analytical tools for managing rock fall hazards in australian coal mine roadways

ACARP PROJECT C14029 PUBLISHED February 2009

ANALYTICAL TOOLS FOR MANAGING ROCK FALL HAZARDS IN AUSTRALIAN COAL MINE ROADWAYS

R Seedsman, N Gordon & N Aziz CIVIL, MINING AND ENVIRONMENTAL ENGINEERING, UNIVERSITY OF WOLLONGONG, SEEDSMAN GEOTECHNICS PTY LTD & GORDON GEOTECHNIQUES PTY LTD

DISCLAIMER No person, corporation or other organisation (“person”) should rely on the contents of this report and each should obtain independent advice from a qualified person with respect to the information contained in this report. Australian Coal Research Limited, its directors, servants and agents (collectively “ACR”) is not responsible for the consequences of any action taken by any person in reliance upon the information set out in this report, for the accuracy or veracity of any information contained in this report or for any error or omission in this report. ACR expressly disclaims any and all liability and responsibility to any person in respect of anything done or omitted to be done in respect of the information set out in this report, any inaccuracy in this report or the consequences of any action by any person in reliance, whether wholly or partly, upon the whole or any part of the contents of this report.

ANALYTICAL TOOLS FOR MANAGING ROCK FALL HAZARDS IN AUSTRALIAN COAL MINE ROADWAYS

ROSS SEEDSMAN1,2

NICK GORDON3

NAJ AZIZ1

1 Civil, Mining and Environmental Engineering, University of Wollongong

2 Seedsman Geotechnics Pty Ltd

3 Gordon Geotechniques Pty Ltd

UNIVERSITY OF WOLLONGONG DISCLAIMER

The information contained in this report is for general information purposes only. The University of Wollongong, its officers, employees and agents make no representations or warranties, express or implied as to the accuracy, reliability or completeness of the information contained within this report. All liability for loss or damage of any kind at all, whether indirect or consequential, against the University of Wollongong, its officers, employees and agents arising from or through the use of the information in this text is excluded. Your use of this text is acknowledgement that the information provided herein is to assist you with undertaking your own enquiries and analysis and that you should always seek independent professional advice before acting on the information contained therein.

SEEDSMAN GEOTECHNICS PTY LTD DISCLAIMER

The information contained in this report is for general information purposes only. Seedsman Geotechnics, its directors and servants make no representations or warranties express or implied as to the accuracy, reliability or completeness of the information contained therein. Seedsman Geotechnics Pty Ltd, its directors and its servants exclude all liability for loss or damage of any kind at all (including indirect or consequential loss or damage) arising from the information in this text or use of such information. You acknowledge that the information provided in this text is to assist you with undertaking your own enquiries and analysis and that you should seek independent professional advice before acting in reliance on the information contained therein.

GORDON GEOTECHNIQUES PTY LTD DISCLAIMER

The information contained in this report is for general information purposes only. Gordon Geotechniques, its directors and servants make no representations or warranties express or implied as to the accuracy, reliability or completeness of the information contained therein. Seedsman Geotechnics Pty Ltd, its directors and its servants exclude all liability for loss or damage of any kind at all (including indirect or consequential loss or damage) arising from the information in this text or use of such information. You acknowledge that the information provided in this text is to assist you with undertaking your own enquiries and analysis and that you should seek independent professional advice before acting in reliance on the information contained therein.

ABSTRACT

This report provides a reference source for the design of ground control measures in coal mine roadways using analytical methods. Analytical methods are based on identifying potential failure modes from knowledge of the geology, followed by simple analyses based on hand calculations, spreadsheets and design charts. It is intended that this report will complement other publications that provide empirical and more intensive numerical approaches.

Coal mining is conducted in layered materials – different rock types, different strengths, different layer thickness, different joint spacing. Underground coal mines openings are rectangular, and at least in Australia, the roadway axes are parallel to the dominant discontinuity sets – bedding and joints. The stress field on development has the principal axes parallel to the roadway axes, and there can be major changes in and around extraction panels.

These geometrical factors require a different approach to the empirical rock mass rating approaches used in metal mines. Fortunately, the comparatively simple geometry allows analytical methods based on identifying collapse modes associated with jointed and bedded material. This ability to analyse the relatively simple geometries allows a more anticipatory approach to ground control management including anticipation of geological conditions, prediction of collapse modes, the design of support or reinforcement system, and the monitoring of ground conditions for exceptions. The approach provides more robust outcomes and limits the demands on the observational method.

Collapse models are provided for roof and rib. The roof models recognise that different collapse modes can apply in different stress fields ‐ high, intermediate, and zero compressive stresses. The rib models draw analogies to rock slope stability and also the impact of high vertical stresses

Methods for determining support or reinforcement requirements are provided. Suspension of collapsed masses is identified as the basis for roof support in both very high and zero compressive stress regimes. Reinforcement of bedding discontinuities is advocated for intermediate compressive stresses. For the ribs, restraint of coal blocks defined by pre‐existing joints or by mining induced fractures is required.

IV

EXECUTIVE SUMMARY

The mining hazard that is addressed in this report is the gravity‐driven fall of a block of ground into a roadway. The block of ground may be defined by pre‐existing discontinuities and/or by mining induced fractures.

The purpose of this report is to provide information to assist the geotechnical officer in identifying potentially unstable roof and rib, and to document design methods that may be considered in the specification of ground support. The report is focused on the stability of the ground in the immediate vicinity of the roadway – that is the immediate roof and rib.

The report has been written for geotechnical engineers, geologists, and mining engineers charged with specifying ground support. Persons occupying the positions of mine manager or technical services manager should find the report of value. The report addresses some of the steps (site characterisation, model formulation, design, implementation, monitoring, and review) that can be found in Clause 48 of NSW regulations for coal, or the rockfall risk management of the Minerals Council of Australia (Potvin and Nedin, 2003).

It is stressed that a full understanding of coal mine ground control is not yet available – hopefully this report represents a step forward. One use of this report can be to challenge the standard interpretations of how a rock mass behaves around openings so that there can be a gradual improvement in the ability to design safe and productive mine roadways. The concepts introduced can be applied at early stages – prefeasibility and feasibility using presumed values based on a review of the geological conditions. As the project progresses to operational phase, the greater the importance to move from presumed to demonstrated values for the various input parameters.

The structure of report is as follows:

• Section 1 serves as an introduction

• Section 2 of the report discusses design in rock masses, highlighting the uncertainties that are intrinsic to design in rock masses which are very difficult to fully characterise. The key message is that any design in rock mechanics will need to be implemented with a commitment to monitor and respond as the geological conditions are revealed.

• Section 3 defines rock falls and makes a distinction between the nature of the hazards in metal mines and coal mines

• Section 4 summarises underground coal mining practices in Australia so that the practical constraints on roadway geometries and excavation methods can be appreciated.

• Section 5 provides a reference guide to the engineering geology of coal measures and includes discussions on discontinuities, rock strength and deformation properties, and the insitu stresses.

• Section 6 examines the redistribution of stresses around a retreat longwall panel and particularly in the immediate vicinity of a roadway. The key point is that our knowledge of stress above chain pillars needs to be interrogated in more detail to understand the stresses that are present at the roof line. Simple numerical analyses for stress about openings in a homogeneous material are presented

• Section 7 presents the key strength properties of the popular support and reinforcement elements in use in Australia and presents models for their reinforcing action.

• Section 8 provides a flow chart for the specification of roof support and a number of analytical tools that can be used to determine bolting densities and bolt lengths.

V

• Section 9 provides a flow chart for rib support design.

Geotechnical engineering requires consideration of the inevitable uncertainties of the properties of rock masses. It is not possible to know all of the properties of a rock mass: conversely it is evident from the range of successful rock engineering ventures that this is also not necessary. It follows that there is a need to identify the key parameters in the rock mass and the ways of dealing with them. Thus geotechnical engineering requires both geological and engineering input.

In an analytical engineering approach, a key step is identifying mode of collapse. For this geotechnical models of behaviour are created. Characteristics of such models are that they capture the key features (efficient), they are amenable to mathematics (practical), and they anticipate conditions to be encountered (predictive). An important aspect of the approach is to identify failure and collapse modes based on observation and only then to seek to analyse the mechanics involved. The identification of failure and collapse modes allows the application of limit equilibrium techniques and the use of factors of safety to assess the level of stability.

The starting model is jointed layered roof with layers parallel to the roadway. This is much simpler than the metaliferous mining problem which involves arched roofs and rock wedges. Depending on the stress condition, this system may fail under tension, compression, or self weight. During roadway development, the immediate stone roof is exposed to high levels of deviatoric stress, which quickly dissipates as the roof deflects. For coal roof, the immediate roof stresses are tensile. At the maingate corner, the deflected “softened” roof means that the increased stresses around the longwall goaf are deflected higher into the roof. At the tailgate, the vertical stresses become dominant and there is relief of horizontal stresses towards the goaf.

A logical framework to assess stability and specify support is provided by way of flow charts for each of the roof and the ribs. The flow charts need to be applied at each stage of mining. Commentaries on pragmatic aspects of the support design are offered for guidance.

VI

TABLE OF CONTENTS

ABSTRACT ........................................................................................................................................................... iii EXECUTIVE SUMMARY ....................................................................................................................................... iv TABLE OF CONTENTS ......................................................................................................................................... vi 1 INTRODUCTION ........................................................................................................................................... 1

1.1 The challenge ................................................................................................. 2 1.2 Target readers ............................................................................................... 2 1.3 Structure of the report .................................................................................. 3 1.4 Limitatons AND OPPORTUNITIES FOR IMPROVEMENTS ............................... 3

2 DESIGN IN ROCK MECHANICS ..................................................................................................................... 5 2.1 Design Process ............................................................................................... 6 2.2 Rock as an engineering material .................................................................... 9 2.3 Analysis tools ................................................................................................. 9

2.3.1 Precedent and practice ........................................................................... 10 2.3.2 Empirical .................................................................................................. 10 2.3.3 Analytical ................................................................................................. 11 2.3.4 Numerical ................................................................................................ 13

2.4 Uncertainty, risk, and the observational method ........................................ 14 2.4.1 Observational Method ............................................................................ 15 2.4.2 Other risk tools ........................................................................................ 16

3 ROCK FALLS ................................................................................................................................................17 3.1 Difference between coal and metal mines .................................................. 17 3.2 Roof collapse ............................................................................................... 18 3.3 Rib collapse .................................................................................................. 22 3.4 Approaches to prevent rock falls ................................................................. 23

4 UNDERGROUND COAL MINING IN AUSTRALIA .......................................................................................25 4.1 Coal seams ................................................................................................... 25 4.2 Mining methods ........................................................................................... 25

4.2.1 Development ........................................................................................... 25 4.2.1.1 Methods .......................................................................................... 25 4.2.1.2 Support Rule Database ................................................................... 27

4.2.2 Extraction ................................................................................................ 28 4.2.2.1 Longwalls ........................................................................................ 28 4.2.2.2 Secondary support .......................................................................... 28

5 ENGINEERING GEOLOGY ...........................................................................................................................31 5.1 Discontinuities ............................................................................................. 31

5.1.1 Description of discontinuities .................................................................. 32 5.1.2 Bedding partings ..................................................................................... 34 5.1.3 Joints in stone .......................................................................................... 35 5.1.4 Joints in coal ‐ cleats ................................................................................ 37 5.1.5 Faults ....................................................................................................... 40

VII

5.2 Lithologies .................................................................................................... 41 5.3 Strength ....................................................................................................... 44

5.3.1 Uniaxial compressive strength ................................................................ 44 5.3.1.1 Stone ............................................................................................... 44 5.3.1.2 Coal ................................................................................................. 46

5.3.2 Tensile strength ....................................................................................... 47 5.3.3 Triaxial parameters .................................................................................. 48 5.3.4 Brittle strength ........................................................................................ 49 5.3.5 Shear strength of discontinuities ............................................................ 50 5.3.6 Mass strength .......................................................................................... 51

5.3.6.1 Stone ............................................................................................... 51 5.3.6.2 Coal ................................................................................................. 52

5.4 Deformation properties ............................................................................... 52 5.4.1 Modulus .................................................................................................. 52 5.4.2 Poisson’s ratio ......................................................................................... 54

5.5 In‐situ Stresses ............................................................................................. 54 5.5.1 Stress in stone ......................................................................................... 54 5.5.2 Stress in coal ............................................................................................ 55 5.5.3 Faulted ground ........................................................................................ 57 5.5.4 Topography ............................................................................................. 58

5.6 Local terminology ........................................................................................ 59 6 MINING INDUCED STRESS CHANGES .......................................................................................................60

6.1 Redistibution about a longwall .................................................................... 60 6.1.1 Maingate corner ...................................................................................... 60 6.1.2 Bleeder/Tailgate ...................................................................................... 62 6.1.3 Tailgate corner ........................................................................................ 62

6.2 Redistribution about a roadway .................................................................. 63 6.2.1 Basic concepts ......................................................................................... 63 6.2.2 Elastic stress redistribution around a rectangular roadway .................... 65 6.2.3 Three‐Dimensional stress changes .......................................................... 70 6.2.4 Non linear stress redistributions ............................................................. 73 6.2.5 Stresses induced within a blocky roof ..................................................... 75

6.2.5.1 Voussoir beams ............................................................................... 75 6.2.5.2 Cantilevers ...................................................................................... 76

6.3 OTHER STRESS REDISTRIBUTIONS ............................................................... 77 6.3.1 Tailgate pillars ......................................................................................... 77 6.3.2 Very high stresses under or above pillars. .............................................. 77

6.4 Compiled models for stress paths in the immediate roof ........................... 78 6.4.1 Stone roof – single seam ......................................................................... 78 6.4.2 Coal roof – single seam ........................................................................... 80 6.4.3 Stress paths in the ribs ............................................................................ 81

7 SUPPORT AND REINFORCEMENT TECHNOLOGIES ..................................................................................82 7.1 Reinforcement action .................................................................................. 82

VIII

7.1.1 Dowel effect ............................................................................................ 83 7.1.2 Friction effect .......................................................................................... 83

7.2 Tendons ....................................................................................................... 85 7.3 Anchorages .................................................................................................. 86 7.4 Straps and panels – skin restraint ................................................................ 87

8 PREVENTION OF ROOF COLLAPSE ............................................................................................................89 8.1 Non‐vertical joints ....................................................................................... 91

8.1.1 Collapse model ........................................................................................ 92 8.1.2 Support design ........................................................................................ 93 8.1.3 Commentary on design to prevent collapse with non‐vertical joints ..... 93

8.2 Compressive failure ..................................................................................... 93 8.2.1 Collapse model ........................................................................................ 94 8.2.2 Support design ........................................................................................ 98 8.2.3 Commentary on design to prevent compressive collapse .................... 101

8.3 Tensile failure ............................................................................................ 103 8.3.1 Collapse mode ....................................................................................... 104 8.3.2 Support .................................................................................................. 105 8.3.3 Commentary on design to prevent tensile collapse .............................. 105

8.4 Delamination failure .................................................................................. 106 8.4.1 Collapse model ...................................................................................... 107 8.4.2 Reinforcement design ........................................................................... 108

8.4.2.1 Driving forces ................................................................................ 109 8.4.2.2 Resisting forces ................................................................................. 1

8.4.3 Densities and patterns .......................................................................... 111 8.4.4 Commentary on design to prevent delamination ................................. 113

9 PREVENTATION OF RIB COLLAPSE ......................................................................................................... 115 9.1 Structure control ....................................................................................... 116

9.1.1 Slides ..................................................................................................... 117 9.1.2 Wedges .................................................................................................. 118 9.1.3 Topples .................................................................................................. 118

9.2 Stress induced rib collapse ........................................................................ 119 9.2.1 Mining induced fractures ...................................................................... 119 9.2.2 Buckling ................................................................................................. 120

9.3 Commentary on rib support design ........................................................... 121 10 REFERENCES ........................................................................................................................................... 123

List of Tables

Table 1 Classification of prediction (after Lambe, 1973) .......................................................................... 5 Table 2 Prediction classes (after Morganstern, 2000) .............................................................................. 5 Table 3 Commentary on recent implementation of SU constitutive models in FLAC ............................ 14 Table 4 Observations and data requirements for roof collapse modes ................................................. 21

IX

Table 5 Australian Longwall statistics (after Cram, 2008) ...................................................................... 28 Table 6 Classification of discontinuity spacing ....................................................................................... 32 Table 7 Classification of discontinuity persistence ................................................................................. 32 Table 8 Roughness of bedding partings .................................................................................................. 35 Table 9 Joint roughness coefficient (JRC) of coal joints .......................................................................... 40 Table 10 Visual and tactile tests (ISRM,1978) ........................................................................................ 46 Table 11 Range of coal unconfined compressive strengths (ACIRL, 1986) ............................................. 47 Table 12 Presumed friction angles and m values ................................................................................... 49 Table 13 Hoek‐Brown parameters for various coals .............................................................................. 52 Table 14 Guidelines for the selection of modulus ratio (after Hoek and Diederichs, 2006) .................. 52 Table 15 Compilation of stress measurements taken from under excavations ..................................... 56 Table 16 Local geological terminology ................................................................................................... 59 Table 17 Roof and side stresses for circles and ellipses ......................................................................... 64 Table 18 Summary of stresses for rectangular roadway (σ1 = 10 MPa) ................................................. 65 Table 19 Summary of Australian bolt types ........................................................................................... 85 Table 20 Summary of tensile strength of USA bolt types ....................................................................... 85 Table 21 Long tendons ........................................................................................................................... 85 Table 22 Data base for brittle failure analyses ....................................................................................... 94 Table 23 Bolt locations to resist BPXS .................................................................................................. 112

List of Figures

Figure 1 Steps in geotechnical design process covered in this report ...................................................... 8 Figure 2 Accuracy of a prediction is a function of the amount of data available and the method used

(after Lambe, 1973). ....................................................................................................................... 10 Figure 3 Logical framework for mine excavation design in massive rock (Brady and Brown, 1985)...... 12 Figure 4 Geometry of a coal mine roadway and its coaxial relationship with discontinuities and

stresses. .......................................................................................................................................... 18 Figure 5 Typical geometry and discontinuity field for a metal mine roadway (Brady and Brown, 1985)

........................................................................................................................................................ 18 Figure 6 General collapse modes for a bedded roof with sub‐vertical joints ......................................... 19 Figure 7 Dipping joints can allow block movement with or without horizontal roof stresses ............... 20 Figure 8 Collapse following delamination .............................................................................................. 20 Figure 9 Collapse related to release along vertical joints ....................................................................... 20 Figure 10 Thin slabs between bolts retained with mesh panels ............................................................ 21 Figure 11 Rib collapse analogies to rock slopes (topples, planar slides, wedges) .................................. 22 Figure 12 Buckling of a coal rib ............................................................................................................... 22 Figure 13 Back surface to a slab formed by a mining induced fracture in a coal rib .............................. 23 Figure 14 Flow chart for the analysis phases for the prevention of collapse of ground (Potvin and

Nedin, 2003) ................................................................................................................................... 24 Figure 15 Typical bolter miner (Source – Joy) ........................................................................................ 26 Figure 16 Typical place change miner and mobile bolter (Source – Joy) ............................................... 26 Figure 17 Summary of Installed Capacity of Primary Support for Gateroads......................................... 27

X

Figure 18 Summary of Installed Capacity of Primary Support for Intersections .................................... 27 Figure 19 Summary of combined minimum maingate primary and secondary support ........................ 29 Figure 20 Summary of combined maximum maingate primary and secondary support ....................... 29 Figure 21 Summary of minimum maingate secondary support ............................................................. 30 Figure 22 Summary of maximum maingate secondary support ............................................................ 30 Figure 23 Difference between true cohesion of an intact sample and apparent cohesion of a rough

discontinuity ................................................................................................................................... 32 Figure 24 Joint Roughness Profiles and Corresponding JRC Values (Barton and Choubey, 1977) ......... 33 Figure 25 Alternative method for estimating JRC from measurements of surface roughness amplitude

from a straight edge (Barton and Choubey, 1977) ........................................................................ 34 Figure 26 Example of joint patterns in sedimentary strata. ................................................................... 35 Figure 27 Joints in undeformed sediments (Fookes et al, 2000) ............................................................ 36 Figure 28 Relationship between joint spacing and bedding thickness (Ji et al, 1998) ........................... 37 Figure 29 Association between joints and faults/folds (Fookes et al 2000) ........................................... 38 Figure 30 Jointed and cleated coal – note very low persistence of cleat (Medhurst and Brown, 1998) 39 Figure 31 An example of rotation of coal joints about thrust faults ...................................................... 39 Figure 32 Cleat and joint orientations in Bowen Basin (Pattison, 1995) ................................................ 40 Figure 33 Distribution of lithologies in fluviatile system (Fookes et al, 2000) ........................................ 42 Figure 34 Distribution of lithologies in a deltaic system (Fookes et al, 2000) ........................................ 43 Figure 35 Plot showing the range in strength and deformation modulus for different lithologies

(Shepherd and Gale, 1982) ............................................................................................................. 44 Figure 36 A variety of UCS sonic relationships for stone ........................................................................ 45 Figure 37 UCS sonic relationships for stone at lower strength .............................................................. 45 Figure 38 UCS sonic velocity relationship for coal .................................................................................. 47 Figure 39 Strength envelopes (dash – Mohr Coulomb, solid – Hoek and Brown) .................................. 48 Figure 40 Brittle rock parameters .......................................................................................................... 49 Figure 41 Strength envelope for a discontinuity .................................................................................... 50 Figure 42 Geological strength index for molasses .................................................................................. 51 Figure 43 Modulus reduction for case of damage index = 0 .................................................................. 53 Figure 44 Simple model for the impact of discontinuity spacing on the deformation modulus and

