures
A major study of frictional properties of fractures within the

shales was undertaken by Rosengren (1968). Less extensive work is
reported in Herget (1968) and Mathews (1970).
Frictional characteristics of fractures are essentially governed by:
(1) the planarity of their surfaces,
(2) nature of the wall rock and the degree of interlock

between the surfaces,
(3) nature of the infilling materials.
On the scale of 54 mm diameter core samples, the majority of fracture

surfaces in Set Nos. 1 - 10 appear to be relatively planar, but minor
undulations and steps (1-5 mrn in depth) are locally apparent (Baczynski,
1974)
However, field investigations suggest that only a few of the more

continuous fracture planes (say, greater than 3 metres) are likely
to be completely planar.
A high degree of interlock is observed between fracture plane surfaces.
About 60 per cent of the fractures display clean surfaces without

any infilling materials along them. The remaining fractures are
infilled with limonite, carbonate, pyrite, chlorite or graphite.
Combinations of the above minerals are also apparent along some planes.
Tests conducted by Rosengren (1968) on small diameter cores suggest

0
that the residual angle of friction varies from about 40 for fresh,
clean surfaces to about 10 for some of the polished, graphitic
surfaces. Cohesion ranges from 0.0 to 1.4 MPa.
43.
On the basis of:
(l) the relative frequency of infilling material types

observed in a sample of 1440 fracture surfaces
(Baczynski, 1977c), and
(2) the frictional properties derived by Rosengren (1968)

for various types of fracture surfaces,
a mean residual angle of friction of 32 and a cohesion of 0.8 MPa

is suggested for the average fracture plane of Set Nos. l - 10.
Furthermore, the relationship noted by Mathews (1970) between
fracture roughness and peak friction angle suggests that this angle
0 0
for the average fracture plane may be about 36 to 40 .
The majority of bedding plane partings (Set No. ll) are relatively
planar on the local scale. However, 0.5 - 2.0 mm deep slickensides,
major and minor undulations, stepped offsets of 1 - 20 mm (caused
by cross fractures) and other surface irregularities are commonly
apparent on the more continuous planes. Complete to almost complete
interlock is observed between most surfaces.
Bedding plane partings may be classed as:
(l) clean-fresh,
(2) polished and graphitic.
An analysis of 1000 surfaces intercepted by diamond drilling suggests

that 43 per cent of the planes are in the second class. Results of
this study are summarized graphically in Figure 12.
In general, the proportion of graphitic bedding plane surfaces

within a sample size of "n" observations may be described statisti-
cally by a normal probability density function with a mean of 0.43
and a variance of 0.73/n.
44 .
. :____f
99.9 I +3
I
l
I
99.5 - ---
99.0
98.0
I
.L
*
+2
+
95.0
I
!
90.0 L
~ 1!
!
+1
.. 80.0
j
i
2~ '
-+--------
if
70.0
"'

~
:.
".,
I
I
'"..,"c.""
60.0 i
".,0' i
"'.,>'"
50.0 _____ ___j_
JL I
I !
-<
....
40.0
;r !
'
I
0 0.
.~
;5
..."',...
k- .
I
....Ol : I
; 0
.
"
30.0
I ! "'"
u
20.0 .1/ j j
~:
!
-1
!
.v
10.0 i
l
05.0 i
:
-2
02.0
i :
01.0
I
I I
:
00.5
I
iI
I
-3
I
00.1 I -t--,--
0.0 0.2 0.4 0.6 0.8 1.0
Proportion of Graphitic Bedding Plane Partings
Cumulative Frequency Distribution on Normal Probability

Paper for Proportion of Graphitic Bedding Plane Partings
in Samples of 10 Partings (Total Sample Size = 1000) .
FIGURE 12.
45.
Available data suggests that the residual angles of friction for

smooth, graphitic and non-graphitic planes are about 10 - 18
and 32 - 34, respectively. This angle approaches 45 for relative-
ly rough surfaces, irrespective of surface coating.
Cohesion and peak friction angles of bedding plane partings are

similar to those noted for fracture Set Nos. 1 - 10.
46.
CHAPTER 4 MODEL FOR SPATIAL DISTRIBUTION OF FRACTURES
One of the parameters affecting rock mass strength and stability is

its inherent blockiness.
The degree of blockiness is largely a function of the number of

fracture sets and their respective orientations, continuities and
intensities within the mass. Progressive ravelling or fretting
away of rock surfaces is commonly observed in mine areas which are
transected by several sets of closely spaced fractures. It is also
difficult to install effective ground support in such areas. Thus,
the assessment of local variability in fracture intensities consti-
tutes an important component in the development of a model for
geological structure.
Fracture intensities within a mass are a function of the spatial

distribution of the various sets of discontinuities within it.
SPATIAL DISTRIBUTION REFERS TO THE POSITION

OF FRACTURE PLANE CENTRES WITHIN A DEFINED
VOLUME OF ROCK MASS,
The aim of the present chapter is to examine and to test the

validity of various alternate hypotheses for spatial distribution
of fractures within rocks. Aspects covered include:
(l) Formulation of a model to account for the observed

distribution of various fracture sets within the
dolomitic shales.
(2) Definition of a field mapping procedure for collection

of the necessary data.
(3) Validation of the proposed spatial model by computer

simulation of fracture plane distribution within rock mass
blocks.
47.
4.1 General Comments
A literature survey indicates that very little information has been

published on this topic. A brief review of the subject is presented
in Baczynski (l978b). The four alternate spatial models that have
been suggested are:
(1) Even
(2) variable
(3) Random
(4) Zonal
4.2 "Even" and "Variable" Models
The "even" spacing model has been suggested by some writers, e.g.,
Hodgson (1961) and Turner and Weiss (1963). However, the lognormal
model derived for spacing between adjacent fracture of a set within
the mine shales indicates that fracture clustering occurs. This
evidence militates against the "even" spacing hypothesis.
A "variable" spacing model has been postulated by others, e.g.,

Nickelsen and Hough (1967) and Moseley (1968). However, such models
are only of a qualitative and descriptive nature.
48.
4.3 "Random" Model
Although not explicitly stated, most geotechnical analyses assume

a random model for spatial distribution of various fracture sets.
The concept of random distribution is also implicitly supported by
Bridges (1975) .
The validity of the model can be most readily tested with the aid
of a simple computer program based on the "Monte Carlo" method.
Aspects of "Monte Carlo" simulation are described in Hanunersley
and Handscornb (1964).
The necessary input data for each set of fractures to be simulated

within a two-dimensional plane consists of;
(1) average orientation of the fracture traces to be

assigned to each member of the set,
(2) a model for continuity of fracture traces,
(3) mean spacing between members of the set as noted

along line samples, and
(4) a suitable model for the number of fractures or the

total length of traces to be generated within the
defined area.
The first three parameters may be derived from existing line sample
data. The last requires some consideration. The various methods
used to generate random fracture distributions and their degree of
success are discussed in Baczynski (1977a) The procedure adopted is
based on the following conceptual considerations.
Assume that the physical dimensions of the area designated for fracture
trace generation are very large with respect to the mean
49.
spacing between fractures. Then, on the average, it may be expected

that the mean spacing between traces encountered along any hypothetic-
al sample line traversing this area in the direction normal to
fracture traces will approach the mean value for the set. This
condition should be satisfied irrespective of the model used for
the spatial distribution of fracture plane traces within the
designated area. Any number of parallel sample lines can be con-
sidered. In the limit, sample line spacing may be assumed to be
infinitely close.
On this basis, the total trace length of all fractures within the
generation area may be simply estimated by the formula:
T D
n
I sm ) X D
p
where,
T total trace length,
D average dimension of the
n
generation area in direction
normal to fracture traces,
D as above, but parallel to
p
fracture traces, and
sm mean spacing between fractures
of the set along sample lines.
This estimate of the anticipated total trace length is used to

furnish the fourth parameter.
A relatively simple computer program was developed by the writer to

test the validity of the "random" model. The basic procedure adopted
for the generation of fracture traces within a designated area con-
sisted of the following three iterative steps:
(1) Random generation of mid-point coordinates for fracture

trace.
(2) Statistical generation of a trace length in accordance

50.
with the fracture continuity model established for the

mine shales.
(3) Determination of x- and y-coordinates for the extremities

of the fracture trace through the designated midpoint.
The iterative procedure is terminated when the cumulative trace

length of generated fractures equals or just exceeds the permissible:
total length.
4.4 Testing of "Random" Model
The coordinate data generated by the above computer program was

output to a drum plotter and the results presented in the form of
fracture traces. The resulting patterns were tested by computer
line sampling and a spacing model deduced.
Figure 13 presents a typical fracture trace pattern generated on

the basis of the "random" model. The most striking feature of this
plot is that:
THE DENSITY OF FRACTURES IS RELATIVELY

HOMOGENEOUS OVER THE ENTIRE AREA,
The resulting cumulative frequency plot for spacing between adjacent

fractures is presented in Figure 14. It is readily apparent from
this plot that:
THE "RANDOM" MODEL FAILS TO REPRODUCE THE

LOGNORMAL SPACING DISTRIBUTION OBSERVED AT
THE MINE,
51.
25 30 35 40
Scale in Metres
Pattern of Fracture Traces Generated on Basis of

"Random" Model for Spatial Distribution of Fracture
Plane Centres.
FIGURE 13.
52.
99.9
99.5 - -~ --- -- - - -~
f r r;--,-
I
~ --,ft..,.
/1
,,~
--
+3
99.0 - -'lji
J

T --
I
l(J~
I~"'
98.0
j ~~~
I~..,.
1.._~
I
i
f- +2
95.0 'I
I
90.0
l'i
80.0
j I
i f- +1
;;
70.0 I /j I
I
g"'
~ ....
c
".,.
QJ
60.0
I
! /J ;
g.
...0.
QJ
"'"
50.0 i _/I
I I
i 0 g.
" i i ....
......,~ 40.0 II <
"'
.-< 11j
! I i ;
"'........
0
"~ 30.0 .,:I
u
20.0
/)f I I
10.0 //
1:-- I
i
i
:
i
-1
05.0
/7
>
/"' I i '
I
I '
i f- -2
02.0
i
01.0
;
00.5 I
I
'
r- -3
00.1
0.01 0.1 1.0 10.0
Spacing in Metres
Cumulative Frequency Distribution on Logarithmic Probabi-

lity Paper for Spacing of Adjacent Fractures Derived on
Basis of "Random" Spatial Model.
FIGURE 14.
53.
On the basis of the above results, the validity of the "random"

model is dismissed for the spatial distribution of fractures within
the dolomitic shales. It is apparent that this model underestimates
the extent of local variability in the intensity of fracturing at
the mine.
4.5 Basis of "Zonal" Model
A "zonal" model for spatial distribution of fractures has been

postulated by several workers over the last 20 years. In all
instances, the model was suggested intuitively on the basis of general
field observations relating to fracture clustering. It was presented
without the firm evidence that is necessary to disprove the commonly
stated converse hypothesis that the observed spatial relationships
may have originated as a result of a "random" process.
The evidence cited in the section above demonstrates that the

"random" model fails to account for the spatial relationship
observed at the mine. This has necessitated the formulation of an
alternative hypothesis which has been termed by the writer as the
"zonal" model.
54.
4.6 Derivation of "Zonal" Model Parameters
4.6.1 Field Investigations
The structural model derived at Mount Isa during the early 1970's
is based on a single line sampling procedure. This data was found
to be inadequate for formulation of two- and three-dimensional
models for distribution of fracture clusters. Further field
investigations were necessary.
The additional work centred on mapping of continuous area traverses

at several location in the mine. The horizontal length of uninter-
rupted traverses ranged from 20 metres to 70 metres. Field of
vertical observation was between 3 metres and 5 metres. Site
investigations were restricted to fracture Set Nos. 1 - 10.
A simple and rapid procedure was adopted for mapping of underground

openings.
Selected underground openings were photographed. The resulting,

suitably enlarged, overlapping photographs served as base maps to
record the traces of all fractures observed on the walls of the
openings. On completion of underground mapping, the data was
transferred to non-distorted maps and fracture traces were assigned
to "sets" on the basis of their orientation. A separate transparent
overlay was compiled for each set.
55.
4.6.2 Evaluation of Data
"Unit Areas" are defined on each transparent overlay in the manner

illustrated in Figure 15. For purposes of the present analysis:
A UNIT AREA IS DEFINED AS AN AREA EQUIVALENT TO

1.0 METRE SQUARED~ MEASURED ON A PLANE PERPENDI-
CULAR TO THE AVERAGE ORIENTATION OF THE FRACTURE
SET~ AND WITH ITS SIDES PARALLEL AND PERPENDICULAR
TO THE STRIKE OF THE SET,
This permits the dimensions of unit areas to be adjusted according to the

angular relationship between the strike of the set and the strike of the
underground opening that is being evaluated. In this manner, the
"normalized" unit area remains the same or constant for each set.
The prime objective of this analysis has been to derive a model for local
variability in the intensity of fracturing between unit areas.
The total trace length of each set of fractures was determined within each
unit area. Conditional probability density functions were then derived to
describe the extent of fracture intensity "zones" in directions normal
and parallel to the average trace of each set. The conceptual basis is
illustrated in Figure 16.
The following generalization is necessary for a 3-D model:
EXTRAPOLATION OF THE 2-D MODEL INTO THE THIRD

DIMENSION ASSUMES THAT THE INTENSITY OF FRACTUR-
ING REMAINS HOMOGENEOUS FOR A UNIT DISTANCE INTO
THE THIRD DIMENSION.
For example, a two-dimensional unit area with an intensity of 8.0 metres

of fracture traces of a set is converted in the three-dimensional
56.
A Central Unit Area (or Unit Volume) with respect to B and c.

B, C "Up Dip 11 and "DOW"n Dip" Unit Areas with respect to A.
D "Zero Intensityu Unit Area.
E "Non-zero Intensity" Unit Area.
D,E "Adjacent" Unit Areas with respect to c.
Unit Areas with respect to a Particular Set.
FIGURE 15.
57.
~
0
.-i
-1-l
u
Q)
h )...!
.-i .-I
Cl
Q)
u ..-!
~ Q)
Q) ..-!
;::l ..-!
..-! l'j
4-1 1-1
.. ~ l'j
H p.,
~.~.
.f" .... .
r
r .
t..
I........ -..-

.-:-:

.,

. . -:-:-:

. :-:-:-:II
.

I
!
:::: :::::::::::::I
.... <<: ::-:-: Influence in
l ....
!'
r.. :::::-:::::::
. . . :.Nonnal Direction
r.. ~
..
..... .

...
i
t- - ---e . ... ,.:_
x-axis
FIGURE 16. Conceptual Relationship: Influence of Immediately

Adjacent Areas on Observed Fracture Intensity
Within a Designated Unit Area.
58.
model to 8.0 square metres of fracturing per cubic metre or

unit domain.
It is apparent that such extrapolation would be very difficult to

justify where relatively large dimensions have been selected for
unit areas or unit volumes. The smaller the dimensions, the more
likely it is that the assumptions are valid. The dimensions of
unit areas selected for the mine shales are approximately 2.5 times
the mean spacing between fracture planes of the most common set
(for set Nos. l - 10) and about 0.7 times their mean trace length.
4.7 Some Statistical Considerations
The objective of the analysis is to formulate a statistical model

to account for the observed extent of fracture "zones" in directions
parallel and normal to the average orientation of fracture planes
in each set. In a two-dimensional, local coordinate system, these
directions of influence correspond to the y-axis and the x-axis,
respectively.
One of the considerations in processing of field data is to examine

various avenues of transforming the results into a form more amenable
to standard statistical analysis. Several statistical models are
available.
The normal probability density function is undoubtedly the most

frequently used of all probability laws (Larson, 1974). This arises
because:
(1) the normal random variable occurs frequently in practical

problems,
59.
(2) it provides an accurate approximation to a large number

of the other probability laws.
Another continuous function useful in many geological situations is

the lognormal density (Krumbein and Graybill, 1965). In its simplest
form, the lognormal distribution is defined as the distribution of
a variate whose logarithm obeys the normal law of probability
(Aitchison and Brown, 1976). Thus, if a transformation is made, such
that "L" is equal to LogbX (where "b" is the base of the logarithm),
the random variable "L" is a normal variable.
Care should be always exercised to ensure that standard distributions

are not assumed where these exhibit a poor degree of correlation
with field data. However, results are in a more manageable form if
such approximations are possible. The computational effort involved
in model testing is also considerably reduced.
Consider a hypothetical example.
Suppose that the intensity of a particular fracture set occurring

within a unit area is influenced by the set's intensities in the
immediately adjacent areas. Assumed directions of influence are
both parallel and normal to the average fracture trace of the set.
The influence in direction parallel to the average fracture trace
is described by a normal probability density function with mean ,ui
. I y
and standard deviation a 1 Values of the two parameters are
y
conditional on "i"; the actual intensity observed in the adjacent
unit area within the specified direction of influence. Similarly,
the influence in direction normal to the average fracture trace is
described by a normal probability density function with mean~;
and standard deviation ~j Again, values of the two parameters
X
are conditional on "j"; the actual intensity observed in the adjacent
unit area within the specified direction of influence.
The fracture intensity within any unit area may then be expressed
statistically by a joint probability density function of the two
continuous random variables Xj and Yi, such as for example illustrated
60.
in Figure 17.
Given that the intensities in the two adjacent areas are "i" and "j",
respectively, then the resulting joint function may be written as:
1 exp 1
-J21f. <T j X
exp
-~(-Yi -~}L~i---~--(-<T~i~/==~=~~--<_x_j_-_;U~~~-))2
__
<T~Jl -~ 2
where,
j i o-j i
?-x'~y' and cr- are as defined earlier, and
X y
correlation coefficient between Xj and Yi

with values in the range -1 :(~ ( +1
(Jj cr i
X y
Expected Value of Joint Distribution XjYi

(Expected Value of Xj x Expected Value of Yi)
The correlation coefficient is equal to zero for independent Xj and

Yi variables.
In terms of the present analysis, the random variables Xj and Yi may

be considered to be essentially independent. Both the fxj (xj) and
fyi (yi) marginal probability density functions are derived independ-
ently. The probability that a particular fracture intensity will be
encountered within an unit area may be readily determined by inter-
gration of the joint density function, where:
61.
'x! l
Sur foe IX. / y)
Joint and Marginal Probability Density Functions of

Continuous Random Variables X and Y (Reproduced from
Ang and Tang, 1975) 0
FIGURE 17 o
62.
f j Yi (u,v) dv.du
X I
In simple terms, the probability corresponds to the volume under

j i
the surface of the joint distribution f j yi (x ,y ) .
X I
There is no doubt that characterization of fracture intensities in

terms of such functions simplifies the model, provided that the
assumed distributions are in reasonable accord with field evidence.
However, it should be appreciated that it is commonly impossible,
because of sampling limitations, to validate models in the tails or
extreme ends of distributions. In fact, discontinuities with the
characteristics inferred by the extreme values may not even exist
in the prototype. A more logical approach in these circumstances
may be to constrain the range of permissible values. The resulting
model will be statistically "untidy", but it may be more represent-
ative of the real situation.
4.8 Spatial Relationship Observed at the Mine
In accordance with the above discussion, the results of the analysis

have been transformed into a form that is more amenable to statistic-
al analysis.
A variable INT is used to represent the intensity of fractures per

unit area (i.e., metres of fracturing per square metre, m/m 2 ). It
is an adjusted value derived from "Observed Field Intensity X" plus
a constant value of 4.0 added. With this constant adjustment, the
distribution of the random variable INT may be approximated by a
lognormal probability density function. Log10 (INT) is a normal
variate. A non-negativity constraint is implied logically because
63.
negative fracture intensities are impossible. Also, on the basis of

field observations, an upper constraint of 14.5 m/m 2 has been imposed
on the range of values that may be assumed by INT. This consists of
a maximum field value of 10.5 m/m 2 and the constant 4.0 m/m 2 .
Typical field results are presented on logarithmic probability paper

in Figures 18 - 20. The first plot summarizes local variability in
fracture intensity between unit areas for Set No.1. The latter two
indicate intensities observed immediately adjacent to unit areas
with a known intensity "i". Results are presented for two, mutually
perpendicular directions, i.e., parallel and normal to the average
orientation of the particular fracture set.
The above cited plots demonstrate that fracture intensities are not
random. They are markedly dependent on intensities which occur in
adjacent areas. This dependence is most pronounced for the direction
of influence parallel to the average orientation of the set.
Similar types of trends were observed for each of the other sets
examined.
Three main conclusions are apparent on the basis of the available

field data:
(1) DISTRIBUTION OF EACH SET OF FRACTURES IS

NOT UNIFORM THROUGHOUT THE ROCK MASS.
The various defined sets, in particular the less

common ones, are only present locally. Accordingly,
the mass may be characterized in terms of "non-zero
intensity" and "zero intensity" unit areas. These
correspond to local occurrence and absence of a
particular set, respectively. Where a set occurs,
intensities may be described by a distribution similar
to that presented in Figure 18. Allowable intensities
are bounded by non-negativity and upper limit con-
straints.
64.
99.9
1 Plotted Ir.tensity = , ;_ ~ +l
: Tru.t? Ir.ter.s1ty + Cor:.star.~ 1 \
99.5
99.0
t----- ----~-~-0 --- ~ ------- ~; Ji__Jl
98.0 l _ __ --------- __.;;:_____ -----~~_t __ _
i .,:: _j_ ,~ L+2
95.0
90.0
l
I
1 --- ---------- :;
""
~:~I
-
.e: --~---J
I
I
:~ J
I~
lE .
_;______ ~: :~ +1
80.0
;; .,.
,;
...,
t>-
1...
1J
70.0
. ...;!
"' ,."'
I ...,
"
g
"c:.:
60.0 ...
It:
..a.
"0" i 1"0
"a.
..."'
:..
c:.:
....>
...,
....,"'
50.0
40.0
J
I
- --------------ffi----.,{
.... ,
>
I I"
~
-+ 0 0.
"'
~-
,..,.....
0
u
"~ 30.0
I!
I
~
z,
20.0
L -1
___ __jI
10.0 ----- ----- ----------.!--------~---~-
I
05.0
--i
~
-2
02.0
01.0
00.5
I
00.1
l ___
i
l 2 3 4 5 10
J20
-3
Intensity of Fractures per Unit Area {~~rn 2 l
Variable INT

Paper for Local Variability in the Intensity of Fracture
Set No.1
FIGURE 18.
65.
Probability Density Functions for Intensities of Fractures

Observed Adjacent to Unit Areas with a Known Intensity "i"
in Direction Parallel to Average Orientation of Set No. 1.
FIGURE 19
66.
99.9 I +3
Plotted Intensity = ,'
I
True Intensity + Constant / /I'"""'
0
(True Intensity+ 4.0) I'?
I
99.5 ~---~---~-~-
I 0'
99.0
I
I~
98.0
I +2
1...
~+
.-<1:>.
95.0
....><1-..t
1.,
~I@
0+'
90.0
IH I
~0 ,c;
rli. I +1
-~Is I"'
80.0
-;; "'
~ I~
~I I ...;
G
c
70.0 ------~-~
IQl
..."'
I>
1'-' ::I
"'::lt>' 60.0 c.
"' ..."'
"'...
r.. 50.0
l.n c.
l..-4
0 c.
......,"'> 40.0
I~ ....<"'
....::l"' l";;l
r:: "'.......
30.0
IJ
..,
lUI
.
0
::I
u ...., 15
20.0
I~ -1
1.-<
I~
10.0 "'
I
I
....c
05.0 I""
l<ll
...
1] -2
10
02.0
I
I. s
01.0
00.5
1:0: "'
II It
-3
""f...,
00.1
1 2 3 4 5 10 20
Intensity of Fractures per Unit Area (mfm2 l
Probability Density Functions for Intensities of Fractures

Observed Adjacent to Unit Areas with a Known Intensity "i"
in Direction Normal to Average Orientation of Set No. 1
FIGURE 20
67.
(2) EXTENT OF "ZONES" IN THE DIRECTION PARALLEL

TO THE AVERAGE ORIENTATION OF A PARTICULAR
SET IS CONDITIONAL ON THE INTENSITY OBSERV-
ED IN AN IMMEDIATELY ADJACENT UNIT AREA,
The probability of encountering the same or greater

intensity is about 0.5. Non-negativity and upper limit
constraints apply.
(3) EXTENT OR WIDTH OF "ZONES" IN THE DIRECTION

NORMAL TO THE AVERAGE ORIENTATION OF A
PARTICULAR SET IS LESS MARKEDLY INFLUENCED
BY THE INTENSITY OBSERVED IN AN IMMEDIATELY
ADJACENT UNIT AREA,
Clear association is only evident for adjacent intens-

ities equal to or less than 1.5 m/m 2 At higher
intensities, the influence is less convincing and may
be approximated by a single function, with little loss
in accuracy.
In view of the latter observation, a simplified structural model has

been assumed for the dolomitic shales. It presumes that the extent
of "zones" in the direction normal to the average orientation of
fracture planes is independent of the intensity occurring in the
immediately adjacent unit area.
The model for local variability in the intensity of fracturing is

expressed in terms of three types of probability functions for each
of Set Nos. 1 - 10:
(1) A binomial distribution function for probability of

fracture set occurrence within unit areas, i.e.,
.probability of "non-zero intensity" unit areas. Probabi-
'
lities for the first ten sets are 0.898, 0.898, 0.464,
0.102, 0.050, 0.117, 0.019, 0.097, 0.062 and 0.050,
respectively.
68.
(2) An adjusted and transformed, normal probability density

function for determination of fracture set intensities
within "non-zero intensity" unit areas. Probability
distributions are similar to that presented in Figure 18.
Permissible range of intensities is (before adjustment) :
2
Set Nos. 1-3 0.1 - 10.5 m/m
2
4-10 = 0.1 - 6.5 m/m
(3) A set of adjusted and transformed conditional probability

density functions for extent of "zones" in direction
parallel to the average orientation of a particular set.
Probability distributions are similar to those presented
in Figure 19. Permissible ranges of intensities (before
constant adjustment) are:
2
Set Nos. 1-3 0.0 - 10.5 m/m
2
4-10 0.0 6.5 m/m
Maximum allowable difference between any two adjacent
unit areas is 4.0 m/m 2 The magnitude of this parameter
is partially based on field observations. Computer
simulation also indicated that intensities had to be
constrained in order to reproduce overall mean intensities
in accordance with field measurements.
69.
4.9 Model Testing
The objective of the testing programme was to confirm that the intensity
and spatial distribution of structural features within rock mass blocks
simulated on the basis of the proposed "zonal" model are in accord with
in situ observations. The validity was ascertained by an examination of
the generated fracture intensities per unit domain and the resulting
spatial relationships between adjacent fractures along sample lines.
A computer program CRACKS-2 was developed by the writer to assist in the

simulation process.
The function of the program is to model fracture plane distributions within

three-dimensional blocks. The program has the capacity to generate fracture
clusters in accord with a given set of input parameters for the "zonal"
model, to output the generated fracture patterns for specified sections
through the block via a corrputer plotter, to line sarry;>le the ensuing
fracture patterns and then to output the resulting statistical distributions
for spacing between adjacent fractures in a number of graphical forms. The
program has been subdivided into several independent sub-programs or modules.
This approach was considered to be necessary to permit a more convenient
manipulation of the potentially large volumes of data that may be generated
statistically. A detailed description and a full listing of the program are
included in Baczynski (1978a)
Six basic steps are employed in the pr9gram for the generation of each set
of fracture planes within a test block:
(1) Selection of dimensions for the test block and the determination
of dimensions for a "master block". The master block refers to a
minimum size, cubic block whose dimensions are such that the block
will always fully enclose the test block, irrespective of the
angular rotation about a common central axis that may be imposed
on either of the two blocks. The concept is illustrated in
Figure 21.
70.
+Y
+Z
-x - - +X
/
/
/
./
./
./
-z/
-Y
EDJ
LSI Test Block
Co!M10n Origin
Conceptual Relationship Between Test-Block and