Poisson’s ratio ................................................................................................................................ 53 Figure 45 Summary of stress measurement data from New South Wales and Queensland coalfields

(Nemcik et al, 2006) ....................................................................................................................... 55 Figure 46 Measured minimum stress in coal as a function of depth (Enever et al, 2000) ..................... 56 Figure 47 Reduced lateral (horizontal) stress associated with the presence of surfaces with lower

frictional resistance based on (a) passive earth pressures, (b) UDEC modeling (Nemcik et al 2006). ........................................................................................................................................................ 57

Figure 48 Effect of topography on stress distribution ............................................................................ 58 Figure 49 Horizontal stresses in hilltop or ridge mining ......................................................................... 59 Figure 50 Pillar design vertical stresses developed above chain pillars assuming a 200m wide panel. . 61 Figure 51 General pattern of vertical and horizontal stress redistribution (Gale 2008) ....................... 61 Figure 52 Concentration of horizontal stress magnitude at the maingate corner as a function of the

angle between principal horizontal stress axis and the roadway direction (after Gale 2008) ....... 62 Figure 53 Stresses above a pillar in a tailgate (Shen et al 2006). ........................................................... 63 Figure 54 Effect of planes of weakness on distribution of roof stresses ................................................ 65

XI

Figure 55 Distributions of deviatoric stresses about a 1.86:1 roadway (σ1 = 10 MPa) .......................... 66 Figure 56 Distribution of horizontal stresses about a 1.86:1 roadway (σ1 = 10 MPa) ............................ 67 Figure 57 Distribution of vertical stresses about a 1.86:1 roadway (σ1 = 10 MPa) ................................ 68 Figure 58 Negative minimum principal stresses induced around a 1.86:1 roadway (σ1 = 10 MPa) ....... 69 Figure 59 Example of shear and normal stresses and bedding parallel excess shear stress assuming a

35o friction angle and K = 2.0 .......................................................................................................... 70 Figure 60 Contours of BPXS as a function of height into distance across the roof (expressed as

MPa/MPa) ...................................................................................................................................... 70 Figure 61 Development of deviatoric stresses at 0.2m above the roof line as the heading is advanced

(K=2) ............................................................................................................................................... 71 Figure 62 Distribution of negative mean stress in the immediate roof of a roadway for K=0.15 .......... 71 Figure 63 Three dimensional view of the distribution of BPXS (ignore negative sign) ........................... 72 Figure 64 Vertical slice through Figure 63 showing how BPXS develops (ignore negative sign) ............ 72 Figure 65 BPXS on a surface 0.2m above the roof line ........................................................................... 73 Figure 66 Stress measurements at Emerald Mine .................................................................................. 74 Figure 67 Simulation of stress redistribution above a roadway using an elastic model ........................ 74 Figure 68 Concept of a stress relieving roadway .................................................................................... 75 Figure 69 Voussoir beam deformations induce compressive stresses at the roof corners and tensile

stresses at the roadway centreline ................................................................................................ 76 Figure 70 Example of a shear surface generated by cantilevering action .............................................. 76 Figure 71 Relaxation of a roof line as a result of vertical deformation in one of the sides .................... 77 Figure 72 Stresses under a rigid footing ................................................................................................. 78 Figure 73 Evolution of deviatoric and negative minor stresses during longwall retreat ........................ 79 Figure 74 Redistributed insitu and induced body stresses about a roadway with K >0.8 once the roof

and floor deflects............................................................................................................................ 80 Figure 75 Redistributed insitu and induced body stresses about a roadway with K <<0.8 once the roof

and floor deflects ............................................................................................................................ 81 Figure 76 Tendons in shear .................................................................................................................... 83 Figure 77 Dowel resistance for 21mm diameter tendons as a function of UCS ..................................... 83 Figure 78 Importance of bolt angle in maintaining closed bedding ....................................................... 84 Figure 79 Frictional shear resistance provided by bolts ......................................................................... 84 Figure 80 Recommended minimum anchorage lengths in coal measure rocks with resin anchorages . 86 Figure 81 Loading of a strap or panel if loaded as a catenary ................................................................ 88 Figure 82 Results of loading 1.5m and 2.0m square mesh panels (Thompson, 2004) ........................... 88 Figure 83 Logical framework applied to coal mine roof support ........................................................... 90 Figure 84 The hazard of parallel non vertical joints. .............................................................................. 91 Figure 85 Non‐parallel joints defining triangular prisms ........................................................................ 92 Figure 86 Relationship between joint dip, joint friction and horizontal roof stress for a stable

symmetric prism ............................................................................................................................. 92 Figure 87 Two analysis of brittle strength factor using Examine2D showing zones of brittle failure and

possible bolting and cable patterns................................................................................................ 94 Figure 88 Relationship between height of strength factor and the K value for a 5.2m by 2.8m roadway

and a major principal stress of 10 MPa .......................................................................................... 96 Figure 89 Height of failure as a function of roof strength index for various K values ........................... 97 Figure 90 Comparison between the results of multiple linear regression and the numerical data ....... 97

XII

Figure 91 Slight changes in failure zones near the excavation with no difference beyond 2m ............. 97 Figure 92 Different isotropy and failure conditions ............................................................................. 102 Figure 93 Negative horizontal stress (K=0.2) for a 5m by 3m roadway ................................................ 103 Figure 94 Height of negative horizontal stress (5.5m by 3.3m roadway) ............................................. 104 Figure 95 Collapse mode for tensile stress regime ............................................................................... 104 Figure 96 Roadways oblique to joint sets will produce better conditions in both headings and

cutthroughs .................................................................................................................................. 105 Figure 97 Loading on mesh panels ....................................................................................................... 105 Figure 98 Generation of compressive stresses and failure in a regime of no imposed horizontal stress

...................................................................................................................................................... 106 Figure 99 Slip and separation in a layered roof rock (Brady and Brown, 1985) ................................... 107 Figure 100 Critical thickness and deflection of voussoir beams as a function of span and rock strength

(E/UCS=250) ................................................................................................................................. 107 Figure 101 Critical thickness and deflection of a 5.5 m span voussoir beam (1m surcharge) .............. 108 Figure 102 Stability of a voussoir beam increases with increasing thickness (1m equivalent surcharge)

...................................................................................................................................................... 108 Figure 103 Examine 3D geometry ........................................................................................................ 109 Figure 104 BPXS at 2.31 m from the face ............................................................................................. 110 Figure 105 Cumulative increase in BPXS with distance from the rib line ............................................. 110 Figure 106 Average BPXS across the roof line for different heights into the roof and different locations

of bolts (area under curves in Figure 107 expressed as uniformly distributed load .................... 111 Figure 107 BPXS as a function of the location and height .................................................................... 111 Figure 108 XS factor for 0.2m into the roof .............................................................................................. 1 Figure 109 XS factor for 0.4m into the roof .............................................................................................. 1 Figure 110 G parameter to account for different friction angles ............................................................. 1 Figure 111 The bolt timing factor ............................................................................................................. 1 Figure 112 Fully grouted bolts are very stiff during initial shear loading (Stjern, 1995) ...................... 111 Figure 113 Comparison between ideal BPXS support capacity at 0.4m into the roof (Kn=2.0, Kp=1.5,

G=1.3) and the Australian database ............................................................................................. 111 Figure 114 Optimum bolt patterns from physical models (Stimpson, 1987) ....................................... 112 Figure 115 Logical framework for support of ribs ................................................................................ 115 Figure 116 Failure modes for rock slopes that can be observed in coal ribs (Hoek and Bray, 1981). .. 116 Figure 117 All intersections are noses .................................................................................................. 117 Figure 118 Planar geometry and required face support for a 3m high rib ........................................... 117 Figure 119 Wedge geometry and required face support ..................................................................... 118 Figure 120 Anchor force to prevent toppling ....................................................................................... 118 Figure 121 Distribution of the cohesive loss component about roadways .......................................... 119 Figure 122 Value of spalling ratio along the spring‐line of the rib for different K ratios and roadway

aspect ratios ................................................................................................................................. 120 Figure 123 Estimation of maximum depth of brittle failure ................................................................. 120 Figure 124 Thickness of slabs that may undergo buckling. .................................................................. 121

XIII

GLOSSARY

Analytical output is dependent on a least one input variable applied to a law of physics

Brittle Complete loss of strength after failure.CHILE Continuous, Homogeneous, Isotropic, Linear Elastic Cohesion The intercept on the shear stress axis of a straight line fitted to a

scatter plot of normal and shear stresses Continuum A material with no gapsDeviatoric stress The difference between the magnitude of the major and minor

principal stresses DIANNE Discontinuous, inhomogeneous, anisotropic, not elastic Discontinuity A feature is a rock mass with zero or negligible tensile strengthEmpirical Based or acting on observation and experiment, not on theory Friction angle The slope of the straight line fitted to a scatter plot of normal and

shear stresses Homogeneous, inhomogeneous Same materialIsotropic ‐ anisotropic Same properties in all directionsJoint An approximately planar natural fracture in a rock mass that is

typically part of a parallel set Linear elastic An elastic body returns to its original form after a displacing

stress is removed. Modulus The ratio between applied stress and resultant strain Numerical Use of computers to solve complex stress redistributions Parting, bedding parting A natural fracture in a sedimentary rock that is parallel to the

bedding texture Plane strain Deformations out of the plane being considered are zero. Poisson’s Ratio The ratio of the normal strain to the transverse strain of a body

under uniaxial strain Principal stress, major, intermediate, minor

The magnitude and direction of the stresses that are normal to planes where the shear stresses are zero. The stress field is 3 dimensional so there are 3 principal stresses

Reinforcement The addition of restraining forces to joints or partings Rib Sides of an excavationRock fall An unplanned fall of ground of any size that causes (or potentially

causes) injury or damage Slickenside Polished or striated surfaces that result from friction along

movement surfaces. Support The addition of tendons to suspend blocks from higher rock units

that are stable Uniaxial compressive strength The strength obtained in a laboratory test where the rock is

loaded uniaxially Voussoir A bock of rock in a masonary or rock arch

XIV

ABBREVIATIONS AND SYMBOLS

a A constant in the Hoek Brown failure criterionα The angle between the roadway direction and the direction of the major principal

horizontal stress Boltfriction The additional shear resistance supplied by a tendon installed across a closed parting BPXS Bedding parallel excess shearBy Yield stress of steel c CohesionC ConstantD Damage constant d diameterDowel The additional shear resistance supplied by a tendon installed across an open parting E Modulus of Elasticity, Youngs modulusFbolt timing A factor in the determination of the BPXS to account for different friction angles on the

bedding partings Ffriciton A factor in the determination of the BPXS to account for different friction angles on the

bedding partings Fu Tensile strength of steel sheetG Independent shear modulus in a transverse anisotropic materialGSI Geological Strength IndexH Roadway height Hf Height of brittle failure JCS Joint Compressive StrengthJRC Joint Roughness CoefficientKn Joint stiffness kT Parameter in toppling equationL, M The ratio of the major and minor principal stresses to the vertical stress L,l Length Lf Horizontal stress concentration around the maingate cornerm A constant in the Hoek Brown failure criterionMv Vertical stress concentration at various stages of miningPLSI Point Load Strength IndexPR Poisson’s Ratio Pv Vertical load on mesh q Sag of mesh q constant with values of 1 for both ends pin‐jointed, and 0.5 for both ends clamped. RSI Roof Strength index s A constant in the Hoek Brown failure criterionT Anchorage length t Strap thickness Tf Anchorage length U Tensile load in bolt UCS, σ1 Uniaxial Compressive Strengthvel Sonic velocity W Roadway width XS A variable used to determine the BPXS from the vertical stress assuming a friction angle

of 35o and bolting 2.31m from the face γ Densityσ1, σ2, σ3 Major, intermediate, and minor principal stressesσn Normal stress τ Shear stress φ Friction angle

1

1 INTRODUCTION

Rock fall – An unplanned fall of ground of any size that causes (or potentially causes) injury or damage1.

Rock falls have been and continue to be unacceptable events in underground coal mines. Over the last 30 years there have been major improvements in the way in which Australian coal mines have managed the rock fall hazard through the application of safe methods of work, technological improvements, alternative mining systems, and risk management concepts.

All Australian underground coal mines now have strata hazard management plans. These plans nominate a person to manage the mine’s ground control regime – depending on the size of the operation they may have tertiary level qualifications or they may hold the position based on their industry experience. In this report, this person will be referred to as the geotechnical officer.

The geotechnical officer usually has mining/civil engineering or geology qualifications, and in many cases a post‐graduate qualification with a component of rock mechanics. Historically, the subject of mining rock mechanics has mostly focused on metal mining and tunneling. Underground coal mining has a number of significant differences from metal mining in terms of the nature of the rock mass, the mining machines and the shape of the openings. This report has been written to specifically address this void in the technical literature.

The purpose of this report is to provide information to assist the geotechnical officer in identifying potentially unstable roof and rib, and to document design methods that may be considered in the specification of ground support. The report is focused on the stability of the ground in the immediate vicinity of the roadway – that is the immediate roof and rib. Specifically it should assist in answering the following questions which the geotechnical officer routinely faces:

• What failure modes are possible for the operation?

• How can the operation anticipate them?

• How can the potential failure modes be identified at the face?

• How do I specify the support pattern in the ground control plan?

• How do I modify support, if necessary?

The report is un‐apologetically about the use of analytical methods: methods by which collapse modes are identified/inferred at the start of the process, driving forces then estimated, and support components specified in response. Others have developed empirical methods for coal mine roof

1 So far in Australian coal mines all the rock falls are gravity driven. This report does not address gas outbursts.

2

support for Australian underground coal mines and these form an invaluable complement to the analytical tools presented in this report.

Some parts of the report may be considered to be controversial and there are some concepts/approaches presented that are yet to be fully validated. It is emphasized that the concepts are consistent with the available data and that there is no reason why they should not be widely applied. It is through this application that validation will be obtained and the appropriate “factors of safety” obtained. In the meantime, appropriate roof and rib stability will be obtained through the application of standard engineering design practice – the use of a combination of independent design tools.

1.1 THE CHALLENGE

Coal mining needs safe and fast roadway development. Rock falls present an obvious hazard to the workforce and, when large, to business continuity. The objective must be no rock falls, both from workplace safety and business perspective.

Currently, the safety hazards in coal mine roadways are primarily associated with the fall of scat from between the bolts2. Based on a survey of conditions in mines in the USA, Molinda (2003) suggests that the hazards are related to the presence of weak roof and unexpected discontinuities in the roof. There is an obligation to reduce the “unexpected” and to have support regimes that intrinsically address the support of the unexpected.

The installation of roof and rib support is currently manually intensive, so it follows that the necessary level of safety may not be simply achieved by installing more bolts. There is a severe business impact of over‐supporting the roof – beyond the level needed to provide safety to the work force. New mines may not be opened, or existing ones closed, if the support is needlessly intense.

1.2 TARGET READERS

The report has been written for geotechnical engineers, geologists, and mining engineers charged with specifying ground support. Persons occupying the positions of mine manager or technical services manager should find the report of value. The report addresses some of the steps (site characterisation, model formulation, design, implementation, monitoring, and review) that can be found in Clause 48 of NSW regulations for coal, or the rockfall risk management of the Minerals Council of Australia (Potvin and Nedin, 2003).

It is emphasised that a full understanding of coal mine ground control is not yet available – hopefully this report represents a step forward. One use of this report can be to challenge the standard

2 This report does not address the design of pillars and the sequencing of coal extraction.

3

interpretations of how a rock mass behaves around openings so that there can be a gradual improvement in the ability to design safe and productive mine roadways.

The concepts introduced can be applied at early stages – prefeasibility and feasibility using presumed values based on a review of the geological conditions. As the project progresses to operational phase, the greater the importance to move from presumed to demonstrated values for the various input parameters.

1.3 STRUCTURE OF THE REPORT

Section 2 of the report discusses design in rock masses, highlighting the uncertainties that are intrinsic to design in rock masses which are very difficult to fully characterise. The key message is that any design in rock mechanics will need to be implemented with a commitment to monitor and respond as the geological conditions are revealed.

Section 3 defines rock falls and makes a distinction between the nature of the hazards in metal mines and coal mines

Section 4 summarises underground coal mining practices in Australia so that the practical constraints on roadway geometries and excavation methods can be appreciated.

Section 5 provides a reference guide to the engineering geology of coal measures and includes discussions on discontinuities, rock strength and deformation properties, and the insitu stresses.

Section 6 examines the redistribution of stresses around a retreat longwall panel and particularly in the immediate vicinity of a roadway. The key point is that our knowledge of stress above chain pillars needs to be interrogated in more detail to understand the stresses that are present at the roof line. Simple numerical analyses for stress about openings in a homogeneous material are presented

Section 7 presents the key strength properties of the popular support and reinforcement elements in use in Australia and presents models for their reinforcing action.

Section 8 provides a flow chart for the specification of roof support and a number of analytical tools that can be used to determine bolting densities and bolt lengths.

Section 9 provides a flow chart for rib support design.

1.4 LIMITATONS AND OPPORTUNITIES FOR IMPROVEMENTS

This report represents an initial effort to formalize an analytical approach to the design of roof and rib support: an approach that has tended to be ignored In comparison to sophisticated numerical design.

The report presents a number of simplified analyses of very complex mechanisms. In particular, the documented stress analyses are for homogenous materials to take advantage of the availability of some software and to demonstrate some basic concepts. More sophisticated numerical codes are

4

available for layered materials, and these could be incorporated within the proposed framework to improve the predictions.

5

2 DESIGN IN ROCK MECHANICS

Tunnel design is different from many other engineering design processes. However, it can be performed on a scientific basis using an intimate blend of engineering geology, precedent, structural analysis and the observational method during construction (Pells, 2002).

The objective of any engineering design is that the constructed entity fulfils its intended function and in a way that is safely, economically, and environmentally acceptable. Engineering design needs to address both the stability and the deformations of the structure and the adjacent area, and at an acceptable cost.

Because we are constructing something now that will perform in the future, all engineering design requires a component of prediction. This is particularly the case in geotechnical engineering because the properties of the geological materials must be inferred from a comparatively small data set.

Good predictions and the associated support designs are necessary but not necessarily sufficient. Performance assurance (both safety and business outcomes) comes from risk management of the design. Having said that, this report is only focused on the design/prediction component of the process.

Predictions, which are forecasts of events in the future, are the essence of geotechnical engineering. Lambe (1973) provides a classification of predictions in geotechnical engineering (Table 1). Morganstern (2000) provides a framework for assessing the quality of predictions (Table 2) – note that achieving an outcome within 15% of a prediction is considered to be good – this is a reflection of the uncertainties that are inherent in geotechnical practice.

Table 1 Classification of prediction (after Lambe, 1973)

Prediction type When prediction made Results at time of prediction

A Before event

B During event Not known

B1 During event Known

C After event Not known

C1 After event Known

Table 2 Prediction classes (after Morganstern, 2000)

Category Accuracy

Excellent 95‐105% (within 5%)

Good 85‐95% or 105‐115% (within +/‐15%)

Fair 75‐85% or 115‐125% (within 25%)

Poor 50‐75% or 125‐150% (within +/‐50%)

Bad <50% or >150%

6

Ideally, all predictions should be Class A and excellent; this is usually unattainable in practice. The goal should be to achieve Class A as much as possible as they do represent he lowest cost and highest level of intrinsic safety. However, in practice we must be satisfied with Class B predictions – and within the excellent category from the safety perspective and within the good category from a business perspective. Class C are of no value and are best considered as calibrations. In the soil mechanics branch of geotechnical engineering, design competitions on Class A predictions reinforce the inability to accurately predict. No similar data exists in underground rock mechanics.

The recognition of the inability to consistently produce good or excellent Class A predictions is the basis of the design process used in geotechnical engineering. The design process implicitly incorporates many of the more recent concepts of “risk management”.

2.1 DESIGN PROCESS

Do the best prediction as possible, accept that ability to get 100% right is limited, and manage risks through observational method.

In civil construction, the design process includes the following steps:

1. Design brief

2. Concept design

3. Detailed design

4. Final design,

5. Design and drawings for construction

6. Specification, Work method statements, Job Safety Analyses, Inspection and Test plans,

Often there is an Independent Verifier to review the detailed and final designs.

Mining differs from civil engineering only in that the economic value is in what is removed, not what is left behind. Until recently, the mine owner was the operator so there were no contractual issues surrounding the construction.

In mining, the concept and detailed design are done in the pre‐feasibility and feasibility stages. The geotechnical design should evolve along with the JORC stages (Haile, 2004). The approach introduced in this report can be applied at early stages – prefeasibility and feasibility using presumed values based on a review of the geological conditions. As the project progresses to operational phase, the greater the importance to move from presumed values to demonstrated.

For steps 4 to 6, the flow chart of Bieniawski (1993) provides the steps that should be followed (Figure 1). These are very similar to the requirement of Clause 49 of the 1999 regulations.

7

Limits on solution ‐ There is always a possibility that there may be practical limits to the design solution that relate to the mine layout and the machinery that are to be used. It is to be hoped that the feasibility study was valid and that the constraints are not severe.

Collect data – This stage refers to collecting and documenting the geological information. It is important to document any assumptions and simplifications made in the engineering geological model. If rock mass classification systems are quoted, the input parameters should be stated.