Master-Block in Computer Program CRACKS-2.
FIGURE 21.
71.
(2) Assignment of fracture intensities to each unit domain of the

master block in accordance with the input statistical parameters
for the "zonal" model.
(3) Generation of fracture planes with a two-dimensional extent in

accordance with the fracture continuity model. The procedure is
terminated when the total intensity assigned to the master block
is satisfied.
( 4) Sorting of fracture planes into a descending sequence of

continuities (i.e., largest planes first) and location of
individual planes within the master block. Clustering of fracture
planes is in accordance with the intensities assigned to "zones"
or groups of domains within the block. All fracture planes are
initially orientated parallel to the y-z coordinate plane of the
master block.
(5) Rotation of the master block about the central axis to permit
each fracture plane contained within it to assume the average
orientation for that set in space. Fracture planes are tested
for intersection with the test block and a permanent computer
file is created for all intersecting planes.
(6) Calculation of the mean fracture set intensity per unit domain
of the master block and the generation of fracture patterns on
selected planes transecting the test block. Determination of
spacing model for adjacent fracture traces intercepted along
specified line traverses.
On the basi~ of several hundred master blocks, each comprised of about 3500
unit domains, an acceptable level of correlation was achieved between the
generated and in situ fracture intensity for each set. The results are
summarized in Table 2.
A typical fracture pattern generated for set No.1 is illustrated in

Figure 22 and the corresponding line sample spacing results are presented
on logarithmic probability paper in Figure 23. Line sampling results for
most of the other sets are included in Baczynski (1978b).
72.
TABLE 2.
Mean Intensity of Fractures per Unit Volume
Fracture Set No. In-situ Simulated
l 2.17 2.15
2 2.17 2.15
3 1.12 1.12
4 0.24 0.27
5 0.12 0.14
6 0.28 0.31
7 0.05 0.06
8 0.23 0.27
9 0.15 0.18
10 0.12 0.14
2 3
Results are expressed in m /m
73.
0 5 10 15
20 25 30 35
Scale in Metres
Typical Fracture Pattern Generated for Fracture

Set No. 1 on Basis of "Zonal" Model.
FIGURE 22.
74.
99.9 ---1" - -
I
+3
----- -~----.1--
J
99.5 ~- ----- ---- -
99.0 I
98.0
I
;
I
+2
I
v
95.0
90.0
~~I +1
80.0
;;;;
G
70.0 / Ul
....
c
"..,'0'"
60.0 / g
"'
"'
'"
"':>'"
.....u
..
.....
50.0
40.0
f.-----------~
--f
/_
I
0 "''"'
p.
It>
....<
"'...,.. .
0
!l"
30.0 ::>
i
Ill
u
20.0
I
I
/ -1
10.0 I
I
I
I
05.0 1
I
I
I -2
I
02.0
I
01.0
Sample Size 353
00.5
Mean Spacing = 0. 4
Median Spacing = .21
-3
00.1 I I
0.01 0.1 1.0 10.0
Spacing in Metres
Cumulative Frequency Distribution on Logarithmic Probabi-

lity Paper for Spacing Between Adjacent Fractures of Set
No.1 Derived on Basis of "Zonal" Spatial Model.
FIGURE 23.
75.
SPACING BETWEEN ADJACENT FRACTURES (FOR EACH SET

TESTED) DISPLAYS A VERY GOOD CORRELATION WITH A
LOGNORMAL PROBABILITY DENSITY FUNCTION,
Fracture patterns generated on six successive planes through a test

block with dimensions of 10m x 8m x 6m are illustrated in Figure 24.
Although these plots are restricted to the first six sets listed in
Table 2, rock mass variability and the contrast between low and high
fracture intensity sub-domains are clearly evident in this figure.
76.
,.. 10 Metres
FIGURE 24. Fracture Pattern for Set Nos. 1 - 6 Generated on Six,

lOrn x 8m Planes Through a Test Block on the Basis of
"Zonal" Model for Spatial Distribution (Depth of Planes
into Test Block is Indicated Beside Each Pattern)
77.
4.10 Spatial Model for Bedding Plane Partings
A detailed statistical study of the two-dimensional extent of bedding

plane parting (Set No. 11) "zones" was not undertaken at the mine.
Investigations are largely restricted to the evaluation of the

average intensity of the partings at various accessible locations in
the immediate hanging-walls of the orebodies. Locally, observations
have been supplemented by diamond drill core data.
This simple approach was necessitated by sampling considerations.
As noted earlier, the mean intensity of bedding plane partings appears

to be stratigraphically controlle~. There are marked differences
between various orebody hanging-walls. Because of this control, the
development of any realistic "zonal" model would have necessitated
an area mapping programme within very specific stratigraphic horizons
of the mine shales. Access to such locations is somewhat limited.
Thus, the proposed investigations would have only yielded a small
sample of data. This sample size would prove inadequate for the
formulation of conditional probability functions similar to those
that were derived for fracture Set Nos. 1 - 10.
THE ADOPTED MODEL FOR BEDDING PLANE PARTINGS IS

BASED ON LOCAL VARIABILITY NOTED IN THE MEAN
INTENSITY OF THE PARTINGS WITHIN OREBODY HANGING-
WALLS,
The model is supplemented with general observations on the persistence

of ground conditions within underground stopes.
The following model has been assumed:
(1) The average intensity of bedding plane partings is

stratigraphically controlled.
78.
(2) Although local variability is apparent, relatively

homogeneous mean intensities may characterize "zones" or
domains within particular stratigraphic horizons. These
may be of considerable area in their two-dimensional
extent.
(3) Local variability within any stratigraphic horizon may be

described by a normal probability density function. The
mean of the function is equal to the average intensity
of bedding plane partings noted for the particular
stratigraphic horizon. The standard deviation is equal to
50 per cent of this mean value. A probability plot for
the intensity of bedding plane partings in the immediate
hanging-wall of Orebody No. 11 is presented in Figure 25.
(4) Variability between segments of a stratigraphic horizon

roughly equivalent in area to the dimensions of a typical
hanging-wall of a sublevel open stope may also be
approximated by a normal probability density function. The
mean intensity is as defined above. The standard deviation
is about 35 per cent of this mean value. A probability
plot for Orebody No. 11 is presented in Figure 26.
79.
99.9 +3
I
I
I
I
99.5 - - - -- . --
I
99.0 /
98.0 I
~ +2
95.0 L
90.0 /
_I_ ,.... +1
;;;
80.0
v
70.0 1
"
~
Q)
"
t1'
...
Q)
60.0
50.0 L
/"
"'
Q)
....> 40.0 - /
...."'"'
u
" 30.0 A
20.0 I
/; -1
'/
10.0
05.0
v t-
-2
02.0
01.0
00.5
-3
00.1
0 10 20 30 40
Intensity of Bedding Plane Partings
Cumulative Frequency Plot on Normal Probability Paper for

Local Variability in the Intensity (per metre) of Fracture
Set No.ll (Bedding Plane Partings) in the Immediate Hanging-
wall of Orebody No.ll.
FIGURE 25.
80.
99.9 7 +3
I
99.5 ----- ~
-- -j' ---- - ~
99.0 ---
98.0 ~
I
+2
95.0
90.0
17
7
+1
80.0 - I
;;
iJ'
70.0 I "'rt
"
"'" 60.0 J
0.
..."'0.
'"
b"
""'~"
50.0 l 0 0.
........ 40.0 I "

<
>-
"'
.....
/ "'>-
rt
0
"u3 30.0
"
!.
1/)
20.0
- .I
-1
10.0
05.0
;
02.0
.J -2
01.0 I
00.5 1/
-3
00.1
0 10 20 30 40
Intensity of Bedding Plane Partings
Cumulative Frequency Plot on Normal Probability Paper for

Variability in the Mean Intensity (per metre) of Fracture
Set No.ll (Bedding Plane Partings) Between Stope Hanging-
walls in Orebody No.ll.
FIGURE 26.
81.
CHAPTER 5 MODEL FOR STRUCTURAL VARIABILITY OF LARGE BLOCKS
A model for structural variability between rock mass blocks of unit

volume (i.e., l.Om 3 ) was developed in the previous chapter. The
purpose of the present chapter is to extend the model to blocks with
larger dimensions. Potential variability in the average structural
conditions within such blocks is examined.
5.1 Model for Each Fracture Set
Any large block of rock mass can be viewed as a collection of unit

volumes. A block with a volume of 100 cubic metres is simply comprised
of 100 unit volumes. Structural variability within the unit volumes
is in accord with the statistical model developed in Chapter 4.
The model for structural variability between large blocks was developed
in the following manner. For each selected block size, geological
structure was generated by assigning and propagating fracture intens-
ities within individual unit domains comprising the block in accordance
with the "zonal" model. The average intensity of a particular set
within a block was derived by:
Average Intensity =
where, fracture set intensity

assigned to the i-th unit
domain,
82.
and n volume of block or number of

unit domains within block.
Blocks with nine different dimensions were evaluated. Their dimens-

ions ranged from 8 to 3375 cubic metres. Each block in the series
was roughly twice the dimensions of the preceeding one. The actual
range of dimensions selected was restricted by limitations inherent
in the computer program CRACKS-2 which was used for the simulation
process.
A sample of 200 blocks was generated for each selected dimension and
for each fracture set.
The resulting ranges of average intensities were approximated by

normal probability density functions. The results displayed a good
degree of correlation with the normal distribution for blocks
comprised of 50 or more unit domains.
The results of the simulations are presented in Figure 27. The

relationship between volume of the block and the "mean" and
"standard deviation" of the normal probability density function
describing variability in fracture intensity is indicated. For
example, the range of intensities of fracture Set No. 1 in blocks
comprised of 100 unit domains may be approximated by a normal
function with a mean and standard deviation of 2.15m2;m 3 and 0.5Sm 2;m 3 ,
respectively.
It should be noted that negative fracture intensities are not possible.

The cumulative area under the probability density function below
zero represents the frequency of blocks in which the particular
fracture does not occur.
83.
--------,
10 100 1000 10000
Volwne of the Block. in Cubic ~tres
1.5 +1.2
!'' ' ' ! ! U .! U [I IT !: [i;T; I >. H ,,

i,;,
+l.l
:':1:
!::': [,if.'l'i 1,:, .: ! :i L' .: ,
.
~
0
+1.0
::: '!'' 1"-r !~_1..:. . '>+_. '';'U'u
'' i :, .' i' I )\ ' ; " 1:[: 'T ' :,, . ~ ,. ,W!,
'. _
. .,_._;
;. ;:
..,
~
~
I ii ~ ; ;: j . !.! ;!J::, ''' 1 ; i I: ~
: .i'' : ,,, ::/'' P:l, . ~. 1 :n, ~
~
::!
z
?
+0.9
10 100 1000
Volume of the !lock in Cuble Ketre
-x x- Mean
Standard Deviation
Approximate Relationship Between Volume of Rock Mass & The

Mean and Standard Deviation of Normal Probability Density
2 3
Functions For Intensity of Fracture Set Nos.l-3 (m /m ) .
FIGURE 27.
84.
1.5
0.2
l.O
o.o
-<>.1
o.s
-0.2
-0.3
o.o -0.4
10 100 1000
Valume of t.he Block in C:\Jbie Metres
1.5
0.2
+0.1
1.0 0,0
-0.1
-<>.>
o.s
-0.3
~
.~
!:l
-o.o~ z:
ll
0.0 -o.s
10 100 1000 10000
Volume o! the Bloek in Cubic Ketrf!:&
-------
Mean
Standard Deviation
Approximate Relationship Between Volume of Rock Mass & The

Mean and Standard Deviation of Normal Probability Density
2 3
Functions for Intensity of Fracture Set Nos.4, 5 & 10 (m jm )
FIGURE 27 (Cont.)
85.
O.~
+0.4
+0. 3
<~-0.2
.l
+0,1
0.0
~
'l
~
-0.1
l::..,
z
?
~
-0.2
lO 100 1000 10000

Vol~ of the Block in C\lhic Metre
..0.6
, , I'"'. , !'.ciT
I
+0. 4
+(). 2
l.O
;:.1 0.0
:i ;I ':1m ; . ["~ .~.. I~C;cjFC;:p'++J"'+H+?f:::;:'ll""!-+cf4~+2J

.
::
'. 'i !':~>!; ''IT. I'll::, '''"'.V i I 1:1" -;:"" '! '' if
. . :, ! ''!i . lii'i . ' .1: ' -0.2
! ::I ' I' i ': I : IbiS : ''! T ;, I ' ,, ' ' ' ... :. ' ' ' IC '
I I:!
0. 5 -0 . .4
..!l
..... 'lc:
ill
"'..,'"
-0.8 ?"'
...
0.0
-------,
lO 100 1000 10000
Volume ot the Blod in CUbic Ket.res
Mean
~-------
------ Standard Deviation
Approximate Relationship Between Volume of Rock Mass & The Mean

and Standard Deviation of Normal Probability Density Functions
2 3
for Intensity of Fracture Set Nos.6 and 7 (m /m ).
FIGURE 27 (Cont.)
86.
1.5
1.0
0.5
0.0
10 100 1000 10000
Vclwne of t..he Block in Cubic Metres
+0. 3
+0. 2
40.1
0.0
-0.1
-0.2
... o.J ..
z
0
:.
-o.
10 100 1000 10000
Volume ot the Block 1n CU::Ioic Metre
--------.-- Mean
Standard Deviation
Approximate Relationship Between Volume of Rock Mass & The Mean

2 3
for Intensity of Fracture Set Nos. 8 and 9 (m /m )
FIGURE 27 (Cont.)
87.
1.5
;;
oll
:
l.o:
.
~ f
0
2.
.
~
0.5
o.s~
I .
] :
i .""'
i
.
~
~
il"'
o.o 0.0
lOCO 10000
Volwne of the Block in Cubie Metrea
----------
____
_., __...__
Mean
Standard Deviation
Approximate Relationship Assumed Between Volume of Rock Mass &

The Mean and Standard Deviation (Expressed as a Ratio of the
Mean Value) of Normal Probability Density Functions for Inten-
2 3
sity of Fracture Set No. 11 (Bedding Plane Partings) (m /m )
FIGURE 27 (Cont.)
88.
5.2 Combined Model for Rock Mass
Variability in the average intensity of any fracture set within the

mine shales may be approximated by means of normal probability
density functions. Thus, the intensity may be considered as a
JL. and o-: 2 are
normal variate, i.e., n(;Ui; o-:-1. 2 ) - variate, where r~1. 1.
the mean and variance of the normal probability density function of
the i-th fracture set.
The average intensity of fracturing of all the sets combined may be

considered in terms of a multivariate probability density function
of "n" random variables x 1 , x 2 , x 3 , ..... , xn, where each random
variable refers to the average intensity of a particular set of
fractures within the rock mass.
The multivariate probability function is some function:
......... , X
n
X
n
< an ) is equal to
;X
n
For purposes of the structural analysis, the random variables

X1 ; x2 ; x3 ; .... , Xn (i.e., the various sets of fractures)
have been assumed to be mutually independent. In other words, at
the geological time of fracture formation, the intensity with which
fractures of a set developed was neither dependent upon nor
influenced by the presence or absence of any of the other sets within
the rock mass. This implies that for all values of x1 ; x2 ; ;X
n
f ( Xl) f (X ) f (X )
; X
n
x1 x2 2 xn n
89.
The average intensity of all fracture sets combined may be deter-

mined by the sum of the independent normal variates x 1 , x 2 , x 3 , .
, X , where each X. has a n ( iJ. . ; cr. 2) distribution.
n 1 I 1 1
Provided that the functions are not truncated by constraints, then

according to Larson (1974), if Y represents the sum of the variates,
i.e., Y is the total intensity of all fracture sets combined,
y X.
1
i=l
then, Y is a normal variate with mean and variance given by:
jLy
n
and cr
y
2
I
i=l
cr.
1
2
The random variable Y may be considered as a n(iA ;CJ" 2 ) -variate.

I y y
Thus, the total intensity of all fracture sets combined is a normal
probability density function with a mean and variance equivalent to
the sum of the means and variances for the various fracture sets,
respectively.
It is evident from Figure 27 and earlier discussions that some of

the functions describing fracture intensities are truncated by non-
negativity and upper limit constraints. Therefore, i t is not
possible to combine these functions in the simple manner detailed
above.
90.
The problem was resolved by:
(1) discretizing each continuous density function into a

number of small, uniformly spaced, discrete class
intervals, and
(2) combining them by a series of cross-multiplications
to yield a single probability density function for total fracture

intensity. The task was expedited by means of a simple computer
program.
The resulting function for variability between unit domains with

respect to fracture set Nos. 1 - 10 combined (i.e., excluding
bedding plane partings; set No. 11) is illustrated in Figure 28.
The distribution is a three parameter lognormal probability density
function, such that Log10 (Derived Intensity) plus a constant value
of 6.0 is a normal variate.
Variability in the intensity of all fracture sets combined in

hanging-walls of Orebody Nos. 5, 7 and 11 is summarized by a series
of functions in Figures 29-31.
91.
99.9
99.5
Plotted Intensity =
Derived Intensity + Constant
(Derived Intensity +
----
6.0
---
)
---1----
l
I
I
I
I
7!
J ~
+3
i~-
I
99.0
I
.~
I ,u
98.0 f--
I I~
95.0
I
I
I
I
j II-<
1<1'
10
,:;
+2
I
I
I I ..
I 12$
I I"
90.0
;r
I
"'!I
"'I IU
+I I" +l
I 1-.-<
80.0 .,I
.... , :j
~
G
70.0
~I
.,,' I 1.,
VJ ,.til
"
"..."'
60.0
"I
HI
f 1_
I<
~
p,
0'
"' 50.0
1
I
~ 1\D
I.
""
p,
r.. -{+ 0 p,
~
11>
~:
..."'
<
.....> 40.0
....rol '7
I"'!
I!:!
.....,.
....
.
"'
.-<
t:! I 0
"!3 30.0
Sl ~-
"
Ill
u
20.0
UJ / l";;j
;..,
I
;I I"'
1<: -1
"'"I t8 '
~ ,,..,
10.0
I
~I I"""'
II 1:'!
"I
,gl 10>
05.0
;~
02.0
1!
-z
IE
::o:
I
:!3
I~
-2
01.0
I I
I
I
00.5
I
I I
I
I I
I I -3
''I I
'
00.1
1 2 3 4 5 10 20 30
Intensity of Fractures per Unit Domain

(m2 /m3)

Paper for Local Variability in the Intensity of Fracture
Set Nos. 1-10 Combined.
FIGURE 28.
92.
99.9 --
I
+3
99.5 --- - ---- ---- ---- ---- -- -

j
99.0
98.0 r
I
r-- +2
;
95.0
90.0
80.0 I +1
-;;
~
70.0 I
":>tt
Q)
60.0 I
f!
"':>
50.0 ~
CIJ
....... 40.0 I I
Ill
.....
:>
9
30.0 I .
u
20.0 i
10.0
;
I -1
05.0
02.0 / -2
01.0 I
00.5 !
I
/ -3
I
00.1 I I
0 1 4 9 16 25 36 49

(n,2;m3 )
Cumulative Frequency Plot on Probability Paper for Local

Variability in the Intensity of Fracture Set Nos. 1-11
Combined within the Hanging-wall of Orebody No. 5 (i.e.,
Square Root of Intensity is a Normal-Variate).
FIGURE 29.
93.
99.9
99.5
99.0
... --
------.---1- . -
98.0 I
I
+2
95.0
/
7
90.0
.. 80.0 I +1
~
70.0 I
.,"'
:;l
.,0'
60.0 I
....,... 50.0 I
:> 40.0 j
..
...,
-<
.....
30.0 I
u
20.0 I
10.0 I -1
05.0 I
02.0
I
I
I
-
-2
01.0
00.5
-3
00.1
0 10 20 30 40 so 60 70

(m2/m3)
Cumulative Frequency Plot on Normal Probability Paper

for Local Variability in the Intensity of Fracture Set
Nos. 1-11 Combined Within the Hanging-wall of Orebody
No. 11.
FIGURE 30.
94.
1.0
2.0 6.0
-~
8
~
>
.:
"'~c ...
. ~1li
1.0 5.0
9
z
2
:>
~z
""'
0.0 4.0
10 100 1000 10000
Volwnc of' tho: Block in Cubic Metres
5.0 13.0
4.0 12.0
~
::
.
8
'2 .:
"'
c
.
.'l
3.0
~

2.0 10.0 z
g
10 100 1000 10000
VolUmO of the Block in C\lblc Hetrs
____
__... ...- Mean
Standard Deviation
Approximate Relationship Between Volume of Rock Mass & the Mean

for Intensity of Fracture Set Nos. 1-10 and 1-11 (within Hanging-
2 3
wall of Orebody No.5) Combined, Respectively (m /m ).
FIGURE 31.
95.
4.0 12.0
).0 ll.S
2.0
1.0 10.5
----,
10 100 1000 10000
Volume of the Block in Cubic Metres
22.5
8.0
7.5
22.0
7 .o
6.5
6.0
5.5
~
g
21.0;:
8
10 100
Volwn. of tho Block in CUbic Hetrea
_,_____.. Mean
------4.-- Standard Deviation
Approximate Relationship Between Volume of Rock Mass & the Mean

for Intensity of Fracture Set Nos. 1-11 Combined within the hanging-
2 3
walls of Orebody Nos. 7 and 11, Respectively (m /m).
FIGURE 31 (Cont.)
96.
CHAPTER 6 - ROCK !"lASS PROPERTIES
The dolomitic shales are structurally complex.
Accordingly, the mechanical properties of the rock mass will be locally

variable and the response of the mass under load will be complex. The
objective of this chapter is to gain a broad, but statistically-based,
appreciation of local variability in mechanical properties of the mine
shales. This is achieved by representing the structural complexity in terms
of a suite of material parameters.
Several interrelated factors will govern the response of complex models:
(1) Range of mechanical properties and their frequency of occurrence

within the mass.
(2) Complexity and magnitude of the stress field around the

opening ( s) .
(3) Probability that the mechanical properties of the mass are

exceeded along "critical paths" or potential failure zones
within the mass .
6.1 Method of Analysis
As noted in Chapter 1, the mechanical properties employed in the present

analysis are largely derived on the basis of published rock mass classifi-
cation systems. The necessary geological input is provided by the
statistical model for structural variability. The basic principles and
97.
applicability of various classifications and other techniques are

discussed in Baczynski (1979). A brief summary of the main results is
included in Baczynski (l980f).
An overview of the task suggests that some aspects of the analysis may
be readily undertaken by means of sil!Y?le graphical transformations between
the model for structure (or other field indices) and various rock mass
parameters. However, the derivation of several other parameters via the
more comprehensive classification systems is a slow, repetitive and very
time-consuming process. The latter arises from the necessity to "sample"
and evaluate a large number of possible structural situations before a
statistical appreciation is gained of the range of parameters.
6.2 Computer Simulation
To expedite the evaluation process, a computer program GEOCLAS was written

for the determination of several classification ratings, including the
RMR-, Q-, RQD- and Modified RMR- indices as proposed by Bieniawski ( 19 7 3,
1976), Barton et al. (1974), Deere et al. (1966) and Laubscher and Taylor
(1976), respectively.
The principle employed in the CO!!Y?uter program is quite simple. The above
indices are determined for rock mass blocks with ground conditions
simulated statistically on the basis of models for the intensity of each
set of geological discontinuities, plus specified inputs for the other
necessary parameters. The procedure is repeated several hundred times to
permit a statistical model to be formulated for variability between blocks
with selected dimensions (or volume) . The minimum number of iterations
employed for each block size in the present assessment is 1000. The
capability of the program also extends to the determination of relation-
ships between the various classification ratings.
98.
A full FORTRAN listing of the program, description of input data

requirements and exarrples of typical output are included in Baczynski
(1979) 0
6.3 Model for RQD-Index
One of the parameters commonly included in rock mass classifications is

the "rock quality designation" (RQD) -index proposed by Deere et al. ( 1966) .
The index is modified core recovery percentage in which all the pieces of
sound core exceeding 0.1 metre in length are counted as recovery.
To permit this parameter to be manipulated in parallel with the statistical

model for geological structure, it was necessary to establish a correlation
basis between fracture frequency and RQD-index rating.
A literature survey on the topic indicates marked differences between

various sources (Baczynski 1979, l980f). The mean trends adapted from
Deere et al. (1966) for a metamorphic rock type, and proposed by Priest
and Hudson (1976) for sedimentary rocks and by Barton et al. (1975) for
essentially igneous rocks, as well as by Kulhawy (1978) from theoretical
considerations, are illustrated in Figure 32. Further discussion on the
subject is pres en ted in Hudson and Priest ( 19 79) and Goodman and Smith
(1980) . It is evident from the results that no single method can be
considered to have universal application.
The mean correlation established at the mine on the basis of diamond drill
core data is also illustrated in Figure 32. The scatter of individual RQD
values about the mean trend is about 15 per cent.
The results are in extremely good accord with those of Priest and Hudson
(1976) , especially for fracture frequencies less than 20 per metre. The
confidence of the trend at higher fracture frequencies is reduced by the
99.
0
0
....- 5629
0 PUBLISHED TRENDS
co
X
w 0
0 ...a
z
t
0
0 ....:t
d
cr:: Hud
0
son f1976;
N
0
--
0 10 20 30 40
AVERAGE NUMBER OF FRACTURES PER METRE
0
0
~
5630
0 DOLOMITIC SHALES:
<X)
MOUNT ISA MINE, QLD.
X 0
w ...a
0
z
t
0
0
.....::t
d
cr:: 0
N
--- ---
0~------~~--------~--------L-------~~~
0 10 20 30 40
AVERAGE NUMBER OF FRACTURES PER METRE
Relationship Between Fracture Frequency and RQD-Index: Mount Isa Mine
FIGURE 32.
100.
lack of appropriate data.
It is interesting to note that both the Mount Isa Mine and the Priest
and Hudson trends have been derived for sedimentary strata. These results
suggest that one of the governing factors in the observed relationships
may be rock type.
6.4 Basis of Model for Rock Mass Strength
6.4.1 In tact Rock Mass
The mean unconfined compressive and tensile strengths of intact rock core
samples are 185 MPa and 20 MPa, respectively. A more detailed discussion
of core strengths is presented in Chapter 2 of this thesis. Considerable
variability is apparent in the results, even for texturally similar test
samples. The variation appears to be further compounded by the anisotropic
effect of sedimentary layering within the dolomitic shales.
It is anticipated that a significant proportion of the observed variability

is a function of test sample dimensions. Therefore, it is unlikely that
the model derived for relatively small test samples can be extended direct-
ly to describe the variation in the strength between larger rock mass
blocks.
A simple example will readily demonstrate the effect of sample size on

variability. Assuming that the core strengths are statistically independent
and that the range of strength values for 54 mm diameter cores may be
101.
described by a normal probability density function with a mean r and

variance a 2 , then the corresponding model for variability between unit
3
volumes (l.Om ) of the mass may also be expressed in terms of a normal
function, but with a mean~ (i.e.,~, suitably adjusted by some scale
factor to account for strength reduction due to intergranular defects)
2
and a variance of about 0. 002 * o-
It is apparent that the variability between unit volumes of intact rock

mass is very small in comparison to that between the test samples. This
suggests that any potential variability in the strength of intact rock
mass blocks is likely to be imposed only by the sedimentary layering of
the shales.
6.4.2 Discontinuous Rock Mass
So far, the discussion has been limited to the strength of intact rock
blocks, whereas the dolomitic shales are characterized by the presence of
geological discontinuities.
Published results of field and laboratory testing, as well as physical

modelling studies of blocky systems, clearly demonstrate that the strength
of rock masses is appreciably reduced by the presence of geological
weakness planes. In instances of highly fractured rock, there is little
relationship between the strength of the intact rock and that of the
discontinuous mass.
Most experimental information is available for coal and, therefore, does

not have direct application to the hard rock environment at the Mount Isa
Mine. A comprehensive discussion of the subject is presented in Lama and
Vutukuri (1978 ) . A number of the techniques reviewed in the present study
are included in Baczynski (1979)
102a.
The dolomitic shales are characterized by the presence of several sets of

geological discontinuities as well as by local variability in their
intensity per unit volume of the mass. Therefore, some degree of local
variability may be anticipated in the strength of the rock mass.
A number of field and laboratory studies that are reported in literature

have attempted to derive a parallel model between intensity of fracturing
and rock mass strength. None of the suggested methods are entirely
satisfactory and i t is also difficult to extend them to derivation of a
statistical model for variability. However, the work of Hansagi (1965,
1974) , and possibly that of Stimpson and Ross-Brown (1979) , appear to be
the most promising with respect to the present analysis.
Hansagi proposed a diamond drill core classification index (essentially

based on the intensity of fracturing) which, when coupled with the
unconfined compressive and tensile strengths of the cores, permits
estimates of rock mass strength. The index has been developed for a rock
mass with broadly similar geological characteristics, apart from sedimen-
tary bedding, to those of the mine shales (i.e., tensile strength of
intact rock cores, rock density, and similar ranges of fracture intensities
as suggested by the values determined by the C-factor index). Therefore,
the method offers a potentially useful approach and accordingly i t has been
adopted in the present study.
The relationship between C-factor index rating and the total intensity
of fracturing within the mine shales is presented in Table 3. Non-tabulated
values may be approximated by linear interpolation.
Estimates of effective rock mass friction angle and cohesion are based on
Bieniawski (1973, 1976). Figure l of the earlier technical paper presents
a relationship between the unconfined compressive strength of intact
rock cores, spacing of joints and the cohesion and friction angle of a
rock mass. Section D of Table II in the latter publication suggests
ranges of friction angles and cohesion values for rock masses in each
of the five RMR-index ratings.
The proposed cohesion values are rather low. They appear to be

characteristic of completely blocky rock masses. The discontinuous nature
102b.
of most weakness planes within the dolomitic shales at the mine

suggests that the mass will only be completely blocky locally. Therefore,
it is anticipated that the rock mass as a whole will possess a greater
degree of cohesion than may be predicted on the basis of Bieniawski
(1973, 1976).
In the present analysis, cohesion has been deduced from a Mohr-circle

representation of the unconfined rock mass strength and a simple, linear
failure envelope incident upon the stress circle. The reduction factor
proposed by Hansagi (1965, 1974) is used to estimate the unconfined
compressive strength of the rock mass. Friction angles are based on the
values suggested by Bieniawski (1976). The procedure employed to determine
cohesion is illustrated graphically in Figure 32A.
It is appreciated that the approach is very simplistic and possibly,

a little unrealistic in some aspects. A linea~ failure envelope is
assumed. Therefore, some cohesion will be predicted in all instances
where the unconfined compressive strength of the mass exceeds zero. It
is also apparent that the magnitude of the derived estimate will not
only be governed by the unconfined compressive strength, but will also
be a function of the friction angle. For example, rock masses with
identical unconfined compressive strengths but different friction
angles would automatically possess different cohesions.
The true shape of the failure envelope can only be defined by an

extensive field testing programme. In the absence of such tests, some
assumptions have to be made. A linear failure envelope is considered
to present a reasonable first-order approximation.
l02c.
' ',
'
c
' ',
'
' ',
'
o-1 /2 o-1 0'
n
CTl = Unconfined compressive strength of

rock mass.
()3 = 0.0 (Confining stress)
(J Normal Stress.
n
f Shear Stress.
cp = Rock Mass Friction Angle.
c = Rock Mass Cohesion.
Derivation of Rock Mass Cohesion.
FIGURE 32A.
103.
TABLE 3.
Relationship Between C-Factor and Intensity of Fracturing per Metre.
Intensity I Metre C-Factor
0.0 1.00
3.0 0.79
5.5 0.65
7.0 0.56
8.5 0.50
9.5 0.45
24.0 0.13
56.0 0.00
6.5 Basis of Model for Rock Mass Modulus
The rock mass is a multi-phase system composed of intact rock and

geological discontinuities. Elastic properties of intact rock are
essentially dependent upon the mineralogy of the rock, whereas the elastic
properties of rock discontinuities depend on their length, tightness,
waviness, degree of interlock and matching between asperities, the elastic
properties of the infilling materials and, commonly, the magnitude of the
confining stress field.
104.
6.5.1 Modulus of Intact Rock
The effective Young's modulus of intact rock may be determined by the

methods described in Lama and Vutukuri (1978 ) . One possible approach is
indicated below:
1 va vc vn
+ + + ... +
E E E
a c n
where, V, Vb, V, . , V are the percentages of the minerals in the

a c n
rock as determined from modal analysis and Ea' Eb, Ec' .. ,En are the
respective Young's modulus values.
An alternative approach is:
V .E
a a
+ + V .E
c c
+ + V .E
n n
where,
(1) if the shape of the grains is spherical, the volumetric

(Va' Vb 1 etc.) percentage of the minerals is taken into
consideration, or
{2) if the shape is not spherical and the particles have a preferred
orientation, the percentage surface area of the minerals along
a section parallel to the direction of application of the stress
is used in calculations. This allows a determination of
modulus of anisotropy in different directions.
According to Lama and Vutukuri (1978 ) 1 the first equation gives the upper
limit whereas the latter gives the lower limit. The actual value is
located somewhere in between the two limits.
As noted in Chapter 2 of the thesis 1 the mean modulus parallel and normal
to sedimentary layering of the shales is 104 GPa and 91 GPa, respectively.
105.
The overall mean value is about 100 GPa.
6.5.2 Theoretical Approach to Rock Mass Modulus
The elastic properties in any direction through a rock mass may be

determined from a consideration of the elastic properties of the intact
rock material in that direction as well as the orientation, spacing and
the normal and shear stiffness of the geological discontinuities within
the mass. For example, in the case of an orthogonally jointed rock mass,
the modulus in the i-th direction may ~e determined by means of relation-
ships presented in Lama and Vutukuri (1978) and Kulhawy (1978):
E.
~
= 1
1 1
E
+ (S . K .)
r 1. n1
where,
E.
~
= elastic modulus of the rock mass in the i-th
direction,
E elastic modulus of the intact rock material,
r
s. ~
spacing of discontinuities in the i-th direction,
K
ni = normal stiffness of the discontinuities in the i-th
direction.
Similarly, the shear modulus in the ij-th plane may be determined by:
G . = 1
~J
1 1 1
+ (S . K .)
+ (S . K ,)
G
r l. Sl.. J SJ
where,
106.
G shear modulus of the intact rock,

r
s. )_
spacing of discontinuities in the i-th direction
K. shear stiffness of discontinuities in the i-th
s~
direction,
s. spacing of discontinuities in the j-th direction,
J
K . shear stiffness of discontinuities in the j-th
SJ
direction.
6.5.3 Rock Mass Modulus
Some estimates of the elastic properties of intact rock, geological

discontinuities and rock masses have been reported in Goodman et al. (1968),
Lama and Vutukuri (1978) and Kulhawy (1978). It is apparent from the
available information that:
(1) Normal and shear stiffness of rock discontinuities may assume

a wide range of values.
(2) Measured stiffness is likely to be dependent upon the

magnitude of the confining stress field.
(3) Estimates of rock mass modulus are very sensitive to the

stiffness values assumed for the discontinuities.
Therefore, the theoretical approach to the determination of rock mass

modulus is only useful if supported by reasonably accurate field tests
undertaken within the rock mass that is being evaluated. In the absence of
such tests, the reliability of estimates derived by the theoretical approach
must be at least as suspect as those determined on the basis of rock mass
classifications. Accordingly, estimates determined on the basis of
classifications may be viewed as providing adequate first order approxima-
tions of the parameter. The approximations are no less accurate than those
107.
resulting from the theoretical approach where the stiffness of the rock
discontinuities is not known but only estimated.
Several rock mass classification approaches are described in literature.

Rock mass modulus may be estimated on the basis of either the relation-
ship between the parameter and RQD-index as derived by Coon and Merritt
(1970), Cording et al. (1971), Orr (1974) and others, or from the relation-
ship with the RMR-index presented by Bieniawski (1975, l978a, 1978b). Both
sets of approaches yield similar mean and ranges of modulus values for
the dolomitic shales (Baczynski, l980f).
There are obvious limitations in each of the two index-approaches and

neither of them can be anticipated to yield anything more than an
approximate estimate. However, since the RQD-index is both easier to
derive and to manipulate in parallel with a statistical model for
variability in the intensity of fracturing, it has been employed in the
present analysis.
The modulus reduction factor (i.e., E

mass
I E.~n t ac t ro ck) for the dolomitic
shales may be approximated by two linear equations (Baczynski, 1979):
Modulus Reduction Factor 0.1 + 0.0014 * RQD

for RQD ~ 70
0.0267 * (RQD- 62.5)
for RQD > 70
The necessary input to the theoretical and the rock mass classification
approaches is provided by the statistical models for local variability in
the intensity of each fracture set and the total intensity of fracturing,
respectively. Where the normal and shear stiffness of fractures or the
relationship between total fracture intensity and RQD-index are known, the
structural model may be sampled iteratively by the Monte Carlo process to
derive a parallel model for variability in rock mass modulus.
It is apparent that the theoretical approach permits evaluation of

directional variability, whereas estimates derived on the basis of the
classification approach are limited to isotropic assumptions.
108.
6.6 Poisson's Ratio Assumed for Rock Mass
Although the conceptual aspects of determining the Poisson's ratio of

composite systems such as soil and rock masses has been discussed by a
number of authors, including Wardle and Gerrard (1972), Harrison and
Gerrard (1972) , Kulhawy (1978) and Lama and Vutukuri (1978 ) , there appears
to be little confirmatory field evidence to support the theoretical
estimates. Most published information is restricted to intact rock cores.
A typical example is that of Kulhawy (1975) who reports that the ranges
for clastic and chemical sedimentary rock types (similar to the dolomitic
shales) are 0.03- 0.46 and 0.04- 0.73, respectively.
Rock masses behave in a complex manner under load and various modes of
local deformation, other than elastic, have been observed. As failure
conditions are approached, deformations may include local slip along
geological weakness planes and dilation caused by block rotation. These
mechanism may operate locally at an early stage of loading, without
necessarily resulting in complete collapse of the system. Therefore, it
is doubtful whether the concept of elastic Poisson's ratio can be applied
in a strict sense to describe the observed behaviour. Conceptually, it
may be a little :rrore appropriate to refer to it as the "lateral deformation
ratio".
The dolomitic shales at the mine are structurally variable. Therefore, the
lateral deformation ratio may also be expected to be locally variable,
unless the stiffness of the discontinuities is identical to that of the
intact rock or the discontinuities are present in the mass with equal
intensities in each of the three coordinate directions.
Since the appropriate deformations have not been measured in the field,
a somewhat subjective approach has been adopted for the determination of
this parameter. It is intuitively anticipated that the ratio will increase
(possibly logarithmically) with increasing intensity of fracturing. This
view is also held by Dr. R.D. Lama of the CSIRD (personal communication)
on the basis of his physical modelling studies. A range of about 0.2 - 0.5
109.
has been suggested for the mine shales.
The general appropriateness of this range has been reaffirmed theoretically

by application of the equations presented in Kulhawy (1978), where a
normal stiffness of about 60 - BO GPa is assumed for discontinuities_, and
fracture intensities are in accord with the statistical model for structure
within the dolomitic shales. Higher ratios are probably unlikely under the
confining loads that are effective at the Mount Isa Mine.
Laboratory evidence cited in Lama and Vutukuri (1978 ) indicates that

Poisson's ratio increases with increasing axial load, possibly as a result
of plastic deformation in test specimens, and that the ratio decreases
with increasing lateral or confining load. However, the inclusion of such
complex relationships is outside the scope of available numerical techni-
ques and i t is assumed in the present analysis that the ratio remains
constant under different load conditions.
6.7 Statistical Model for Material Properties
6.7.1 Definition of Domains
Two types of structural domains are recognised within the dolomitic shales.
They are the essentially barren shale "hanging-wall" and the economically
mineralized "orezone". In the case of each orebody studied, the two
domains are distinguished by different mean intensities of bedding plane
partings (fracture set No.ll). The mean intensity of each of the other
fracture sets appears to be the same in both domains.
110.
6.7.2 Material Types
EIGHTEEN(l8) ISOTROPIC MATERIAL TYPES ARE

EMPLOYED TO CHARACTERIZE MECHANICAL VARIABILITY
WITHIN THE DOLOMITIC SHALES,
The range is intended to represent the full spectrum of local

variability in ground conditions that are encountered at the mine.
It extends from intact rock to highly fractured rock at the two
ends of the scale. Each of the material types is defined in Table 4.
The derivation of "mean" and "+1 standard deviation" material pro-

perty types for the hanging-wall and orezone domains of Orebody
Nos. 5, 7 and 11 is based on corresponding set of structural
conditions postulated for these areas. The relevant material types
are indicated in Table 5. Since the ground conditions in Orebody
Nos. 5 and 7 are very similar, only data for the former is presented
in the table.
6.7.3 Frequency of Occurrence
The relative frequency with which each of the material types occurs
within the rock mass parallels the model for local variability in
geological structure.
The material model derived for unit volumes (l.om3 ) within the
hanging-wall and orezone domains of Orebody Nos. 5 and 11 is
I
t
presented on probability paper in Figure 33.
l
TABLE 4
Rock Mass Material Properties
Material Type Modulus (GPa) Poisson's Friction Angle Cohesion (MPa) Tensile * Unconf.Compressive
Ratio (Degrees) Strength (MPa) Strength (MPa)
1 95 0.20 45 37.0 8.4 178
2 80 0.22 44 34.0 7.7 165
3 71 0.24 43 31.0 6.6 146
4 63 0.26 42 27.0 5.5 124
5 45 0.28 41 22.0 4.6 104
6 37 0. 30 40 19.0 4.0 87
7 35 0.32 40 18.0 3.6 78
8 33 0.33 39 16.0 3.2 69
9 31 0.34 39 14.0 2.8 59
10 29 o. 35 38 12.5 2.4 52
11 27 0. 36 37 10.8 2.0 44
12 25 0.38 36 8.4 1.2 33 .....

.....
13 23 0.40 35 5.8 1.0 22 .
.....
TABLE 4 (CONTINUED)
Rock Mass Material Properties
Material Type Modulus (GPa) Poisson's Friction Angle Cohesion (MPa) Tensile* Unconf.Compressive
Ratio (Degrees) Strength (MPa) Strength (MPa)
14 21 0.41 34 4.8 0.8 19
15 19 0.42 33 3.9 0.6 15
16 17 0.43 32 3.0 0.4 11
17 15 0.44 31 2.1 0.2 7
18 13 0. 45 30 0.8 0.05 3
* Tensile strength deduced on basis of Hansagi (1965), but NOT used in VISPLAS analysis.
1-'
1-'
IV
TABLE 5
Average Material Types Used in Modelling
Orebody Domain Material Property Types

-1 Standard Mean +1 Standard
Deviation Deviation
5 Hanging-wall 9 8 7
Orezone 7 6 6
11 Hanging-wall 13 12 9
Ore zone 8 7 6
t-'
t-'
w
114.
99.9 +3
99.5
99.0
98.0
+2
95.0
90.0
+1
80.0
...
G
<:
70.0
"'g...
"'D'
:I 60.0 0.
."' 50.0
..."'0.
"'"'> 0 0.
"'
-< 40.0 .,,.....<
"'Ill
..... ....
u
30.0 .
0
::s
20.0
-1
10.0
05.0
-2
02.0 (1) ~ 5 Orebody Hanging-wall
e e (2) ~ 5 Orebody Orezone
01.0
00.5 'f' Y (4) 11 Orebody Orezone
-3
00.1
1 2 3 4 5 6 7 8 9 10 u 14 16 18
MATERIAL TYPE
Cumulative Frequency Distributions Plotted on Probability

Paper for Occurrence of Material Types within Unit Volumes
(1.0 m3 ) of the Rock Mass.
FIGURE 33.
l
115.
CHAPTER 7 STABILITY ANALYSIS
Three different approaches are used in this chapter to assess the

stability of unsupported stope hanging-walls and abutment areas:
(1) Rock Mass Classifications
(2) Two-dimensional, Continuum, Finite Element Method
(3) Probability of Failure Analysis.
THE PARAMETRIC STUDY IS LIMITED TO MODELLING OF

THE SEQUENTIAL EXTRACTION OF A SINGLEJ RECTANGULAR-
SHAPED STOPE.
Stopes within Orebody Nos. 5 and ll are considered. Results for

Orebody No. 5 reflect the anticipated mining behaviour within Orebody
No. 7. Similar geological conditions are indicated in both areas.
The objective of this chapter is to briefly introduce each of the

three approaches used in the analysis and to summarize the main
results. Where applicable, implications of the work are discussed.
116.
7.1 Rock Mass Classifications
7.1.1 Classifications Considered
Four main classifications are reported in technical literature for

assessing the stability of unsupported stope openings:
(1) RQD-Index: Deere et al. (1966)
(2) RMR-Index: Bieniawski (1973, 1976)
(3) Modified RMR-Index: Laubscher and Taylor (1976)
(4) Q-Index: Barton et al. (1974)
The RQD-index is the simplest parameter to derive of the four listed

above. However, a literature review indicates considerable differences
of opinion as to estimates of unsupported spans. The spans also appear
to be extremely conservative in comparison to past experience at the
mine. Therefore, the present analysis has been restricted to the

latter three classifications.
Details of the three classifications and associated input data require-

ments are presented in Baczynski (1979)
1.
e
117a
7 .1.2 Estimates of Stable Hanging-wall Spans
The range of stable hanging-wall spans is based on statistical input

for geological structure.
In the present assessment, the hanging-wall is considered to be of

unit thickness in the third dimension. This approach provides a basis
for comparison between the results of rock mass classification estimates
and those of 2-D finite element and 2-D probability of failure analyses.
This assumption is necessary because structural variability between
rock mass blocks is a direct function of block volume (see Figure 27
on pages 83-87). The greater the block volume, the smaller the variability
between the mean structural conditions in the blocks. The various
approaches employed in-the analysis of hanging-wall stability assume a
unit thickness in the third dimension.
The first two metres into the hanging-wall are considered to be of the
greatest importance from the stability point of view. This arises
because hanging-wall dilution in the silver-lead-zinc stopes at the
mine is commonly associated with slabbing from this region of the wall.
Depending on the thickness of the ore zone, even 2.0 metres of
collapse may constitute significant dilution. In the narrow orebodies,
tipified by the model adopted for Orebody No.ll, such collapse may
represent a dilution of 50-70 per cent.
In terms of the assessment undertaken by Bieniawski's RMR geomechanics

classification, structural variability between stope hanging-walls is
based on the statistical model for a rock mass block with a volume of
25.0 m3 A 50.0 m3 volume is the basis for analyses employing
Laubscher and Taylor's (1976) and Barton et al.'s (1974) classifications.
Structural conditions in each model are generated by means of "Monte

Carlo" sampling of the probability density functions for the intensity
of each fracture set. Model for variability between the hanging-walls
is in accordance with Figure 27 of this thesis. Pre-mining stresses,
unconfined compressive and tensile strengths of intact rock cores, and
I
ll7b.
other input parameters are discussed in the body of the thesis and in
Appendix I. The results of the analysis are summarized in Table 6 and
in Figures 34-39. Each of the cumulative frequency distributions pre-
sented in the diagrams is based on a sample of 1000 analyses of
structurally feasible hanging-walls. The results in Table 6 are presented
in terms of the mean value and a range within two(2) standard deviations.
The cumulative frequency distributions are plotted on probability paper.
Bieniawski (1973, 1976) designated upper and lower limits of application

for span estimates. These were based on case histories. Spans smaller than
the specified lower limit were stable irrespective of ground conditions.
Those within the lower-to-upper limit range also represented stable spans.
However, depending on the classification rating, they required some ground
support. Spans outside the upper limit were considered unstable without
ground support. Results in Figures 34-35 are presented in terms of the two
defined limits, as well as an average trend. Estimates in Table 6 are based
on the frequency distribution for the upper limit of application.
A relationship between cavability and the extent of undercut area (expressed

in terms of "hydraulic radius") is presented in Table VI of Laubscher and
Taylor (1976). The hydraulic radius is defined as the area of the undercut
surface divided by the length of the perimeter around the undercut area.
Thus, the hydraulic radius is not only a function of the extent of the
undercut area, but also of its shape (geometry). It is further stated by
the two authors on page 125 of their paper that " large open stapes can
only be mined in competent ground and the stope should have hydraulic radius
20% less than required for caving a rock mass . ". In terms of the hanging-
wall geometry in open stapes at the mine, the hydraulic radius ranges from
about 0.5 times the span in narrow stapes to about 0.25 times the span for
stapes where the hanging-wall is roughly equidimensional in its strike and
down-dip extent. Conservatively, the hanging-wall span in the present analysis
has been assumed as twice the hydraulic radius, with a resulting maximum
permissible span of 60 metres. The results of the assessment on the basis
of Laubscher and Taylor's classification are summarized in Figures 36-37.
One of the key parameters in the Q-index rating developed by Barton et al.
(1974) is the "stress reduction factor" or SRF. Its value is largely
117c
dependent on the unconfined compressive and tensile strengths of

intact rock cores, in-situ stresses and mean rock mass blockiness.
A number of frequency distributions for various values of SRF are
presented in Figures 38 and 39. Estimates in Table 6 are based on
SRF values of 0.75 and 2.0 for Orebody Nos.5 and 11, respectively.
MARKEDLY DIFFERENT DIMENSIONS ARE SUGGESTED

FOR SPANS OF UNSUPPORTED HANGING-WALLS BY
EACH OF THE CLASSIFICATIONS ,
These differences highlight the limitations that are inherent in the

rock mass classification approach to stability analysis.
The mean spans determined for Orebody Nos. 5 and 11 range from
11 to 28 metres and 6 to 15 metres, respectively.
118.
TABLE 6.
Hanging-Wall Spans (in Metres) on Basis of Rock Mass Classifications
Rock Mass Classification System

Orebody Bieniawski Laubscher & Taylor Barton et al.
No. Mean I Range
I
95% Mean
:I Range
95% Mean I
I 95%
: Range
I
I I I
I I I
5 11 I
7-16 25 I 17-33 28 I 16-48
I I
I
I
! L
I
I
I
I I
I
11 6 I 3-15 15 I 7-32 11 I 4-26
I I I
I I
'
119.
(Span)~ in Metres
99.9 +3
99.5
99.0
98.0
95.0
===l-f-~--'4-~-l-t
90.0
80.0 --f-LL!- - - - - l - +1
G 7o.o +----+----+~---1+----+t,---- ... --~ [

c
~ 60.0
i ~ ~
0'
~"' 0 '
------l---...-..,f-l__+-1----+-=i:_________j i
50.0
0
"'
.~
..., 40.0
...."' 7 i ~ ' ~
t
:;l
30.0
9
u
n j ] l g
20.0
+---T-4.--1--/~1---+-----
-1
10.0
05.0
I T
02.0 r 1 -2
01.0 1 I
00.5 I L
00.1
.. ~ I r I I
-3
0 1 5 10 15 20
Span in Metres
Orebody No.5: Cumulative Frequency Distributions for Stable

Hanging-wall Spans plotted on Probability Paper: Estimates
Based on Bieniawski (1973).
FIGURE 34.
120.
121.
99.9 --
+3
99.5
99.0
98.0
+-- - ----
--7r .- ---------------------
I
+2
95.0
90.0 I
80.0
/. +1
""
), 70.0 I
u
c
Q)
"0'
60.0 I I
...
QJ
ILo 50.0 i
QJ
:>
........ 40.0 I
"'
.-<
"!3 30.0 I
u
20.0 I
10.0 I -1
05.0 I --
02.0 I r- -2
01.0 I
00.5
j
00.1
0
I
10
I I
20
I
30
I
40
I
50
I
60 70
-3
Span in Metres
Orebody No.5: Cumulative Frequency Distribution for Stable

Based on Laubscher and Taylor (1976)
FIGURE 36.
122.
Logarithm10 (Span} in Metres
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

I I
99.9
I I I
' -f- +3
99.5
99.0
98.0 -
-- ---~- -- --
/
1-
I
+2
95.0
90.0
;

80.0 I +1
;;-
~
70.0 I ,.
Ill
""...
0'
60.0 I I i
~
..
c.
r.." 50.0 .I 0
"c.
c.
">
.-<
<>
40.0 / "'<....
........,.
...."'
"9 30.0 I ".
0
u
20.0 1 L -1
10.0 I
05.0 T
02.0 I -2
01.0 I
00.5 7
00.1
I
5
I I
10
I
20
I I I
30
I
40
I
50 100
-3
Spa!' in Metres
Orebody No.ll: Cumulative Frequency Distribution for Stable

Based on Laubscher and Taylor (1976}
FIGURE 37.
123.
Logarithm 10 (Span) in Metres
Arnt-
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
1------___l__ I I I
99.9
+3
I I I I
I I I
I I I
99.5 -~-~ ---- ~---~----------
{--/--./.___ .
I /
I
I
99.0 -
I I I I
98.0 f----
Ll/ I
I I
'
I
,' I
I1 I I
I I
I
1
I
r-- +2
95.0
II' ,' I II I '
l/
I
I I I
90.0
I I1 I I I I
I I I
I 1 1
+1
;-
80.0
),1 I
I
1
1I
I
II
I
I
c 70.0 ----------------- ' (/)

....
!I II II I g
"
Ill I I I
"
0'
Ill
60.0
50.0
I I
I II I I
1 / I
.......
0.
0.
"':>'"' I--
I;,' , ,' 0 0.
Ill
....<"'
.,.;
.j..l
"'
.-<
40.0
II
I I
I
1
1
I
1 I
I I ..."....
0
u
"9 30.0
0
'-'

/ I I I
'
I
"
II>
20.0
10.0
Z
tfJ
/i
II / /
I
/ ,'
I
I
I
I
I
1
t
I
I
/
I
-1
~ Ov, v, 1 I
rv~ ...?C: ~
05.0
Cl]
}J7
"'"'
I
I /
I 1,
I
'-'
/
--
I I I -2
I 1 I
02.0 -
,,,' I I ,'
I I I
I
I
01.0
z/'
II II I I
00.5 I I I I
I I I I
I I I' ' -3
00.1
1/// T
I I I I I
10 20 30 40 50 60 100
Span in Metres
Orebody No.5: Cumulative Frequency Distributions at Various

"Stress Reduction Factor" Levels for Stable Hanging-wall
Spans plotted on Probability Paper: Estimates Based on
Barton et al. (1974).
FIGURE 38.
124.
Logarithm 10 (span) in Metres
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