Identify collapse modes – Domains should be defined based on depth, variations in the immediate roof lithologies, and the presence of faults. In longwall mines, there may be changes in the collapse modes related to changes in the stress field developed during the extraction phases. It is necessary define these collapse modes as they form the basis of the analyses and hazard recognition later in the process.

Analysis – In the context of this report, this step relates to analyses in excess of empirical rock mass classification systems. Where possible, a number of alternatives should be analysed. Adequate details of calculations should be provided to allow it to be duplicated. All input data should be listed, together with relevant output.

Evaluation – Identification of the preferred solution, possibly involving an assessment of the risks of implementing the candidate design. A comparison with precedent/practice and empirical systems should be presented. There should be documentation of the limits of acceptable deformations and design of monitoring programs.

Implement – This includes the drawing of support rules, preparation of TARPS, training in hazard identification, and a workplace risk assessment.

Monitor hazards – Proactively check geology assumptions, collapse modes, and confirm acceptable deformations. Reports on falls of ground should be in standardised documented format.

All of this should be in the form of a written report. A complete record of the context of the design exercise, the data used, and the procedures followed makes review and extension of the design to similar work in the future much easier. The report should be written as part of the activity and referred to in the support rules and TARPS (Triggered Action Response Plans).

8

STAGE ACTIONS NSW REGULATIONS 1 DEFINITION Who commissioned report

Context (prefeasibility, feasibility, operations)

2 CONSTRAINTS Roadway geometry, machines, consumables 1f, 3 GEOLOGY Seam geometry, roof and floor conditions

(strength, bedding, joints), faults, stresses 1a, 1i

4 GEOTECHNICAL MODELS Collapse modes, selection of design values, domains of different rock classes

1c

5 ANALYSIS Support of detached blocks, reinforcement of discontinuities

1c

6,7,8 SYNTHESIS, EVALUATION, OPTIMIASATION

Identification of preferred solution, comparison with precedent/practice and empirical tools, Documentation of acceptable deformations and design of monitoring

1c

9 IMPLEMENTATION Support rule, TARPS, training in hazard awareness, workplace risk assessment.

1b, 1d, 1e,1h, 1k, 1l, 2a‐e,

10. MONITORING Check geology assumptions, confirm acceptable deformations, report falls of ground.

1g, 1j

Figure 1 Steps in geotechnical design process covered in this report

9

2.2 ROCK AS AN ENGINEERING MATERIAL

Discontinuities – features in a rock mass with zero or negligible tensile strength.

The defining characteristic of rock as an engineering material is the presence of discontinuities. Discontinuities include weak bedding planes, joints, cleats, faults, and shears. The term parting can be used to distinguish bedding planes from bedding textures that are often seen in sedimentary rocks.

For the rockfall problem in coal mining, the scale of the roadway compared to the spacing of discontinuities is such that we are typically dealing with either intact rock or a few blocks defined by three sets of discontinuities – bedding partings and two set of joints (orthogonal).

As a consequence of the presence of discontinuities, rock masses should always be presumed to have zero tensile strength. A practical implication of this presumption is that a rock mass should never be put into tension. Brady and Brown (1985) state – “The important point is that a rock mass in compression may behave as a stable continuum. In a destressed state, small imposed or gravitational loads can cause large displacements of component rock units”.

In addition to the presence of discontinuities, there are a number of other inherent complexities in rock mechanics. Included in these are the influence of groundwater and the deterioration of freshly exposed rock to weathering. In the context of rockfalls, there is significant uncertainty as to how rock fails under compression (Brady and Brown (1985). This issue is highlighted in the report whereby research reported in the early 1990s is used as the basis for assessing compressive failure around roadways.

Recognising the role of discontinuities and the variability with different lithologies, rock masses are discontinuous, inhomogeneous, anisotropic, and not elastic (DIANNE3). DIANNE materials have not been, and possibly still are not, amenable to routine engineering analysis and hence routine design. Numerical codes such as UDEC and 3DEC allow the analysis of DIANNE materials but these are best considered as specialised consulting and research tools.

To enable routine design, rock masses are often ascribed modified properties based on an assumption of equivalent continuous, homogeneous, isotropic, and linear elastic materials (CHILE). CHILE assumptions are required particularly when continuum numerical codes such as FLAC and Phase2 are used.

2.3 ANALYSIS TOOLS

Previous sections have discussed the integration of a number of steps into the design methodology. At the analysis stage, it is good engineering practice to use more than 1 design tool. The decision on

3 discontinuous – presence of discontinuities (not a continuum), inhomogeneous – not all the same rock type, anisotropic – different properties in different directions, not elastic – same increment or decrement of stress may produce a different increment or decrement of deformation (this basically means that rocks fail)

10

which tools to use depends to a large degree on the amount of data that is available. Lambe (1973) proposed that for the same amount of data, the accuracy of a prediction may decrease if an overly sophisticated method is used or if a simple method is employed when more data is available (Figure 2). In this figure, it is proposed that the method spectrum runs from rock mass ratings, through limit equilibrium analytical tools, to 2 dimensional and then 3 dimensional numerical codes.

Figure 2 Accuracy of a prediction is a function of the amount of data available and the method used (after Lambe, 1973).

2.3.1 PRECEDENT AND PRACTICE

There is no doubt that reference to precedent and current practice is an essential component to ground support design in ongoing operations. If there is confidence that the geotechnical conditions will remain the same, then it is reasonable to argue that a demonstrated successful support in immediately adjacent areas can continue to be applied. Given the inherent variability of all geological materials including coal measures, the maximum geographic extrapolation is perhaps in the order of one hundred meters, and will be less if the geology is observed to change.

2.3.2 EMPIRICAL

Empirical : “based or acting on observation and experiment, not on theory”.

In the face of the complexity presented by DIANNE rock, classification schemes for rock masses have been developed in the past. In these schemes, numerical values are assigned to parameters that are considered to influence the behaviour of a rock mass and then these values are combined to give a single numerical rating value. These rating values are then used to interrogate databases of case histories. Two of the early and popular rating systems are the RMR system of Bieniawski (1976) and

11

the Q system of Barton et al (1974). The Coal Mine Roof Rating (CMRR, Mark and Molinda, 2005) is a derivative of the RMR system that the authors propose to be more applicable to sedimentary rock masses.

The classification approach is valid if the rating value can be reliably and repeatedly obtained from the geological information, the mechanics of the system is well understood, and the data base includes case studies from similar geological and mining system. Attaining this goal may not be as simple as the rock mass classification systems would suggest.

Brady and Brown (1985) state “… Although the use of this approach is superficially attractive, it has a number of serious shortcomings and must be used only with extreme care. The classification scheme does not always fully evaluate important aspects of a problem, so that if blindly applied without supporting analysis of the mechanics of the problem it can lead to disastrous results.” This report argues that if the mechanics of the problem are understood, the simplicity of the coal mining geometry and modern stress analysis tools makes the empirical approach unnecessary.

There are recommendations for ground support as a function of both RMR and Q values, particularly for metal mines and tunnels. Workers in horizontally bedded rocks have reported failures in the use of RMR and Q methods to provide appropriate ground support recommendations. (Pells and Bertuzzi, 2007).

For underground coal mines, the empirical relationships based on CMRR should be used, but once again with extreme care. The geomechanics of extended cut mining is substantially different from close‐face bolting and this means that the relationship between the CMRR and the support designs will be different. It is inevitable that extended cut‐mining will have lower support densities for the same rating.

Palmstrom and Broch (2006), in discussing rock mass classification systems and particularly the Q system, concluded that classification systems are useful for planning and less useful for the prescription of rock support during construction. Pells and Bertuzzi (2007) support rating systems for communicating rock conditions but not for the specification of support. Extending this observation further, rating systems are useful to incorporate consideration of rock conditions in the design of another aspect – an overall mine layout, or the dimensions of a chain pillar (Colwell, 1998).

2.3.3 ANALYTICAL

Analytical ‐ “output is dependent on a least one input variable applied to a law of physics”

The most accessible analytical tools are based on the limit equilibrium method, whereby a failure mode is identified, driving forces are estimated at the limit of stability, and resisting forces determined to provide equilibrium of forces ‐ with the application of a safety factor to account for inherent variability and error. The same case history data used for developing the empirical methods is used to check the validity of method. Limit equilibrium methods require a smaller data base than the empirical methods, recognising that the data base must be based on the same failure mechanism.

Limit equilibrium methods are considered to be very suitable for the rockfall design problem as the objective is ultimately the prevention of gravity driven collapse of blocks of rock. This is particularly the case in coal mines as the geometry of the block is comparatively simple, being rectangular prisms in many cases.

12

While very well developed in soil mechanics, there are comparatively few limit equilibrium methods available in rock mechanics. This report presents a number of such methods for the underground rockfall problem in coal mines.

Experience and training is needed to correctly identify and anticipate the failure modes and this may also require more detailed knowledge of the rock mass. The arithmetic in quantifying the driving and resisting forces may be intimidating. Analytical tools do not have to be applied in isolation of numerical methods – in fact it is often useful to use computer codes to estimate the driving stresses for complex geometries and then to separately apply those stresses to the identified failure modes.

This report seeks to present a simplification of the arithmetic, allowing more focus on examining the rock mass and the failure modes that can develop. By having simple arithmetic tools, easier back analysis of observed collapses will allow a better understanding to develop.

In many respects, this report develops the proposition of Brady and Brown (1985) that a rock mass traversed by one or two persistent structural features can be considered massive, opening‐up the ability to apply a range of simple design tools in a structured flow chart (Figure 3). In a later edition of the book, Brady and Brown noted that the logical framework can also be applied to moderately jointed rock.

Figure 3 Logical framework for mine excavation design in massive rock (Brady and Brown, 1985)

All factors of safety need to be developed and applied with care. The choice of an appropriate value depends on considerations relating to the confidence in the material properties, the simplifications of the behaviour models, and the assumptions made regarding various inputs. The best way to determine the values to use in design is to back analyse failure conditions and other experience bases. Since many of the users of this report will have been exposed to the concept of a factor of safety in empirical pillar design it is important to highlight that there is a subtle but nonetheless important

13

difference. In empirical design methods where there has been a statistical analysis of a data base, factors of safety can be selected based on multiples of standard deviations to give probability of safety. Such a statistical approach is not possible in the early stages of the development of an analytical design tool.

2.3.4 NUMERICAL

Numerical – use of computers to solve complex stress redistributions

In a numerical analysis, the key parts of the problem are simplified and a variety of computer programs are used to solve the equations of elasticity based on the imposed stress and the geometry of the openings. It is possible to extend the analysis into the non‐linear domain by implementing yielding. The engineer inputs the elastic deformation and failure properties of the materials being modeled and, in more sophisticated programs, the deformation parameters and strength parameters after failure (elasto‐plastic).

The programs basically calculate stresses. Numerical models have the superficial attraction that they do not require knowledge about the failure modes. However, this is far from the case. Their ability to identify failure depends totally on the decision on what code to use, the nature of the simplifications, and particularly the failure criteria and plasticity parameters that are decided upon. Our knowledge of the deformation and failure of rock masses (e.g. the CHILE assumption) and particularly about how rock masses deform after failure is extremely limited. The danger with numerical methods is that they can downgrade the importance of observations underground, replacing them with reliance on unverified assumptions on how rock behaves and simplifications made by the analyst remote from the operational problem.

The Geological Strength Index (GSI, Hoek and Brown, 1998) has been developed as the basis to modify the engineering properties of laboratory samples to give equivalent continuum values for the rock masses – this formalizes the DIANNE to CHILE transition. This report highlights the fact that around the excavation boundary itself this may not be the appropriate strategy and that consideration of brittle failure (Martin et al, 1991) is more appropriate.

Numerical methods are the subject of intensive ongoing research. However, it has yet to be demonstrated that they are suitable for the design of ground support. Major simplifications of the real conditions are still required (best made by the geotechnical officer rather than by a numerical analyst) and judgment is needed in selection of the software (continuum or discrete elements codes). Continuum codes are unrivalled for the analysis of the far‐field where multiple discontinuities such that the rock behaves as a homogenous material with reduced properties (Hoek and Brown, 1997) or in homogenous rock masses with no discontinuities. Discrete codes would appear to be better for the rock fall problem, but such codes are not as well developed and are slower to run.

Given the overwhelming dominance of continuum codes (FLAC, Phase2) in the market place, there are a number of points that can assist in conducting and/or commissioning numerical analyses for rock fall investigations:

• Finite difference and finite element codes are just different ways of solving the equations of elasticity. With more complex models, the shorter run times of finite difference codes can have commercial advantages.

14

• A continuum with equivalent properties must have zero tensile strength.

• Bedding partings and joints must have a cohesion value of very close to zero.

• To address the uncertainties in numerical analysis, plots of the distribution of stresses, separate from the distribution of displacements and failure zones, should be examined. Any model with zones of negative minimum principal stress but no corresponding failure should be challenged.

• If bolts are explicitly modelled, plots of failures zones with and without bolts should be examined.

• Models that use plasticity are of particular concern as the state of the art in the plasticity of rock masses is poorly developed.

With the failure criteria available in continuum models, there is an ability to include transverse anisotropy (the impact of bedding partings – via the selection of a low value for G, the shear modulus) and there is the ability to model post‐failure behaviour of isotropic materials; but there is no validated model that allows both. There is a trend to seek to model transverse anisotropy and yield together through the use of strain softening ubiquitous joint (SU constitutive models) models in FLAC. These models contain algorithms for material behaviour, the parameters of which are claimed to be validated by independent material property tests. Zipf (2005) presents the state of the art in SU models. Table 3 presents a summary of the material property assumptions and a commentary on the possible impact and highlights the fact that there are a number of key parameters in the SU constitutive model for which there is limited if any justification in the published technical literature.

Table 3 Commentary on recent implementation of SU constitutive models in FLAC

Parameter Common FLAC assumption Commentary Cohesion Mohr Coulomb assumed.

Lab UCS values reduced by 0.56 before calculating cohesion. Cohesion decreases to 10% over 5 millistrain post failure

Other failure criteria are possible Basis for strength reduction may need to have the friction angle assumed to be zero No experimental support for the assumed rate of cohesion loss.

Friction angle 21‐36o , constant with deformation Lower friction angles for the sheared condition well established in the literature

Dilation 10o decreasing to 0o over 5 millistrain of post‐failure shear strain

No experimental or independent literature support for initial or rate of loss

Tensile strength

Tensile strength ranges from about 10% of UCS, to zero over 1 millistrain post failure

Zero tensile strength of rock should be assumed with bedding discontinuities and joints

2.4 UNCERTAINTY, RISK, AND THE OBSERVATIONAL METHOD

Uncertainty is a characteristic of all geotechnical ventures and the management of these uncertainties has been and continues to be an essential aspect of geotechnical engineering practice. The elegance or simplicity of any design tool does not reduce the uncertainty.

With uncertainty comes risk. Whilst this report seeks to reduce these uncertainties to some degree, it is essential to remember at all times that we are providing tools to better understand the uncertainties. Any perceived sophistication of the tools is more a reflection of the current limited alternatives to design and not necessarily a material reduction in the uncertainties.

15

Morgenstern (2000) discusses the sources of uncertainty in geotechnical engineering:

• model uncertainty – this results from gaps in scientific knowledge such that simplified models need to be created to allow for subsequent engineering analysis. This is explicit in the analytical approach that requires the formulation of models of ground behaviour; importantly it is also implicit in empirical and numerical approaches as well.

• parameter uncertainty – this results from the difficulties in ascribing values to the necessary input variables. Parameter uncertainty tends to increase with increasingly more complex models, for example with FLAC models that invoke plastic (post‐failure) properties of rock

• human uncertainty – The quality of the workmanship applied to the analysis and also to the implementation must always be controlled.

In analytical design, the first and second uncertainties can be addressed to some extent through the selection of factors of safety. Judgment and experience is used to increase driving stress or reduce restraining stresses in analytical methods. In empirical methods, the factors of safety may be derived from regression statistics.

In some respects, the observational method (Peck, 1969) evolved to address these uncertainties in soil engineering and has since been applied to rock engineering as well. It is stressed that this method was developed primarily as a way of managing contractual risk and is not necessarily sufficient to manage safety and other business risks.

2.4.1 OBSERVATIONAL METHOD

Peck (1969) defines the observational method in the context of managing construction risk on major projects. The steps in the observational method as applied to the ground support problem in an underground coal mine would be:

• Exploration to establish the general nature of mine.

• Assessment of most probable conditions and most unfavourable conceivable based on the known geology.

• Establishment of a design based on a working hypothesis of behaviour under the most probable conditions.

• Selection of variables to be observed and calculation of their anticipated values on the basis of working hypothesis.

• Calculation of same for most unfavourable condition.

• Selection in advance of a course of action or modification of design for every foreseeable deviation, with particular reference to whether time is available to react.

• Measurement of quantities to be observed.

• Modification of design to suit actual conditions.

Some of the steps are made by the design engineers and some are made by face workers.

It should be apparent that the observational method is not an alternative to basic engineering analysis and design. There is a need to have assessed the range of hazards and the implications – if anything the observational method requires a greater design effort.

It is important to note that monitoring (for example through the use of roof extensometers) is only one component of the observational method. In the complete application of the observational method, a full range of hazards will have been anticipated and candidate solutions already reviewed such that there is confidence that they can be implemented in time and with available equipment.

16

In the Australian coal industry, TARPS are used to give face workers direction on how to react. In developing TARPS it must be recognised that not all conditions may have been anticipated and that crews need to be given tools to allow them to determine if unanticipated geology has been encountered. There is also a danger that, in the absence of detailed design, the TARPS will require unnecessary additional support to be installed to the detriment of the business and without increasing safety.

Perversely, mining is often considered an ideal application of the observational method because of the lack of contractual restraints, although these restraints were the fundamental reason for its initial development. Implicit in the application of the observational method is the availability of time and flexibility to change the support – mining may have less flexibility in this regard and particularly mass mining techniques such as longwall. The method may not be successful if there are brittle failure modes where indications of collapse are very short, or where the result is an unacceptable delay (for example coal flow off a longwall).

2.4.2 OTHER RISK TOOLS

Monitoring of the ground, involving inspections complemented with deformation monitoring as appropriate, is essential to the maintenance of a safe workplace. The greater the vigilance the more likely it is that the workforce can be evacuated from an unsafe position.

By itself, monitoring is not sufficient to manage safety or other business risks associated with geological and geotechnical uncertainty.

Workplace Risk And Consequence forums (WRAC) are not suitable for assessing design and business risks. They should be used to assess the hazards and risk of implementing a specific support rule.

Of the risk tools that are currently available to the coal industry, those based on identification of failure modes are probably more appropriate for business risks. Tools such as FEMCA should provide the opportunity to identify the weaknesses in the geological and geotechnical models.

17

3 ROCK FALLS

This report is concerned with the rock fall hazard and specifically the way to prevent the gravity driven fall of blocks of rock. In some cases the blocks of rock may be defined by pre‐existing surfaces and the prevention of rock fall is by way of support. In other cases, the rock mass may be reinforced so that some or all of the necessary surfaces can be prevented from forming (reinforcement).

As will be seen later, rock falls do not necessarily mean elevated horizontal stresses are present – in fact they can often indicate the opposite – the lack of confinement.

Rock fall hazards can relate to local collapse of the strata (skin effects, scat) or more general collapse of the roof span. Both have to be managed.

3.1 DIFFERENCE BETWEEN COAL AND METAL MINES

The geometry of coal mine roadways is substantially different from that typically assumed for metal mines and tunnels. In typical Australian coal mines, the horizontal axis of the roadway is effectively in the same plane as the dominant bedding discontinuity (Figure 4). Roadways are typically confined to the seam and the dips are typically less than 5o. One of the principal stresses is vertical and this means that on development, the roadway axis is parallel to the plane of two of the principal stresses. On retreat, different stress directions may develop. This geometry should be compared to that presented as the basis for metal mining and tunnelling (Figure 5) whereby the excavation is equi‐dimensional, with an arched roof, and with joint sets and stress axes that are neither horizontal or vertical.

In coal mining, the major rock fall hazards in development roadways relate to delamination along bedding in the roof, the slide of thin slabs of coal along steeply dipping joints in the sides, and the onset of failure of the coal in the sides. There are some important additional hazards that will be discussed later in the report related to the onset of compressive failure. In metal mines, the major rock fall hazard in the advancing roadways is the fall of wedges and in some circumstances (deep mines) the onset of compressive failure. All mines and tunnels are exposed to the problem of highly broken rock within a fault zone.

For longwall coal mines, the roadways at the maingate corner undergo increases in both horizontal and vertical stresses in the former and possibly vertical stress increases and horizontal stress decreases in the case of tailgates.

18

Figure 4 Geometry of a coal mine roadway and its coaxial relationship with discontinuities and stresses.

Figure 5 Typical geometry and discontinuity field for a metal mine roadway (Brady and Brown, 1985)

3.2 ROOF COLLAPSE

As an initial simplifying assumption, a coal mine roof can be considered as a jointed bedded beam made up of material with a finite compressive strength. Depending on the horizontal stress and vertical surcharge loading applied to the beam, the beam will collapse by either vertical shear along the joints, bedding‐parallel shearing leading to delamination and the subsequent failure of thin layers, or the onset of compressive failure through the material causing fractures to develop from both ribsides until a block is defined above the installed support (Figure 6). In addition, there can be the local collapse of thin veneers of rock between zones where there is some tensile strength on the discontinuities or between bolts once they are installed (often referred to as scat).