I --
99.9 +3
99.5
99.0
98.0
+2
95.0
90_0
+1
80.0
;;-
iJ'
70.0 ,.tn
r::
Ql ' ~
60_0
::>
0' "'"'
G!
..." 50.0 "' ""'
0 0.
G!
.......,
> 40_0 .....:"'
....Ill::> "',.....
0
30.0
u
~ ""'
20.0
-1
10.0
05.0
-2
02.0
01.0
00.5
-3
00.1
5 10 20 30 40 so 60 100
Span in Metres
Orebody No.ll: Cumulative Frequency Distributions at Various

"Stress Reduction Factor" Levels for Stable Hanging-wall
Spans plotted on Probability Paper: Estimates Based on
Barton et al. (1974).
FIGURE 39.
125.
7.2 Finite Element Modelling
7.2.1 Choice of Computer Program
A finite element program suitable for the envisaged two-dimensional

analysis should exhibit the capability to:
(1) model elasto-plastic behaviour,
(2) allow several sets of material properties in each model,
(3) model sequential excavation of the opening.
Several generally available programs were considered to be potentially

suitable for the analysis. A brief listing of the programs, their
capabilities, and a discussion of program testing and difficulties
encountered in the study are presented in Baczynski (1980b)
The program selected for the analysis is VISPLAS. This elasto-visco-

plastic program was developed by Dr. J. Meek at the University of
Queensland. Further details of the program are also presented in
the above cited reference. In terms of the present study, it is
sufficient to note that the continuum was discretized into eight(8)-
node, isoparametric, quadrilateral elements. The advantages of
employing isoparametric elements are discussed in Desai and Abel
(1972) and in Desai and Christian (1977). Plane-strain conditions,
the generalization of the Mohr-Coulomb yield criterion proposed by
Drucker and Prager (1952), and an 'associated' plastic strain
function are also assumed in the analysis. The viscous component is
not involved.
126.
Further details of the finite element method are available in a

number of standard text-books, e.g., Zienkiewicz and Cheung (1968),
Desai and Abel(l972), Gudehus (1977) and Desai and Christian (1977).
Capabilities of VISPLAS were extended by the writer to permit a

greater number of material types to be used, to reduce output to
selected elements and nodes and to present it in a more useful
format (i.e., in terms of principal stress and displacement vectors)
and to create files for subsequent computer plotting of the vectors.
Source listing of modifications included in VISPLASNRB-version of
the program, as well as input data requirements are presented in
Baczynski (l980a) .
All numerical analyses were undertaken on the UNIVAC 1100 computer

system located at Mount Isa Mine.
127.
7.2.2 Corrplementary Programs
Several minor complementary computer programs were written to facilitate

data-file preparation, mesh modifications and mesh-plotting, as well as
for plotting of stress and displacement vectors. To some extent, these
programs are structured in a format that is only suitable for the present
analysis.
A source listing of the programs and details of input data requirements

are included in Baczynski (l980a).
7.2.3 Geometry of Model
The stability of sub~level open stope hanging-walls and abutments appears

to be largely a function of the strike or horizontal span (length) of the
hanging-walls. This arises because the dimensions of these stapes in the
plane of bedding are considerably greater down-dip than in the horizontal
direction. Arching occurs across the shorter distance and therefore the
response of the wall is controlled by its horizontal dimension.
To study the effect of progressively extracting a stope and thereby

increasing the hanging-wall span in the horizontal direction, the finite
element mesh used in the analysis represents a sub-horizontal section
through the stope. The section is a plane which strikes in a north-south
0
direction and has a 25 easterly declination from the horizontal. This
coincides with a plane normal to the sedimentary layering within the
dolomitic shales. The modelled section is illustrated in Figure 40.
~N
... .. .
~ ..
~N
~
..c.
0
l'luJte !le"'at M:,.lu liaagt"a-WaJJ Do,_l.,
o..tu..e ot !!Zid.,.8J'ou,a Ope., Stope
C.ross-secti.on Of Un"'e
SUb-level Open Ston ~..~ n.rg.roUna
ea by r
~n~te
...-e -nep.resent-
Element Moael. J...,.
&
129.
7.2.4 Mesh Formulation
There are several basic requirements which need to be examined and

satisfied in any finite element analysis. These include:
(1) The continuum should be adequately discretized into elements,

bearing in mind that finer subdivision is necessary in regions
where stress concentrations are expected.
(2) The mesh represents an idealization of an infinite body by a

finite model. Therefore, boundary conditions should be carefully
considered to minimize their effect on the results of the
analysis.
(3) A balance should be maintained between increased accuracy and

the computational effort that is necessary to achieve it.
A reduction in computational effort may be achieved in several
ways, including reducing the number of elements in the mesh,
minimizing bandwidth, reducing the number of iterations for
acceptable convergence of solutions and reducing the number of
excavation stages by which mining of a stope is simulated.
A discussion of the basic considerations and the overall problem is present-

ed in Desai and Abel (1972).
One further consideration is imposed by the scope of the present analysis.

To permit nodelling and assessment of the effect on stability of hetero-
geneous rock mass properties, it is necessary to formulate a mesh with a
sufficient number of elements in the vicinity of openings. Furthernore,
since the variability within the rock mass is assumed to be a function of
element size (i.e., larger elements display lower variability), the
effectiveness with which local variability can be modelled is governed by
the extent to which it is possible to discretize the hanging-wall and
abutment areas in the immediate vicinity of the stope. A larger number of
small elements permits the model to reflect in a more realistic manner the
inherent variability that exists within the mass.
130.
7.2.5 Details of Base Mesh
The base mesh and fixity of the boundary are illustrated in Figure 41.
Several aspects of the mesh need to be noted:
(1) The mesh presumes a north-south axis of symmetry for the stope.
This has the immediate benefit of considerably reducing the
computational effort because the number of elements is halved.
The use of this approach is justified on the basis that the
prime objective of the analysis is to evaluate hanging-wall
and abutment stability. Stability of these areas is important
because failed ground is likely to be dislodged under the
action of gravity, thus contributing to dilution of ore within
the stapes. The scope of the study does not extend to footwalls.
However, it should be appreciated that the assumed symmetry
automatically implies identical or 'mirror image' responses
in the footwall of the stope.
(2) There are two principal geological domains defined in the mesh,
namely, the hanging-wall and orezone. Different statistical
models for local variability in material properties are associat-
ed with each domain.
(3) The continuum in the vicinity of the opening has been discretized
into a regular array of small, square elements. The detail present
in this zone is justified in view of the model that is assumed
for statistical variability in rock mass properties. The depth
of this regular array into the hanging-wall and abutments is
considered to be adequate for realistic modelling of potential
failure zones around the stope. Since failed ground may be
anticipated to dilate considerably (50 per cent is not an
unrealistic estimate) after being dislodged, the maximum depth
of potential collapse from a hanging-wall will be limited by the
width of the stope. Accordingly, the depth of the regular array
has been extended to roughly twice the width of the stope.
~~'11-~"!,,y...lii~~,~~~ ~-,..,.."""'~~ ~-~---~~---~~--~-~w~~ -~~- -------
~---~----~------
OUTER MESH West
Ul z
g 0
li
s: rt
::r
East
Node

~ Node with total fixity in x- and y-direction
I@ Node with fixity in x-direction

......
w
.
......
FIGURE 41 Base Mesh.
Node with fixity in y-direction
INNER MESH
~65/ ~ ,....__....
t
-43
I~ -/
-49
r-- l~~/ -~ 97
/ ~ 113/
....
127./
1-41
:X:
~
G)
so 51 52 66 67 68 82 83 M 98 99 100 114 115 116 lZB 12J 130 H
z
G)
~44
I
~
53 S-4 55 69 70 71 85 86 87 101 102 103 117 118 119 131 132 133 1-42/'
-45 -46 56 57 58 n 73 74 88 89 90 104 105 106 120 121 122 134 135 136 1-43 144
I
-48 59 60 61 75 76 77 91 92 93 107 108 109 123 12-4 137 138 145 0
47 1-46 GJ
V~62 140~~ 1:\r.:\.
tJj
63 64 78 79 eo 94 95 96 110 111 112 125 126 139 0
1'::\ ~
,... r-.. ,.... ,.,... ,..,... ./":'\ /":'\ _/":'\ .r.!>. /.:'\ 1'::\ /::\ _/::\
~
_.::.::;...:...:: ..:::::;.. -~-~~-~-~-~-~-~-~-~-~~~-~- -~- ~ ::::.~-
~ 5 -'
.,.. 4 .,--
..!~
3 J...
'T 2 ~I~
"r 1 -to!
EXCAVATION SEQUENCE
I-'
FIGURE 41 (Continued) Base Mesh. w
1\.)
133.
(4) The north-south and east-west distances from the edge of the
opening to the boundary of the mesh are approximately 8 and
25 times the corresponding semi-diameter of the opening.
(5) Boundary conditions are illustrated in Figure 41. The four

corner nodes defining the extremities of the mesh are completely
constrained against movement (i.e., x- andy-displacements are
zero) whereas all other nodes on the boundary of the mesh,
except for those defining the excavation surface itself, are
partially fixed. A residual stress field is assumed in the
model before mining. To simulate excavation of the stope, equal
stresses but acting in the opposite direction are applied
along the surface of the excavation.
(6) The stope is excavated in five (5) sequential stages, with

mining progressing from north to south. The small number of
excavation stages chosen to simulate extraction of a stope
may be somewhat unrealistic in the case of Orebody No.5 as
it implies that the stope is mined in 15 metres deep cuts.
However, the writer considers that the five stages are
adequate for purposes of the analysis.
7.2.6 Mesh Types
Several versions of the base mesh described above were employed to model
stapes in Orebody Nos. 5 and ll. Modifications to the dimensions or the
deletion of elements from the base mesh were irnplernented by means of
cornplementary programs described in Baczynski (l980a). Each mesh type is
illustrated in Figure 42.
134.
A Hanging
Wall
B Orebody
DIMENSIONS (METRES)
MESH TYPE OREBODY No. ~ B D E
1 5 260 10 590 20 110

2 11 260 2 590 20 110
3 11 52 118 4 22
. 1"-,_:./
... . 1'-;../ . . . l"'.' r---.. /"' "-~/
. ., . 1"'.--:-/ ... ... !'-;../
... '" ...
. . " . ,. " . . . ... . ... '" ... ... "' ~
... ... "' .. /
.. .. " "' .. " .. " . ., ... ... ... "' ... ...
Hanging Wall
D
"""" ... ... .. ...

.. . " ... .. . ". .. ." .. ...... ... ... ...... ......
n
~
B
n
..."'
r-
~~
.. Orebody
V" " " "' "'
-+- '-..j.-
Dimensions of Various Mesh-Types Used in Analysis.
FIGURE 42.
135.
7. 2. 7 Accuracy of Mesh
Theoretically, the present plane-strain analysis belongs to a class of

problems that involve continua for which all the boundaries are not de-
fined. However, because an infinite body cannot be included in the
assemblage, it has to be limited or made finite. This approach is feasible
because the effect of the loading decreases with increasing distance from
the point of its application. Moreover, it is possible to determine the
significance and the extent of error which arises from the finite
boundaries.
The adequacy of the base mesh, especially the effect of the fixed finite
element boundaries on displacements in the vicinity of the opening, was
tested on the CSIRO CYBER 7600 system by means of the boundary integral
program BITEM developed by Crotty and Wardle (1978}. Infinite boundaries
are assumed in this analysis. The comparison was undertaken on the basis
of Mesh-type 1 and mean rna terial properties for the hanging-wall and
orezone domains of Orebody No.5.
The results of mid-span hanging-wall displacements for various. stages of

excavation are illustrated in Figure 43. Both sets of results are in good
accord, especially within a 40-50 metres wide zone around the opening.
This zone coincides with areas of more immediate interest in terms of the
present analysis.
136.
0
II)
.
0
..... ....,f:
:&
.......::
...
~
0
"' 1,.
.....::
.,.
~
0
Ll'l
:.:
.s
......::
..
u
...~
......
0
....
Q
0
"'
----- - - - Boundary Integral Program BITEM.
Finite Element Program VISPLAS.
Mid-span Displacement of Hanging-wall at Various Excavation

Stages: Comparison between Boundary Integral and Finite
Element Programs.
FIGURE .43.
137.
7.2.8 Stress Field Assumed in Analysis
Rock load conditions have been derived on the basis of past stress
measurements undertaken at the mine (as reported in Baczynski, 1977c) as
well as discussions with mine personnel during late 1977 (Mr. M.F. Lee,
personal communication).
The stress field assumed for the present analysis is indicated in

Table 7. It corresponds with residual rock load conditions at 13-15 levels
of the mine. The relative magnitude of these values has been subsequently
reaffirmed in Bridges (1979b).
However, to test the sensitivity of the analysis to different pre-mining

stress fields, several additional computer analyses were also undertaken.
These runs were considered desirable to assess the response (i.e.,
displacements and stresses around the stope) of the rock mass under the
range of different load conditions that are likely to occur at the mine.
Assumed pre-mining stresses are indicated in Table 8.
7.2.9 Extent of Finite Element Modelling Programme
Modelling programme was comprised of 51 finite element computer runs.

Brief summary of the work is indicated in Table 9. A comprehensive
discussion of each model, including an indication of the spatial distribu-
tion of various material types within the mesh, is presented in
Baczynski (1980b).
138.
TABLE 7.
Magnitude of Pre-Mining Stress Field Assumed in Modelling
Direction
Stress Type Magnitude (MPa) Field Model
Principal - o-1 36 090/30* (normal to cr

y
bedding)
Intermediate - o-2 26 270/60 (down-dip of <J
z
bedding)
Minor - o-3 16 ooo;oo (along strike cr
X
of bedding)
* Expressed as dip direction I dip angle from horizontal.
TABLE 8.
Sensitivity Testing: Different Initial Stresses
Load Type No. Magnitude of Stress (MPa)
(J'"y
1** 36 26 16
2 26 26 16
3 16 26 16
4 12 26 16
5 12 16 26
** Stress field assumed for mine at 13-15 levels.

lf!!~;tM
0;,o/c\;c''"'~~~,,"'""' _ _ _ _ _ _ _, . , . . . . . , . ._ _ _ . , . . . . . . . . _ _
TABLE 9.
Finite Element Modelling Programme Using VISPLAS.
Maximum Accuracy
Ore body Mesh Material No. of Analysis No. of Achieved
No. Type* Types# Runs Type Iterations (%)
--- --
5 1 Mean 5 Elastic: Sensitivity Testing 1 N/A
Mean 1 Elasto-plastic: Sequential mining 15 3-5
-1 Standard Deviation 1 "
II
"
+1 Standard Deviation 1 II II
Statistically Heterogeneous 16 II
11 2 Mean 1 Elasto-plastic: Sequential Mining 15 10-15

-1 Standard Deviation 1 II II
"
+1 Standard Deviation 1 II II
"
3 Mean 1 II II II
-1 Standard Deviation 1 II II
"
+1 Standard Deviation 1 II
" II
Statistically Heterogeneous 9 II
" II
II II
3 50 2-7
Modified 3 II
5 Elastic: Single-stage Mining 1 N/A e:
0.0
" 4 Elasto-plastic: Single-stage mining 15 2-16
* Figure 42 indicates Mesh types. # Material type parameters are defined in Tables 4 and 5.
140.
7.3 Results of Finite Element Modelling
The extent of plastic yielding, zones of tensile stress, and

principal stress and displacement vectors for the five stages of
excavation are presented in Baczynski (1980b, 1980c) for each finite
element run undertaken. The present discussion summarizes the main
trends.
To facilitate the description of the results, areas of the finite

element mesh in the immediate vicinity of the stope have been de-
signated as "hanging-wall" and "abutment" domains. The extent of
these from the surface of the opening into the rock mass is illustrated
in Figure 44.
7.3.1 Extent of Plastic Yield
Results of the analysis are summarized in Tables 10 - 13 and in

Figures 45 - 48.
Figures 45 and 46 present frequency distributions for the number of

elements yielding plastically within the hanging-wall domain of
stopes. Results are for each stage of excavation. A star-symbol is
used to designate models with homogeneous material properties. The
code "M", "+1" and "-1" denotes models with "mean", "+1 standard
deviation" and "-1 standard deviation" material properties within
hanging-wall and orezone domains. Large dots are used to indicate
results for heterogeneous models with statistically variable material
properties.
Figures 47 and 48 summarize the relative frequency with which

particular elements around the stope yield plastically. A legend
,%JlQ..-, '~ ~~,.-1. & .~~\"-"'ftllf''!Hk d . iJM &J;.t.:~l!'l.,~~)W!.t .. J. !!!it#!- -~~,.f.?li., f. W gz )hJ;;ti;,,\f.FO.kl&lW~~ A~--------
2-0 FINITE El.F.MmT MOI::E.IllNG Of SHALE HANGlNG-WAJ.LS LEAD-ZINC Oru:BOot5, MOUNT ISA MINt, QUU?lSLAND
KEY DIAGRAM INDICAnNG EU.MENTS 1HAT ARE CONSIDERED roBE LOCA'JED Wl1HIN HANCING-WI.U{x) AND ABUTMENT(@>) DOMAINS.
Euavatiooa Stage No.1 Euavatioo Stage No.4
I~\
I
I
~ \,_/
['-.
r-....,./ ['._/ !'-,./ !'...!.../ ',.!-/
X
X
X
X
X
X
I
X
X
X
X
X
X
~
1/ /
~
'
\
!"\_/I"-..!_/ ',!../ 1"\,!-./ "-.;../ "-.!.../
X
X
X
X
X
X
X
X
X
X
X
x
X
X
X
X
X
x
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X X
X
X
1/ /
-------
~"
Ill>@> (jl)@>
~ ~
v @>Ill> Ill> l/ @>@> @> @>"
Excavatiaa Stage No.2 Excovation Stage No.5
.x<:ayaticm Stage No.3
~\
r-------__
\ I',_/ 1"-.r-/ ["'-~/ 0-~/ ;"-.~/ r'....~/
X X X X X X X X X
I 17
X X X
FIGURE 44.
....
.....t>.
'
x x x /
"'
X X X X X X X X X
X X X X X X X X X X X X
1/
Ill>@>
@>@>
._@
@>
.;,
I
~---------------,.,...,_---------''"""' .,
TABLE 10 - Stability of 5 Orebody Stope with Uniform (Homogeneous) Material Properties
Material Type Hanging-wall Abutments
Mean: 40m strike span stable, but local failure is initiated Stable when hanging-wall span
at both edges of the wall when its strike span is extended is 25m, but extensive failure is
to SSm. Depth of failed ground is Sm. suggested when the span is ex-
tended to 40m.
+1 Standard 55m strike span is stable, but local failure is initiated As above.
Deviation: at one edge of the wall when its span is extended to 70m.
Depth of failed ground is Sm.
-1 Standard 25m strike span is stable, but failure is initiated at As above.

Deviation: both edges of the wall when its span is extended to 40m.
Depth of failed ground is lOrn.
1-'
~
N
@:A.,-L4..J?,_-.)L....Q)H,-Gh !t .3 a ~.t')Hl!Oj;-vJJtl$1 . " 1!1 1-!fffL"t"",;s tv !'Mii#Wi%fii':;,.M? l="'tltltll!!\.!W_!!!k }M!iif4M4iJ --~mr~"""'-- _m,M.t;!Q!Q!J., A4A$.-3 J
TABLE 11 - Stability of 5 Orebody Stope with Heterogeneous (Statistically Variable) Material Properties
Strike Span of Suggested Probabi1itl of Failure on Basis of 16 Runs

Hangins-wall (m) Hansing-wall Abutments
Local Failure Extensive Failure Max. Depth (m) Local Failure Extensive Failure Depth(m)
10 5% - 5 Stable Stable N/A

25 30% - 5 55% 5-10% 5
40 80% 5-10% 15 100% 95% 5
55 100% 40-50% 20 N/A 100% 5
70 100% 95% 20 N/A 100% 5
.....
.1:>
w
'J4_ .. $~.8(1 G.. 1. 14 . ie .~It tQ il!iiM\il$ -->HP "'""" '~
TABLE 12 - Stability of 11 Orebody Stope with Uniform (Homgeneous) Material Properties
Material Type Hanging-wall Abutments
Mean: A 2m strike span is stable, but local failure extending Stable when hanging-wall strike
lm into the wall is initiated at both edges of the wall span is 2m. Local failure occurs
when its strike span is extended to Sm. Extensive failure when the span is extended to Sm
of 2-4m into the wall is apparent when the span of the and becomes extensive when the
wall increased from 8 to 14m. span is increased to Sm.
+1 Standard A Sm strike span is stable, but local failure extending Stable when hanging-wall strike
Deviation: lm into the wall is initiated at both edges of the wall span is Sm, but extensive failure
when its span is extended to Sm. Failure appears to be is suggested when the span is
confined to the edges of the wall as its strike span is extended to Bm.
extended to 14m.
-1 Standard The minimum 2m span that was modelled is unstable. Stable when hanging-wall strike
Deviation: span is 2m, but extensive failure
is apparent at greater spans.
I-'
.e.
.e.
TABLE13 - Stability of 11 Orebody Stope with Heterogeneous (Statistically Variable) Material Properties
Strike Span of Suggested Probability of Failure on Basis of 9 Runs

Hanging-wall (m) Hanging-wall Abutments
Local Failure Extensive Failure Max.Depth(m) Local Failure Extensive Failure Depth (m)
2 70% 10% 1 75% 45% 1

5 99 70% 4 90% 90% 2
8 100% 99% >4 N/A 100% 2
11 N/A 100% >4 N/A 100% 2
14 N/A 100% >4 N/A 100% 2
f--'
"'"
U1
146.
.... I
"'
...
tt.O
1
I
STAGE 1.
1
I
ST.IIiGE 2.
I
I
I
I ..
.... '! 10,0 I
I
I
'* .... *-r

I
....
I
t; lO.D t' 'MI.O
l I
i 1
.
60.0
l ia.O
~
! 50,0 -~
I J
:
!. ! ....
I
20.0
I
10.0 * I
I
*
02.0 02.0
Ol,O
....
..
91,5
/
' STAGE 3.
n.
''"
STAGE 4.;
,
I /
/
I
' /
I I
* _, !. -
.,
t;
; ....
'MI,CI
! .... -*. I
~ o.o I
/
.,. I - ~ lO.O
'
! so.o
~
60.0
.to,o
.
..
'
*''
/
,.
L
-
l .... l "'
!' I
)0,0 *I 10.0
* i I
llJ,O
" OS.CI
01.0
I
I
L "
....
.... ..
~.s
STAGE s.
I
r * +1,M,-l = +1 Standard Deviation, Mean,
-1 Standard Deviation Uniform.
.... I e =Statistically Heterogeneous Properties .
I
;
*'' ;. Orebody No.5: Cumulative
f -~ Frequency Distributions (plotted
I
'HioO
to.O
;
f
::
i ....
M.D
. e7
-1 t
~
on Probability Paper) for Number
! ....
I
I t of Yielding Elements in Hanging-
.;10.0
10.0
* ;. wall of Stope with Uniform and
os.o I Heterogeneous Material Properties.

02.0
/
I
....
01,0
I
I FIGURE 45.
_,
....
I
147.
,-
1--- 7 STAGE 1. STAGE 2.
-----;L
--
I I
+--- I I
!-
I /
1 I
I
!*.. ,; *
~-
. . I
!
! "'
1 .o.o
!0.0
-*. /
/
~
t
I
I
~
! "')0,0
., /e t
to.o
1*1
I *
I
I I
I
I OLO '
I
I ' I
02.0
I I
Ol.C
I
.... -
,
STAGE 3. STAGE 4. ,,
/ "' ,
I I
, . o
I
I
I
L I
..
.. '
; * ''-' *'
~ 70.0 I --- t" '70.0
1'
I 1-
,.
; 40.0 j; 60.0
E so.o I !
, "' 50.0
c ~
.... I '
J
i
....
* ;. I
*''
I I
I
I
I
/
I
I ,
I I
_LI
. .... /
"
I
STAGE 5.
I
I
* +l,M,-1 +1 Standard Deviation, Mean,
-1 Standard Deviation Uniform
Material Properties.
I Statistically Heterogeneous Properties.
I
1 "'* orebody No.ll: Cumulative Frequency
t" "10.0 I
-~ --
i o.o Distributions (plotted on Probability
! so.o '!.
~ o.o
.
I Paper) for Number of Yielding Elements
I
1"' in Hanging-wall of Stope with Uniform
*'' - and Heterogeneous Material Properties.
I
/
I FIGURE 46.
.... ''
'
''
.
2-D FINIT EILMENT MOOUlNC OF SHAlL HANCINC-WALLS IF.AD-'ZINC ORE.IIOD!S, MOUNT !SA MINE, Q\J'I'DISLAND
ExeawtioD Sbage No.1 . Excawtion Stage No.4
~\
~
\ I I/
i'-_/1\-_/ 1'-,_/ ,_/ 1'.,_/1!'--.r-/
1\. /
X
v f-
f-"
Excavation Sbge No.2 Excavation Stage No.5
Excavation Stage No.~

x,x X corresponds to yielding elements in 1-24%,
25-49%, 50-74% and 75-100% of finite element
runs, respectively.
5 OREBODY: EXTENT OF PLASTIC YIELDING AROUND STOPE WITH ......