In the face of this simplification, there is a particularly serious collapse mode based on the presence of joints that dip at less than about 65o‐70o (Figure 7). At this orientation, any compressive horizontal

Stress field

19

stresses can induce shearing along the dipping joints leading to collapse. Conversely, the lack of any compressive stress allows shearing along the vertical joint driven by the dead weight of the blocks and any vertical surcharge.

Examples of the types of roof collapses are shown in the following figures. Figure 8 shows a general roof collapse that resulted from delamination along bedding partings. Note the evidence of bedding parallel shear on the roof bolt still anchored in the roof. Figure 9 shows a local collapse by shear along vertical joints. In this case the roof material is thickly bedded coal. Figure 10 shows the role of mesh panels in controlling the hazard of rock falls associated with scat.

Figure 6 General collapse modes for a bedded roof with sub‐vertical joints

Low stress levels High stress levels

Mas

sive

roc

k

Linear elastic response with little or no rock failure.

Spalling and crushing initiates at points of high stress concentration at the roof /rib corners.

Bed

ded

rock

Bedded rock subjected to low horizontal stress in the roof line. Voussoir beam action develops and if the beams are too thin, compressive failure develops at the roof corners and joints open near the centreline.

Bedding results in higher induced stresses in the roof. Failure can develop higher into the roof than for massive rock.

Join

ted

rock

If roof stresses are very low, jointed-bounded blocks may fall. Cantilever action in any blocks that remain in-situ may result in crushing near the roof/rib corner.

Spalling and crushing initiates at points of high stress concentration at

the roof /rib corners (similar to massive rock.

20

Figure 7 Dipping joints can allow block movement with or without horizontal roof stresses

Figure 8 Collapse following delamination

Figure 9 Collapse related to release along vertical joints

21

Figure 10 Thin slabs between bolts retained with mesh panels

The various collapse modes can be anticipated in the analysis phase and identified during mining operations (Table 4). An important point to note is that observations may not conclusively identify the collapse mode. In particular, it is noted that rib line guttering may develop in all general collapse modes, as it reflects only the localised concentration of high compressive stresses and not necessarily elevated insitu stresses.

Table 4 Observations and data requirements for roof collapse modes

Collapse mode Observations Geological controls Information sources Delamination Rib line guttering,

Centreline cracking Spacing of bedding partings Shear strength of bedding partings

Lithologies Geotechnical logging

Compressive failure Rib line guttering Unconfined compressive strength

Laboratory tests Geotechnical logging Sonic velocity logs

Joint shear Open joints, joint‐ bounded collapse Rib line guttering

Kinematically acceptable joint blocks

Joint orientation and spacing

Dipping joints Open joints, joint‐ bounded collapse Rib line guttering

Joints dipping less than 45 plus friction angle/2 Joints striking within 20o of roadway trend

Joint orientation and spacing

Scat Loading on mesh Thin rock slabs between straps

Closely spaced bedding partings

Lithologies

22

3.3 RIB COLLAPSE

From one perspective, ribs are simply vertical rock slopes and hence are exposed to the same collapse modes – planar slides, wedge slides, and toppling as seen in rock slope engineering (Figure 11).

Planar Wedge Topple

Figure 11 Rib collapse analogies to rock slopes (topples, planar slides, wedges)

In addition, there are additional collapse modes related to the vertical stresses that are applied to a coal rib and which are not present in a rock slope. Theses elevated stresses can induce buckling of joint‐bounded slabs (Figure 12) and, in massive coal, the onset of mining induced fractures that define blocks that simply topple into the roadway (Figure 13).

Figure 12 Buckling of a coal rib

23

Figure 13 Back surface to a slab formed by a mining induced fracture in a coal rib

3.4 APPROACHES TO PREVENT ROCK FALLS

Rock falls are gravity‐driven collapse of blocks. There are therefore two approaches to prevent rock falls. Firstly, there is a support strategy whereby it is accepted that the block will form or is already present so that the approach is to support the block so that its gravity fall into the roadway is prevented. Secondly, there is the reinforcement approach based on preventing the block from forming in the first place, based on the recognition that blocks require a minimum of five surfaces before the kinematics allow the fall and that some of these surfaces may already exist in the form of bedding partings or joints. Referring back to the logical framework (Figure 3), support is required if compressive or tensile failure is induced, and reinforcement is required if there is a hazard of slip on discontinuities.

Distinguishing between support and reinforcement can sometimes be an academic exercise, for example the case of shear translation of pre‐existing blocks along non‐vertical surfaces. From a pragmatic design perspective, it may be preferable to take a step back from the collapse concept and instead distinguish between structurally‐controlled failures and stress induced‐failures. Figure 14 is the flow chart proposed by Potvin and Nedin (2003) for metaliferous mines. It is apparent that Figures 14 and 3 present very similar ideas.

24

Figure 14 Flow chart for the analysis phases for the prevention of collapse of ground (Potvin and Nedin, 2003)

25

4 UNDERGROUND COAL MINING IN AUSTRALIA

If you can’t mine it without hurting people and you can’t mine it without making a profit it is just another black rock

4.1 COAL SEAMS

In 2008, underground coal mining in Australia is being conducted in New South Wales and Queensland with one operation in Tasmania. Both coking and thermal coal are mined by underground methods.

For underground extraction, the coal seams are in excess of 1.8m thick, although thinner seams have been considered with a lower bound of about 1.2m ‐ 1.3m. The maximum seam thickness is in excess of 7m (depending on the definition of economic coal) and the maximum roadway height is 3.6m to 3.8m. In the thick seams, the better coal is most often found in the lower portions of the seam so a coal roof is common. Where the seam allows it, coal floor is often left to improve trafficability on low strength clayey floors.

The dips are relatively flat (say less than 5o) and the depths of cover range from as low as 50m to in excess of 550m.

There are a number of accesses to the working seam – either shafts, declines, or final highwalls.

4.2 MINING METHODS

4.2.1 DEVELOPMENT

4.2.1.1 METHODS

All development mining uses continuous miners and hence the openings are rectangular with heights ranging from 1.5m to 3.8m, and widths ranging typically from 4.8m to 5.5m (an exemption by the NSW mining regulator is required for widths in excess of 5.5m).

Presently, the most popular system for heading development in Australia is integrated cutting and bolting (bolter miners): there are some place‐changing system (cut and flit, extended cut mining)

Integrated cut and bolt systems are more popular in Australian mines and in part this may be a reflection of the higher frequency of faulting in our coal measures compared to seams in the USA where place changing is more common. The integrated system is efficiently implemented with two heading gate roads, and rectangular pillars with headings and cutthroughs at right angles. In the integrated system, roof bolting is carried out using miner‐mounted hydraulically‐operated units (Figure 15). Generally, the integrated system is used in conjunction with wide‐head continuous miners and this enables the installation of support in the roof at a distance of about 2.5m ‐ 3m behind the face. Most mines use shuttle cars for coal clearance behind the continuous miner.

26

The Place Changing method involves using high capacity continuous miners alternating with purpose‐built high capacity multi‐bolter roadway support machines (Figure 16). Narrow head continuous miners mostly suit the system, for frequent flitting operations. Typical machine heads are 3.5m to 4m wide. The length of the cut is typically in the order of 8m‐10m.

The fall of even thin slabs of roof can interrupt the system, so the place changing method requires the ability for unbolted roof to stand for about up to four hours without either local or general collapse. Lithologies with close‐spaced bedding partings and jointed roof areas near faults may not have the required level of stand‐up time.

Figure 15 Typical bolter miner (Source – Joy)

Figure 16 Typical place change miner and mobile bolter (Source – Joy)

27

4.2.1.2 SUPPORT RULE DATABASE

A database of primary and secondary support installed in Australian longwall coal mines as of 2007 is discussed below. The database includes both the primary and secondary support installed in the gateroads. Data have been sourced primarily from site visits. An additional source of data has been obtained from ACARP project C15005 which tabulates the primary and secondary support rules for each of the mines in the development project database. In this database, the installed capacity is calculated as the number of bolts per square metre of roadway times the ultimate strength of the bolts (40 tonnes/m2 is equivalent to about 6 X grade bolts per metre of roadway advance).

The installed capacity of the primary support for both the gateroads and intersections has been plotted as a function of depth in Figures 17 and 18 respectively. Stone roofs have been separated from coal roofs. The figures highlight that depth (as a proxy for stress) is not the sole determinant of roof support, with other issues dominating: these issues may include the strength of the rock, operational decisions to install support required for longwall retreat on advance, and decisions based on precedent and practice. In Figure 17, the high density support (40 t/m2) at depths of 100m‐200m is from mines with very low strength roof strata. A linear regression line is included for diagrammatic purposes only. In general, there is no additional primary support routinely installed in the intersections prior to longwall retreat.

0

10

20

30

40

50

60

70

80

90

100

0 100 200 300 400 500 600

Prim

ary

Sup

port

Cap

city

(ton

nes/

m2 )

Depth (m)

Stone roofCoal RoofLinear (Stone roof)

Figure 17 Summary of Installed Capacity of Primary Support for Gateroads

0

10

20

30

40

50

60

70

80

90

100

0 100 200 300 400 500 600

Depth (m)

Prim

ary

Sup

port

Cap

caity

(ton

nes/

m2 )

Stone roof

Coal Roof

Figure 18 Summary of Installed Capacity of Primary Support for Intersections

28

4.2.2 EXTRACTION

4.2.2.1 LONGWALLS

The typical Australian longwall layout utilizes two‐heading gateroads, mainly as a result of the choice to use bolter miners. Where the coal seam is gassy, the mine usually implements methane gas drainage programs, involving pre and post drainage systems of inseam and inter‐seam drilling as well as drilling from the surface. Multiple headings have been required in some mines to control exposure of the workforce to heat.

Currently (2008) there are a total of 29 operation faces in Australia, producing around 79 million tonnes (Table 5). The average longwall face width is 233 m with face lengths vary between 150 m and 400 m. The panel length runs between 520m and 4800m. The range of the extracted seam thickness varies between 2.0m and 4.5m (average 3.44m). The depth of cover varies between 50 m and 500 m.

Table 5 Australian Longwall statistics (after Cram, 2008)

Mine Longwall production

(t)

Seam Longwall face width

(m)

Extracted thickness (m)

Panel1 length (m)

Angus Place 3 248 500 Lithgow 250 2.9 2997 Appin West 1 947 600 Bulli 250 3.05 1400, 520 Ashton 2 904 800 Pike gully 205 2.35 1700, 1830 Austar 1 505 600 Greta 150, 220 5.75 1500, 1300 Baal Bone 1 922 300 Lithgow 240 2.25 1500, 1000 Beltana 7 849 500 Lower Whybrow 264 3 3 x 3300 Broadmeadow 3 559 600 Goonyella Middle 200 4.25 2 x 2100 Crinium 4 145 400 Lilyvale 270 3.4 2494, 2400 Bundoora 1 218 000 German Creek 225 2.8 2000 Dendrobium 3 619 000 Wongawilli 236 3.5 1955, 1615 Integra 2 955 900 Liddell 246 2.55 2213, 2331 Grasstree 3 843 000 German Creek 300 2.65 2400, 2400 Kestrel 4 449 000 German Creek 250 3.15 3186, 3579 Mandalong 4 767 300 West Wallarah 150 4.2 3020, 2766 Metropolitan 1 484 000 Bulli 154 3.2 1167, 1208 Moranbah North 4 548 000 Goonyalla Middle 300 4.15 1900, 2500 Newlands North 4 894 900 Upper Newlands 300 4.5 4800 Ravensworth 1 096 800 Pikes Gully 250 2.35 1500, 1800 Newstan 2 708 100 West Borehole 220 3.4 3090 North Goonyella 2 426 300 Goonyella Middle 300 4.2 3000 Oakey Creek No. 1 6 256 500 German Cr 300, 200 2.5 3338, 3489 Oaky North 5 340 800 German Cr 250 3.05 3500 Springvale 3 002 400 Lithgow 305 3 3443, 3443 Tahmoor 1 924 900 Bulli 275 2.0 2180, 988 Ulan 3 365 100 Ulan 400 3.05 3018, 1200 United 3 403 100 Woodland Hill 200 3.1 3300, 3120 Wambo North 1 478 100 Wambo 250 2.1 3650 West Cliff 3 372 700 Bulli 300 2.65 3227, 3500 West Wallsend 830 000 West Borehole 165 4.8 1667, 1208

4.2.2.2 SECONDARY SUPPORT

Unlike the primary support, which is relatively consistent along the length of a single longwall block, a wide range of maingate secondary support patterns are used. For example different patterns are

29

used for intersections, gateroads, structured zones etc. The primary support capacity has been included in these figures to capture the variation between gateroads and intersections e.g. many mines install a 4 or 6 bolt primary support with no secondary support mid pillar. In order to compare support capacities both minimum and maximum values have been plotted in Figures 19 and 20 – the calculations assume a 50 tonne capacity cable. Similar to primary support, there is not a strong relationship with depth.

0

10

20

30

40

50

60

70

80

0 50 100 150 200 250 300 350 400 450 500 550 600

Depth (m)

Com

bine

d P

rimar

y an

d M

inim

um S

econ

dary

Sup

port

Cap

acity

(ton

nes/

m2 )

Stone roofCoal Roof

Figure 19 Summary of combined minimum maingate primary and secondary support

0

10

20

30

40

50

60

70

80

0 50 100 150 200 250 300 350 400 450 500 550 600

Depth (m)

Com

bine

d P

rimar

y an

d M

axim

um S

econ

dary

Sup

port

Cap

acity

(ton

nes/

m2 )

Stone roofCoal Roof

Figure 20 Summary of combined maximum maingate primary and secondary support

30

The minimum and maximum secondary support only is plotted in Figures 21 and 22. Overall, there is no relationship between depth of cover and secondary support intensity. There are many mines that do not install any routine secondary support (Figure 21). Most longwall mines install between 8 to 16 tonnes/m2 (Figure 22). Cable length ranges from 4.1m to 8.1m.

0

5

10

15

20

25

30

35

40

0 100 200 300 400 500 600

Min

imum

Sec

onda

ry S

uppo

rt C

apac

ity (t

onne

s/m

2 )

Depth (metres)

Stone roofCoal Roof

Figure 21 Summary of minimum maingate secondary support

0

4

8

12

16

20

24

28

32

36

40

0 100 200 300 400 500 600

Depth (metres)

Max

imum

Sec

onda

ry S

uppo

rt C

apac

ity

(tonn

es/m

2 )

0

1

2

3

4

Num

ber o

f 50

tonn

e ca

bles

/met

re o

f roa

dway

Stone roofCoal Roof

Figure 22 Summary of maximum maingate secondary support

31

5 ENGINEERING GEOLOGY

The geotechnical engineer should apply theory and experimentation but temper them by putting them into the context of the uncertainty of nature. Judgment enters through engineering geology. (Terzaghi, quoted in Palmstrom and Broch, 2006)

The report has already highlighted the defining characteristic of rocks – the discontinuities. In coal measure rocks, the main discontinuities are joints and bedding partings. Bedding textures such as cross bedding are not necessarily discontinuities. This section starts with a discussion of discontinuities, then moves onto strength and deformation properties and then insitu stresses.

It is emphasised that the report is concerned with roof and rib support, whereby the scale is set by the roadway width – say less than 6m. Discontinuities with spacings less than this may impact rock performance.

5.1 DISCONTINUITIES

Rock mass performance is controlled by discontinuities – defined as features in a rock mass with zero or negligible tensile strength. Joints and bedding partings are the major discontinuities in coal measure rocks. There is nothing fundamentally different about the engineering properties of bedding partings compared to joints – they are both discontinuities albeit with different roughness and persistence and the only defining characteristic is the dip of the surfaces.

The reader should be aware that this definition of a bedding parting as a discontinuity is consistent with the RMR system of Bieniawski (1976), the Q system of Barton et al (1974) and the ISRM Commission (1978) but conflicts with the use of the same term in the Coal Mine Roof Rating (CMRR) system (Molinda and Mark, 1994) where discontinuities are allocated a cohesive strength rating. A discontinuity with negligible tensile strength will also have negligible “true” cohesion when tested in shear. A discontinuity may still have an apparent cohesion intercept related to the roughness of the discontinuity surface (Figure 23), however this apparent cohesion is not the same as true cohesion related to the cementation of particles. The difference in definition is particularly stark in the more recent recommendations on the CMRR that promote the use of an index to diametric point load tests instead of the index to discontinuity spacing in some circumstances (Mark and Molinda, 2005).

32

Figure 23 Difference between true cohesion of an intact sample and apparent cohesion of a rough discontinuity

5.1.1 DESCRIPTION OF DISCONTINUITIES

The description of discontinuities is discussed by the ISRM Commission (1978), and also included in the field geologist handbook (AusIMM, 1995). Aspects covered include orientation (dip, dip direction, strike), spacing, RQD, persistence, roughness, aperture, and filling.

Table 6 Classification of discontinuity spacing

Term Spacing (mm)

Extremely close <20

Very close 20‐60

Close 60‐222

Moderate 200‐600

Wide 600‐2000

Very wide 2000‐6000

Extremely wide >6000

Table 7 Classification of discontinuity persistence

Term Length (m)

Very low <1

Low 1‐3

Medium 3‐10

High 10‐20

Very high >20

The spacing of bedding partings can be expected to vary over short distances reflecting the changes in lithology. Shales may have extremely close partings, and sandstones and conglomerates may have extremely‐wide spaced partings. In general, bedding partings will have high to very high persistence, and joint persistence may range from very low to very high

Apparent cohesion

True cohesion

Normal stress ‐ σN

Shear stress ‐ τ τ

σN

τ

roughness angle ‐ i

σN

33

Joints and bedding are not perfectly flat planes. On the relevant scale for a coal mine roadway, the surfaces may be planar, undulating or stepped. The Joint Roughness Coefficient (Barton et al, 1974) can be used to quantify this shape. The Joint Roughness Coefficient (JRC) can be estimated as per Figure 24. In determining the JRC value, not only is the profile of the discontinuity surface important but also length of the profile measured, as shown in Figure 25 – this figure provides a convenient way of determining the JRC in a mine roof.

Figure 24 Joint Roughness Profiles and Corresponding JRC Values (Barton and Choubey, 1977)

On the smaller scale the surface may be classified as rough, smooth, or slickensided.

• rough – sandpaper • smooth – table top • slickensided ‐ polished

34

Figure 25 Alternative method for estimating JRC from measurements of surface roughness amplitude from a straight edge (Barton and Choubey, 1977)

5.1.2 BEDDING PARTINGS

Bedding discontinuities are often referred to as partings, this terminology has the advantage of emphasising the difference from bedding textures. Persistence of bedding partings is typically high or greater. Spacing will vary within the same lithology and in response to changes in lithology. There is no easy way of assessing the spacing of partings, and certainly not from the current generation of geophysical logs. Core drilling is required. Some inference can be made from knowledge of the lithology, but the lack of precision makes this approach unsuitable for design.

The results of JRC measurements on bedding partings during this project are presented in Table 8. The average JRC for partings was 9, with a minimum of 3.

35

Table 8 Roughness of bedding partings

Sample Type Lithology JRC

(10cm interval)

Bedding Plane Siltstone, coaly fossils 12

Bedding Plane Mudstone, coaly laminations 4

Bedding Plane Silty sandstone, minor coaly wisps 6

Bedding Plane Siltstone, 3

Bedding Plane Fine grained sandstone/siltstone, micaceous 5

Bedding Plane Fine grained sandstone, coaly wisps 7

Bedding Plane Fine grained sandstone, coaly fossils 9

Bedding Plane Mudstone 11

Bedding Plane Siltstone 8

Joint Fine grained sandstone 5

Bedding Plane Siltstone 20

Bedding Plane Fine to medium grained sandstone, carbonaceous laminae 5

Bedding Plane Siltstone, coaly fossils 20

Bedding Plane Fine grained sandstone/siltstone, coaly fossils 12

5.1.3 JOINTS IN STONE

There is a very strong trend in sedimentary sequences, and particularly in coal measures, for the development of 2 joint sets at right angles to each other and orthogonal to bedding (Figure 26).

Figure 26 Example of joint patterns in sedimentary strata.

Hobbs (1967) was one of the first to discuss the possible mechanics behind the observed linear relationship between bed thickness and the spacing of joints (Figure 27). In relatively undeformed strata, the spacing of joints tends to follow a log‐normal distribution with the ratio of the median joint spacing to bedding spacing being approximately 1.0 (Figure 28). The implication of Figures 27 and 28 is that as a first approximation, the stone roof of coal mine roadways is composed of cubes of rock.

36

Figure 27 Joints in undeformed sediments (Fookes et al, 2000)

37

Figure 28 Relationship between joint spacing and bedding thickness (Ji et al, 1998)

The data in the literature are limited to beds with thicknesses of less than about 3m. Observations by the authors of highwalls and cliff faces suggest that the general trend may be present to at least about 20m thickness.

Determining the orientation of joints during exploration programs is now a relatively easy task through the use of the acoustic scanner, and in many cases in the Bowen Basin from immediately adjacent highwalls. General trends can also be inferred from a knowledge of the regional geology, in particular the orientation of fold axes (Figure 29). Different joint orientations will be present in the vicinity of faults or dykes.