II:>
co
I
STATISTICALLY VARIABLE MATERIAL PROPERTIES.
FIGURE 4 7.
2-D ftNITE ELMl'NT MODEI.l.lNC C*" SHALE. HANCJNc-WALLS: LEAD-ZINC ORDIODIES: MOUNT !SA MINE, QUUNSLAND
Exc.av:ation Stage No.1 Excavation Stage No.4
E.xcantion Stage No.2 Excantlon Stage No.5
1~\\
r-----_ ...
I 1/
1'-r-- /I'- r--/ 1"-r--/ ""'-/ 1"'-,_ /I'-_/
.. .
X X
!"'- )(
xX
)(
X XI)C .
)(
/
~ c-:-'\..
)(
v X X
Excavaticm Stage No .. 3
x X )( corresponds to yielding elements in 1-24%,
25-49%, 50-74% and 75-100% of finite element
runs, respectively.
11 OREBODY: EXTENT OF PLASTIC YIELDING AROUND STOPE WITH ~

~
STATISTICALLY VARIABLE MATERIAL PROPERTIES. w
FIGURE 48.
150.
accompanies each diagram. The intensity of shading is proportional

to the observed yield rate.
Modelling of stopes in Orebody No. 5 suggests that:
(1) hanging-walls should remain stable for strike spans

not exceeding 25 metres, but some local failures of the
wall and abutment areas is anticipated,
(2) considerable instability of hanging-wall and abutment

areas is envisaged at strike spans of 40 metres or
greater.
The results for Orebody No. 11 suggest that:
(l) -there is a likelihood of hanging-wall collapse and

abutment failure even in stopes with a hanging-wall
strike span of 2 metres.
(2) extensive failure is likely around stopes with larger

strike spans.
It is evident from the analysis that:
MODELS WITH STATISTICALLY HETEROGENEOUS

MATERIAL PROPERTIES CONSTITUTE A MORE REALISTIC
APPROACH TO STABILITY ANALYSES.
Such models are extremely useful in demonstrating the potential

extent of ground failure around openings and the excavation stage
at which failure is initiated. A typical example of the contrast
between models with homogeneous and statistically heterogeneous
material properties is illustrated in Figures 49 and 50.
These plots illustrate the extent of plastic yielding. Various codes

are used to designate incidence of yield, failure mode (i.e.,
tensile or compressive) and zones of tensile stresses. An appropriate
2-D FlNrrE. EW.m,JT MOa:.LLD<lC CE SHAlL H.ANCINC...WA1..U1 lLAO.ZlNC ORDOIJ[[$: MOUNT ISA MIN'. Cj?\JE.D'!SLA.'lt)
1\UN No.7: 11 OIU!ODY MESH-mEl: MEAH MAnJUAL I'ROPD.TIES.

bc:a-ao. b p No.I
LEGEND
~\ I/~ 1 =
Element yielded In compression
during present &tage of excaV1ltion.
Element yielded tn tension during
1----- I' / '-r-1 i',_ I' '-,_/'-,

2 =
present stage of excavation.
Element yielded during a.n ea.rHer stage

/ o! the excavation sequence.
I
2* ~ Elem.en.t yielded during an earlier stage
~ ~ of the excavation sequence and is
currently in tension.
Element has not yielded but is
&c. nib Sb1 No.2 currently in tension.
Extent of Plastic Yield

in Model with Mean,
Uniform (Homogeneous)
Material Properties:
~v-do.St-pNo,3 U:eanti01111 Sbsa 1 .. 5; Elaml!:llt Matertal Tnea
Orebody No.ll
FIGURE 49.
.....
Ln
.....
ill _w rnmnl!!w .,
2-D FINITt El..[.)r.{Do~T MO~UIHG OF SHALE HANCINC.WAU.S1 UAO..ZlNC OR.UODIS1 MOUNT ISA MENE., QUT..D:fSL.AND
J\lJN No.lZ1 11 OR.DOOY1 MSHeTYP S1 STAnmCAL MATFJUAL PROP'tRllES,

bundola Staptl6.1
~\ I// ~
1 = Element yielded in compre:ssioo. during

present stage of excavation.
------ I" " /

/"-,...... " /' /
1* = Element yielded l.n tension during present
stage of excavation.
/
I 2 = Element yielded during an earlier st.go o!
the excavation sequence.
~
I
I I lj 2* = Element yielded during an earlier stage of
the excavation sequence and is cunently
la tension.
Dccarttim h1e:Pio.Z b.eavstiOD St:11eNo.S * = Element has a.ot yielded but is cutTently
in tension.
Typical Example of the

Extent of Plastic Yield
in Model with Statistically
~Sc:apNo.J
Heterogeneous Material
bea,..tioa St:11u 1 - Sl D.ata~t M.aWU.l Type~
Properties: Orebody No.ll
FIGURE 50.
Ill sl t 1141131 3 t t l16l14llOI141UilSI17It2IUI 7IUI111 7ltoht
tNH;Ji , I I-'
V1
N
153.
legend accompanies each diagram.
Rigorous statistical assessment of the finite element results is

limited by the small number of runs undertaken with respect to each
orebody. However, on basis of available evidence, the response of
the two model types is significantly different at the 95% level
of confidence.
MODELS WITH STATISTICALLY HETEROGENEOUS MATERIAL

PROPERTIES PREDICT:
(1) GREATER EXTENT OF PLASTIC YIELD
(2) FAILURE MAY BE INITIATED AT

AN EARLY STAGE OF EXCAVATION,
Distribution functions for the number of yielding elements in the

two model types are presented in Figures 45 and 46.
7.3.2 Displacements
The mid-span hanging-wall displacements are illustrated graphically

in Figures 51 and 52. The five stages of excavation within the two
orebodies are represented.
The curves indicate the mean displacement of the walls. They are
derived on the basis of models with statistically heterogeneous
material properties. Variability between models is expressed in
terms of the standard deviation which is stated as a percentage of
the plotted mean value. For example, a mean value of 60 mm and a
standard deviation of 4 per cent means that the range of displace-
ments observed at that location within the hanging-wall may be
5 OREBODY - Statistical Material Properties
0
t;
('!)
0'
0
~
z
.
0 ..
QJ
\.11 ..."
:;: ...
QJ
Ei
.-4
1-'
P.o ....
.-4
X
....c so :=::unr~ ;g: lHi fh~ il~ m;~ :~: ~rr ~~~ g;~ Hi:!~ :~ .: ~;:~ :~~: ~~: ~~f~ :;;; n;l ::n lW H~J jj EH m; u~ ~;: ( !~~ lf: i~ ~~~ :~~ r:;;
I
til 11H
kj
'0
PJ ... l!m m n~~ ;h[ ~!Ji HE ;H~ iF! g;~ l~:~ H:1K i~f tE~ ~~~ ~~! : ~: ;!:: :~~; ~:~~ ~~:: ~) ~~;; i~Jfth~ 1Ul ;!il .rg F~~ t~M ;;; ;;: ~~ji ::~ ;~~ ~~~~ [.~ fi~
~: :: :~;! ~~~~!~;~~iii i:!! ~~tim~ ~i~ i~~ P~~ ~i~i ~i:: ~;-~~ ~u ~~J ~!~ :~;~ :~:~ ~~~~ ~~;:
H
G)
t:l
1!" :ttf # 1:: i!~ ~!f= t} :~l~ ~~~~ ;~~~ m~ m~ r~n HfJ~ :;: r~~ ~~~
c:: ::1: QJ
4o ~{f.~ ih~ ;r!! ii~i lH; ni~~~ ~g; i!r~ ~~~~ :g! I~! 1i~~ ~l::~: ~i~: :;~: ::~ :i;~ ~~~i ~~~~ ;:~: ~!H iiil n~ Hi! ~til t~l ~~li ~ii: L~i ~~L ~:: :;!: ~~~ ~~~ ~~~! l~~ ~~;:
~ PJ ""'
..c.
-~~~0:;, l~ ~ ;~~: ~~~ ~ t~ ~;: ~l i ~~ ;~~~ :J-+ ".: ~ : ;. ::t; m: rt ~ ~1 r1; f}f : ~ ~.; il:: :l :_, ~ - -~:~;~ ~ ~
t:l .-4
\.11 ....
I.Q
....
1-' t:l 0
I.Q
~
I
.
.-4
.... 30 i~!~ iU; :~ ::~: ~n~ ~!i~ ~:if~~~~ Et; !~! ~lE~ ::~: f~:: ::~. !:< :::i ~:~! :::: :~: :R ::~ ~n i!!; lU~ ;~;~ g;~ I~:~:;~~:::~~~ !~~; :~~: :.:= ~:~i ~;;~ ;~
~m ~;;; r:~: ,r;:~ ~ ~:;: :::; :m ~:: ;i; j;: :;:; !ffi~ti~ : : : : .::.: m: ; :~ ,;;; :~~ ~:, ,::: :::: ,: ; :::: ~= .::; ;;~ ,:: ~::
3:
PJ I
1-'
1-' ....""'
'"' "" :: t'; IT~:~~~;:~: :l ii: lli; .::: :; ; :;.~~ ::;, :: : !i!! illi lUi H !::; :;; ~~ ,.:: :~ :::: :;;
""'..
!lt;,~m.";W ; ;: ;;:~:!l~: :;'(~ :::.> . :~:/Ir:0::;: 1 ;:;:~~:! ;:::0: :':. ~:~;:~
t:l
1-' :>:
til
'0 ~
20 (,%,
1-'
PJ .
c. rl't ;~;~S: :~ :~: ~~;; !J!-; 'L' !~;~ : ; ::: ;:Jc.;: 5~ . :: :. . ::: !;ii ::: !:~: :;;~ !i!! !;i]t;.;.;.. ~ i'~ !iii~;; ii~ :r:
ill_tl!! i1H ~;ii ~ g !~ii ~t: ;;!~ ~f:_ ~ ~ ~;~: ~!~~ ~~~ /i :;:: ::: ::.~b :-:~ ~d! ;;:~ !i!i !i!i ;;;i fj ~~~: i~~~ ii=! ~~~~ ::: :~: ~::
0 I
..."'X :t:: ;;;; :::: :;:: (4")
:. :. : '~~ : :~:!: :~~~ : : : :: ~. .: :: ": : .: j; :J : ~ ~ :::

('!)
~~[;JB lj;, :!m~ .: : B~~

~
!:j 10
rt"
til
1-'
0 ilii !~lD!illJF.i m; r:n ;;:, ::'' .,,, ,., .. ,. .,, ,, ' '"' "" !' :;:. :. ,, :::: I, :p:: ,. ,, . ::: :::: "'' :: .. ,,. ::,, <7") U'l
10 20 30 40 50 60 70 8 90 100 110
"'"
Distance into Hanging-Wall in Metres
l'f}ijl;d{,__JMJiJLf~iJ~~,$l_M,AJ(S~JiJlilt4ZfL"!C'6J-i r: --.-;g~~.l.Il, r ;z- AW.mL&~~-M\$$1~#1 '" \. ., ___;_._" ,. _ ";_$$U)ik~~-W!tst"-"'~!!"&- a, ~~-n, ,-~~-,3 ;,-;r;t~XWili41Mil&f;;;;::w;;_, _ _-:J t wn"m~=~:r <;~
11 OREBODY - Statistical Material Properties

Legend: . Srn =Mean displacement curve for hanging-wall span of 5 metres;
0 (7\) = Standard deviation of normal distribution expressed as a percentage of Mean Value.
li
(1)
tr
0
%
z
.0 .,.,
1-'
..
1-' ...k
....~
.....
~- .........
0.. :
I
rn .....:
t"lj
'tl
OJ ...c:
H ::l
Gl
c:: ::r:
.,0~
~ OJ
::s "'0.
.....
....c.,
"1~::~. ::~~m.~~ ~~~

<0
1; : : .r,: .~ :;' " , :~;: !:: :. :: ~ :~ :: :~: j\ :i~ ~ ~: ~;

U1 1-'
N ::s
<0 .....
.....
I
:e:
OJ
"'
)
I
: . ..:. :: .. '::n::
1-' "'
.: FIE 11~~ iiHUih~:N :1:! ~!:; i!~: :Iii U1~ ~H: ~i::h :~:: ;:g ~~~: :::.: ;::: ~) ;; i~i ~;~: ;~~: 1;~; :!P ~~1 i:!! ii ;~!~ i:~; 1~;; ~~:~ ~~~~ :~~~ :~~f i~~
I-'
~~,~ :~~ ;;;~0~: ;~;~ ::~ G: :~:r ~:! ;~;.~ :::~2: ,: : : :~ m: m; :i:: m~ ;;: :;:: :: m:::; i :~:; ~:: :;t :t: ~~; :m
.-<
tl'
0 ~ 10 8m
1-' :c
rn
'tl !ac. !,n~::~ :!!: ~;:.:::::;iii i;H n!t EU ~~ :::~ ~~~: ~i:~ ~: :) ~~:: !~::~ :::! ;;;~ l~~i ~ii; ~~ ~~; :li ~:: ;i~: ~;~~ ;1;~ :i~i i~~ ~~~~ ~::;
I-'
OJ . I
ol , , , , ( , 1 ~ :; 1' :0 I I o- I o 1 .. , . ' I' ,.,.
2 ."'..,
:
----.
8::I
I
(7")
I
rt -00
(Jl .
(~)
.
li!ii(!i~8Tl 112!11 I-'
0 f;, [? I;;"FI 'i""i"'r:: J::i:J:;:~I':' I --tts"' VI
U1
0 2 4 E a 10 12 14 1! 'a 21 22
Distance into Hanging-Wall in Metres
156.
MID-ABUTMENT DISPLJ\CEMENTS AT VARIOUS DEPTHS (0.0, 2.5, 5.0, 10.0, 1~.0,

25.0, 35.0, 80.0, 130.0 METRES) INTO THE ABUTMENT OF 1\ 20.0 METRES WIDE
S'roPE: S OREBODY -MEAN MATERIAL PROPERTIES.
(ON BASIS OF "BI TEM" ProGRAM)
-6
0 10 20 30 40 50 60 70
Hanging-Wall Span in Metres
FIGURE 53.
157.
represented by a normal probability density function with a mean

of 60 mm and standard deviation of 2.4 mm.
Estimates of hanging-wall displacements for spans between those

illustrated may be derived by interpolation between the curves.
The mid-abutment displacements for the five stages of excavation

of a stope within Orebody No. 5 are presented graphically in
Figure 53. Mean material properties have been assumed for the
hanging-wall and orebody domains of the mesh. The curves represent
the displacements at various distances into the abutments. The
distance indicated adjacent to each curve corresponds to the dis-
tance (in metres) from the free face of the opening to the selected
depth in the rock mass.
Examples of typical displacements at other locations around the

model stope are summarized for each of the orebodies in terms of
vector plots in Figures 54 and 55. The length and direction of
vector corresponds to the magnitude and direction of displacement.
7.3.3 Stresses
TyPical maximum and minimum principal stress vectors for each stage
of excavation in models with homogeneous material properties are
illustrated in Figures 54 and 55. Tensile stress vectors are repre-
sented by double lines.
Stresses range from about -10 MPa to 110 MPa. Magnitude depends upon
location, stage of excavation and the spatial distribution of material
types around the opening.
IIIUW 0. 1 [lUVAIIOIN SlAG( NO. I i=I:UN HC I [IOVA'ION 5HC( NO I
D15.PLAC(P\("'1 W(ttORS ~[CIDR UNCIH 0'" t.O Cn [0UUS 5C Ml'l OtSf'lAC(I'I(Nf I'IU ANO I'IIN P~lNClf'lL STR(SS{S HCTO'I ltiCH! Of I.!) Cf'1 0\JALS SO!"'" A.
t t t t t t + t
t t
t ',, t
f t
t t
t
t
t
t
t
t
t
+trttt
t tti++tt t
rn't t
t 1
'I:UN NO 1 [J:CA.YAfiCN SIAC( NO. 2 1/l:UN NO. I [XCUAliON SIAC( NQ. 2

DISPl.I.CfH[NT Y[CIOJ!S 'ECTOR UNCTH or 1.0 CPI DUALS 50 ..., OISPU.C(I'I(Nf ,AX. UIO PHil. J"~INCIP''l
~TA(SSf~ II[C"IOA UNCht Cf:" 1.0 C1'4, [0\JALS 50,.,.,,
t t t + + + t t
t +
t t
1 r ' ' t t
I . . . :'rlr!lrr'r'( I t ++ t
ClUN NO. I [XCUAl ION SUC( NO. 3 . A.UN NQ. I [XCAVAl!ON SHC( NO. 3
O!SPLAC(H[NT Y(CTORS Y(CIOR UNCIH OF" 1.0 CH [DUAlS SO H'1 Ol!iPLAC(HfNf HA"IC. AND 1'4,l1rL PRINCIPAL SIA.tSS5 ~[C10R LNCIH or 1.0 Ctt 0UALS ~0 f'IC'l
t t f f i \ \ t
t +
t t f d;/-->,\H 1
t t + //~-- ''-\
;-"' t +
t t t t
FIGURE 54 Orebody No.5: Model with Uniform Material Properties: Principal Stress and 1-'
VI
())
Displacement Vectors.
~UM o. I (I'CUHION Sfa.C( NO. 4 'tUN NO. I [lCCUPION SlAG( NO
OISPUtH(IrU 'I[CIORS ~(CfOR HNCIW Of 1.0 CH (DUllS SC 1'1'1 DISrUC[t'I[NI I'U.I. UO MIN P!UNC!Pll SlRSSS Y[CTOf\ trNCTH or 1.0 C"' (OU .. lS SO f1Pi.
/ I t f f I + \ \ t
t +
t t t~Jj~/-=-'~\\~~ t
\
, 1r'rl:l.'ldflrl\lrlrlr'r
+ t ~'/ / ~ : ; ; ; : ~'-<it +
+ t + II I + t
'tUN NO. I [l(CAVATION SUC( NC. 5 RUN NO. 1 XCAW.I.TION SYAC( NO. 5
DISPL~C[M[NI HCTO'lS i(CIOR LfMGTH ::)r I 0 C~ 0U .. \.S 50 H., DISPLAC(H(Nf f'IU. ~'10 Mllol P 1:UNC1ru StRESSES tECTOIIt. lfNG1H !)f I 0 '" E'CI.ULS ~0 Jltfl' ...
\ I I I f I I X "'- \ \ t
I
t t
t +
+ +
Oli1Jb:!1 Seal. -- I c~
FIGURE 54 (Continued) Orebody No.5: Model with Uniform Material Properties: Plot of
Principal Stress and Displacement Vectors.
......
c..n
1.0
1l!it.i1!!!1l@J!!l!!'ji''iiW !il!ll_~tlf'JM'T!itlf -. iff ff7 -~ 1 w'!Jijllf'Slla'll '~W !!U_ Sf lM lf t U ;;; - t'. J!lll!l,,!$1&- tn''lfiFf]Pifi-tlOO .-IW-Iif~~~~~MU- Ai$$i8-fiL1_ e-lf!'1f!MrnTi\l!n-sr:w.w,'ffl'-.,.i!o.U p;:! i oii;;#-...._=o=~
II:.UN NO. 1 0:CAU110rt Sfii.C{ NO I ~:.JN 1110. -, [X(U)1J("I~ $1AG( NO. I

Dl!tf'LA([f1fNl \f(CTOq,$ II'(( 1M Lrtrtf.ll1 Or 1 .D Ct1 [OUJ.LS '0 f1"1 O!SPUC[11[tilt "'ll 1,~[1 1'11N PC:INC\rll SIA::fSU'S. ([(lOIIi. LfNC.Ht ot 1 0 C" [CUI.l!. ~0 f'tl'~
+ + t + + + t t
t t
t t t + t +t +t t +t
+
+tt~P+
r, 1
t t t t I
t + + + + tn'+ +
'tUN NO. 1 [X(UA.ftC-. SIAC[ NO. 2 ~UN 1110- :r [XCI.VAT!ON SfJ.C[ triO. 2
01Sr-LA.C(I1[NT 'IECJOitS t[CIM LfNCTH :lf 1.0 Cl't [OUUS ~0 ..,., DIS~U.C1'\[IrU !'I.I.X- AND 11[lrl P~INCIP~l Stft[SS[S t[C10"- l[NPH Of I .0 Ct1 (DUALS ':10 ~A
t t t + + + + t
t +
1
t
t
rrt r,
: frrrrI:
I
[
t
t
t t +
+ +
t + + -
+
/;--...\
+t
W'Ht'\W t
t
~ON Ill). 7 [XCAVllUlN SUC[ NO. ] Ki;Utl tiD. 1 (XCAVHION SHC[ NO. l
OISPLI.C(I1[NT Yf:CIORS- t'[CTOR LENGTH OF" 1.0 Cl'l 0UHS :!0 11~ 01$PLAC(I1(NT MX. AND MIN ,.RI14CIPI,L SfR.CS$(S lf[Cf01t LfNGll-1 0" 1.0 C'l'l COU-'LS 50 f1'1'~,
t t f f -;. \ \ +
r
t t
t t t I w~;~::;~u ~ t
1
1
' ' r1r r1r r' r r''
: 1r r r 1 r r r l : : l t
t t
+u_____
'/ ~ . . . '" t
~ I I + t
FIGURE 55 Orebody No.ll: Model with Uniform Material Properties: Plot of Principal .....
0'1
Stress and Displacement Vectors. 0
I;:UN 1110
1 O:CUHION StAG[ NO. 4 IWH NO 7 {X(HHlO" SIA~( NC
'1((10~ L[N(.1H :;~r 1 C CH (DUALS 20 H"' OISr'LAC[H[tH ~!riD or
D!SPUC[I1[Nf y[(TOR'io
,-
HU: l"'.lt.l PC:!NC!r'Al STR(SS[S I'[( IQI<. trlriC.Tiol 1.0 CM [O'JllS SO f1ll.l.
t f I I ... \ \ t
!
t +
1
t I t
I '
t t
RUN NC. 7 D'CAYATIO"' STU;( NO- ~ '\UN NO. "l [H1'1'PI0"4 5HG[ WC ~
C!Srl.I.C(t'l(tH Y((TOI\S I'[ClO!\ LfNGTH ~r- 1-0 C"' (ClJA.LS ;!:l 11"1 OISPUC[I"',(NT MU . .1,"10 MIN rctii~Ciql STI\ESSn "ECTOR lfN(lH :Jf 1-0 Cl"l [DUALS '50 f'!f'A
! + I I X
" \. \
I + +
, t I r
t t
t t
Ori~l Sc:.a\e. - - 1 =-
FIGURE 55 (Continued) Orebody No.ll: Model with Uniform Material Properties: Plot
of Principal Stress and Displacement Vectors.
1-'
C5'
1-'
162.
MAGNITUDE AND ORIENTATION OF TENSILE STRESSES

WITHIN THE HANGING-WALL DOMAIN ARE MARKEDLY
DIFFERENT IN MODELS WITH HOMOGENEOUS AND
STATISTICALLY HETEROGENEOUS MATERIAL PROPERTIES.
Under the pre-mining stress field assumed in the analysis, tensile

stresses act largely in a direction parallel to the strike of the
hanging-wall (i.e., in the x-axis direction) for models with
homogeneous material properties. In statistically heterogeneous
models, a significant proportion of elements in the mesh are
associated with relatively high tensile stresses acting near-normal
to the strike of this wall (i.e., in y-axis direction). Examples of
the observed variation are indicated in Figure 56.
In the case of the dolomitic shales at the mine, these stresses

would contrihute to separation of bedding planes within the walls
of the stope. This result suggests a potential failure mechanism
for hanging-walls. Unstable blocks of rock mass may be isolated
from these walls by the combined process of failed abutments and
separated bedding planes.
MODEL RESPONSE IS EXTREMELY SE~SITIVE TO THE

MAGNITUDE OF THE PRE-MINING STRESS FIELD,
Tensile stresses were not apparent within the hanging-wall of models

tested under the stress conditions assumed in a sensitivity analysis:
as described in Table 8. Hanging-wall and abutment displacements were
also markedly different. The results demonstrate the need for the
stress parameter to be defined as accurately as possible. Otherwise,
it is difficult to validate the response of numerical models with
field observations.
~Uif MO. 7 [XCU.I.1\0N S.f.I.C[ HO 5 RUN NO. t't [XCAYATION SlAG( IWO. '5
hU. AND 111N. I"~IHCU'Al SfR[S'SofS tHIOR l(HClH Of 1 ,Q Ct1 0UUS 50 H"A tUX. i.NO MIN f'F:INCI'"Al S1R[!.'HS W(CTO~ l.h1C111 or I .Q (', (OUAlS ~0 ttrJ.,
f I I X ..,._ \
\ t + I / J( ..,..
' \ +
t t t t
t + f~lj~~~:~~~~~~~~ t + fth;/~----: .. :/f,~,\~~ t
t fl- '/" . . . . . . . . . '"' + + t + /,~~::~~:~::~\ ~ +
t . I . Mean Properties+ + + Statistical Properties t-
RU~ NO. 12 [XCAVATION SUC( NO. ~ f::UH NO. J.J f)':CA.VJ..l!ON STA.e( ~0 5
~AX. loNO MIH P"RINC'IPAL STRSS[S, W{'CIOR lfNCHt or 1.0 Cl'l fOUALS 50 I'I"R. tuX. 4NO KIN PI\INCIPA.l STR(SS(S V((TOf:. lfNCTM Of i .0 CN [OU.I.LS 50,., ....
f I I X '1.. \ \ + f I I X ..... \ \ +
t + t +
I I~~/;:< --:- , :-:"\ ~ 1
+ r// t
I + L~.::.;,:., .:o,o.:,:.
t / ~ : : : : : ~ ~ '"'\~\ + +
t I 1 I 1 /II Statistical Properties t
II:UN NO. 15 fXCAUliON $UC[ NO. '5
MU. UO MIN. f'IUNrJf'lol SlJI:(SSfS: t'(CYOR LrNClli or 1.0 CN (0VAI..S SO I'I"A. Compressive Stress
f I I X " \. \ t Tensile Stress
t . t
/ - - '~~
Original Scale -.___..1 em.
t1
I / /..,..."' _.. , . ....._ ' ' -
+ +1~'"''"''"' '"""'"'''"
t f./~::.::~:~~~:: +
+ +
......
0'1
w
FIGURE 56 Example of Typical Principal Stress~s Observed in Finite Element Models
with Uniform and Statistically Heterogeneous Material Properties.
164.
7.4 Probability of Failure Analysis
7.4.1 Description of Technique
The rock mass is viewed as a heterogeneous continuum with statistical-

ly variable material properties. The technique assesses the probability
of local ground collapse around openings. Basis of the analysis con-
sists of:
(1) _ a stress field,
(2) a statistical model for local variability in the strength

of the mass.
The following parameters are necessary for the analysis:
(1) Principal stresses at specified locations around the

opening.
(2) Failure envelopes for each material type involved in the

analysis.
(3) Probability of occurrence of each material type. This

probability is specified for each recognised structural
domain of the model, such as the hanging-wall and orezone.
Principal stresses may be determined by means of an elastic finite

element or boundary element analysis. The latter two parameters are
a function of sample size. In terms of the present analysis, the
rock mass has been characterized by an assemblage of unit volumes.
It is a system of domains, each with a volume of l.Om 3 variability
165.
in rock mass properties and the probability of occurrence of

particular material types are dependent on domain dimensions. For
example, large domains will display lower inherent variability
between them. Therefore, a different model for probability of
occurrence of various material types will apply in such instances.
The probability of local failure is estimated by a simple graphical

procedure. The technique can be readily computerized to expedite
the analysis.
The following steps are involved for each location that is evaluated
within the continuum:
(1) Construction of a Mohr-circle for the observed principal

stresses. This circle is superimposed upon the plot of
failure envelopes for the various material types.
(2) Determination of material types that will yield under

the prevailing stress field. This simply involves the
identification of material failure envelopes that are
cut by the Mohr-circle for stress.
(3) Derivation of the cumulative probability of occurrence

for all yielding material types. Probability may be
determined either from a "Material Type" distribution
function such as illustrated in Figure 33 for the
appropriate structural domain, or from a look-up table
listing the relative frequency of occurrence for each
material type.
(4) Contouring of failure probabilities.
THE DESCRIBED METHOD PRESUMES THAT THE SPATIAL

DISTRIBUTION OF VARIOUS MATERIAL TYPES WITHIN
THE CONTINUUM IS RANDOM,
166.
The procedure involved in the analysis of a hypothetical example

is illustrated in Figure 57.
7.4.2 Assessment of Probability
A probability of failure analysis was undertaken for each excavation

stage of stapes located within Orebody Nos. 5 and 11.
BITEM, the boundary element program developed by Crotty and Wardle

(1978) was used to determine input stresses. The analysis was extend-
ed to openings with the alternate, pre-mining, stress fields
indicated in Table 8.
A simple computer program was written to expedite the analysis. It

processes the principal stress data from BITEM and outputs probabili-
ties of failure for specified locations around the model opening.
Results are summarized in Figures 58 - 63. They are represented in

terms of contours for the probability of local ground failure. The
probability of failure within unit domains of the rock mass system
is indicated.
The likelihood of more extensive failure can be readily deduced

from these plots. The method employed is illustrated schematically
in Figure 64. The boundarydefining the potentially unstable zone
is subdivided into "unit" distances and the probability of segment
failure is determined. It should be noted that segment length is
dependent upon the dimensions of domains used to characterize the
rock mass system. In the present analysis, segment length is 1.0
metre.
167.
100
dP
i X
75
:>...
u
s::Q)
::s
tf
Q) 50
j
H
I
"'Q)"' MATERIAL TYPE
:>
-rl DISTRIBUTION
-1-l 25
co FUNcriON
.-l
::s
s::s
u
0
l 2 3 4 5 6
Material Type
l
LOOK-UP TABLE
Material Probability
2
Type of Occurrence
(P. = %)
--~
l 15%
3
2 40%
3 15% 4
4 15% 5
5 10%
6
6 5%
~~ cT;:::::::::::::::::::::::::::::::::::::::::::
\ ..... a a I .. f 4 e. 0 0 .... . ......... f ,'
CTl
' . . . . . . . . . . . . . . . . . :/
";::::::::::::::::::::::::::::::::::::::::
...................
................. .
.. ,
', ............... ' .................. -" ~
'
'-~: ::: :::::::: ::: ~~ :~: :~ ;: >-
Probability of Failure 15% + 15% + 10% + 5%

45%
FIGURE 57. A Hypothetical Example of Probability of Failure

~alysis.
CONTOURS OF PROBABILITY FOR OCCURRENCE OF LOCAL .GROUND FAILURE.
5 Orebody: Stage 1 of Excavation Sequence with Hanging-wall Span of 10 metres.
Stress normal to hanging-wall = 36 MPa; Stress parallel to hanging-wall 16 MPa.