5.1.4 JOINTS IN COAL ‐ CLEATS

In this report, the terms cleat and coal joints are used interchangeably. In a pure sense, cleat refers to the small scale (very low persistence, extremely close spacing) fractures, particularly within bright coal (non‐persistent cleat, Figure 30). From a geotechnical perspective, there is more interest in the features with greater persistence. The latter are more correctly called joints, but common usage in the mines often refers to them as cleats as well.

38

Figure 29 Association between joints and faults/folds (Fookes et al 2000)

39

Figure 30 Jointed and cleated coal – note very low persistence of cleat (Medhurst and Brown, 1998)

Jemeric (1985) presents a classification system for cleats and highlights that they may have a number of different origins.

• Endogenous Cleavage – mainly oriented perpendicular to the bedding planes. Related to the drying and shrinkage of organic material associated with its compaction and release of volatile matter. Endogenous cleat typically has low persistence.

• Exogenic Cleavage – formed by external forces related to tectonic events. This cleat would tend to be aligned in the direction of the major principal stress active at the time of formation. This is the probably the origin of the persistent cleat or coal joints. This definition may explain how the orientation of coal joints may vary either side of a thrust fault (Figure 31).

It could be expected that cleat sets in coal and joints in stone may be parallel. Information from the Bowen Basin suggests that this may be generally true (Figure 32), but that there can be important deviations. Since bedding surfaces can be the locus for major tectonic horizontal movements, it is reasonable to anticipate rotations as well as translations along such surfaces. The result would be different discontinuity sets in the coal compared to the stone.

Figure 31 An example of rotation of coal joints about thrust faults

40

Figure 32 Cleat and joint orientations in Bowen Basin (Pattison, 1995)

The results of initial testing into the roughness of coal joints indicates that the JRC is greater than for bedding partings (Table 9).

Table 9 Joint roughness coefficient (JRC) of coal joints

Sample Type Direction of Testing with respect to bedding

JRC

(10cm interval)

Cleat Perpendicular 9

Cleat Parallel 3

Cleat Parallel 12

Cleat Parallel 4

Cleat Parallel 4

Cleat Parallel 4

Cleat Parallel 12

Cleat Parallel 4

Cleat Perpendicular 14

Cleat Parallel 4

Cleat Parallel 12

Cleat Parallel 8

Cleat Parallel 8

5.1.5 FAULTS

The eastern Australian coal fields have had a complex tectonic history, with the result being the overprinting of several fault regimes. The state of the art is such that on a mine scale it is not possible to predict in advance the full suite of fault styles and directions that may be encountered. Close‐spaced drilling and seismic surveys can provide information on the major faults in a block of coal but

41

are limited in their ability to resolve small scale faults, fault zones with little net throw, or bedding‐parallel faults. The application of drilling and reflection seismics is to identify ground that should not be mined, but sesimics in particular has limited application in providing geotechnical details of the ground that is to be mined.

Once the overall fault regime is known, it is possible to anticipate the jointing patterns (Figures 27 and 29) – and it is the jointing pattern that is of concern from a ground control perspective. There is likely to be an increase in the density of bedding partings in proximity to faults – it would appear that the tectonic stresses tend to convert bedding textures to partings.

It is highlighted that faults are characterised by broken ground and this implies that the ground has a less ability to accommodate high deviatoric stresses. It is more likely that stress magnitudes will be locally reduced in the vicinity of faults. In fact, the poorer ground conditions in proximity to faults are more likely to be related to the higher density of joints and partings defining pre‐existing small blocks.

Based on the authors’ experience, the following should be anticipated for the 3 typical fault types encountered:

Normal faults

• 70 o ‐ 90o dip.

• Gouge less than 20mm wide.

• Parallel joints for 5m either side.

• Increased bedding parting intensity 10m either side.

• Increased permeability.

Thrust faults

• 15 o ‐ 40o dip.

• Gouge zone of 0.5m to 1.5m in width.

• Very close and close spaced shears and slickensides.

• Undulating surfaces defining lozenge shaped blocks ranging from 10cm upwards.

• Increased permeability.

Bedding parallel faults

• 10mm ‐ 250mm thick.

• Very close spaced slickensides.

5.2 LITHOLOGIES

Knowledge of the lithologies in the roof is important as it gives the key piece of information on which to anticipate the spacing of bedding partings. There is no relationship between lithology and rock strength.

Coal measures have been deposited in a range of fluviatile (Figure 33) or deltaic environments (Figure 34). The fluviatile system can include meandering rivers depositional environments such that the roof of coal seams can range from thick mudstones (flood plain) through interbedded sandstones and mudstones (crevasse splays and levees) to cross‐bedded and planar‐bedded sandstones (channel

42

sands). Of these, the crevasse splay and levee environments tend to produce the closer‐spaced and laterally persistent bedding partings – e.g laminites in the Southern Coalfield. The spacing of bedding partings in the channel sands can be highly variable.

Figure 33 Distribution of lithologies in fluviatile system (Fookes et al, 2000)

43

Figure 34 Distribution of lithologies in a deltaic system (Fookes et al, 2000)

The braided channel systems tend to have coarser grained sediments in thicker beds. The cross‐bedding tends not to develop as a parting, so very‐wide and extremely‐wide spaced bedding partings are often present. The roof types associated with deltaic systems tend to have closer spaced partings when compared to fluviatile systems.

44

An important aspect of Figures 33 and 34 is the scale. Lateral variations in lithologies can be on the scale of tens of metres. The implication of this is that even drill hole spacings of 100m cannot be used to predict the lithologies that will develop above every roadway.

5.3 STRENGTH

Depending on the scale and geometry of the excavation, the strength of the rock mass around the excavation may be controlled by either the discontinuities or by the properties of the rock substance between them.

Strength is an all encompassing term, and covers the unconfined strength, confined strength and tensile strength of intact samples and the shear strength along discontinuities.

5.3.1 UNIAXIAL COMPRESSIVE STRENGTH

5.3.1.1 STONE

For coal measure rocks, the unconfined compressive strength (UCS) of the roof and floor stone as measured in the laboratory ranges from as low as 5MPa ‐ 10MPa to in excess of 100 MPa (Figure 35) – there is no relationship between lithology and rock strength. Igneous materials, such as dykes or sills, may be less than 5 MPa (for example weathered or chilled margins) or in excess of 300 MPa (fresh dolerite).

Figure 35 Plot showing the range in strength and deformation modulus for different lithologies (Shepherd and Gale, 1982)

45

Due to the time and cost in testing core, Australian coal mines often utilise a site‐specific sonic velocity to UCS correlation. Core is selected for laboratory testing from drillholes in which sonic velocity logs have been run. Figure 36 shows a range of relationships with examples from both QLD and NSW minesites and also the McNally (1987) formula for stone. The figure highlights the variability in the relationships and suggests that there are other factors besides rock strength that control the sonic velocity. Other relationships include consideration of density but even these do not cover all factors.

From a ground control perspective, it is the low strength rocks that are of particular interest. The difference in the various relationships at low strengths is highlighted in Figure 37 where 10MPa has a range of 1300m/s in its associated sonic velocity. It is generally accepted that the original McNally database included many samples that had been allowed to dry out before testing with the result that anomalously high strength values were obtained (particularly for the lower strength range). The continued use of the McNally equation is not recommended.

0

20

40

60

80

100

120

140

2000 2500 3000 3500 4000 4500Average Sonic Transit Time (metres/second)

UC

S (M

Pa)

Mine 1 Mine 2Mine 3 Mine 4Mine 5 Mine 6Mine 7 Mine 8Mine 9 McNally (stone)

Figure 36 A variety of UCS sonic relationships for stone

0

5

10

15

20

25

30

2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 3000Average Sonic Transit Time (metres/second)

UC

S (M

Pa)

Mine 1 Mine 2Mine 3 Mine 4Mine 5 Mine 6Mine 7 Mine 8Mine 9 McNally (stone)

Figure 37 UCS sonic relationships for stone at lower strength

46

There are a number of practical constraints to the accuracy of any strength/sonic correlation. These include matching the depth of the samples with the sonic log, allocating a sonic velocity value, preserving the water content prior to testing (especially important for rocks less than about 20MPa), sampling and testing the lower strength rocks, the reproducibility of laboratory testing, and the limitations of the regression analyses (typically limited to standard spreadsheet functions).

The conclusion to be drawn from this discussion is that it is good practice to use existing relationships only with great care, and to develop a relationship for each mine using the highest level of control on the sampling and testing procedures. Given the greater significance of low strength materials, it is recommended that the relationship is created from a data base biased to this end of the spectrum. Having said that, a relationship that the authors have found to be reasonably valid across a range of coal fields is:

UCS (MPa) = 5785 e (‐17374/vel) Equation 1

where vel =sonic velocity in m/s.

The use of Point Load Strength Index testing (PLSI) to estimate UCS needs to be considered carefully. The PLSI can be related to the rock strength using similar statistical methods as for sonic velocity. The test suffers from a high degree of variability and a large number of data points are required to obtain a reliable estimate of the UCS. It is known from civil engineering that the reliability of the PLSI estimate for values less than about 20MPa is very low. The sonic velocity approach to estimating UCS is preferred.

For very low strength range, and for situations where the thickness of the band does not allow laboratory testing, good estimates of strength can be obtained from standard visual and tactile tests (Table 10).

Table 10 Visual and tactile tests (ISRM,1978)

Grade Term UCS (MPa)

Field observations

RO Extremely weak 0.25‐1 Indent with fingernail

R1 Very weak 1‐5 Peel with pocket knife, crumbles under firm blows with point of geological hammer

R2 Weak 5‐25 Peeled with difficulty with pocket knife, shallow indentation with firm blows with point of geological hammer

R3 Medium strong 25‐50 Cannot scrap or peel with knife, specimen fractured with single blow from the blunt end of a geological hammer

R4 Strong 50‐100 More than 1 blow from the blunt end of a geological hammer

R5 Very strong 100‐250 Many blows from the blunt end of a geological hammer

R6 Extremely strong >250 Can only chip the sample with a geological hammer

5.3.1.2 COAL

Because of its cleat and jointing, the UCS of coal is a relatively difficult parameter to determine in the laboratory. A compilation of the results of testing coal samples, mainly by ACIRL (1986), is presented in Table 11 where it can be seen to range between 10MPa for Middle Goonyella Seam (coking) and 35MPa Great Northern Seam (thermal). Medhurst and Brown (1998) suggest that the UCS of coal is related to the brightness of the coal and will range from 35 MPa for dull coal to 9.7MPa for high brightness coals – this is consistent with the data in Table 11.

47

Table 11 Range of coal unconfined compressive strengths (ACIRL, 1986)

Seam UCS (MPa) Modulus (GPa)

Harrow Creek 8.7 1525

Bulli 28.9 4675

Tongarra 17.9 3403

Katoomba 33.1 4013

Mt Arthur 21.9 3146

Bulli 20.2 3191

Wallarah 27.4 3838

Moura D Seam 11.4 2488

Lithgow 26.4 4096

Wongawilli 16.0 3080

Victoria Tunnel 28.0 4458

Great Northern 35.6 4560

Middle Goonyella 10

Newlands bright bottoms 14

German Creek (Southern) 10

A different sonic velocity/strength relationship exists for coal (Figure 38).

y = 0.07x - 148.57R2 = 0.5178

-10

0

10

20

30

40

50

60

2000 2100 2200 2300 2400 2500 2600 2700 2800

Sonic velocity (m/sec)

Labo

rato

ry U

CS (M

Pa)

Figure 38 UCS sonic velocity relationship for coal

5.3.2 TENSILE STRENGTH

It cannot be stressed enough that the presence of discontinuities in a rock or coal mass means that the tensile strength of the mass is zero.

The tensile strength of an intact rock can be assumed to be in the order of 10% of the UCS.

48

5.3.3 TRIAXIAL PARAMETERS

Increased confinement of samples results in an increase in the stress that will cause failure. Conventionally, the relationship between failure stress and confining stress (Figure 39) is simplified to a straight line (Mohr Coulomb failure criterion) or as a curve (Hoek Brown failure criterion).

Figure 39 Strength envelopes (dash – Mohr Coulomb, solid – Hoek and Brown)

For the Mohr Coulomb criterion,

σ1 = σc + σ3 tanψ Equation 2

where σ1 and σ3 are the maximum and minimum effective stresses4 at failure and σc is the uniaxial compressive strength, and the relationship between uniaxial compressive strength, cohesion (c), and friction angle (φ) is:

σc = 2 c cosφ/(1‐sinφ), or

c = σc (1‐sinφ)/(2 * cos φ), and

tanψ = (1 sin φ)/(1‐sin φ), or

sin φ = (tanψ – 1)/(tanψ + 1).

For the Hoek Brown criterion,

σ1 = σ3 + σc(m σc σ3 + s)a Equation 3

4 effective stress = total stress – pore pressure. Coal seams are aquifers and have significant pore pressures prior to mining. At the mining face the pore pressures are zero so it follows that there have been significant effective stress changes ahead of mining.

σc

σ1

σ3

ψ

49

where m is a parameter related to lithology (somewhat analogous to friction angle), a = 0.5 for rock and 0.65 for coal, and s= 1.0 for intact rock. As will be discussed, m and s can be varied to reflect rock mass strength.

The σc parameter is calculated from the curve fitting and may not have the same value in the two criteria and may not have the same value as the UCS value.

As a good approximation, the friction angle φ and m values can be related to lithology (Table 12)

Table 12 Presumed friction angles and m values

5.3.4 BRITTLE STRENGTH

Recent research suggests that neither the Mohr Coulomb nor the Hoek Brown criterion is suitable for the behaviour an excavation (Kaiser and Kim, 2008). It appears that cohesion and friction are not mobilised at the same time, and that initial rock failure relates only to a loss of cohesion. One of the early papers on brittle failure (Martin et al, 1999) concentrates on massive igneous rocks, but a closer examination reveals that one of the early case studies involved the Donkin‐Morien tunnel in the coal measures of the Sydney Basin of Nova Scotia. The research provides the technical explanation for the early observations that around openings the mass strength appears to be 0.5 the laboratory measured strength.

This concept can be incorporated in a failure criterion either as σ1 ‐ σ3 = σc/3 to σc /2 (φ =0.0), or m = 0.0, and s =0.11 to 0.25 (Figure 40). There is no tensile cut‐off with the Hoek Brown criterion so, if tensile roof stresses are possible, it may be preferable to use the Mohr‐Coloumb criterion with a zero tensile strength cutoff.

Figure 40 Brittle rock parameters

Lithology Friction angle mi Siltstones, fine sandstones 30 10 Mudstone 25 7 Medium and coarse sandstones 35 15 Claystones 20 4

50

This brittle rock concept is applicable to estimating the zone of fracturing about excavations and hence is particularly relevant to the design of ground support systems – the subject of this report. For coal measure rocks, s = 0.11 has been found to be appropriate (see later).

Full implementation of the brittle rock concept also requires the definition of the spalling limit. As a concept, the spalling limit is not well defined and it needs to be determined empirically (Diederichs et al, 2004). The height of roof falls on development suggests that the spalling limit for coal measure rocks may be in the order of 3 ‐ 5 assuming isotropic deformation parameters. This is lower than that inferred in the work on igneous and metamorphic rocks but may be an artifact of the assumption of isotropic deformation properties. The inferred spalling limit increases to a value of about 10 if the independent shear modulus (G) is reduced to simulate the impact of bedding anisotropy. For coal in the ribs, the spalling limit appears to be greater – about 10 (see later).

5.3.5 SHEAR STRENGTH OF DISCONTINUITIES

The shear strength of a discontinuity can be expressed in terms of the Mohr Coulomb criterion:

τ= c + σn tan φ Equation 4

where τ = the shear strength, c= cohesion and σn = the normal stress on the surface (Figure 41).

For perfectly flat surfaces, the cohesion is equal to zero. As discussed above, real world discontinuities are rough and uneven and the effect of this can be considered either in terms of an effective cohesion or in terms of an apparently high friction angle at low normal stresses.

Figure 41 Strength envelope for a discontinuity

The Barton Bandis shear strength criterion for discontinuities incorporates the contribution of surface roughness (Barton, 1973):

τ= σn tan [JRC log10(JCS/σn)+φ] Equation 5

where JCS= Joint Compressive Strength.

c

τ

σn

φ

Φ + i

51

For bedding partings with a JRC value of 3 in a mudstone of 40 MPa, this relationship gives an increase of 6o to the friction angle for a normal stress of 300 kPa.

5.3.6 MASS STRENGTH

The strength of the rock mass when considered as a continuum is of limited value to the rockfall problem. This section is included to give an introduction to the technical literature.

By definition, the tensile strength of any rock mass is zero.

5.3.6.1 STONE

The Geological Strength Index (GSI), Hoek and Brown (1997) provides the latest and most comprehensive approach to determining the equivalent continuum properties of a rock mass when the scale being considered results in more than 2 or 3 sets of discontinuities. The GSI is an extension of the Rock Mass Rating system of Bieniawski (1976) and can be estimated readily from a number of charts. Of the available charts, the one created for molasse (Hoek et al, 2004) is perhaps the most applicable to coal measures (Figure 42). The surface conditions for coal measures are best described as fair – reflecting the smooth nature of the bedding partings. The suggestion by Hoek at al is that quoting to the nearest 5 points is appropriate. A GSI of about 50 should be typical for coal measures.

Figure 42 Geological strength index for molasses

The GSI is used to modify the m and s parameters in the Hoek Brown criterion:

52

mb =mi exp((GSI‐1000/(24‐14D),

sb=exp((GSI‐100)/(9‐3D)), and

a= 0.5 + 1/6*(exp (‐GSI/15)‐exp(‐20/3))

where D= damage index. For coal mine roadways excavated with continuous miners, D can be taken as 0.0.

Brown (2008) suggests that the GSI reductions should not be applied for GSI values greater than 70 or less than 30, not for rocks with a UCS less than 15 MPa.

5.3.6.2 COAL

Medhurst and Brown (1998) and Medhurst (1999) provide a method for estimating the mass strength of coal based on the Hoek and Brown failure criterion (Table 13). The m and s values are related to the vitrinite reflectance values, seam brightness logging, and the σc value is 35 MPa. For coal, the a value was found to be 0.65.

Table 13 Hoek‐Brown parameters for various coals

Vitrinite reflectance

mi mb S

C5 C4/C3 C3/C2/C4 C2 C1

0.9‐1.1 20‐16 3.0‐2.4 0.085 0.075 0.07 0.065 0.06

1.1‐1.30 16‐12 2.4‐1.8 0.08 0.07 0.065 0.0625 0.0575

1.3‐1.5 12‐10 1.8‐1.5 0.075 0.675 0.0625 0.06 0.055

Care is needed when using these relationships, particularly for high brightness, highly cleated, low strength coals as there is a possibility to underestimate the strength. Medhurst (pers comm.) also advises that determining the laboratory strength of coal is difficult and is perhaps best estimated by testing at very low confining pressures and then extrapolating.

5.4 DEFORMATION PROPERTIES

5.4.1 MODULUS

Modulus values for the range of coal measure rocks are given in Figure 35. Ratios of the modulus to the UCS have been found to be reasonably consistent for each rock type (Table 14).

Table 14 Guidelines for the selection of modulus ratio (after Hoek and Diederichs, 2006)

Lithology Modulus/UCS Conglomerates 300‐400 Sandstones 200‐350 Siltstones 350‐400 Claystones 200‐300 Shales 150‐250

Reductions in the laboratory values (Ei) to rock mass values (Erm) can be made using the GSI value (Hoek and Diederichs, 2006) whereby:

53

Erm = Ei * [0.02 + (1‐D/2)/(1+exp((60+15D‐GSI)/11))].

The reduction factor is shown in Figure 43 for the case where D = 0.0: it can be seen that intact or massive rocks have a value close to unity, and for typical coal measures rocks (GSI = 50) the reduction factor is in the order of 0.3.

00.10.20.30.40.50.60.70.80.9

1

0 10 20 30 40 50 60 70 80 90 100

GSI

Mod

ulus

redu

ctio

n fa

ctor

Figure 43 Modulus reduction for case of damage index = 0

Brady and Brown (1985) analyse the simple case of parallel discontinuities to model transverse anisotropy in a continuum model. They provide equations to estimate the equivalent deformation modulus and Poisson’s ratio values as a function of the spacing and properties of the discontinuities (Figure 44). The trend is for a reduction in the equivalent modulus and Poisson’s ratio normal to the discontinuity as the spacing reduces.

11.11.21.31.41.51.61.71.81.9

2

0 1 2 3 4 5

Spacing of discontinuites (m)

Equi

vale

nt m

odul

us (G

Pa)

0.13

0.154

0.178

0.202

0.226

0.25Eq

uiva

lent

Poi

sson

s ra

tio

ModulusPoisson's ratio

E=2 GPa,Kn = 10 GPa/mPR=0.25

Figure 44 Simple model for the impact of discontinuity spacing on the deformation modulus and Poisson’s ratio

The previous discussion has assumed the material is an isotropic continuum. Coal measures are characterised by the presence of bedding not only as a texture but also as the dominant discontinuity. It follows that a better assumption would be a transverse anisotropic continuum with the major additional parameter being the independent shear modulus (G). There is little guidance for the selection of the independent shear modulus for coal measures. Through back analyses using

54

continuum codes, the principal author has found values of between 30MPa and 250MPa give reasonable simulation of fall cavities.

5.4.2 POISSON’S RATIO

There is little information about the variations in Poisson’s ratio for different rock types, mainly because its value has little impact on 2 dimensional stress models. As will be seen, this is not the case when in the case of 3 dimensional stress analyses.

Figure 44 shows the decreasing trend in Poisson’s ratio as the spacing of discontinuities decreases. For rock masses, it can be anticipated that the Poisson’s ratio is less than the laboratory values, and that it increases as the mean stress level increases.

5.5 IN‐SITU STRESSES

It is not possible to predict the state of stress in the ground from a knowledge of the depth of cover. The state of stress at any point will vary depending on the depth of cover, the presence of faults, the nature of any discontinuities, and the stiffness of the local rock types.

The following general rules need to be applied with care. Ideally, the stress state will be measured at each mine site. However, such a measurement is unlikely to be representative of all of the mine, especially when faults are traversed. Furthermore, stress measurements are difficult and expensive. The approach should be to apply regional knowledge of the general stress field to point measurements at the mine site, and particularly to observations of how excavations behave underground. In the latter case, it is essential not to jump to the paradigm that all roof falls are due to elevated horizontal stresses; in fact this is unlikely to be the case once the roof is supported at the densities typical of current Australian coal operations.

5.5.1 STRESS IN STONE

Data from New South Wales and Queensland (Nemcik et al, 2006) show that the ratio of the horizontal stress to the vertical stress is typically between 1.0 and 2.5 times at typical mining depths, with even higher values at shallow depths (Figure 45). Hillis et al (1999) report that 80% of the data from the Bowen Basin has the vertical stress being the lowest of the 3 principal stresses, and 17% with the vertical being the intermediate stress; in the Sydney Basin 90% have the vertical being the minimum stress.

Within the limitation of any model for stress magnitudes and directions, the following considerations have proved to be useful:

• The high horizontal stresses are most likely related to Tertiary and Recent Age migration of the Australian plate to the north‐north‐east.

• At depths in excess of about 200‐250m the major principal horizontal stress is likely to be oriented north‐north‐east.

• At depths less than about 200m‐250m, stress relief to the surface along bedding will mean that the minor principal horizontal stress will be parallel to the dip direction and hence the major principal horizontal stress will be parallel to the strike of the coal measures.

55

• The presence of bedding as the dominant discontinuity means that the vertical stress is likely to be a principal stress.

• The vertical stress (in MPa) can be estimated as the depth of cover (m) times 0.025

• Large deviations in the general stress pattern will be encountered around large faults and fold structures.

• A reduction in stress magnitudes should be anticipated near faults.

• As a starting position, the magnitude of the major principal horizontal stress should be assumed to be twice the vertical stress, and the minimum principal horizontal stress should be assumed to be 1.5 times the maximum principal horizontal stress.

Figure 45 Summary of stress measurement data from New South Wales and Queensland coalfields (Nemcik et al, 2006)

5.5.2 STRESS IN COAL

The stress field in coal is substantially more complex than in stone. Enever et al (2000) have reported consistently low values of the minimum horizontal stress based on step tests during coal seam methane exploration (Figure 46). All the data indicate that the vertical stress is either the intermediate or the major stress. No definite explanation has been provided.

The situation is further complicated as overcore measurements conducted from mine openings into thick coal seams have revealed a different stress field again (Table 15). There are not many test results available and the few that are do show a consistent pattern. The field data gives a stress field that is not perfectly aligned to the vertical and horizontal. It is not known if this is measurement error or a reflection of the true stress orientation. The key observations are that the major stress is approximately vertical and significantly less than the vertical stress that would be expected from the depth of cover. The horizontal stresses are lower than the measured vertical stresses – and as low as 25% in many cases.

56

Figure 46 Measured minimum stress in coal as a function of depth (Enever et al, 2000)

Table 15 Compilation of stress measurements taken from under excavations

Depth (m)‐

Nominal vertical overburden stress

(MPa)

“Vertical” stress (MPa)

Greater “horizontal” (MPa)

Lesser “horizontal” (MPa)

Seam A 170 4.3 1.3 0.6 0.3

Seam B 165 4.1 1.9 1.2 0.5

200 5.0 2.3 1.0 0.4

200 5.0 2.6 1.1 0.6

220 5.5 1.3 0.46 ‐0.27

Seam C 220 5.5 3.3 4.1 1.3

Seam D Tests unsuccessful (Enever and Doyle 1996)

Seedsman (2004) has speculated on the mechanisms that may be acting to produce this measured stress field in coal. He noted that the ratios of the vertical to horizontal stresses are similar to those expected from the vertical loading with lateral restraint (Poisson’s ratio effect) and that the vertical stresses are lower than those anticipated from depth of cover. He proposed a model based on shrinkage of coal ahead of the mining face as the groundwater (and gas?) pressures are reduced by the mining face. The overlying stone does not shrink and as a result there is a decoupling between the stone and the coal. The coal may undergo some sort of passive failure at this stage (as yet not defined). Closer to the mining face and as a result of a broadening of the dewatering – shrinkage zone, the overlying stone sags and loads the coal vertically. These re‐imposed vertical stresses generate horizontal stresses.

The timing of the reloading needs to be considered. The discrepancy in the vertical stress suggests that separation has developed and this must disappear over time. It is possible that there are lower vertical stresses at the face compared to outbye. This matches observations underground where there is outbye and /or time dependent deterioration of the ribs.

57

If the reloading model is accepted, it is possible to extend the analysis to consider the impact of intense cleating in the coal. For the simple case of a coal mass containing a single set of parallel discontinuities, it can be shown that the closer the spacing of the discontinuities, the lower is the Poisson’s ratio of an equivalent transversely anisotropic continuum (Figure 44). This lower Poisson’s ratio would result in a lower induced horizontal stress in more jointed coal as the coal is reloading.

5.5.3 FAULTED GROUND

In general, stress magnitudes should be lower in faulted ground.

Elevated deviatoric stresses generate failure initially and this failure may be seen as faults, more often than not with associated sub‐parallel joints. Once the rock is broken, the maximum deviatoric stresses within the broken rock are controlled predominantly by the frictional resistance of the surfaces generated by the faulting. Indications of the scale of this effect can be obtained by considering passive earth pressures (Herget, 1988) assuming a cohesion of zero (Figure 47). Nemcik et al (2006) produced similar results through numerical methods. Brady and Brown (1985) discuss the application of specific tests for the mechanical acceptability of a stress state with respect to equilibrium on pervasive planes of weakness.

0

100

200

300

400

500

600

700

0 5 10 15 20 25 30 35 40 45

Magnitude of horizontal stress at passive failure (MPa)

Dep

th (m

)

Vertical stress

5°

10°

15°

20°

25°

30°

35°

Figure 47 Reduced lateral (horizontal) stress associated with the presence of surfaces with lower frictional resistance based on (a) passive earth pressures, (b) UDEC modeling (Nemcik et al 2006).

In Figure 47, it can be seen that the maximum horizontal stresses decrease as the friction angle decreases – for example as would happen in the presence of slickensided surfaces. This observation is important for 2 reasons: (1) it would suggest that horizontal stresses are lower in proximity to thrust faults in clayey rocks than in sandstone rocks, and (2) the poor roof conditions that are typically encountered in the vicinity of thrust faults are not the result of elevated horizontal stresses but are a consequence of the presence of broken rock about the faults.

58

5.5.4 TOPOGRAPHY

For most practical purposes, the influence of topography on the stress field is relatively small. Some simple continuum models of a 100m deep valley with valley walls at 62.5o and a 2:1 horizontal to vertical stress field have been conducted to demonstrate this (Figure 48). The models show that there are concentrations of horizontal stresses immediately beneath valleys along with the obvious reductions in vertical stresses leading to increases in deviatoric stresses. However, the impact of the valley is negligible within about 1.5 times the depth of the valley.

Hill‐top or ridge‐top mining (Figure 49) will be different with potential low horizontal stresses normal to the ridge line and much higher values parallel to the ridge line.

2.00

6.00

10.00

14.00

Sigma 1Field stress: gravityGround surface elevation: 1000 mUnit weight of overburden: 0.025 MN/m3Total stress ratio (horizontal/vertical in-plane): 2Total stress ratio (horizontal/vertical out-of-plane): 1.5Locked-in horizontal stress (in-plane): 1Locked-in horizontal stress (out-of-plane): 1Stress Analysis

Ground Elevation : 1000.00

4.00

6.00

8.00

Deviatoric stressField stress: gravityGround surface elevation: 1000 mUnit weight of overburden: 0.025 MN/m3Total stress ratio (horizontal/vertical in-plane): 2Total stress ratio (horizontal/vertical out-of-plane): 1.5Locked-in horizontal stress (in-plane): 1Locked-in horizontal stress (out-of-plane): 1Stress Analysis

2.00


10.00

15.00

Horizontal stressField stress: gravityGround surface elevation: 1000 mUnit weight of overburden: 0.025 MN/m3Total stress ratio (horizontal/vertical in-plane): 2Total stress ratio (horizontal/vertical out-of-plane): 1.5Locked-in horizontal stress (in-plane): 1Locked-in horizontal stress (out-of-plane): 1Stress Analysis

5.00


Figure 48 Effect of topography on stress distribution

59

Figure 49 Horizontal stresses in hilltop or ridge mining

5.6 LOCAL TERMINOLOGY

Over time, different coalfields have evolved different expressions for various geological features (Table 16). Some of the terms are colourful. There are some significant differences between the usage of the term “cutter” – the USA uses the term for stress induced failures at the rib line while in Australia the term is used to describe the trace of basically vertical joints.

There is a tendency in Australia to use the term “gutter” as a general term for all roof cracking.

Table 16 Local geological terminology

Draw rock low strength shale/coal that often drops before bolts

installed Stack rock, catalogue rock laminiteHead coal Top coal, coal topsRash Carbonaceous shales and coalKettlebottoms Pot arses – fossilized tree trunksHorsebacks Roof falls that develop between bolts Hill seam Dilated vertical joints parallel to the slope Greasy backs Slickensided surfaceCutter roof (USA) GutteringWater cracks Open joints Gutter Shearing of the roof at the roof/rib corner that tends to

hade at a steep angle over the roadway. Trench Joint bounded rectangular cavities in the roof either

between bolts or over full width of roadway Channel roof Arch shaped cavities that develop prior to bolting. Cutters (Australia) Small scale normal faults or joints

Reduction in horizontal stresses,related to proximity to free face of valley

60

6 MINING INDUCED STRESS CHANGES

In this section, both the vertical and horizontal stress changes induced about coal mine roadways and also the immediately adjacent extraction areas are discussed. The focus is on the stresses that develop in the immediate vicinity of the roadway, and not necessarily on the stresses developed elsewhere in the system.

There can be no doubting that considerations of the stress redistribution is complex, involving interactions between the in‐situ stresses and the large longwall goaf, stress redistribution around the roadway itself, and any body stresses induced by the deformation and movement of the immediate rock and coal mass.

6.1 REDISTIBUTION ABOUT A LONGWALL

Both the vertical and horizontal stresses are concentrated at the maingate corner (Figure 50). Vertical stresses in coal have been measured in the context of pillar design (Mark, 1990, Colwell, 1998). Our knowledge of the redistribution of horizontal stresses about a longwall is based on extensive stress monitoring work conducted by Dr Gale and his associates over the last 20 years. In these latter studies, stresses and stress changes have been measured using hollow inclusion cells installed in stone above the chain pillars and about 5m ‐ 10m into the roof. As will be discussed, these are not the stresses at the roof line.

6.1.1 MAINGATE CORNER

Standard chain pillar design methods provide an estimate of the vertical stress expressed as a uniformly distributed load. The actual stress magnitudes depend on the depth of cover and the width of the pillars. Typically the vertical stress is doubled at the maingate corner (Figure 50). Similar results can be obtained from a simple 2D elastic stress analysis. At the face/tailgate corner the verticla stress has increased fourfold.

Measurements of horizontal stress concentrations and reductions about the maingate corner have been conducted by SCT (Gale, 2008) and others. There is a concentration of horizontal stresses above the chain pillar at the maingate corner and a reduction in the horizontal stresses above the pillar behind the face adjacent to the goaf (Figure 51). There is also an increase in the vertical stress observed above the pillar behind the face.

61

0

10

20

30

40

50

60

Virgin Maingate Bleeder Tailgate Double goaf

Vert

ical

str

ess

(MPa

)

0

1

2

3

4

5

6

Vert

ical

str

ess

as a

pro

port

ion

of th

e vi

rgin

st

ress

40m pillar, 500m depth (MPa)

25m pillar, 250m depth (MPa)

40m pillar, 500m depth

25m pillar, 250m depth

Figure 50 Pillar design vertical stresses developed above chain pillars assuming a 200m wide panel.

Figure 51 General pattern of vertical and horizontal stress redistribution (Gale 2008)

62

The magnitude of the concentration of the major principal horizontal stress depends on the angle between the stress axis and gateroad direction (Figure 52), with the possibility that there is a doubling of the magnitude at a 45o angle. A simple 2D elastic analysis produces the same trend with respect to the angle between the stress direction and the roadway but the magnitudes are higher – this indicates that some failure may be developing in rock mass in the horizontal plane. This failure is likely to be bedding‐plane shear.

As will be discussed, the simultaneous increases in both the horizontal and vertical stress are significant in terms of the stresses induced in the immediate roof of an excavation. Depending on the concentration factor that applies to the horizontal stress, it is possible that the vertical stress acting in the maingate may become the major principal stress.

Figure 52 Concentration of horizontal stress magnitude at the maingate corner as a function of the angle between principal horizontal stress axis and the roadway direction (after Gale 2008)

6.1.2 BLEEDER/TAILGATE

Inspection of Figures 50 and 51 reveals that there are additional vertical stress increases on the chain pillar behind the longwall face (this will be referred to as a bleeder roadway) and reductions in the principal horizontal stress. The horizontal stress reduction is greater than that indicated from a 2 D elastic analysis. Tarrant (2006) possibly provides the mechanism for this greater stress relief and also for the lesser stress concentration at the maingate corner – shear along bedding in the direction of the goaf.

The combination of an increase in vertical stress with a decrease in horizontal stress would appear to be inconsistent with the so‐called Poisson’s ratio effect. This effect cannot be used to explain the generation of stresses near excavations. Jaeger and Cook (1979) stress that the horizontal to vertical stress ratio is related to the Poisson’s ratio only in the case of complete lateral restraint (uniaxial strain). Such complete lateral restraint is not available in proximity to an excavation.

6.1.3 TAILGATE CORNER

Monitoring of horizontal stresses above pillars in tailgates has rarely been reported. The vertical stresses within the pillar have been comparatively better studied. Shen et al (2006) instrumented a

1

1.25

1.5

1.75

2

2.25

0 20 40 60 80Angle between direction of major principal horizontal stress

and gateroad (degrees)

Con

cent

ratio

n of

the

mag

nitu

de o

f the

pr

inci

apl s

tress

tang

entia

l to

the

long

wal

l exc

avat

ion Range of values

63

tailgate at Ulan at a depth of about 200m and showed that overall the vertical stress increased by 3‐4 MPa while at the same time the horizontal stress decreased by up to about 1 MPa (Figure 53). The overall pattern in the tailgate is consistent with increase in vertical stresses implicit in the pillar design models (Figure 50) and further reduction in horizontal stresses resulting from more lateral translation of the roof into the now two longwall voids.

Figure 53 Stresses above a pillar in a tailgate (Shen et al 2006).

6.2 REDISTRIBUTION ABOUT A ROADWAY

6.2.1 BASIC CONCEPTS

Pragmatically, stresses and failure state are better understood in 2 dimensions – a vertical stress and one horizontal stress normal to the roadway axis. In the following discussion, both these stresses are conveniently taken to be principal stresses. Whilst this is a reasonable assumption for the vertical stress on initial roadway development (because the seam dips are so flat), it is not strictly valid for the horizontal stress if the roadway is not aligned parallel to a principal stress. Also note that the assumption of a vertical principal stress may not be valid once the stresses are redistributed about a longwall goaf. A simplifying assumption needs to be made such that the magnitude of the horizontal principal stress in the following 2 D models can be calculated from the following relationship:

K=0.5(L+M) – 0.5(L‐M)cos 2α Equation 6

Increases in vertical stress

Decreases in horizontal stress

64

p

Kp

W

H

where K = ratio of horizontal to vertical stress in plane geometry

L = ratio of major principal horizontal stress to vertical stress

M = ratio of minor principal horizontal stress to vertical stress

α = angle between roadway axis and direction of the major principal horizontal stress.

The reader is encouraged to download Examine2D from the Rocscience web site. This allows analysis of elastic stress redistribution about any shaped roadway to complement the readily available solutions for a circular excavation. Whilst the following elastic analyses give invaluable insight, it must always be remembered that rock and coal masses are not elastic and that even the smallest deformations can result in major changes in the way the stresses are subsequently redistributed.

Elastic solutions are available for the stresses induced about circular and elliptical holes (Table 17). Note that the stresses are independent of size of the hole. For a circular hole, the stresses at the roof centerline are tensile for K values of less than 0.3; for ellipses with a width/height ratio of 2:1, the critical K value is 0.5.

Table 17 Roof and side stresses for circles and ellipses

Centreline Circle Ellipse Roof or floor P(3K‐1) P(K‐1+2KH/W) Sides P(3‐K) P(1‐K+2W/H)

These simple equations provide some insight into how coal mine roofs may behave. For the general stone stress model, where the horizontal stresses are greater than the vertical and hence K is greater than 1, the implication is that roadway roofs are under compression. Importantly the coal stress model, where the horizontal stresses are less than the vertical (Table 15), implies that the roof stresses may approach the onset of tension. In the bleeder roadway and particularly the tailgate, the higher vertical stresses together with reduced horizontal stresses will mean a reduction in the K value and the trend to reduced roof stresses if not the onset of tensile stress5.

Brady and Brown (1985) extend this elastic analysis of stresses to the case of a flat lying feature (bedding parting?) located towards the top of the excavation (Figure 54). Should there be slip on such a surface, the result is a reduction in stresses acting in the crown of the excavation. If such a surface is located above the excavation, the result is an increase in the stresses acting in the crown.

5 If it is assumed that the rock or coal mass in the immediate roof has zero tensile strength because of the presence of jointing, the actual stresses cannot become tensile because of the onset of failure.

65

Figure 54 Effect of planes of weakness on distribution of roof stresses

6.2.2 ELASTIC STRESS REDISTRIBUTION AROUND A RECTANGULAR ROADWAY

The elastic stress redistribution about rectangular roadways can be readily assessed using Examine2D. It is emphasized that this simple model assumes no subsequent yielding of the rock. This should not be used as a reason to dismiss the simple elastic codes in favour of more sophisticated ones because, whilst the latter may have the numerical ability to handle yielding, the lack of validated constitutive equations and input parameters means that they are still in the research domain.

Figures 55 to 57 present the results of the analyses of a typical rectangular roadway with a number of stress fields. In these analyses the roadway is 2.8m high and 5.2m wide. Six stress ratios have been considered – K= 0.2, 0.5, 1.0, 1.4, 1.7, and 2.0, with the major principal stress being 10 MPa. Three stress components are presented: deviatoric stress (σ1 – σ3) which is the driver for compressive/shear failure (Figure 55), σh – the horizontal stress which, if tensile, would allow the onset of shear along vertical joints (Figure 56), and σv‐ the vertical stress, which is significant in terms of rib behaviour (Figure 57). A summary of the results is presented in Table 18.

Table 18 Summary of stresses for rectangular roadway (σ1 = 10 MPa)

K Horizontal stress at 0.2m into roof at centerline

(MPa)

Horizontal stress 0.1m from rib and 0.1m into roof

(MPa)

Maximum deviatoric stress 0.1m from rib and 0.1m into roof

(MPa)

Vertical stress at top of rib at 0.2m into rib

(MPa)

0.2 ‐5 8.3 19.5 19 0.5 ‐1 13.5 21 22.5 1.0 6 24 27 20.5 1.4 8 22 24 14 1.7 9 22 21 10.5 2.0 10 21 21 9 When the K value is greater than 1, the contours of deviatoric stress tend to form an arch over the roof line but this does not develop when the K values are less than 1 (Figure 55). The highest values are at the roof/rib corner with the magnitude ranging from about 30 MPa for a K value of 1.0 to about 21 MPa for higher or lower K values. It is this concentration at the roof corners that is one of the sources of stress guttering.

Stress reductions

Stress increases

66

From Figure 56, the centreline of the roof has tensile horizontal stress for K values less than 0.5, and the roof stress becomes increasingly compressive as the K value increases. Near the rib line the horizontal stresses are compressive, even for low K values. The vertical stresses imposed on the rib vary from 22.5 MPa to 9 MPa (Figure 57). When these values are considered in the context of percentage increases, the variation is relatively small – 225% to 180%. It is of value to highlight that elastic theory proposes that the stresses in the roof of an excavation may reduce and become tensile in the dominantly vertical stress field. This result is the direct opposite to that proposed by Colwell and Frith (2006) whereby the Poisson ratio effect is invoked in order to induce elevated horizontal stresses that are required for their failure mechanism. As mentioned above, the Poisson ratio effect is not applicable adjacent to excavation voids because the requirement for absolute lateral restraint cannot be met.