Scale: 25.7 mm = 10 metres.
....
0'1
(X)
FIGURE 58.
CONTOURS OF PROBABILITY FOR OCCURRENCE OF LOCAL GROUND FAILURE.
Stress normal to hanging-wall = 36 MPa; Stress parallel to hanging-wall = 16 MPa.

Hanging-wall
FIGURE 58 (Continued) ~
0'1
\.0

Hanging-wall
1-'
-J
FIGURE 58 (Continued) . 0

5
10
Hanging-wall
~
--.!
f--'
FIGURE 58 (Continued).
1

20 30
Hanging-wall
1-'
FIGURE 58 (Continued). -...J
N
Stress normal to hanging-wall = 36 MPa1 Stress parallel to hanging-wall = 16 MPa.

Scale: 12.9 mm = 1 metre.
FIGURE 59. ......

-...J
w

.....
~
"""
FIGURE 59 (Continued) .

Scale: 12.9 rom = 1 metre.
~
FIGURE 59 (Continued) . -..J
U1
r~~tz:z :s- -1 ;ii<@WWf'A'"fifMil~!AA'tU~r'i'~- lllf!liWbiftiS)l~Uf31Uf$t'llli'liiillii:"r--.;tf -'Jil''t:i'J:~..w
wm * L _!iiWFJM
, _ _ _ -Mmtt"" '-&! "1 et rttf!tPJidi 1'Mill1tT
tliil1iJiWil*tlit:ltiiif"l'tCzr~cgtiMll t -wwrr t'U'IT WfiStMnras!'wtttwaa'n r er trvcwT m noon J:t:t&!i ''w

:..::::::=: 60 40
70
Hanging-wall
,....
-.I
FIGURE 59 (Continued) . C1'l
11 Orebody: Stage 5 of Excavation Sequence with Hanging-wall Span of 14 metres

---- 60 ----
70
80
Hanging-wall
1-'
FIGURE 59 (Continued) .
.
-....!
-....!
5 Orebody: Stage 5 of Excavation Sequence with Hanging-wall Span of 70 Metres.

Hanging-wall
f-1
-..]
FIGURE 60. CXl
'/f@'>dt:iof*f;t}Zfff1tj(f!jfilf*1Y'~ #WT!ilr'rt'~~me '"Wt;#WW? ~-"" 0 .1l .. ?IM~ <<-7-rr -r "-'"''rAttiiif"ttH1_ ill I -t ---------

Scale: 25.7mm = 10 metres.
Hanging-wall
FIGURE 61. """'

-..1
1.0

Scale: 25.7 rnrn = 10 metres.
~ Hanging-wall
t-'
FIGURE 62. co
0

Scale: 25.7mm = 10 metres.
Hanging-wall
1-'
FIGURE 63. 00
1-'
182.
60% 50%
60%
50%
1
Unit Segment
Zone of Potentially Unstable Ground
Failure Probability for Zone
0.80*0.70*0.65*0.60*0.65
0.14 or 14 per cent.
FIGURE 64 Schematic Representation of Probability

of Failure Analysis for Extensive Zones.
183.
Since the model assumes a random spatial distribution for material

types, the probability of failure for the delineated block is
simply equal to the product of failure probabilities determined
for the segments. For example, likelihood of zone instability that
is delineated by five unit segments with associated failure proba-
bilities of p 1 , p 2 , p 3 , p 4 and p 5 may be determined by:
Probability of Failure
Where the probability of failure of each of "n" segments is the

same, the likelihood of extensive instability is given by pn.
7.4.3 Advantages and Limitations-of Method
There are several advantages in this method of approach:
(1) Technique is relatively simple.
(2) Cost of analysis is negligible in comparison to the

elasto-plastic finite element approach.
(3) Only a single, two- or three-dimensional, elastic stress

analysis is necessary for each particular test case.
(4) various alternate statistical models for local variability

in material properties can be quickly evaluated. Analysis
can be expedited by the use of simple computer programs.
(5) There is no limit to the complexity of the model for

rock mass strength (apart from properties that change
with time or are displacement dependent such as strain
184.
hardening or softening) . Directional properties may be

easily included.
(6) Significance of alternate premining stress fields can

be readily evaluated by means of a parametric study.
However, there are also three main disadvantages:
(1) Stress pattern around openings is based on an analysis

of a homogeneous model. A single set of elastic para-
meters is assumed for each structural domain (e.g.,
hanging-wall or orebody) of the model. The present study
has demonstrated that marked differences exist in stress-
es between models with homogeneous and statistically
heterogeneous elastic properties. This aspect is
completely ignored in the "probability of failure"
analysis.
(2) The method is a static analysis. It does not consider

load transfer, whereby the load previously sustained by
a yielding element is redistributed to the remaining
intact elements. It also fails to include gravity as a
component of loading. In practical situations, this
component becomes progressively more important as
yielding increases.
(3) The method presumes that the spatial distribution of

various material types is random within the continuum.
This assumption is at variance with the "zonal" model
developed for spatial distribution of fracture sets in
rock masses. Complexity of the technique could be
extended to include the "zonal" concept.
In spite of these limitations, the described technique is no less

accurate ~han the slip-circle analysis employed in soil mechanics
to assess slope stability. It serves to provide first-order
estimates of the probability that local failures will occur around
openings.
185.
The approach may be extended to approximately simulate redistribu-

tion of excess loads. This would involve the undertaking of a series
of related stress analysis runs. At each successive stage, the new
boundary of the opening would coincide with some contour for the
probability of failure (either for individual components within the
system or for more extensive zones) The selected contour would
correspond to some level of probability that the analyst considers
to be unacceptable from the stability point of view.
Changes in the geometry of the opening will result in a new stress

pattern. Assessment of this model may indicate further areas where
the probability of failure is above the acceptable level of risk.
The analysis is iterated until the extent of cumulative instability
is assessed.
186.
CHAPTER 8 COMPARISON BETWEEN MODEL AND PROTOTYPE RESPONSE
This chapter compares modelling predictions with mining experience

in the silver-lead-zinc orebodies at Mount Isa.
The aim of this component of the project has been to assess past
performance within sublevel open stapes and to derive a statistically
based model for strike spans of stable hanging-walls. The incidence
of local and extensive ground failures is also considered.
8.1 Practical Difficulties
A preliminary survey of records available at the mine suggested

that it may be feasible to undertake the envisaged study. Apparent
data-banks were provided:by Mine Survey's, Mine Planning's, Mine
Geology's and Mining Research Section's files. Discussions with
long-term employees of the company also suggested a potential
source of information.
Some aspects related to the proposed study were already assessed

and documented, e.g., Brady (1975, 1977), Bridges (1979a, 1979b,
1980), Lang (1980), Lee and Bridges (1980).
However, closer examination of the task revealed certain practical

difficulties:
(1) Volume of "potentially" useful records is considerable.

These are sometimes in a form that is not readily
accessible.
187.
(2) Quality and detail of records is not consistent. Some

discrepancies exist between various data sources.
Personal bias is apparent in the description of ground
failure severity. The relative importance of events is
commonly proportional to stope dimensions (i.e., the
larger the stope, the greater must be the volume of
failed ground before it is considered as significant)
(3) Some records are not available and there are gaps in
the information.
(4) Quality and reliability of geological data is variable.

This is largely a reflection on the diminished import-
ance that was attached to mapping of geological
structures and core logging for rock mechanics purposes
prior to the 1970's.
(5) The geology of some stope hanging-walls is more complex

than assumed in the present study. Several rock types
and structural environments are apparent along so~e
walls. These include the presence of silica--dolomite,

uniformly bedded shales, major folding, transverse
faulting and strike shearing.
(6) Poor stoping practise, such as severe undercutting of

hanging-walls, has contributed to some failures.
(7) S~bility of some stapes was affected by other factors

apart from geological structure, e.g., increased loads
due to proximity of other mine workings.
(8) Unsupported hanging-walls were exposed by several

different mining methods. Sublevel open stoping, long-
hole benching, shrink-stoping and other methods have
been used. Combinations of these techniques were employed
in some stopes. The effect of each mining method on
hanging-wall stability may be different.
188.
(9) A large number of stope hanging-walls have similar

dimensions. This arises because the orebodies are common-
ly extracted on a regular pattern of stapes and pillars.
Only limited information can be deduced in regard to the
maximum dimensions for stable spans.
(10) Stapes have been mined at various depths below surface.

Therefore, a common pre-mining stress field is not
associated with all of them.
(11) Other factors are likely to modify the overall rock mass
response. These include pillar dimensions, stope extract-
ion sequence, blasting practice and stope-filling
history. These contribute to some of the variability
observed in hanging-wall stability.
8.2 Scope of Analysis
In view of the limitations suggested above, the depth of the study

that was undertaken does not permit a rigorous back-analysis. How-
ever, some broad comparisons can be made between prototype and
model response.
The investigation of Orebody Nos. 5 and 7 is restricted to stapes

extracted between 7 and 15 levels at the mine. Stapes in Orebody
No. 11 are located between 7 and 9 levels. These levels correspond
to depths below surface of between 270 metres and 730 metres.
Considerable stoping was also undertaken above 7 level. However,

records describing their behaviour are generally not available.
Furthermore, the proximity of these stapes to the surface places
189.
them in an environment that is markedly different to that which was

modelled, i.e., lower stresses and greater degree of weathering of
the rock mass.
8.3 Observed Response
A total of forty-five(45) stopes were examined in Orebody Nos. 5

and 7 combined, and seven(?) within Orebody No.ll.
Only 30 per cent of_hanging-walls in the former two orebodies are

located in uniformly bedded dolomitic shales. A further 20 per
cent may be classed as a predominantly similar geological environ-
ment (i.e., less than 25 per cent of their surface area is associated
with major fold zones, transverse faulting or the more massive
"silica-dolomite" rock type). Therefore, about 50 per cent of the
stapes have hanging-walls that are broadly similar to that modelled
in the present analysis.
All stapes examined in Orebody No. 11 are located in uniformly

bedded dolomitic shales or predominantly similar geological environ-
ment. However, it should also be noted that most are in close
proximity to the contact with the silica-dolomite rock type.
The mean dimensions deduced for the horizontal or strike span of

hanging-walls and for widths of stopes are summarized in Figures
65 and 66 These indicate that the excavated strike span of hanging-
walls is generally greater than 25 metres. It is commonly about
30 - 40 metres.
Although not represented graphically, the down-dip extent of stapes

is in the range of about 80 to 120 metres.
190.
50
l!l
...z
0 45
0
f>l
"'0
I'<
40
-
~ 35
-

~c
~
.t.
z
...0
I'<
u
::1
30 -
-\;,

..
....
0
z
H
25 - .,

.""
"'~ 20 -
I'<
!;! )(
z 15
....
f>l
"'f?
"'
10
-
-

..
)( )( @I'

"'
0
(!l t;
p @
:c
!:::
....
3:
5

0
15 20 25 30 35 40 45 50 55 60 65 70 75 eo
HORIZONTAL OR STRIKE SPAN OF HANGING-WALL IN METRES
" 11 Orebody Stopes

5 Orebody Stopes
)( 7 Orebody Stopes
0 Stopes with predominantly Uniformly Bedded Shale Han<Jing-walls
NCP Hanging-wall(H), Crown(C) and Pillar(P) ground failure.
Orebody Nos. 5, 7 and 11: Relationship Between Strike Span of

Hanging-walls and Width of Stapes.
FIGURE 65.
191.
99.9
99.5
99.0
- - - - - - i--~ I
-- - -
+3
98.0
95.0
90.0
r
;
--
- +2
j +1
. BO.O
70.0 1 !/)
G
.,
.:
".,0'
60.0 I ./ rt
g
..
0.
.,:>'"'
r.. 50.0 I 0
"
0.
0.
.....
.....
40.0 I ".<.
.....
...."'
~ 30.0 .I rt
0
::>
"'
u
20.0 I
10.0 I --
-1
05.0 I I
02.0
/.
T
! -2
01.0
I
00.5
I
I
I -3
I
00.1 T I I I I I I I I
10 20 30 40 50 60 70 80 100 200
Strike Span of Hanging-walls

(Metres)
Orebody Nos. 5 and. 7 Combined: Cumulative Frequency

Distribution plotted on Probability Paper for the Strike
Span of Hanging-walls.
FIGURE 66.
192.
Geometry of individual stapes is rarely uniform across the opening.

Variation is most pronounced in their width normal to bedding.
Stope dimensions commonly reflect local changes in the thickness of
the economic mineralization. Observed local variations are summarized
in Table 14.
The stability of hanging-walls and abutments of stopes with various

strike spans are indicated in Table 15. The results suggest that
the walls and abutments are stable in stopes with hanging-wall
strike spans not exceeding 30 metres. The probability of major
collapse increases considerably as the span is extended to 40 metres.
Data for stapes with spans exceeding 40 metres is limited. The

available results do not appear to be in accord with the trend de-
duced for stopes with smaller hanging-wall dimensions.
Some additional information is provided by the trial stoping

experiment undertaken within Orebody No. 7 between 11 and 13 levels
of the mine (Brady, 1975, 1977). The hanging-wall of the two stopes
involved remained stable for spans of 31 metres and 39 metres,
respectively. The pillar between the stopes deteriorated as its
dimension in the north-south direction was progressively reduced to
about 14 metres. Eventual failure of this pillar exposed a hanging-
wall with an unsupported span of 80 metres. Considerable overbreak
from this wall was recorded.
Hanging-wall and footwall displacements have been monitored in a

number of stapes. Results of two case studies undertaken in parallel
with the present research project are reported in Lang (1980) and in
Bridges (1979a). Monitoring sites are located in Orebody No. 5
between 8 and 10 levels, and in Orebody No. 8 between 13 and 15
levels, respectively.
A mid-span hanging-wall displacement of about 25 mm is suggested

in the Orebody No. 5 stope prior to failure of the wall. A mid-span
footwall displacement of 10 mm is suggested at a strike span of
75 metres. Collars of instrumentation are located 35 metres and 45
metres from the respective walls. The observed displacements are
193.
TABLE 14.
Case Study: Variations in the Width of Stapes (in Metres).
Orebody No. I-1ean Standard Maximum Minimum

Deviation
5 23 8 42 4
7 12 7 27 3
11 8 4 17 2
TABLE 15.
Case Study: Stability of Hanging-walls & Abutments.
Probability of Failure (%)
Hanging-wall Span Orebody Nos.5 and 7 Orebody No .11

(Metres)
Hanging-wall* Abutment# Hanging-wall* Abutment#
Less than 30 0 0 0 0
30-35 9 25 33 0
36-40 33 50 67 33
Greater than 40 25? 0? 0? 0?
* In uniformly bedded dolomitic shales only.

# Irrespective of hanging-wall geology.
194.
expressed relative to these positions. This indicates that the total

displacement relative to an "infinite" boundary will be much greater.
In the Orebody No. 8 study, total displacements of about 25 - 35 mm

and 50 - 60 mm were indicated for the hanging-wall and footwall,
respectively, at a strike span of 80 metres for the stope. These
values have been adjusted to remove the effect of instrumentation
collar proximity to the opening. A comprehensive discussion of the
pre-mining stress field, observed displacements and their interpre-
tation is presented in Bridges (1979a) Total deformation is inter-
preted in terms of elastic and non-elastic components.
It should be noted that the hanging-wall of both groups of stopes

is geologically complex. It is locally associated with the more
massive silica-dolomite rock type and with major folding. The foot-
wall in both instances is comprised entirely of shale. Significant
hanging-wall collapses were encountered in both orebodies. Failures
were "dish-shaped" and were dislodged from shale portions of walls.
There was virtually no collapse from the silica-dolomite.
8.4 Rock Mass Classification Predictions
THERE ARE DIFFICULTIES IN DIRECT APPLICATION

OF PUBLISHED ROCK MASS CLASSIFICATIONS TO
ESTIMATES OF SPANS FOR UNSUPPORTED OPENINGS,
Estimates of stable spans for hanging-walls of sublevel open stopes

in the silver-lead-zinc orebodies at the Mount Isa Mine appear to
be conservative. This is particularly apparent with respect to the
RMR-Geomechanics classification system proposed by Bieniawski (1973).
195.
However, it must be appreciated that the RMR-system was basically

designed for the evaluation of near surface structures. It was never
intended to embrace mining situation. Its principal applications
are to permanent engineering structures within rock masses that are
commonly subjected to relatively low stresses. With respect to
these, the classification provides an excellent system which is no
more conservative than any of the others (Baczynski, 1980f).
The following three main factors confound the application of the

RMR-index to mining situations:
(1) The system does not provide for the incorporation of

stress effects on stability.
(2) Although the system attempts to determine stand-up time

which could then be indirectly related to stability in
mining situations, the proposed times are conservative
with respect to past mining experience at the mine.
(3) There is an upper limit of 20 metres for the maximum

permissible span dimension. This value is extremely
conservative.
The modified RMR-system proposed by Laubscher and Taylor (1976) and

the Q-index developed by Barton et al. (1974) were proposed for, or
extended to embrace, mining situations. Both classifications yield
similar estimates for stable spans.
The determined values again appear to be conservative. However, the

results are considerably better than those derived on the basis of
Bieniawski's classification. The derived spans for Orebody Nos. 5
and 7 appear to be within 20 - 30 per cent of the average values
suggested by past mining experience.
There is a marked discrepancy between field observations and

stability estimates for stopes in Orebody No. 11. However, it should
be appreciated that the structural model employed in the analysis
196.
was formulated on the basis of investigations undertaken between

11 and 15 levels of the mine. Described mining experience relates
to stopes above 9 level, where detailed structural information is
not available. These areas are presently inaccessible.
Experience in cut-and-fill stopes between 11 and 15 levels suggests

a mean stable span of about 3 - 5 metres. These estimates are tenta-
tive because of the difficulty associated with the extrapolation of
experience gained in cut-and-fill stopes where hanging-walls are
progressively disturbed during mining to sublevel open stoping
situations.
THE MODIFIED-RMR RATING AND THE Q-INDEX ARE

CONSIDERED TO BE POTENTIALLY USEFUL TOOLS IN
PROVIDING FIRST-ORDER ESTIMATES FOR STABLE
HANGING-WALL SPANS,
The two classification systems may be tailored to suit more closely

local mining conditions.
8.5 Finite Element Predictions
Material properties used in the finite element analysis have been

estimated on the basis of various rock mass classification systems.
The results of the analysis may be anticipated to reflect the
limitations inherent in these classifications.
However, there is a marked difference between the rock mass classi-

fication and finite element approaches. In the latter, the continuum
around the opening is discretized into finite elements. This permits
a definition of critical areas around the stope and enables
197.
the assessment to be expressed in terms of the severity of ground

collapses. Also, in contrast to the classification approach, the
rock mass is modelled as a dynamic system. Excess loads in yielding
elements are redistributed to surrounding areas.
THERE IS REASONABLE CORRELATION BETWEEN THE

RESPONSE OF FINITE ELEMENT MODELS AND IN-SITU
OBSERVATIONS,
Both sets of results are presented in Tables 10 - 13 and 15.
Considerable instability is suggested in hanging-walls and abutments

of stapes within Orebody Nos. 5 and 7 at strike spans exceeding 40
metres. Stapes with spans of 25 - 30 metres are generally stable,
apart from minor, local instabilities.
As noted earlier, a marked discrepancy is apparent between model

and prototype for stapes in Orebody No. 11.
Monitored hanging-wall and footwall displacements do not provide

sufficient data to validate rock mass modulus values assumed in
the analysis. Evidence is contradictory.
However, it should be noted that the geology of monitored hanging-

walls differs from that of the modelled stapes. The more massive
silica-dolomite rock type is present within the walls. This imparts
a greater stiffness to them and may account for the smaller deflect-
ions that have been measured in-situ. Displacements are also extremely
sensitive to variations in the magnitude of the pre-mining stress
field. Although inferences made without supporting field evidence
can only be regarded as speculative, this variation alone can explain
the observed differences in displacement.
It is interesting to note that Bridges (1979a) suggested a 50 - 60 rnrn

displacement for the shale footwall of the stope located in Orebody
No. B. The pre-mining stress field of the area is complicated by the
shadow or shielding effect of nearby mining. Estimates of the stress
198.
acting normal to bedding range from 20 MPa to 35 MPa. The stress is

also non-uniform across the stope. In view of the uncertainties,
the predicted displacement of 75 - 80 mm (as extrapolated from
Figure 51 for a hanging-wall span of 80 metres at a stress of 36
MPa acting normal to bedding) compares favourably with field observa-
tions. Full agreement is indicated for a normal stress of 26 MPa.
8.6 Probability of Failure Analysis Predictions
The analysis suggests that the probability of local failure_ and its
extent within the hanging-wall and abutment areas of stapes in
Orebody Nos. 5 and 7 are low for strike spans not exceeding 25 metres.
Failure probabilities increase markedly as spans are extended to
40 and 55 metres.
These results are in good accord with field observations.
The significance of different pre-mining stress fields on the

probability of failure around a stope are demonstrated in Figures
60 - 63. These plots clearly highlight the importance of the stress
parameter in stope stability analyses and in the interpretation of
past mining behaviour.
199.
CHAPTER 9 SUMMARY AND DISCUSSION
The thesis presents a systemmatic approach for the structural

characterization of rock masses, estimation of material properties
and the stability assessment of engineering structures within
heterogeneous continua.
This chapter summarizes and discusses the conceptual approach used,

basic principles, limitations and possible improvements to the model.
9.1 Background Considerations
Most rock masses are complex physical systems. The objective o any
stability analysis is to achieve an understanding of their response
under a set of specified external conditions.
A number of factors dictate the approach adopted to analyses. Two

of the more important ones are:
(1) Physical dimensions of the rock mass block.
(2) Extent to which the rock mass block is transected by

geological discontinuities.
Where block dimensions are relatively small and critically orientated

weakness planes completely penetrate the mass, the response o the
systems under load will be largely governed by the shear and stiff-
ness characteristics of these discontinuities. Simple graphical
techniques, similar to those applied by Miller and Barrett (1979) to
200.
slender pillars, may be employed in the stability analysis. However,

care should be exercised to ensure that the failure mode will be
by movement along these planes and that it will not involve signifi-
cant rotation and loss of interlock between the sub-blocks delineated
by them. The latter mode of deformation may considerably reduce
estimated shear strength of the mass (Ladanyi and Archambault, 1969).
Larger blocks with several sets of non-penetrating and non-uniformly

distributed sets of weakness planes constitute more complex systems.
Adequate characterization of such systems is not an easy task. It

demands the ability to quantify large volumes of structural data
into a meaningful and useful form. Sufficient detail should be
preserved in the model to reflect the inherent local variability in
the system.
However, it must be borne in mind that the analysis of complex

systems is difficult. Even where such analyses are numerically
possible, the computational effort involved in reaching a solution
may not be justified economically.
The tendency has been to reduce the number of components in systems.

This simplifies the model. The approach has invariably resulted in
the rock mass being characterized by a relatively homogeneous
continuum or discontinuum. Directional properties common to all
elements of such models are sometimes included.
Since the mid-1960's, large numbers of such parametric studies have

been reported in technical literature. Modelling includes both
physical and numerical techniques.
201.
9.2 Basis of Adopted Model
In the present study, the rock mass is characterized as a hetero-

geneous continuum with statistically variable properties. The
material model is a function of the inherent structural variability
within the mass.
The basis of the model rests on the contention that any large block
of rock mass may be represented by an equivalent system of smaller
sub-blocks or modules.
9.3 Derivation of Model Parameters
A model for structural variability within the rock mass is derived

on the basis of area-mapping of underground openings. Assessment of
data involves an examination of the variability that exists between
"unit areas".of the rock mass. Some assumptions are necessary to
extend the model into the third dimension.
Several sets of weakness planes are defined and a statistically-based

model is formulated to account for their spatial distribution in the
mass.
The work undertaken at the Mount Isa Mine demonstrates that the
spatial distribution of each set of fractures is not random. They
tend to occur in zones or clusters. The extent of the zones in
directions parallel and normal to the orientation of the average
fracture plane of a set may be described statistically.
202.
The use of standard semi-variogram techniques, such as presented in

Clark (1979), to describe the extent of fracture zones was considered.
However, the multi-component nature of the model and characteristics
of the statistical distributions describing local variability in
structure indicate that this approach would not offer any simpler
solution than the one adopted.
The assumed structural model is validated by a computer program

developed for the simulation of fracture distributions within rock
mass blocks. Good correlation was observed between model and proto-
type. The model is extended to permit assessment of structural
variability between rock mass blocks of any dimensions.
A number of published rock mass classifications are used to derive

estimates of material properties. A computer program was developed
to expedite the assessment. The analysis centres on "Monte Carlo"
sampling of probability density functions for the intensity of each
fracture set as well as other parameters and the derivation of
appropriate classification ratings for several systems.
In theory, there is no limit to the complexity of the material model

that may be assumed for individual modules of the rock mass. Direct
field testing of large samples offers the most reliable solution.
Alternatively, the results of physical modelling tests on blocky
systems, such as described in B!own (1970a, 1970b), Einstein et al.
(1969) , Rosenblad (1970) and many others, may be employed to derive
more realistic directional and stress-dependent continuum properties.
However, there are practical constraints. These largely arise from

inherent limitations of the numerical techniques used to perform
stability analyses. Although three-dimensional techniques have been
available for several years, their widespread use is limited by
cost considerations. Most analyses are performed on two-dimensional
idealizations of problems. Currently available programs only permit
relatively simple anisotropic materials to be assessed. It is
anticipated that some of these restrictions may eventually disappear
with the development of multi-laminate material models, such as
203.
described by Zienkiewicz and Pande (1977) for a two-dimensional

analysis. However, limitations are present at the moment and they
dictate the complexity that may be assumed for material properties.
9.4 Stability Analysis
Three different approaches are used to assess the stability of

stopes:
(l) Rock Mass Classifications
{2) Two-dimensional, Finite Elements, and
{3) Probability of Failure Analysis.
Results indicate a reasonable correlation with prototype response.
Certain difficulties are associated with each of the three analysis

techniques employed.
None of the published rock mass classification systems is completely

satisfactory. Estimates are conservative. However, in some instances,
they are within 20 - 30 per cent of the values suggested by past
mining experience in the silver-lead-zinc orebodies at Mount Isa Mine.
Thus, a number of the classifications are considered to be potentially
useful in providing first-order estimates. These systems may be
tailored to suit more closely local mining conditions.
Limited evidence suggests a significant difference in the response

between finite element analyses of models with homogeneous and
statistically heterogeneous material properties. Differences are
204.
apparent in the extent of plastic yielding and the distribution of

stresses around the opening.
It is acknowledged that some of the observed differences may have

arisen purely from the deficiencies that are inherent in numerical
techniques, such as the "peculiarities of the discretization process"
referred to in Baecher and Ingra (1979) . These include the shape,
size and orientation of elements in the mesh. However, because of
the extent of discretization and the isoparametric nature of elements
employed in the present analysis, it is unlikely that any signifi-
cant portion of the observed difference can be attributed to this
cause. The differences arise from the variation in local stiffness
and strength of the continuum, and the redistribution of excess
loads in yielding components of the mass.
Application of finite elements to analysis of statistically

heterogeneous models is limited by cost considerations. These are
prohibitive even when a relatively small number of models with
elasto-plastic behaviour are involved. This problem is compounded
in three-dimensional analyses.
The cost factor militates against the use of statistically-based

finite element modelling approaches.
An alternative approach to assessing the potential extent of ground

collapse around underground openings is offered by the "probability
of failure analysis". This method has many advantages and i t is a
powerful tool for gaining a first-order appreciation of rock mass
response. It can be readily employed in parametric evaluations of
ground behaviour.
.205.
9.5 Conclusions
(l) All stability analyses are conceptual processes. They involve

the idealization of real rock mass systems by means of
simplified models.
(2) Most rock masses are complex and locally variable physical
systems. They cannot be adequately represented by homogeneous
models.
(3) The complexity of a rock mass system is reduced by the

subdivision of the mass into an equivalent system of smaller
modules or subsystems. This approach offers a sound framework
for the statistical characterization of rock masses.
(4) The response of the whole system is equal to the total sum
of the responses of individual subsystems that are located
along "critical paths" within the system.
(5) Spatial distribution of fracture plane centres is not random

in at least one rock mass. Members of each set of geological
structures appear to occur in zones. However, the spatial
distribution of "high intensity" zones is assumed to be random.
(6) "Area mapping" of underground openings offers a suitable

method for the determination of statistically-based model for
local variability in the intensity of geological structures.
(7) The commonly recommended method of single "line sampling"

does not furnish sufficient data for the 2-D or 3-D
characterization of rock masses.
(8) There are considerable difficulties associated in the

estimation of rock mass properties such as unconfined
compressive and tensile strengths, cohesion, friction angle,
206.
modulus of deformation and Poisson's ratio.
(9) In-situ or laboratory testing of large rock mass samples

offers the only reliable method for the determination of
material properties.
(10) Reasonable, first-order estimates of some rock mass properties

may be deduced for the dolomitic shales at the Mount Isa Mine
on the basis of various published rock mass classification
systems.
(11) Aspects of the following classifications appear to be

potentially useful in the determination of rock mass properties:
Strength: (i) Hansagi's C-Factor in conjunction with