K = 2.0 K = 1.7

K = 1.4 K = 1.0

K = 0.5 K = 0.2Figure 55 Distributions of deviatoric stresses about a 1.86:1 roadway (σ1 = 10 MPa)

67

K = 2.0 K = 1.7

K = 1.4 K = 1.0

K = 0.5 K = 0.2

Figure 56 Distribution of horizontal stresses about a 1.86:1 roadway (σ1 = 10 MPa)

68

K = 2.0 K = 1.7

K = 1.4 K = 1.0

K = 0.5 K = 0.2

Figure 57 Distribution of vertical stresses about a 1.86:1 roadway (σ1 = 10 MPa)

Figure 58 presents a number of analyses for low K values, such that the horizontal stress magnitude is significantly less than the vertical stress magnitude (10 MPa). In this case the contours are the minimum principal stress and only the negative (tensile) values are shown. The tensile zone is weakly developed at a K value of 0.5 and it increases in height with decreasing K. Whereas the height of the tensile zones does not change with increasing stress magnitudes, the magnitude of the stresses within the envelope does.

69

Figure 58 Negative minimum principal stresses induced around a 1.86:1 roadway (σ1 = 10 MPa)

Shear stresses are developed about the roadway, and of particular interest are those that are aligned parallel to bedding surfaces and normal to the roadway centreline. Shear displacements may be induced along bedding if the shear stresses are in excess of the frictional restraint that can be developed in response to the normal load across the bedding. The bedding parallel excess shear stress (BPXS) is simply calculated from elastic models as:

BPXS = τ – σn tan(φ) Equation 7

where τ = bedding parallel shear stress σn =stress normal to bedding, φ = friction angle.

Figure 59 provides an example of how the shear, normal, and bedding parallel excess shear stresses develop along a surface 0.2m into the roof of a 5.2m wide roadway. The values are normalized to the magnitude of the far‐field vertical stress. It can be seen there are no shear stresses at the centre of span and both the normal and shear stresses increase towards the ribline. The bedding parallel excess shear stress peaks about 0.4m from the rib for a surface 0.2m into the roof and this position trends towards the centreline higher into the roof (Figure 60). Integrated across the roadway, the maximum BPXS is developed about 0.5m into the roof.

K=0.2

K=0.5

K=0.4

K=0.3

70

0

50

100

150

200

250

300

350

400

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5

Excess bedding parallel shear stress (kN/M

Pa)Stress (M

Pa/M

Pa)

Distance across roadway (m)

Normal stress

Shear stress

Excess shear stress

Figure 59 Example of shear and normal stresses and bedding parallel excess shear stress assuming a 35o friction angle and K = 2.0

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5Distance across roadway (m)

0.5

1

Hei

ght i

nto

roof

(m)

Figure 60 Contours of BPXS as a function of height into distance across the roof (expressed as MPa/MPa)

6.2.3 THREE‐DIMENSIONAL STRESS CHANGES

The analyses to date have been in 2 dimensions. In the mine, the coal face itself provides restraint to the immediately adjacent excavation and this results in changes in the way the various stress components develop. The following plots are from a 3D elastic boundary element code (Examine3D) and show the stresses developed on a surface 0.2m above the roof line.

For the case of the deviatoric stress, the maximum values are near the face line and along the rib line (Figure 61). In this analysis, the roadway is aligned at 30o to the direction of the major principal horizontal stress and the bias in the stress magnitudes towards one side of the road near the face line is apparent.

71

Figure 61 Development of deviatoric stresses at 0.2m above the roof line as the heading is advanced (K=2)

For a stress field where the vertical stress is the major principal stress, there is no bias in the development of tensile roof stresses and the peak of the tensile stresses is distant from the face (Figure 62).

Figure 62 Distribution of negative mean stress in the immediate roof of a roadway for K=0.15

For bedding parallel excess shear, most of the shear stresses develop close to the face (Figures 63, 64, 65). Note that in this 3D analysis, the shear stress that is considered is the bedding parallel shear stress normal to the roadway centreline. A key observation that must be emphasized is the rate at which BPXS develops away from the face – at a distance of 2m, some 75‐80% of the final BPXS has

72

developed. This observation is critical to understanding how roof bolts are loaded in shear – the bolts are only exposed to the shear stresses that are induced after the bolts are installed.

Figure 63 Three dimensional view of the distribution of BPXS (ignore negative sign)

Figure 64 Vertical slice through Figure 63 showing how BPXS develops (ignore negative sign)

73

0 1 2 3 4 5 6 7 8 9 10

Distance from the face (m)

-2

-1

0

1

2

Dis

tanc

e fro

m c

entre

line

(m)

Figure 65 BPXS on a surface 0.2m above the roof line

6.2.4 NON LINEAR STRESS REDISTRIBUTIONS

It is well established that coal mine roofs deform into the excavation as the roadway is advanced. This deformation zone is routinely identified with roof extensometry and is loosely referred as the “softened zone” or “the height of softening”. Whilst the term is somewhat misleading as the observations are simply of vertical movement and the origin of the movement may be dilation across bedding or the onset of compressive failure, the use of the term “softening” does have the advantage of focusing attention on what the impact may be in terms of the immediate roof stresses. Softening implies a lower modulus of deformation, which should mean that there is less of an ability to bear stresses compared to stiffer units nearby.

The scale of roof movement and the associated stress redistribution has been demonstrated recently by Mark et al (2007). These authors were able to show that even at less than 20mm of roof movement; the horizontal stresses were already redistributed into an arch over the roadway (Figure 66). To quote from the paper “The orientations of the principal stresses increases imply that the immediate roof of the crosscut yielded or “softened” and was not capable of transmitting additional horizontal stress…….The measurements made during the study showed that the additional stresses were redirected above the immediate roof of the crosscut even before significant roof deformation had occurred”. This redirection continued as the longwall retreated and the roadway was exposed to maingate stress concentrations.

The same general mechanism can be demonstrated in a 2D continuum numerical model whereby the softening is simulated by the onset of compressive failure. Figure 67a show onset of brittle failure above a rectangular roadway. If the excavation shape is changed to reflect the failure zone, it can be seen not only the stress trajectories are redirected around the failure zone but also how there may be additional failure at the crown of the simulated excavation (Figure 67b).

74

After development, prior to longwall After longwall

Figure 66 Stress measurements at Emerald Mine

(a) Failure zone and stress trajectories developed above a rectangular roadway

(b) Failure zone and stress trajectories developed if the failure zone acts as a fully softened zone.

Figure 67 Simulation of stress redistribution above a roadway using an elastic model

75

Further evidence in support of the non‐linear redistribution of stresses can be found in the data on stress relieving roadways. Gale and Matthews (1992) reported that substantial relief from lateral stress can be obtained if roadways are driven close to an earlier driven roadway (Figure 68). The quantification of roof softening was not available at the time, but experience in the Southern Coalfield is that a bolted roof – not an overall collapse – is adequate to generate the stress relief. However, the key point is that Newton’s third law requires that the stress relief is also present within the existing roadway.

Figure 68 Concept of a stress relieving roadway

This non‐linear stress redistribution is particularly problematical when considering the stresses in the roof during the formation of an intersection. The second roadway is driven in a completely different stress field to that of the first roadway such that the stresses in the intersection roof at the point of breakthrough will be less than those encountered during the straight driveage.

There is no doubt that this behaviour is a key feature of underground roadway. Unfortunately, with the current state of the art, it is difficult to incorporate it in numerical analyses, and certainly not in routine design using numerical approaches. This limits the applicability of the numerical tools and requires the application of more judgment by the design engineer.

6.2.5 STRESSES INDUCED WITHIN A BLOCKY ROOF

This discussion on mining‐induced stress so far has been based primarily on continuum concepts (linear and non‐linear). As discussed earlier, a rock or coal mass is not a continuum and it is possible that its behaviour as a discontinuous medium can significantly modify the stress around an opening. There is value in considering a block model for the immediate roof and how stresses may be induced by the reorientation of the blocks in response to the formation of the opening. The simplification to rectangular blocks that is possible with coal measures allows consideration of two simple analogues.

6.2.5.1 VOUSSOIR BEAMS

The bedded nature of coal measures allows the ready application of the voussoir beam model (Brady and Brown, 1985). The concept here is that under situations of no applied lateral force, the incipient

76

rotation of the voussoirs induces a lateral thrust in the beam (Figure 69). The magnitude of this induced lateral thrust depends on the span, density and thickness of the beam. At the point of failure of voussoir beam, the compressive stresses at the roof/rib corner approach the magnitude of the UCS of the rock. An important point to note is that the result of voussoir action is the possible development of compressive stresses at the roof/rib corner and tensile stresses at the roof centreline. An underground observer may observe the development of a “stress gutter”. The voussoir beam model will be discussed in more detail in later chapters.

Figure 69 Voussoir beam deformations induce compressive stresses at the roof corners and tensile stresses at the roadway centreline

6.2.5.2 CANTILEVERS

If the roof line is exposed to the onset of tensile stress and there is sufficient relaxation such that the joints dilate, it is possible that a cantilever will develop (Figure 70). The failure of a cantilever in this situation will be through elevated shear/compressive stresses at the roof/rib corner. Once again there is the possibility of generating compressive failure in a situation of no imposed horizontal stress at the roof line.

Figure 70 Example of a shear surface generated by cantilevering action

0.0530.0530.0530.0530.0530.0530.053

Sigma 1[MPa]

-0.50

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

4.50

-10

12

8 9 10 11 12 13 14

77

6.3 OTHER STRESS REDISTRIBUTIONS

6.3.1 TAILGATE PILLARS

If there is differential movement between the 2 sides of a roadway, the result is an increase in the bay length of the roof line (Figure 71, after Diederichs and Kaiser, 1999) and a consequence reduction in the stress acting across the roof. The scale of this effect is in the order of 1MPa to 2MPa for a differential compression of 100mm, with a greater reduction for roofs with higher modulus values. The scale of stress reduction is significant when it is recalled that the stresses at the roof line after the formation of the roadway may be already low as a result of the stress relief into the goaf and non‐linear effects discussed above. A potential location for such differential movement is when the roadway is bounded by a yielding pillar or coal fender and this is considered to be the basis of the relationship between chain pillar design and tailgate roof support discussed by Colwell (1998).

‐3

‐2.5

‐2

‐1.5

‐1

‐0.5

0

0 10 20 30 40 50 60 70 80 90 100

Stress change in ro

of (M

Pa)

Compression of ribside (mm)

5000 MPa

10000 MPa

15000 MPa

Modulus

Figure 71 Relaxation of a roof line as a result of vertical deformation in one of the sides

6.3.2 VERY HIGH STRESSES UNDER OR ABOVE PILLARS.

For multiple seam operations, and particularly when the seams are close together, there may be a significant stress footprint about the existing pillars or goaf edge. The shape of the footprint has not been measured, but there are indications from mining operations that the significant stress changes extend further than a simple analogy to foundations engineering would suggest.

Insight into the stress footprint can be gained by considering the stress under a rigid footing for isotropic and transversely isotropic materials. When a high level of transverse anisotropy induced by joints or bedding partings is considered (modelled by using a low value for the shear modulus), the distribution of vertical stress changes significantly from that for an isotropic material (Figure 72).

78

11 11User DataSigma YY

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

2.25

2.50

2.75

3.00

11 11User DataSigma YY

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

2.25

2.50

2.75

3.00

Figure 72 Stresses under a rigid footing

The pragmatic response should be to anticipate elevated vertical stresses when mining within say 30m of overlying or underlying pillars. When combined with what is known about the non‐linear distribution of horizontal stresses about voids, it would be prudent to consider the implications of a very low K value for the stress field and the possibility of induced tensile stresses in the roof (Seedsman, 2003).

6.4 COMPILED MODELS FOR STRESS PATHS IN THE IMMEDIATE ROOF

Even allowing for depth variations, there is not a unique stress field acting in a coal mine roof. The stress conditions vary depending on the lithology that is present, the stage of mining, the dimensions of the pillars, and the presence of workings in other seams. Furthermore, the stress field may change in proximity to faults and other major geological features. Care needs to be taken when interpreting point measurements at one mine site as they may be more related to small‐scale body stresses induced by block rotations and not indicative of the overall stress field.

The situation is not hopeless – it is possible to formulate some general models. These modes provide the basis from which deviations can be identified and uncertainties identified.

6.4.1 STONE ROOF – SINGLE SEAM

A model for the stress path for a stone roof exposed to an initial stress field of 6 MPa vertical and 9 MPa horizontal is discussed below for initial development, the maingate corner and the tailgate corner (Figure 73). Deviatoric stresses and minor principal stresses are considered.

At the point of excavation, the reaction to the overburden load is removed and the vertical stresses at the roof line vanish. Until the roof deforms, the horizontal stresses are not yet redistributed. High deviatoric stresses and bedding parallel shear stresses develop immediately. These stresses may induce stress guttering at the roof corners, either by compressive failure of low strength rock or by the incipient deflection of the roof beams.

As the mining face advances, say to in excess of roadway width, the roof will have deformed to a ‘final state”, or in some situations failed if overall compressive failure develops. In either case, the

Shear modulus = 250 MPa Shear modulus = 8333 MPa

79

horizontal stress in the immediate roof will have decreased to very low values and a “stress arch” developed higher in the roof. Deflections of the bolted roof will generate body stresses associated with the formation of a voussoir beam in the immediate roof (Figure 74).

At the maingate, increases in both the vertical and horizontal stresses around the retreating goaf will alter the stress arch above the roadway and this may lead to an increase in the height of softening. There is no direct increase in the horizontal stress acting in the immediate roof because that roof has already deformed and “softened’. The extra “softened” material will be an additional surcharge loading on the bolted beam which will then cause an increase in the body stresses within the voussoir arch and an indirect increase in the horizontal stresses at the roof line. This may lead to the onset of stress guttering.

DEVELOPMENT MAINGATE TAILGATE Vertical stress : Horizontal

stress = 6:9 Vertical stress : Horizontal

stress = 12:10 Vertical stress : Horizontal

stress = 25:7

Figure 73 Evolution of deviatoric and negative minor stresses during longwall retreat

At the tailgate end of the face, the imposed stress field is now dominantly vertical with a reduction in the horizontal stress due to the presence of the goaf on one side and also behind the faceline. The stresses within the stress arch decrease and there are some very high vertically‐oriented deviatoric

guttering ? Increased size of softened zone ?

Loosening of rock in crown of softened zone?

80

stresses at the roof/rib corner. The horizontal stresses in the immediate roof remain relatively constant at low magnitudes, unless the chain pillar yields. In the latter case, reaction to the body stresses is lost and the horizontal stresses vanish. Localised compressive stresses may develop if the roof structure allows the formation of cantilevers.

Figure 74 Redistributed insitu and induced body stresses about a roadway with K >0.8 once the roof and floor deflects.

6.4.2 COAL ROOF – SINGLE SEAM

The stress field in coal just ahead of mining should be considered to have all three components compressive but with the vertical stress dominant – K values of about 0.2 should be assumed, with even lower values in the presence of closely spaced joints.

This stress field results in very low horizontal stresses in the roof of the excavations and possibly the onset of tensile stress. The size of the tensile zone increases away from the face which introduces the risk of falls developing outbye. If the model of the coal stress field being related to depressurization of the coal and reloading by a gradually deforming overburden applies, there is also the possibility the vertical stress magnitude will increase outbye, but that should not alter the height of the tensile zone unless there is a reduction in Poisson ratio with the higher stress that develop outbye.

Destressed roof may show signs of elevated body stresses due to block rotations which may mask the overall behaviour (Figure 75). This is more likely if there are only a few large blocks – multiple small blocks may not be able to adequately interact. It should also be noted that a destressed coal roof may have dilated joints and cleats such that resin loss is possible.

At the maingate corner the vertical and horizontal stresses increase, but if K value stays the same the result would be no substantial change in the roof stress conditions.

With the lack of horizontal roof stresses, coal roofs are particularly exposed to any stress relaxation in the tailgate that could be introduced with a yielding tailgate pillars design.

Softened zone Vertical dead weight acting on bolted rock beam

Stress arch induced in rock beam

Insitu stresses redirected around softened zone

81

Figure 75 Redistributed insitu and induced body stresses about a roadway with K <<0.8 once the roof and floor deflects

6.4.3 STRESS PATHS IN THE RIBS

The stresses in coal are dominantly vertical throughout the longwall mining cycle. This means that the ribs are exposed to progressively increasing deviatoric stresses dominated by very high vertical stresses and negligible horizontal stresses.

Destressed roof due to onset of tensile conditions

Stress arch induced in coal beam

82

7 SUPPORT AND REINFORCEMENT TECHNOLOGIES

The terminology for the various support and reinforcement elements is rather loose, and is not consistent between mining and civil. In the following discussion the generic term “tendon” will be used to highlight the fact that these elements work in tension.

There are 2 basic approaches to the prevention of roof and rib collapse – support or reinforcement.

Support implies that the ground did not have, or has lost most if not all, any self‐supporting capacity in the stress regime that is active at the time. The term “support” may have it origin in the use of props and beams that provided support from underneath the roof. Nowadays, the full cross‐sectional area of the roadways is required and the “support” is most often in the form of suspension from a zone in the roof that is considered to be stable. Recognising that the objective is suspension, the key aspect of the installed tendons is tensile loading.

Reinforcement involves conserving or improving the overall rock mass properties and, in the context of coal measures, this becomes the prevention of shear along bedding. During the reinforcement action, the tendons should be loaded in tension.

The various technologies can be used in either a support or reinforcement mode. Essentially the key aspect is the tensile capacity of the tendon.

In mining, some deformation is acceptable and they are implicitly if not explicitly included in the design of openings. These deformations can come from the dilation of a yielding rock/coal material, from shear along bedding, or from rotation of discrete blocks. The implication of this is that the tendons must have a relatively large degree of tolerance of deformation. The analytical tools presented in this report are based on a balance of driving and restraining stresses; deformations are not considered. Validation of the tools has been through back analysis of existing technologies. The implication of any differences in the post‐yield capacity of tendons, as compared to H and X grade steel, needs to be carefully considered. Higher strength, but more brittle steel may leads to major reductions in roof conditions.

7.1 REINFORCEMENT ACTION

In a support function, the tendons can be assumed to be loaded simply by dead weight. The imposed loading is tensile controlled by the mass of the block that is being suspended and the capacity of the anchorage.

As defined above, the reinforcement mode relates to shear along bedding discontinuities. The reinforcement action can be direct shear of the tendon (guillotine), the dowel effect where the shear resistance is controlled by a bearing failure in the rock, or the friction effect (bolt) where the frictional restraint provided by a closed bedding parting can be exploited (Figure 76). Coal measure rocks are not strong enough to generate a guillotine effect.

83

Figure 76 Tendons in shear

7.1.1 DOWEL EFFECT

The dowel effect applies when the bedding surface is open and results in the controlling factor being the compressive strength of the rock. There are several different equations that describe this effect and all are similar. Bjurstom (1974) proposes that the dowel resistance (Figure 78) is given by:

Dowel = 0.67 d2 √(UCS.By) Equation 8

where By = yield stress of the steel and d = diameter of tendon.

0

1

2

3

4

5

6

7

8

0 10 20 30 40 50 60 70 80 90 100

Dow

el re

sist

ance

(ton

nes)

Rock UCS (MPa)

X grade -650 MPa

S grade -400 MPa

Mild steel - 300 MPa

Figure 77 Dowel resistance for 21mm diameter tendons as a function of UCS

7.1.2 FRICTION EFFECT

The friction effect requires closed bedding partings. Noting that the BPXS has developed to a large degree by the time any tendons are installed, there is a high likelihood that bedding partings will be open during the bolting. Installing fully grouted dowels can lock in the bedding parting open. The pragmatic ways of closing any open bedding parting is either to pretension the bolt against a point anchor, or to jack‐up the roof by reacting against the floor.

Guillotine Bolt Dowel

84

There is also the requirement that the bedding stays closed during any deformation – the trend of the plunge of the bolt needs to be vertical or in the opposite direction to the sense of subsequent shear (Figure 78). For coal mine roadways this means either vertical or fanned outwards.

Figure 78 Importance of bolt angle in maintaining closed bedding

For rational bolt orientations and shear displacements, the frictional resistance of a bolt (Figure 78) is given by:

Bolt Friction = U*tan(φ) Equation 9

where U is the tensile load in the bolt.

0

5

10

15

20

25

0 50 100 150 200 250 300 350

TENSILE LOAD IN BOLT (kN)

SHEA

R R

ESIS

TAN

CE

(TO

NN

ES)

20

25

30

35

Figure 79 Frictional shear resistance provided by bolts

Comparing Figures 79 and 77, it can be seen that the same X grade tendon can provide either say 5 tonnes (as a dowel in 45 MPa rock) or 24 tonnes (as a bolt across a closed parting with a 35o friction angle).

Increase in shear resistance due to greater frictional restraint and tensile stress in tendon

Decrease in shear resistance as a result of loss of frictional restraint

85

7.2 TENDONS

For completeness, a listing of the most common bolts used in Australia is provided below along with a table of typical bolts used in USA.