Bieniawski's RMR-index.
Modulus: (i) Deere's RQD-index, or
(ii) Bieniawski's RMR-index.
(12) There are serious limitations inherent in the rock ~s
classification approach. Most of these centre on the simplicity

of the ensuing material model and the validity of extrapolating
relationships established at other sites to the mass under
investigation.
(13) More complex material properties may be deduced for rock mass
subsystems with specified structural configurations on the
basis of numerous physical and numerical modelling studies of
blocky systems already reported in technical literature.
(14) Practical constraints are commonly imposed on the complexity

of the material model. These arise from limitations that are
inherent in the numerical techniques used to perfoxm
stability analyses.
(15) The results of stability analyses undertaken in the present

study suggest a reasonable degree of correlation between
prototype and model response.
207.
(16) None of the published rock mass classification systems

appears to be completely satisfactory with respect to the
determination of stable hanging-wall spans. Estimates are
generally conservative by at least 20-30 per cent.
(17) Some classifications may be tailored to suit local mining

requirements. The most promising classification systems are
Barton's Q-index and Laubscher and Taylor's Modified RMR-index.
(18) The limited evidence that is available suggests that there

is a statistically significant difference between the response
of finite element models with homogeneous and heterogeneous
material properties.
(19) Stability analyses are extremely sensitive to variations in

the pre-mining stress field. Therefore, any correlation
between model predictions and actual field observations will
remain in the realm of speculation until field stresses are
defined with some confidence.
(20) The scope of any finite element analysis involving statistical

assessment of heterogeneous models is markedly limited by the
enormous computational effort associated with such studies.
Even limited analyses involving 2-D models are outside the
scope ~f routine assessments. 3-D analyses are prohibitive.
(21) Assessment of past mining behaviour of stapes at the Mount Isa

Mine is confounded by various practical difficulties. These
are detailed in Paragraph 8.1 of the Ph.D thesis.
(22) The "Probability of Failure" analysis offers a relatively

simple, inexpensive and statistically-based method for
first-order estimates of rock mass stability. Very few
limitations are imposed on the complexity of the material
model that may be assumed for rock mass properties.
208.
(23) The "Probability of Failure" analysis approach may be

extended to permit an approximate assessment of the extent
of progressive collapse around underground openings.
(24) All phases of data assessment, statistical model formulation

and verification, and stability analysis may be greatly
expedited by the routine use of computer capabilities.
9.6 Recommendations
(1) The procedure of "Area Mapping" should be introduced as part

of routine geotechnical investigations. This will permit the
formulation of statistically-based models for the 2-D and
3-D variability in geological structure within rock masses.
However, because of the associated work-load, this mapping
should be restricted to selected areas of mine openings or
surface excavations.
(2) Some form of the "Probability of Failure" analysis technique

should be adopted for first-order evaluation of potential
rock mass response. Such analyses will be markedly expedited
by the development of the appropriate computer software. This
will permit the procedure to be employed on a more routine
basis as an aid to mine planning.
(3) Published technical literature on numerical and physical

simulation of blocky systems should be reviewed critically
and a model bank describing the response under load of
subsystems with particular structural configurations should
be established. This will furnish better material properties
for statistically heterogeneous models of real rock masses.
209.
(4} Modelling predictions can only be validated by field

observations. The art of predicting unsupported spans in
rock masses will only be advanced by monitoring,
documenting and interpreting the response of the mass during
mining. This work should be undertaken on a more routine
basis and the results should be made available in the form
of a general data-bank.
-
210.
REFERENCES
Ackoff, R.L. (1974). Towards a Sys tern of Sys terns Concepts.

in System Analysis Techniques, edited by J.D. Couger
and R.W. Knapp, John Wiley & Sons, 509p.
Aitchison, J. and Brown, J.A.C. (1976). The Lognormal Distribution:

with Special Reference to Its Uses in Economics.
Cambridge University Press, 176p.
Ang, A.H-S. and Tang, W.H. (1975). Probability Concepts in

Engineering Planning and Design: Volume I: Basic
Principles. John Wiley & Sons, Inc., 409p.
Baczynski ,_ N.R. (1973). 5 Orebody Below 13 Level - Part II,

Mount Isa Mines Limited, Memorandum NRB/2.6/RES MIN
1.0.8, dated 14th November 1973 (Unpublished).
Baczynski, N.R.P. (1974). Structure and Hanging Wall Stability

In Lead Orebodies at Mount Isa Mine. M.Sc. Thesis,
James Cook University of North Queensland
(Unpublished) .
Baczynski, N. R. ( 1977a) Generation of Fracture Patterns.

Report to Department of Mining & Metallurgy,
University of Melbourne, dated 15th August 1977,
17p.
Baczynski, N.R. (1977b). Modelling of Shale Hangingwalls:

Details of the Proposed Investigation Programme.
University of Melbourne, dated 26th August 1977, 20p.
211.
Baczynski, N.R. (1977c). Geological Characteristics of the Rock

Mass: Dolomitic Shales: Mount Isa Mine, Q1d. Report
to Department of Mining & Metallurgy, University of
Melbourne, dated lOth November 1977, 79p.
Baczynski, N.R. (1978a). Computer Program CRACKS 2: Procedures

Manual. Report to Department of Mining & Metallurgy,
University of Melbourne, dated 15th June 1978, 344p.
Baczynski, N.R. (l978b). A Three-Dimensional Model for the

Spatial Distribution of Extension Fractures Within
the Dolomitic Shales at the Mount Isa Mine, North-
Western Queensland. Report to Department of Mining &
Metallurgy, University of Melbourne, dated 30th July
1978, 22p.
Baczynski, N.R. (1979). Estimate of Rock Mass Properties and

Stable Hanging Wall Spans on Basis of RoCk Classifi-
cation: Dolomitic Shales: Mount Isa Mine, Queensland.
University of Melbourne, dated 6th March 1979, 312p.
Baczynski, N.R. (l980a). Finite Element Modelling: Listing of,

and Input Data Requirements for Complementary Programs
and Modifications to VISPLAS. Report to Department
of Mining & Metallurgy, University of Melbourne,
dated 21st January 1980, 56p.
Baczynski, N. R. (198Gb) . Finite Element Modelling of Stope

Hanging-walls with Heterogeneous Material Properties.
University of Melbourne, dated lst February 1980,
233p.
Baczynski, N.R. (l980c). Finite Element Modelling: Confirmatory

Reruns. Report to Department of Mining &
Metallurgy, University of Melbourne, dated 28th
February 1980, 36p.

212.
Baczynski, N.R. (1980d). Bibliography of Technical Papers on

Physical Modelling. Report to Department of Mining &
Metallurgy, University of Melbourne, dated 24th March

1980, 38p.
Baczynski, N.R.P. (1980e). Zonal Concept for Spatial Distribution

of Fractures in Rock. Proc. Third Aust. N.Z. Conf.
Geomech., Wellington, May 12-16, Vol.2, pp.29-33.
Baczynski, N.R.P. (1980). Application of Various Rock Mass

Classifications to Unsupported Openings at Mount Isa,
Queensland: A Case Study. Proc. Third Aust. N.Z. Conf.
Geomech., Wellington, May 12-16, Vol.2, pp.l37-143.
Baecher, G.B. and Ingra, T. (1979). Equivalent Parameters in

Soil Profile Modeling ..Proc. Specialty Con.
Probabilistic Mechanics and Structural Reliability,
Tucson, Jan. 10-12, pp.330-337.
Barton, N.; Lien, R. and Lunde, J. (1974). Engineering Classifi-

cation of Rock Masses for the Design of Tunnel
Support. Rock Mech., Vo1.6, pp.l89-236.
Barton, N.; Lien, R. and Lunde, J. (1975). Estimation of Support

Requirements for Underground Excavation. Proc. 16th
Sy~. Rock Mech., University of Minnesota, Sept. 22-24,
pp.l63-177.
Bertalanffy, L. von (1969). Chance or Law. in Beyond Reductionism,

edited by A. Koestler and J .R. Smythies, New York:
Hutchinson, p.l48.
Bieniawski, Z.T. (1973). Engineering Classification of Jointed

Rock Masses. Trans. s. Afr. Instn. Civil Engrs.,
Vol. 15, pp.335-343.
213.
Bieniawski, Z.T. (1975). Case Studies: Prediction of Rock Mass

Behaviour by the Geomechanics Classification. Proc.
Second Aust. N.Z. Conf. Geomech., Brisbane, July 21-25,
pp. 36-41.
Bieniawski, Z.T. (1976). Rock Mass Classification in Rock

Engineering. Proc. Symp. Exploration for Rock Engng.,
Johannesburg, Nov. 1-5, Vol.l, pp.97-106.
Bieniawski, Z.T. (1978a). A Critical Assessment of Selected

In-situ Tests for Rock Mass Deformability and Stress
Measurements. Proc. 19th u.s. Symp. Rock Mech.,
Stateline, May 1-3, pp.523-529.
Bieniawski, Z. T. (1978b). Determining Rock Mass Deformability:

Experience from Case Histories. Int. J. Rock Mech.
Min. Sci. & Geomech. Abstr., Vol.lS, pp.237-247.
Brady, B.H.G. (1975). Rock Mechanics for Excavation Design in

the Racecourse Orebodies. Mount Is a Mines Limited,
Technical Report No. RES MIN 45, dated November
1975, 22p. (Unpublished).
Brady, B.H.G. (1977). An Analysis of Rock Behaviour in an

Experimental Stoping Block at the Mount Isa Mine,
Queensland, Australia. Int. J. Rock Mech. Min. Sci. &
Geomech. Abstr., Vol.l4, pp.59-66.
Bridges, M.C. (1975). Presentation of Fracture Data for Rock

Mechanics. Proc. Second Aust. N.Z. Conf. Geomech.,
Brisbane, July 21-25, pp.l44-148.
Bridges, M. C. ( 1979a) . Rock Mechanics of M665 Triplet, 8 Orebody,

lSB to 13 Level. Mount Isa Mines Limited, Technical
Report No. RES MIN 55, dated October 1979, 135p.
(Unpublished)
214.
Bridges, M. C. (1979b) . Virgin Stress Field for the Northern Mine

Area. Mount Isa Mines Limited, Merrorandum
MCB/7.5/ RES MIN 3.0, dated 28th November 1979, 12p.
(Unpublished) .
Bridges, M.C. (1980). Geomechanics of Lead-Zinc Mining Below

15 Level. Mount Isa Mines Limited, Technical Report
No. RES MIN 56, dated February 1980, (Unpublished).
Brown, E.T. (1970a). Strength of Models of Rock with Intermittent

Joints. J. Soil Mech. Fdns. Div. Am. Soc. Civ. Engrs.,
Vol.96, pp.l935-1949.
Brown, E.T. (1970b). Modes of Failure in Jointed Rock Masses.

Proc. Second Int. Rock Mech. Congr., Belgrade, Vol.2,
paper 3-42, 6p.
Christian, J.T. and Wong, I.H. (1973). Errors in Simulating

Excavation in Elastic Media by Finite Elements.
Soil and Foundations, Vol.l3, pp.l-10.
Clark, I . (1979). Practical Geostatistics. Applied Science

Publishers Ltd., London, l29p.
Coon, R.F. and Merritt, A.H. (1970). Predicting In-situ Modulus

of Deformation Using Rock Quality Indices. in
Determination of the In-situ Modulus of Deformation
of Rock, Amer. Soc. Testing and Mater., STP 477,
pp. 154-173.
Cording, E.J.i Hendron, A.J. and Deere, D.U. (1971). Rock

Engineering for Underground Caverns. Proc. Symp.
Underground Rock Chambers, Pheonix, January,
pp.567-600.
Couger, J.D. and Knapp, R.W. (1974). system Analysis Techniques.
John Wiley & Sons, 509p.

215.
Crotty, J.M. and Wardle, L.J. (1978). User's Manual for Program
BITEM - Two Dimensional Elastic Stress Analysis for
Piecewise Homogeneous Solids. Aust. CSIRO Div. Appl.
Geomech. , Development of the Boundary Integral
Equation Method Applied to Underground Mining,
Project Report No.2
Cruden, D.M. (1977). Describing the Size of Discontinuities.

Int. J. Rock Mech. Min. Sci. & Geomech. Abstr.,
Vol.l4, pp.l33-137.
Cundall, P.A. (1971). A Computer Model for Simulating Progressive

Large-scale Movements in Blocky Rock Systems. Rock
Fracture: Proc. Int. Symp. Rock Mech., Nancy,
paper II-8, l2p.
Davis, G.B. (1974). Management Information Systems: Conceptual

Foundations, Structure, and Development.
McGraw-Hill Kogakusha Limited, 482p.
Deere, D.U.; Hendron, A.J.; Patton, F.D. and Cording, E.J. (1966).
Design of Surface and Near-Surface Construction in
Rock. Proc. Eighth Symp. Rock Mech., University of
Minnesota, pp.237-302.
Desai, c.s. and Abel, J.F. (1972). Introduction to the Finite

Element Method: A numerical Method for Engineering
Analysis. Van Nostrand Reinhold Company, 477p.
Desai, c.s. and Christian, J.T. (1977). Numerical Methods in

Geotechnical Engineering. McGraw-Hill Book Company,
783p.
Drucker, D.C. and Prager, W. (1952). Soil Mechanics and Plastic

or Limit Design. Q. Appl. Math., Vol.lO, pp.l57-165.
216.
Einstein, H.H.; Nelson, R.A.; Bruhn, R.W. and Hirschfeld, R.C. (1969).
Model Studies of Jointed Rock Behaviour. Proc. 11th
Syrup. Rock Mech., University of California, pp.83-103.
Goodman, R.E.; Taylor, R.I. and Brekke, T.L. (1968). A Model

for the Mechanics of Jointed Rock. J. Soil Mech.
Fdns. Div. Am. Soc. Civ. Engrs., Vol.94, pp.637-659.
Goodman, R.E. and St.John, c. (1977). Finite Element Analysis

for Discontinuous Rocks. in Numerical Methods in
Geotechnical Engineering, edited by C.S. Desai and
J.T. Christian, McGraw-Hill Book Corrpany, Chapter 4,
pp.l48-185.
Goodman, R.E. and Smith, H.R. (1980). RQD and Fracture Spacing.
J. Geotech. Engng. Div. Am. Soc. Civ. Engrs.,
Vol.l06, pp.l91-193.
Gudehus, G. (1977) . Finite Elements in Geomechanics. John Wiley &

Sons, 5 73p.
Hammersley, J.M. and Handscomb, D.C. (1964). Monte Carlo

Methods. Methuen & Co. Ltd., London, 178p.
Hansagi , I ( 1965) Numerical Determination of Mechanical

Properties of Rock and Rock Masses. Int. J. Rock Mech.
Min. Sci. & Geomech. Abstr., Vol.2, pp.219-223.
Hansagi, I. (1974). A Method of Determining the Degree of

Fissuration of Rock. Int. J. Rock Mech. Min. Sci. &
Geomech. Abstr., Vol. 11, pp.379-388.
Harris, J.F.; Taylor, G.C. and Walper, J.L. (1960). Relation of

Deformational Fractures in Sedimentary Rocks to
Regional and Local Structure. Bull. Am. Assoc.
Petroleum Geologists, Vol. 44, pp.l853-1873.
217.
Harrison, W.J. and Gerrard, C.M. (1972). Elastic Theory Applied

to Reinforced Earth. J. Soil Mech. Fdns. Div. Am. Soc.
Civ. Engrs., Vol.98, pp.l325-1345.
Herget, G. (1968). The Stress Field in the Urquhart Shales at

Mount Isa (Queensland, Australia) Based on
Structural Investigations. Rock Mech. Engng. Geol.,
Vol.6, pp.l90-200.
Hodgson, R.A. (1965). Regional Study of Jointing in Comb Ridge

Navajo Mountain Area, Arizona and Utah. Bull. Am.
Assoc. Petroleum Geologists, Vol. 45, pp.l-38.
Horibe, T. (1970). Abstract of the report from "Committee

Concerning the Method of the Measurement of the
Strength of Rock". in M.M.I.J. Rock Mech. in Japan,
Vol.l, pp.29-31.
Hudson, J.A. and Priest, S.D. (1979). Discontinuities and Rock

Mass Geometry. Int. J. Rock Mech. Min. Sci. &
Geomech. Abstr., Vol.l6, pp.339-362.
Kiraly, L. (1969). Statistical Analysis of Fractures (Orientation

and Density). Geologische Rundschau, Vol. 59, pp.l25-
l51.
Krumbein, W.C. and Graybill, F.A. (1965). An Introduction to

Statistical Models in Geology. McGraw-Hill Book
Corrpany, 4 75p.
Kulhawy, F.H. (1975). Stress Deformation Properties of Rock and

Rock Discontinuities. Engng. Geol., Vol.9, pp.327-
350.
Kulhawy, F.H. (1978). Geomechanical Model for Rock Foundation

settlement. J. Geotech. Engng. Div. Am. Soc. Civ.
Engrs., Vol. 104, pp.211-227.
218.
Ladanyi, B. and Archambault, G. (1969). Simulation of Shear

Behaviour of a Jointed Rock Mass. Proc. 11th Sy~.
Rock Mech., University of California, pp.105-125.
Lama, R.D. and Vutukuri, v.s. (1978). Handbook on Mechanical

Properties of Rocks: Testing Techniques and Results-
Volumes II, III and IV. Trans Tech Publications,
Clausthal, Germany.
Lang, A.M. (1980). Final Monitoring Report: L685 Stope, 5 Orebody.

Mount Is a Mines Limited, Memorandum AML/7. 5/RES MIN
5.2.1, dated 2nd April 1980, 4p. (Unpublished).
Larson, H.J. (1974). Introduction to Probability Theory and

Statistical Inference. John Wiley & Sons, Inc.,
430p.
Laubscher, D.H. and Taylor, H.W. (1976). The Importance of Geomechanics

Classification of Jointed Rock Masses in Mining
Operations. Proc. Symp. Exploration for Rock Engng.,
Johannesburg, Nov. 1-5, Vol. 1, pp.ll9-128.
Leask, W. (1977). Review of Some Tests of the Physical Properties

of Rock Types in the 1100 Orebody Area. Mount Isa
Mines Limited, Memorandum MCB/2.6/RES MIN 6.147,
dated 4th February 1977 (Unpublished).
Lee, M.F. and Bridges, M.C. (1980). Rock Mechanics of Crown

Pillars Be tween Cut-and-Fill Stopes at the Mount Is a
Mine. Proc. Conf. Application of Rock Mech. to
Cut-and-Fill Mining, University of Lulea, June 2-4,
Vol.3, Preprint.
MacLeod-Carey, I.R. (1969). Rock Mechanics Problems Associated

with the Extraction of Highly Inclined Parallel
Orebodies. Ph.D Thesis, University of Newcastle-upon-
Tyne, 173p. (Unpublished).
219.
McKavanagh, B.M. and Lee, M.F. (1977). Ultrasonic Velocity

Measurements in the 8 Orebody Crown Pillar, 11 Level,
Mount Isa Mine. Aust. CSIRO Div. Appl. Geomech.,
Support and Stabilization of Stapes, Project Report
No. 16, August 1977.
McMahon, B.K. (1967). Indices Related to the Mechanical

Properties of Jointed Rock. Status of Practical Rock
Mechanics: Proc. Ninth Syrnp. Rock Mech., Colorado
School of Mines, April 17-19, pp.ll7-133.
Markland, J. (1974). The Analysis of Principal Components of

Orientation Data. Int. J. Rock Mech. Min. Sci.,
Vol. 11, pp.l57-l63.
Mathews, K.E. and Ed~ards, D.B. (1969). Rock Mechanics Practice

at Mount Isa Mines Limited, Australia. Proc. Ninth
Cornmonwe al th Min. & Metal!. Congr . , London,
May 5-24, Mining and Petroleum Technology Section,
paper 32 , 68p.
Mathews , K. E. (19 70) . Excavation Design in Hard and Fractured

Rock at the Mount Isa Mine, Australia. M.E. Thesis,
University of Queensland (Unpublished).
Miller, D. R. (1978). Strength and Deformability in Triaxial

Compression of Rock Materials from the Crown Pillar,
11 Level, 8 Orebody, Mount Isa Mine, Queensland.
Aust. CSIRO Div. Appl. Geomech., Support and
Stabilization of Stapes, Project Report No. 24,
March 1978.
Miller, D.R. and Barrett, J.R. (1979). Jointed Rock Mass

Characteristics and Their Influence on Slender
Pillars. Proc. 20th u.s. Symp. Rock Mech., Austin,
June 4-6, Preprint.
220.
Moseley 1 F. (1968). Joints and Other Structures in the Silurian

Rocks of the Southern Shap Fells 1 Westm:>rland.
Geology Journal, Vol. 6, pp.79-96.
Nickelsen, R.P. and Hough, Van N.D. (1967). Jointing in the

Appalachian Plateau of Pennsylvania. Bull. Geol.
Soc. Amer., Vol. 78, pp.609-629.
Orr, C.M. (1974). The Geological Description of In-situ Rock

Masses as Input Data for Engineering Design.
S. Afr. CSIR, Report No. ME 1274, 133p.
Piteau, D.R. (1973). Characterizing and Extrapolating Rock Joint

Properties in Engineering Practice. RoCk Mech.,
Suppl. 2, pp.S-31.
Price, N.J. (1966). Fault and Joint Development. Pergamon Press,

London, 1 76p.
Priest, S.D. and Hudson, J.A. (1976). Discontinuity Spacings

in Rock. Int. J. RoCk Mech. Min. Sci. & Geomech.
Abstr., Vol. 13, pp.l35-148.
Robertson, A. MacG. (1970) . The Interpretation of Geological

Factors for Use in Slope Theory. Proc. Symp.
Theoretical Background to the Planning of Open Pit
Mines with Special Reference to Slope Stability,
Johannesburg, Aug. 29 - Sept. 4, pp.SS-71.
Rosenblad, J. L. ( 1970) Failure Modes of Models of Jointed RoCk

Masses. Proc. Second Int. Rock Mech. Congr.,
Belgrade, Vol. 2, paper 3-11, 7p.
Rosengren, K.J. (1968). Rock Mechanics at the Black Star Open

Cut, Mount Is a. Ph.D Thesis, Australian National
University (Unpublished).
221.
Secor, D.T. (1968). Mechanics of Natural Extension Fracturing at

Depth in the Earth's Crust. in Research in Tectonics,
edited by A.J. Baer and O.K. Norris, Geol. Surv.
Canada, Paper 68-52, pp.3-48.
Skempton, A.W.; Schuster, R.L. and Petley, D.J. (1969). Joints

and Fissures in the London Clay at Wraysbury and
Edgware. Geotechnique, Vol. 19, pp. 205-217.
Stimpson, B. and Ross-Brown, D.M. (1979). Estimating the Cohesive

Strength of Randomly Jointed Rock Masses. Min. Engng.,
Vol. 31, pp.l82-188.
Taha, H.A. (1976). Operations Research: An Introduction.

MacMillan Publishing Co. Inc., 648p.
Turner, F.J. and Weiss, L.E. (1963). Structural Analysis of

Metamo;phic Tectonites. McGraw-Hill Book Company,
Inc., 545p.
Wardle, L.J. and Gerrard, C.M. (1972). The "Equivalent"

Anisotropic Properties of Layered Rock and Soil
Masses. Rock Mech., Vol. 4, pp.lSS-175.
Wuerker, R.G. (1959). The Shear Strength of Rocks. Min. Engr.,

Vol. 11, pp.l022-1026.
Zienkiewicz, o.c. and Cheung, Y.K. (1968). The Finite Element

Method in Structural and Continuum Mechanics.
McGraw-Hill Publishing Company Ltd., 274p.
Zienkiewicz, O.C. and Pande, G.N. (1977). Time-Dependent

Multilaminate Model of Rocks - A Numerical Study of
Deformation and Failure of Rock Masses. Int. J.
Numerical and Analytical Methods in Georoech., Vol. 1,
pp. 219-247.
222.
ADDITIONAL REFERENCES
This section includes additional references that have been cited in

Baczynski (1977c, 1978b, 1979 and 1980b); technical reports submitted
to the University of Melbourne as part of the Ph.D project.
Awaji, H. and Sato, S. (1978). A Statistical Theory for the

Fracture of Brittle Solids Under Multiaxial
Stresses. Int. J. Fracture, Vol. 14, pp. Rl3-Rl6.
Baczynski, N.R. (1971). Frequency of Bedding Plane Breaks in

the Hanging Wall of the Racecourse Orebodies,
11-13 Levels. Mount Isa Mines Limited,
Merrorandum NRB/2.6/RES MIN 1.0.8, dated 2nd November
1971 (Unpublished).
Barton, C.M. (1974). Simplified Procedures for the Vector

Summation and Statistical Analysis of Spherically
Distributed Point Clusters. Aust. CSIRO Div. Appl.
Geomech., Technical Report No. 20, 24p.
Barton, N. R. ( l976a) Unsupported Underground Openings. Rock

Mechanics Discussion Meeting, BeFo, Stockholm,
lOth February 1976, pp.61-94.
Barton, N. (1976b). Recent Experiences with the Q-system of

Tunnel Support Design. Proc. Syrrg;>. Exploration for
Rock Engineering, Johannesburg, Nov. 1-5, Vol. 1,
pp.l07-ll7.
223.
Bennett, E.M. (1965). Lead-Zinc-Silver and Copper Deposit of

Mount Isa. in Geology of Australian Ore Deposits,
edited by J. McAndrew, Australasian Institute of
Mining & Metallurgy, pp.233-246.
Bielenstein, H.U. and Eisbacher, G.H. (1969). Tectonic

Interpretation of Elastic-Strain-Recovery
Measurements at Elliot Lake, Ontario. Canada,
Department of Energy, Mines & Resources, Mines Branch,
Ottawa, Research Report No. R210, 64p.
Bieniawski, Z.T. (1967). Mechanism of Brittle Fracture of Rock.

S. Afr. CSIR, Report No. MEG 580, 226p.
Bieniawski, Z.T. (1974a). Geomechanics Classification of Rock

Masses and Its Application in Tunnelling. Proc. Third
Int. Rock Mech. Congr., Denver, Sept. 1-7, Vol. II-A,
pp. 27-32.
Bieniawski, Z.T. (1974b). Discussion: Engineering Classification

of Jointed Rock Masses. Trans. s. Afr. Instn. Civil
Engrs., Vol. 16, pp.239-254.
Brady, B. H. G.; Friday, R.G. and Alexander, L.G. (1976). Stress

Measurement in a Bored Raise at the Mount Isa Mine.ISRM
Synp. Investigation of Stress in Rock: Advances in
Stress Measurement, Sydney, Aug. 11-13, pp. 12-16.
Brekke, T.L. and Howard, T.R. (1972). Stability Problems Caused

by Seams and Faults. Proc. First Nth. Arner. Rapid
Excavation and Tunnelling Conf., Chicago, June 5-7,
Vol. 1, pp. 25-41.
Bridges, M.C. (1974). Graphical Statistical Analysis of

Orientation Data. Report to Department of Geology &
Mineralogy, University of Adelaide, 29p.
(Unpublished)
224.
Brito,s. (1974). Correlations Between Rock Quality and

Construction Characteristics of Four Tunnels. Proc.
Second Int. Congr. Engng. Geol. IAEG, Sao Paulo,
Aug. 18-24, Vol. 2, paper VII-14, pp. l-11.
Broadbent, C.D. and Rippere, K.H. (1970). Fracture Studies at the

Kimbley Pit. Proc. Symp. Theoretical Background to
the Planning of Open Pit Mines with Special Reference
to Slope Stability, Johannesburg, Aug. 29 - Sept. 4,
pp. 171-179.
Broch, E. and Franklin, J.A. (1972). The Point-load Strength

Test. Int. J. Rock Mech. Min. Sci., Vol. 9,
pp. 669-697.
Broili, L. (1969). Fundamentals of Rock Mechanics: Course and

Lectures. Int. Centre for Mech. Sciences, No.8,
pp. 29-43.
Burt, N.J. and Dougill, J.W. (1977). Progressive Failure in a

Model Heterogeneous Medium. J. Engng. Mech. Div.
Am. Soc. Civ~ Engrs., Vol. 103, pp. 365-376.
Cecil III, O.S. (1970). Correlations of Rockbolt-Shotcrete

Support and Rock Quality Parameters in Scandinavian
Tunnels. Ph.D Thesis, University of Illinois,
Urbana, 414p. (Unpublished) .
Cecil, o.s. (1975). Correlations of Rockbolt-Shotcrete Support

and Rock Quality Parameters in Scandinavian Tunnels.
Proc. Swedish Geotech. Institute, No. 27, 275p.
Chang, c.s. and Soong, T.T. (1979). A Probabilistic Approach to

Consolidation Analysis. Proc. Int. Conf. Applications
of Statistics and Probability in Soil and Structural
Engineering, Sydney, Jan. 29 - Feb. 2, Vol. 2,
pp. 487-496.
225.
Chang, C.Y. and Nair, K. (1972). A Theoretical Method of

Evaluating Stability of Openings in Rock.
Woodward-Lundgren & Associates, Final Report, dated
12th April 1972, (Advanced Research Projects Agency
of the Department of Defence, U.S. Bur. Mines
Contract No. H 0210046), 144p.
Chang, C.Y. and Nair, K. (1973). Development and Applications of

Theoretical Methods for Evaluating Stability of
Openings in Rock. Woodward-Lundgren & Associates,
Final Report submitted to U.S. Bur. Mines under
Contract No. H 0220038, l3lp.
Coates, D.F. (1964). Classification of Rock for Rock Mechanics.

Int. J. Rock Mech. Min. Sci., Vol. l, pp.421-429.
Cook Jr., N.E.; Ko, H.Y. and Gerstle, K.H. (1978). Variability
and Anisotropy of Mechanical Properties of the
Pittsburgh Coal Seam. Rock Mech., Vol. 11, pp. 3-18.
Cording, E.J. and Mahar, J.W. (1978). Index Properties and

Observations for Design of Chambers in Rock.
Engng. Geo1., Vol. 12, pp. 113-142.
Deere, D.U. (1963). Technical Description of Rock Cores for

Engineering Purposes. Rock Mech. Engng. Geol.,
Vol. 1, pp.16-22.
Deere, D.U. and Miller, R.P. (1966). Engineering Classification

and Index Properties for Intact Rock. U.S. Air Force
Weapons Laboratory, New Mexico, Technical Report
No. AFNL-TR-65-116.
Deere, D.U. (1968). Geological Considerations. in Rock Mechanics

in Engineering Practice, edited by K. Stagg and
o.c. Zienkiewicz, John Wiley & Sons, Inc., Chapter 1,
pp. 1-20.
._-
226.
Edwards, D.B. (1971). The Influence of Ground Control on Mining

Operations in the Racecourse Lead Orebodies.
Mount Isa Mines Limited, Technical Report No.
RES MIN 26, (Unpublished).
Einstein, H.H. and Hirschfeld, R.C. (1973). Model studies on

Mechanics of Jointed Rock. J. Soil Mech. Div. Am.
Soc. Civ. Engrs., Vol. 99, pp. 229-248.
Ewing, R.D. and Raney, E.M. (1976). Users Guide for a Computer
Program for Analytical Modelling of Rock I Structure
Interaction. Agbabian Associates, Report No.
U-7638-3-4183, (U.S. Bur. Mines Contract No.
HO 262020), 295p.
Fitzgerald, D.J. (1976). ROCSAP: Two-dimensional and Three-

dimensional Finite Element Stress Analysis for
Geornechanical Problems. Research Report to New Broken
Hill Consolidated Limited, Department of Mining
Engineeri?g, University of Melbourne, dated
January 1976, 34p.
Franklin, J.A.; Broch, E. and Walton, A. (1970). Logging the

Mechanical Character of Rock. I~erial College Rock
Mechanics Research Report, No. Dl4, dated July 1970,
pp. 1-11.
Geological Society Engineering Group Working Party (1970}. The

Logging of Rock Cores for Engineering Purposes.
Q. Jl. Engng. Geol., Vol. 3, pp. 1-24.
Glynn, E.F.; Veneziano, D. and Einstein, H.H. (1978). The

Probabilistic Model for Shearing Resistance of Jointed
Rock. Proc. 19th U.S. Sy~. Rock Mech., Stateline,
May 1-3, pp. 66-76.
227.
Granstrom, S.A. (1969). Special Discussion: Paper by K.S. Bawa on

"Design and Instrl.llnentation of an Underground Strata
for Washington Metro System". Proc. Int. Synp. Large
Permanent Underground gpenings, Oslo, Sept. 23-25,
p. 117.
Grossmann, N. (1966). Discussion of Theme 1. Proc. First Int.

Rock Mech. Congr., Lisbon, Sept. 25- Oct. 1, Vol. 3,
pp. 232-234.
Haimson, B.C.; Doe, T.W. and Fuh, G.F. (1977). Geotechnical

Investigation and Design of Annular Tunnels for
Energy Storage. Rockstore 77: Proc. First Int. Synp.
Storage in Excavated Rock Caverns, Stockholm,
Sept. 5-8, Vol. 2, pp. 275-282.
Harr, M.E. (1979). Probability of Failure in Geotechnical

Engineering. Proc. Int. Conf. Applications of
Statistics and Probability in Soil and Structural
Engng., Sydney, Jan. 29- Feb. 2, Vol. 2, pp. 680-685.
Hawkes, I. and Mellor, M. (1970). Uniaxial Testing in Rock

Mechanics Laboratories. Engng. Geol., Vol. 4,
pp. 177-285.
Herget, G. (1968). Investigation of Fracture and Weakness Planes,

Excluding Transverse Faults, and Their Effects on the
Strength of the Rock in-situ at Mount Isa. Mount Isa
Mines Lirni ted, Technical Report No. RES MIN 10,
(Unpublished)
Hodgson, R.A. (1961). Classification of Structures on Joint

Surfaces. Arner. J. Sci., Vol. 259, pp. 493-502.
Heg~ K. and Murarka, R.P. (1974). Probabilistic Analysis and

Design of a Retaining Wall. J. Geotech. Div. Am.
Soc. Civ. Engrs., Vol. 100, pp. 349-366.
228.
Hoek, E. (1965). Rock Fracture Under Static Stress Conditions.

S. Afr. CSIR, Report No. MEG 383, 228p.
Hofmann, G.W. (1973). Engineering Geological Techniques Used in

Damsite Investigations by the Geological Survey of
Queensland, Geol. Surv. Qld., Report No. 76, 19p.
Holland, C.T. (1973). Pillar Design for Permanent and Semi-

permanent Support of the Overburden in Coal Mines.
Proc. Ninth Canad. Rock Mech. Sy:rrp. , Montreal,
December, pp. 187-209.
Hoskins, E. (1967). Field and Laboratory Experiments in Rock

Mechanics. Ph.D Thesis, Australian National
University (Unpublished).
Houghton, D.A. (1976) The Role of Rock Quality Indices in the

Assessment of Rock Masses. Proc. Syrnp. Exploration
for Rock Engng., Johannesburg, Nov. 1-5, pp. 129-135.
Hustrulid, W.A. (1976). A Review of Coal Pillar Strength

Formulas. Rock Mech., Vol. 8, pp. 115-145.
International Society for Rock Mechanics (1972). Suggested

Methods for Determining the Uniaxial Compressive
Strength of Rock Materials and the Point-Load
Strength Index. ISRM Committee on Standardization
of Laboratorx Tests, Document No. 1, dated October, 12p.
Jennings, J .E.; Brink, A.B.A. and Williams, A.A.B. (1973). Revised

Guide to Soil Profiling. Trans. S. Afr. Instn. Civil
Engrs., Vol. 15, pp. 3-12.
John, K.W. (1969). Civil Engineering Approach to Evaluate Strength

and Deformability of Regularly Jointed Rock. Proc.
Eleventh Rock Mech. Synp., University of California,
pp. 69-80.
.229.
John, M. (1971). Properties and Classification of Rock with

Reference to Tunnelling. S. Afr. CSIR, Report
No. MEG 1020.
Kim, H.S. and Major, G. (1978). Application of Monte Carlo

Techniques to Slope Stability Analyses. Pnoc. 19th
u.s. Syrrp. Rock Mech., Stateline, May 1-3, pp. 28-39.
Kraft Jr., L.M. and Mukhopadhyay, J. (1977). P robabi li.s tic

Analysis of Excavated Earth Slopes. Proc. Ninth Int.
Conf. Soil Mech. Foundation Engng., Tokyo, July,
Vol. 2, pp. 109-116.
Kuznetsov, A.P. (1973). Attaining Statistical Reliability in

Models of Random Properties in a Continuous Medium.
Sov. Mining Sci., Vol. 9, pp. 573-575.
Laubscher, D.H. (1977). Geomechanics Classification of Jointed

Rock Masses - Mining Applications. Trans. Instn.
Min. Metall., Section A, Vol. 861 pp. 1-8.
Lohnes, P.R. and Cooley, W.W. (1968). Introduction to Statistical

Procedures: With Computer Exercises. John Wiley &
Sons, Inc., 280p.
Lundborg, N. (1967). The Strength-Size Relationship of Granite.

Int. J. Rock Mech. Min. Sci., Vol. 4, pp. 269-272.
Mahtab, M.A.; Bolstad, D.D. and Kendorski, F.S. (1973). Analysis

of the Geometry of Fractures in San Manuel Copper
Mine, Arizona. U.S. Bur. Mines, Report of Investiga-
tions No. RI 7715, 24p.
Maier, G. and Hueckel, T. (1979). Nonassociated and Coupled Flow

Rules of Elastoplasticity for Rock-like Materials.
Int. J. Rock Mech. Min. Sci. & Geomech. Abstr.,
Vol. 16, pp. 77-92.
230.
Marek, J.M. and Savely, J.P. (1978). Probabilistic Analysis of

the Plane Shear Failure Mode. Proc. 19th U.S. Syrnp.
Rock Mech., Stateline, May 1-3, pp.40-44.
Matsuo, M. and Ueno, M. (1979). Prediction of Slope Slide by

Probability of Failure. Proc. Int. Conf. Applications
Engng., Sydney, Jan. 29- Feb. 2, Vol. 2, pp. 449-
458.
Merritt, A.H. (1968). Engineering Classification of In-situ Rock.

Ph.D Thesis, University of Illinois, Urbana
(Unpublished)
Merritt, A.H. (1972). Geological Predictions for Underground

Excavations. Proc. First Nth. Amer. Rapid Excavation
and Tunnelling Conf., Chicago, June 5-7, pp. 115-132.
Obert, L. and Duvall, W.I. (1967). Rock Mechanics and the Design
of Structures in Rock. John Wiley & Sons, Inc., 650p.
Payne, J .G. (1976). An Assessment of the Use of Rock Classifi-

cation Systems in Practice. M.Sc. Thesis, Imperial
College, London, 112p. (Unpublished).
Peck, R.B.; Deere, D.U.; Monsees, J.E.; Parker, H.W. and Schmidt, B.
(1969). Some Design Considerations in the Selection of
Underground Support Systems. Department of Civil
Engineering, University of Illinois, Urbana, Report
to U.S. Dept. Transportation under Contract No.
3-0152, l08p.
Fells, P.J.N. (1975). Discussion of Paper " Engineering

Classification of Rock Masses for the Design of
Tunnel Support" by N. Barton, R. Lien and J. Lunde.
Rock Mech., Vol. 7, pp.246-248.
-
231.
Perkins, w.c. (1972). Results of Confirmatory Drilling:

Racecourse Orebodies, 13 Level to 15 Level.
Mount Isa Mines Limited, Technical Report No. GE0-112
(Unpublished)
Piteau, D.R. and Martin, D.C. (1977). Slope Stability Analysis

and Design Based on Probability Techniques at Cassiar
Mine. Canad. Inst. Min. Metal!. Trans., Vol. 80,
pp. 51-62.
Protodyakonov, M.M. (1964). Methods of Evaluating the Cracked

State and Stre_ngth of Rocks in-situ. Proc. Fourth
Int. Conf. Strata Control & Rock Mech., Columbia
University, May 4-8, Addendum.
Protodyakonov, M.M. and Koifman, M.I. (1964). The Scale Effect

in Investigations of Rock and Coal. Proc. Fifth Conf.
Int. Bur. Rock Mech., Deutsche Akademie der
Wissenschaften, Berlin, No.3, pp. 97-108.
Rechitskii, V.I. and Gaziev, E.G. (1978). Methods of

Probabilistic Estimation of the Stability of Rock
Slopes. Hydrotechnical Construction, No. 3, pp. 277-
282.
Rechnitzer, G. (1978). Linear and Non-linear Finite Element

Analysis of Underground Structures: Including Case
Studies of Flagstaff Station and Running Tunnels,
Melbourne Underground Rail Loop. M.Eng.Sc. Thesis,
University of Melbourne (Unpublished).
Roberts, A. (1977). Engineering in Earth Materials. in Geotechnology:

An Introductory Text for Students and Engineers,
Chapter 1, Pergamon Press, pp. 1-16.
Salamon, M.D.G. (1967). A Method of Designing Bord and Pillar

Workings. s. Afr. Inst. Min. Metal!., Vol. 68,
pp. 68-78.
232.
Sheinin, V.I. (1979). Computation and Analysis of Probabilistic

Characteristics of Stresses Near Underground Opening
in Stochastically Inhomogeneous Rock Mass. Proc.
Third Int. Conf. Numerical Methods in Georrech.,
Aachen, Apr. 2-6, Vol. 2, pp. 663-669.
Sivandran, c.; Chiev, K. and Balasubramanian, A.S. (1979).

Application of Probability Theory to the Finite
Element Method in Predicting Settlements in Soft
Bangkok Clay. Proc. Third Int. Conf. Numerical
Methods in Geornech., Aachen, Apr. 2-6, Vol. 3,
pp. 1025-1032.
Sorensen, W.K. and Pariseau, W.G. (1978). Statistical Analysis

of Laboratory Coll1Pressive Strength and Young's
Modulus Data for the Design of Production Pillars
in Coal Mines. Proc. 19th u.s. Symp. Rock Mech.,
Stateline, May 1-3, pp. 30-37.
Stapledon, D.H. (1968). Discussion of D.F. Coates paper on

"Classification of Rock Substances". Int. J. Rock
Mech. Min. Sci., Vol. 5, pp. 371-373.
Stoyan, D.; Forster, W. and Weber, E. (1979). On the Probability

of Failure of Slopes. Proc. Int. Conf. Applications
Engng., Sydney, Jan. 29- Feb. 2, Vol. 2, pp. 459-465.
Torrent, R.J. (1978). The Log-normal Distribution: A Better

Fitness for the Results of Mechanical Testing of
Materials. Materials and Structures, Vol. 64,
pp. 235-245.
Vanmar.cke, E.H. (1977). Probabilistic Modelling of Soil Profiles.

J. Geotech. Div. Am. Soc. Civ. Engrs., Vol. 103,
pp. 1227-1246.
233.
Voight, B. (1968). On the Functional Classification of Rocks for

Engineering Purposes. Proc. Int. Symp. Rock Mech.,
Madrid, pp. 131-135.
Weibull, W.A. (1939). A Statistical Theory of the Strength of

Materials. Proc. R. Swedish Acad. Engng. Sci.,
Vol. 151.
Wickham, G.E.; Tiedemann, H.R. and Skinner, E.H. (1972). Support

Determinations Based on Geological Predictions.
Proc. First Nth. Aruer. Rapid Excavation and Tunnelling
Conf., Chicago, June 5-7, Vol. 1, pp. 43-64.
Wickham, G.; Tiedemann, H. and Skinner, E. (1974). Ground Support

Prediction Model RSR Concept. Proc. Second Nth. Amer.
Rapid Excavation and Tunnelling Conf., San Francisco,
June 24-27, Vol. l, pp. 691-707.
Wilson, A and Ashwin, D (1972) . Research Into the Determination

of Pillar Size. Min. Engr., No. 141, pp. 409-429.
Worotnicki, G. and Denham, D. (1976). The State of Stress in the

Upper Parts of the Earth's Crust in Australia
According to Measurements in Mines and Tunnels and
from Seismic Observations. ISRM Symp. Investigation
of Stress in Rock: Advances in Stress Measurement,
Sydney, Aug. 11-13, pp. 71-82.
Al
APPENDIX I ROCK MASS CLASSIFICATION INPUT PARAMETERS
Al.l Bieniawski's RMR Method
(i) Unconfined Compressive Strength of Intact Cores: 185 MPa

(ii) Rock Quality Designation (RQD): determined on basis of
total fracture intensity in accordance with relationship
presented in Figure 32 of the thesis.
(iii) Spacing of Joints or other Discontinuities: determined on
basis of statistical model presented in Figure 27. A weighted
rating was determined on basis of the relative frequency
(i.e., intensity) of each fracture set.
(iv) Condition of Joints:
(a) Bedding Plane Partings: relative frequency of
graphite coated surfaces determined on basis of
statistical model presented in Paragraph 3.3.4 of
the thesis. These surfaces are classed as
"slikensided surfaces" with a rating of 6. Non-graphitic
surfaces were classed as "slightly rough surfaces,
separation less than 1 nun, hard joint wall rock" with
a rating of 20.
(b) Other Fractures: same classification as adopted for
non-graphitic bedding plane partings.
A weighted average was determined on basis of the
relative frequency of each discontinuity type.
A2
(v) Orientation of Joints: each set of discontinuities as

defined in Table 1 on page 33 of the thesis were classed
individually as indicated below:
Set No. Class
1 Very Favourable
2 Fair
3 Very Unfavourable
4 Very Unfavourable
5 Very Unfavourable
6 Unfavourable
7 Very Unfavourable
8 Unfavourable
9 Fair
10 Fair
11 Very Unfavourable
A weighted average rating was determined on the basis of

relative frequency of each discontinuity set.
(vi) Groundwater: it was assumed that water was essentially
absent, i.e., rock mass was "completely dry".
Al.2 Laubscher and Taylor's Modified RMR Method
(i) Unconfined Compressive Strength of Intact Cores: as in

Paragraph Al.l above.
(ii) Rock Quality Designation (RQD): as in Paragraph Al.l above.
(iii) Spacing of Joints or Discontinuities: as in Paragraph Al.l
above.
(iv) Condition of Joints:
(a) Bedding Plane Partings: frequency of graphite coated
surfaces as in Paragraph Al.l above. Graphitic surfaces
A3
were classed as "(A) straight, (B) polished, (D) fine

soft-sheared", whereas non-graphitic surfaces were
assumed as "(A) straight, (B) smooth".
(b) All Other Fractures: same classification as adopted
for non-graphitic bedding plane partings.
A weighted average rating was determined on the basis
of the relative frequency of each discontinuity type.
(v) Groundwater: as in Paragraph Al.l above.
(vi) Adjustments:
(a) Weathering: It was assumed that RQD and unconfined
compressive strength were not affected (i.e., adjustment
ratio of 1.0), whereas condition of joints was slightly
affected (i.e., adjustment ratio of 0.95).
(b) Field and Induced Stresses: affect of stress is assumed
to be dependent on the orientation of discontinuties
within the rock mass. The following adjustment ratios
were applied to the sets:
Set No. Adjustment Ratio
1,2 1.2
9,10,11 l.O
4,5,7 0.9
3 0.8
6,8 0.76
A weighted average was applied to the "condition of

joints" parameter on the basis of the relative frequency
of each set of discontinuities within the mass.
(c) Changes in Stress: It was not possible to evaluate the
effect of this parameter on the basis of available
information. A null adjustment factor of 1.0 was applied.
(d) Strike and Dip Orientation of Fractures: all discontinuities
apart from set No.1 were assumed to be inclined away from
vertical. The appropriate adjustment was determined with
respect to the number of fracture sets that were assigned
to the hanging-wall.
(e) Mining Technique: A classification just below "Good
Conventional Blasting" was assumed for the mine, with
A4
an adjustment ratio of 0.9. This ratio was applied

to the RQD and Condition of Joints parameters in
accordance with the maximum possible, i.e., each
of the two parameters was reduced by 0.7 and 1.26
per cent, respectively.
Al.3 Barton's Q-Method
{i) Rock Quality Designation (RQD): as in Paragraph Al.l above.

{ii) Joint Set Number: all discontinuities with an average
intensity exceeding 1.0 m2;m 3 were considered as sets,
otherwise they were assumed to be random. No other adjustments
{e.g., as for portals or tunnel intersections) were applied
to the rating.
{iii) Joint Roughness Number: the following classifications and
ratings were assumed for each fracture set:
Set No. Classification
1-10 (A) Discontinuous joints

11 (a) graphitic (G) Slikensided, planar
(b) non-graphitic {A) Discontinuous joints
A weighted rating was determined on the basis of the relative

frequency of each set.
{iv) Joint Alteration Number: all fracture sets and non-graphitic
bedding plane partings were classed as "(B) unaltered joint
walls, surface staining only", whereas graphitic bedding plane
partings were assumed as "(E) softening or low friction clay
mineral coating, i.e., kaolinite, mica, or chlorite, talc,
gypsum, graphite, etc. "
{v) Joint Water Reduction Factor: Class(A) was assumed, i.e.,
"dry excavations or minor inflow ".
AS
(vi) Stress Reduction Factor:

(a) Orebody Nos. 5 and 7: Items J and K of classification
table were assumed to apply, i.e., the value of SRF
is located in the range of about 0.75- 2.0
(b) Orebody No. 11: Items F and G of classification
table were assumed to apply, i.e., SRF is in the
range of 2.0 - 5.0
(vii) Excavation Support Ratio (ESR) :an ESR value of 5.0 is
assumed to be representative of sub-level open stapes.

Rock Mass Characterization

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Rock Mass Characterization

Загружено:

Авторское право:

Доступные форматы

Minerva Access is the Institutional Repository of The University of Melbourne

Terms and Conditions:

NORBERT RICHARD PRZEMYSLAW BACZYNSKI

a..6pa.Jl.:U.al. nuln-Ume.n:t on :the. Jte.qu-Ur.e.me.nt-6 no.l!..

The research study centres on the stabiZity assessment of

A systematic approach is adopted for characterization of

MechanicaZ properties of the mass are derived on the basis

SeveraZ techniques are employed to assess the significance

Some consideration is given to broadly reviewing past

Results of analyses demonstrate that a greater depth of

Information derived from the published or unpublished work of others

The work was basically funded from a research grant awarded by

Funding of the work was supplemented via a single award of the

Paragraph Subject Page No.

1.1 Definition of Problem and Project

CHAPTER 2 MODEL FOR INTACT ROCK COMPONENT OF ROCK MASS 18

2.1 Strength of the Intact Rock 18

Paragraph Subject Page No.

CHAPTER 3 PHYSICAL CHARACTERISTICS OF GEOLOGICAL

3.1 Source of Data 28

CHAPTER 4 MODEL FOR SPATIAL DISTRIBUTION OF FRACTURES 46

4.1 General Comments 47

Paragraph Subject Page No.

CHAPTER 5 MODEL FOR STRUCTURAL VARIABILITY OF

5.1 Model for Each Fracture Set 81

CHAPTER 6 ROCK MASS PROPERTIES 96

6.1 Method of Analysis 96

Paragraph Subject Page No.

CHAPTER 7 STABILITY ANALYSIS 115

7.1 Rock Mass Classifications 116

Paragraph Subject Page No.

CHAPTER 8 COMPARISON BETWEEN MODEL AND PROTOTYPE

8.1 Practical Difficulties 186

CHAPTER 9 SUMMARY AND DISCUSSION 199

9.1 Background Considerations 199

ADDITIONAL REFERENCES 222

APPENDIX I ROCK MASS CLASSIFICATION INPUT PARAMETERS Al-AS

Al.l Bieniawski's RMR Method Al

Table No. De scription Page No.

1 Characteristics of Fracture Sets. 33

2 Mean Intensity of Fractures per Unit

3 Relationship Between C-Factor and Intensity

4 Rock Mass Material Properties. 111

5 Average Material Types Used in Modelling. 113

6 Hanging-wall Spans (Metres) on Basis of

7 Magnitude of Pre-Mining Stress Field

8 Sensitivity Testing: Different Initial

9 Finite Element Modelling Programme Using

10 Stability of 5 Orebody Stope with Uniform

11 Stability of 5 Orebody Stope with

Table No. Description Page No.

12 Stability of 11 Orebody Stope with Uniform

13 Stability of 11 Orebody Stope with

14 Case Study: Variations in the Width of

15 Case Study: Stability of Hanging-walls &

Figure No. Description Page No.

1 Typical Cross-Section Through Northern

2 Diagrammatic Layout of a Sublevel Open

3 Formulation of a Geotechnical Model. 7

4 Conceptual Basis for Rock Mass Characterization. 9

5 Relationship Between Core-Bedding Angle and

6 Strength of 61 Specimens of Mineralized Shale

7 Principal Fracture Sets. 32

8 Cumulative Frequency Plot on Logarithmic