Table 19 Summary of Australian bolt types

Grade DSI X DSI H DSI S DSI AVH Jennmar JX Bolt Diameter (mm) 23.2 23.2 23.2 22.5 24 Core Diameter (mm) 21.7 21.70 21.7 20.7 21.7 Yield Strength (MPa) 650 400 300 775 650 Yield strength (kN) 240 150 110 260 240 Ultimate Tensile Strength (MPa) 920 680 475 935 890 Ultimate Tensile Strength (kN) 340 250 175 315 340 Standard elongation (%) 15 22 35 14 19 Uniform elongation (%) 8 10 12 3

Table 20 Summary of tensile strength of USA bolt types

40 60 75 90 Yield load (kN) 340 480 585 690

Reference Diameter 5 18.1mm 88 124 151 178 6 21.8mm 126 178 217 256 7 25.4mm 172 243 296 349 8 29.0mm 225 317 386 456 9 32.6mm 284 401 489 577

Ultimate load (kN) 550 720 790 930 5 18.1mm 142 186 204 240 6 21.8mm 204 268 294 346 7 25.4mm 278 364 400 470 8 29.0mm 363 476 522 614 9 32.6mm 460 602 660 777

Table 21 Long tendons

Type HiTen Flexibolt

TG Cable Bowen Cable SSB Megabolt

Features Plain strand cable resin capsules

Internal breath tube

Bulbed strand along full length

Strand with bulbed

anchorage

Straight indented wires Internal tube

Diameter (mm) 23.4 28 21.8 Bulbs: 32‐38

21.8 Bulbs : 32‐18

27‐39

Hole diameter (mm) 28 38‐42 42‐45 35‐55 35‐45 Yield strength (kN) 500 560 525 525 480‐870 Ultimate strength (kN) 580 630 590 590

N.B. Capacity may be limited by the barrel and wedge strength

86

7.3 ANCHORAGES

Overwhelmingly in the Australian coal industry, the anchorages are resin‐based to allow rapid installation. Post grouting is typically done with cementitous grouts.

The design of the anchors can be based on the ground anchor approach in civil engineering (Littlejohn, 1993). Recognising that the strength of the resins is in the order of 70 MPa, which is stronger than many of the coal measure rocks (Figure 35), failure will take place at the resin/rock interface, assuming that the steel/resin interface is adequately rough. The pull out capacity of an anchor(Tf) is given by:

Tf = c * 3.14159*d*L Equation 10

where c = UCS*(1‐sinφ)/(2*cosφ), L= length, and d = diameter of hole.

It is recommended to apply a relatively high factor of safety for cartridge based systems to account for uncertainties with mixing and gloving. Civil engineering would have a minimum of 2.0 for cement grout systems, and for mining this value would be appear to be appropriate. Figure 80 provides a simple design chart for the length of grouted anchorage for each tonne of bolt load.

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

10 15 20 25 30 35 40

Unconfined compressive strength (MPa)

Leng

th o

f anc

hora

ge @

fact

or o

f saf

ety

of 2

.0 (m

/tonn

e)

283032354555

Hole diameter

friction angle = 30o

Figure 80 Recommended minimum anchorage lengths in coal measure rocks with resin anchorages

It is noted that this approach indicates that large holes provide shorter anchorages. It is essential to note that this assumes ideal resin mixing – which may not be possible with a large annulus that could result if standard diameter bolts are used in large holes.

Equation 10 can also be used in reverse, to estimate the UCS from the results of short encapsulation length pull tests, once an assumption is made regarding the friction angle.

87

7.4 STRAPS AND PANELS – SKIN RESTRAINT

Australian coal mines use straps or mesh panels, not shotcrete. The task of straps or panels is to protect the operators from the fall of scat and to provide restraint to the roof if there is a tendency for the roof to collapse between the bolts. From a pragmatic operation perspective, straps and mesh also assist in the correct location of the bots. The straps or panels transfer vertical loads to the bolt collars. The mechanics of this transfer are complex as the loading is orthogonal to the orientation of the strap or panel.

If the roof is under compression and has not suffered overall failure, the loading on the strap or mesh is relatively low and is related to the collapse of thin slabs of the immediate roof (say in the order of 0.5 tonnes for bolts spaced 2m apart). If the roof has failed under either compressive or particularly tensile loading conditions, the loads on the skin may be significantly higher (readily in excess of 5 tonnes).

Failure of the strap or mesh will be at the bolt locations and may be due to vertical shear through a rigid membrane or tensile failure if deformations are allowed to develop – the latter is considered the likely mechanism for steel straps or panels.

Coates (1970) provides an analysis of the loading of straps or panels on the assumption that it forms a catenary. The tension (T) induced in the mesh can be estimated by:

T = Pvs2/(8q) Equation 11

where Pv is vertical pressure on mesh, s is bolt spacing, and q is the sag of the mesh.

A design chart based on this equation is given in Figure 81. Inspection of this chart shows that very high loads can be readily induced in a strap or mesh panel especially if large deflections do not develop. The more the membrane deflects, the lower are the induced loads.

There are no data on the loading and failure of coal mining panels, but there are data on simple mesh panels used in the metal mines (Figure 82). For the mesh of 5mm diameter wires, the ultimate capacity is about 2 – 3 tonnes after about 100mm to 300mm deflection. Note that each weld is typically certified at 0.85 tonnes, so the ultimate load represents only about 3 welds – this is not surprising once it is recognised that ultimately the load on the mesh is transferred to 2 wires reacting against a bolt or cable collar.

For W straps, reference can be made to the bearing capacity of bolts in sheet steel, whereby the bearing resistance per bolt, Br, can be computed in accordance with the following equation;

Br = C d t Fu Equation 12

where C = 3 for this geometry, d = nominal bolt diameter, t = sheet thickness, and Fu = tensile strength of steel sheet.

For 350 grade 1.9mm thick steel sheet, this represents 4.2 tonnes. This suggests that W straps could be twice as effective a mesh panels in this loading condition.

88

0.1

1

10

100

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Applied load (tonnes)

Indu

ced

load

in m

embr

ane

(tonn

es)

50mm deflection, 0.5m bolt spacing50mm deflection, 1m bolt spacing50mm deflection, 2m bolt spacing100mm deflection, 0.5m bolt spacing100mm deflection, 1m bolt spacing100mm deflection, 2m bolt spacing200mm deflection, 0.5m bolt spacing200mm deflection, 1m bolt spacing200mm deflection, 2m bolt spacing

Figure 81 Loading of a strap or panel if loaded as a catenary

Figure 82 Results of loading 1.5m and 2.0m square mesh panels (Thompson, 2004)

89

8 PREVENTION OF ROOF COLLAPSE

The previous sections have outlined the range of geological and strength conditions that may be present in the roof of a coal mine and have discussed the range of stresses that can be applied or are induced about a coal mine roadway. This section provides specific steps in a framework for the specification of support or reinforcement to prevent roof collapse.

The basis for the framework is the “logical methodology for mine excavation design” presented in Brady and Brown (1985) and subsequent editions. The logical methodology was initially proposed for massive rock cut by one or two discontinuities, but in later editions its application was extended to moderately jointed rock. Brady and Brown also emphasize that the presence of failed zones around a mining opening is common; the mining problem is not necessarily to prevent failure but certainly to prevent uncontrolled displacement of rock from the boundary of the excavation.

There are 4 fundamental steps in the framework that lead to 5 different roof support designs (Figure 83). Because of the complex stress path that a coal mine roof can undergo during longwall extraction, the steps in the framework may need to be applied at least 3 times – initial roadway formation, maingate corner, and tailgate corner. The steps are:

1. Check for the presence of non vertical joints – the operational concern is that such joints can define “key blocks” that are unstable in any stress regime and may be difficult to support with the support hardware and drill rigs on standard continuous miners. Such joints are more likely to be present around faults, but should be anticipated to be present anywhere.

2. Check for the possibility of overall compressive failure. This condition is the major hazard at the maingate. It may also develop on initial roadway formation, particularly at low depths of cover where weathering/alteration effects have reduced rock strength. It would appear that current mining depths in Australia (less than 550m) do not induce such failures in “standard” roof rocks.

3. Check for the possibility of tensile horizontal stresses in the roof. This condition should be anticipated for coal roofs, tailgates where the chain pillar is designed to yield, under pillar and goaf edges in multiple seam mining, and when mining close and parallel to valleys and highwalls.

4. Assess whether large unsupported spans can be formed. It will be shown that a spanning unit needs to be only 0.25m ‐ 0.3m thick. While intrinsic geological variability does not allow the assumption that such roof is always present, it may be possible to test its presence by taking an extended cut (unsupported) longer than the roadway span.

5. If such a test cannot be made on each cut, the roof design must be based on the assumption that thinly bedded units are present that need to be reinforced to stop bedding parallel shear. Because the behaviour of bedded roof is non‐linear and there are major stress redistributions in the roof after only very small deflections, this reinforcement needs to be considered only at the time of initial roadway formation.

90

NON-VERTICAL JOINTS ? Y

N

COMPRESSIVEFAILURE ? Y

N

TENSILEFAILURE ? Y

N

CHANGE LAYOUTSUSPENSION

STABLE OVEREXTENDED

CUT ?Y

N

REINFORCEMENTSKIN RESTRAINT

SUSPENSIONCROSS MEMBER

SUSPENSION LONG TENDONSCROSS MEMBER

SKIN RESTRAINT

Figure 83 Logical framework applied to coal mine roof support

The analyses contained within the 4 steps utilise relatively simple elastic stress models applied to a number of pre‐identified failure modes. A limit equilibrium approach, balancing driving and restraining stresses, is used to quantify the support or reinforcement requirements. In common with good engineering design practice and where possible, calibration/verification is sought by reference to actual mine experiences and this is used to provide guidance on the initial selection of factors of safety.

91

8.1 NON‐VERTICAL JOINTS

For the mining geometries and roof bolters that are used in Australia, the presence of a joint or a series of parallel joints dipping at less than a certain threshold and normal to the roadway centerline is a major hazard in any roof stress regime (Figure 84). For a compressive stress regime, shear can develop along the joints; conversely for a tensile regime the blocks can fall under gravity. Furthermore, a standard roof bolt pattern is unlikely to intersect the joints and provide adequate reinforcement. Note that loading on the skin restraint (straps or mesh) could readily exceed the available capacity.

Figure 84 The hazard of parallel non vertical joints.

The model in Figure 84 is only in 2 dimensions. There is a need to consider the trend or strike of the joint with respect to the roadway axis. Direct analogy with rock slope design practice would suggest that the hazard is reduced if the strike of the joints is greater than 20o from the trend of the roadway. The authors are not aware of any specific work on the appropriate angle for roof geometries so recommend this analogy should be validated by mine site observations and not relied on without challenge.

Non‐parallel joints can form wedges or triangular roof prisms that will simply fall from the roof (Figure 85). Whilst this hazard cannot be ignored, in some ways these represent a lesser hazard in the coal mining sector as they will tend to collapse prior to the installation of the support. Furthermore this condition will most likely be related to the presence of faults and the support regime will have already been modified.

92

Figure 85 Non‐parallel joints defining triangular prisms

8.1.1 COLLAPSE MODEL

The system shown in Figure 85 collapses when shear develops along the joints.

There is a need to consider the stress conditions that are active in the roof. Close to the mining face, before the roof has deflected, the stress field near the roof line with a stone roof will be characterized by a high horizontal compressive stress. For coal roof, the horizontal roof stresses may already be very low. Back from the face, and also in the tailgate, the roof may relax and the horizontal stresses vanish. In that case, the major stress will be vertical – related to the self weight of the block.

For the compressive stress regime, the critical dip angle in the immediate roof is 90 ‐ φ, where φ = friction angle on the joint. For typical roof lithologies and joint roughness, an appropriate critical minimum acceptable dip is 65o. This means that the hazard could develop if the joints dip at less than 65o.

Figure 86 is based on the method of Brady and Brown (1985) for symmetric triangular prisms and can be used to estimate the likely roof stresses that were present prior to a joint‐bounded fall. Note that for these triangular prisms a relatively low horizontal stress stabilizes the roof. The reference provides the basis for more detailed analyses of different prisms.

45

50

55

60

65

70

75

80

85

90

0.01 0.10 1.00 10.00

Horizontal stress in the roof (MPa)

Sta

ble

dip

of jo

ints

form

ing

a sy

met

rical

tria

ngul

ar ro

of

pris

m

25 degrees35 degrees45 degrees

1m high prism,density = 2.5 t/m3

Figure 86 Relationship between joint dip, joint friction and horizontal roof stress for a stable symmetric prism

93

8.1.2 SUPPORT DESIGN

It is not possible to give specific recommendations on how this sort of ground can be supported.

The orientation of the joints and the available bolt angles makes it unlikely that the joints can be reliably intersected and then reinforced with bolts.

One approach could be to invoke suspension from more stable ground. The mass of potential joint bounded blocks can be readily determined from knowledge of the joint geometries and bolt/cable density and length assessed.

An alternative is to strap or mesh at the roof line across the joints. In this approach would be necessary to consider the loads that may be induced on the membrane. A more closely‐spaced bolt pattern will probably be needed to achieve this outcome.

8.1.3 COMMENTARY ON DESIGN TO PREVENT COLLAPSE WITH NON‐VERTICAL JOINTS

Non‐vertical joints in the roof are a particular hazard that should not be ignored. The actual conditions that a mining crew may encounter may be difficult to characterize in advance. Furthermore, the signs of the presence of such joints may be subtle and difficult to identify. It is essential that the crews are made aware of the hazard and the need for close observations.

Many zones of non‐vertical joints will be found to be related to fault structures. If they can be anticipated in advance, it would be good practice to seriously consider the alignment of roadways so that they are at least 20o away from the fault trends.

Recognising the severity of the hazard, trigger action response plans (TARPS) for the face crews should be based on either:

• Stop mining and seek advice if dipping joints are trending within 20o of the roadway trend, or

• Install a relatively intense pre‐determined support pattern.

The presence of non‐vertical joints is the defining feature of metaliferous mines (Figure 5). By recognizing the difficulties in securing wedges, it is perhaps understandable why that sector has tended to use rock mass rating systems as the basis of their roof support design. Equally important is the recognition of the advantages of the drill and blast mining system in such rock masses – extended cuts, high blast vibration to dislodge joint blocks, and a greater operational tolerance of overbreak.

8.2 COMPRESSIVE FAILURE

In this failure mode, there is an overall compressive failure of the rock mass induced due to high deviatoric stresses compared to the compressive strength of the rock. The failed volume of rock is then assumed to undergo a gravity‐driven collapse. Roof collapse is prevented by the suspension of the rock mass from unfailed material higher up.

94

To maximize the ease of use of this compressive failure mode, the rock mass will be considered as an elastic isotropic continuum with brittle failure properties. The validity of these simplifications are tested against two published case studies. The simplifications allow this failure mode to be readily examined using freely available software (Examine2D). This report presents some simple equations that can be used to rapidly assess the likelihood of compressive failure. A more accurate analysis could be obtained using more sophisticated codes such as Phase2 or FLAC that allow layering of different materials. Such effort will produce a fuller understanding of the mechanics of roof condition but may produce only marginal improvements in the support design due to the limitation of being able to characterise the materials to enable more sophisticated plasticity approaches.


Figure 87 presents the results of two Examine2D models for a brittle material with a UCS of 10 MPa. Contours of the strength factor are presented. The classic representation of stress guttering can be seen whereby there are zones of more failure (low strength factors) at the roof corners. In the first case, it is possible that roof bolts could be used to suspend the fall mass, in the second case longer tendons would be required.

Figure 87 Two analysis of brittle strength factor using Examine2D showing zones of brittle failure and possible bolting and cable patterns.

A series of analyses of a 5.2m wide, 2.8m high roadway in 10MPa rock have been conducted with the major principal stress of 10 MPa, and the two other stress components equal (Table 22). In general, low values of the strength factor define zones at the corners and higher values extend across the roadway. A series of analyses with different K values allowed the production of Figure 88 which shows the maximum height of the various contours of the strength factor for different K values. The surface defined by these 3 variables is distinctively non planar.

Table 22 Data base for brittle failure analyses

Horizontal stress (MPa)

Vertical stress (MPa)

Strength factor Maximum height (m)

K RSI @UCS=10MPa

10.00 5.00 0.525 6.37 2.00 3.81 10.00 5.00 0.5 5.57 2.00 4.00 10.00 5.00 0.45 4.39 2.00 4.44 10.00 5.00 0.4 3.42 2.00 5.00 10.00 5.00 0.35 2.54 2.00 5.71 10.00 5.00 0.3 1.33 2.00 6.67 10.00 6.67 0.65 6.44 1.50 2.31 10.00 6.67 0.6 5.65 1.50 2.50 10.00 6.67 0.55 4.90 1.50 2.73 10.00 6.67 0.5 4.21 1.50 3.00 10.00 6.67 0.4 2.92 1.50 3.75

95


Vertical stress (MPa)

Strength factor Maximum height (m)

K RSI @UCS=10MPa

10.00 6.67 0.35 2.24 1.50 4.28 10.00 6.67 0.32 1.44 1.50 4.69 10.00 7.14 0.7 6.50 1.40 2.00 10.00 7.14 0.6 5.21 1.40 2.33 10.00 7.14 0.5 4.03 1.40 2.80 10.00 7.14 0.4 2.86 1.40 3.50 10.00 7.14 0.35 2.16 1.40 4.00 10.00 7.14 0.33 1.63 1.40 4.24 10.00 7.14 0.32 1.46 1.40 4.38 10.00 6.25 0.6 6.19 1.60 2.67 10.00 6.25 0.5 4.47 1.60 3.20 10.00 6.25 0.4 3.05 1.60 4.00 10.00 6.25 0.35 2.32 1.60 4.57 10.00 6.25 0.31 1.36 1.60 5.16 10.00 8.33 0.8 6.05 1.20 1.50 10.00 8.33 0.7 5.28 1.20 1.71 10.00 8.33 0.6 4.47 1.20 2.00 10.00 8.33 0.5 3.61 1.20 2.40 10.00 8.33 0.4 2.62 1.20 3.00 10.00 8.33 0.35 1.98 1.20 3.43 10.00 8.33 0.34 1.77 1.20 3.53 10.00 10.00 1.2 6.44 1.00 0.83 10.00 10.00 1.1 6.08 1.00 0.91 10.00 10.00 1 5.72 1.00 1.00 10.00 10.00 0.9 5.30 1.00 1.11 10.00 10.00 0.8 4.86 1.00 1.25 10.00 10.00 0.7 4.39 1.00 1.43 10.00 10.00 0.6 3.88 1.00 1.67 10.00 10.00 0.5 3.23 1.00 2.00 10.00 10.00 0.4 2.39 1.00 2.50 10.00 10.00 0.35 1.73 1.00 2.86 10.00 5.55 0.55 6.10 1.80 3.28 10.00 5.55 0.5 4.97 1.80 3.60 10.00 5.55 0.45 4.05 1.80 4.00 10.00 5.55 0.4 3.26 1.80 4.50 10.00 5.55 0.35 2.43 1.80 5.15 10.00 5.55 0.31 1.36 1.80 5.81 10.00 7.52 0.7 6.02 1.33 1.90 10.00 7.52 0.6 4.92 1.33 2.22 10.00 7.52 0.5 3.88 1.33 2.66 10.00 7.52 0.4 2.80 1.33 3.32 10.00 7.52 0.35 2.11 1.33 3.80 10.00 7.52 0.325 1.55 1.33 4.09 10.00 7.69 0.75 6.36 1.30 1.73 10.00 7.69 0.7 5.83 1.30 1.86 10.00 7.69 0.6 4.80 1.30 2.17 10.00 7.69 0.5 3.84 1.30 2.60 10.00 7.69 0.4 2.76 1.30 3.25 10.00 7.69 0.35 2.07 1.30 3.71 10.00 7.69 0.325 1.52 1.30 4.00 8.33 10.00 1.1 4.87 0.83 0.91 8.33 10.00 0.9 4.11 0.83 1.11 8.33 10.00 0.8 3.75 0.83 1.25 8.33 10.00 0.7 3.38 0.83 1.43 8.33 10.00 0.55 2.62 0.83 1.82 8.33 10.00 0.5 2.42 0.83 2.00 8.33 10.00 0.45 2.11 0.83 2.22

96

0

1

2

3

4

5

6

7

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2

Strength factor - brittle failure

Max

imum

hei

ght o

f abo

ve ro

of li

ne (m

)

2.001.801.601.501.401.301.201.000.83

Horizontal/vertical stress ratio (K)

Figure 88 Relationship between height of strength factor and the K value for a 5.2m by 2.8m roadway and a major principal stress of 10 MPa

The simplicity of the brittle rock parameters (only a single compressive strength parameter, as the friction angle is zero) combined with an elastic analysis allows a number of easy parametric studies as it can be shown that the contours in Figure 89 are constant for the same strength/stress ratios.

The roof strength index (RSI) is defined as the ratio of the unconfined compressive strength of the rock to the vertical stress. Plots of RSI against height of failure for different K values (Figure 89) indicate that they may define a surface that is close to planar, at least for heights of greater than 1.5m. A multiple linear regression gives the following relationship for the height of failure defined by a strength factor = 1.0 (Hf):

Hf (m) = 1.25 + 6.12K ‐ 1.94RSI (r2=0.97), Equation 13

only for heights greater than 1.5m.

The equation is extrapolating for K values of less than 0.83, and should not be used for K values less than 0.6. A comparison of the actual and fitted data shows that there is a systematic error in the fit, with a potential error of 0.5m, which is judged to be acceptable. A brief study of the impact of different roadway heights suggests that there are small differences in the strength factor contours near the roof/rib corners but no significant differences in the contours for heights above 2m (Figure 91).

The roof strength index is defined in this way as it is readily calculated from standard geophysical logs – the UCS can be calculated from the sonic velocity logs and the vertical stress can be calculated from the density log. The RSI has the same formulation as the Competence Factor (Muirwood, 1972).

97

1.5

2

2.5

3

3.5

4

4.5

5

5.5

6

0 1 2 3 4 5 6 7

Roof strength index

Max

imum

hei

ght o

f fai

lure

(m)

2.00

1.80

1.60

1.50

1.40

1.30

1.20

1.00

0.83

Horizontal/vertical stress ratio (K)

Figure 89 Height of failure as a function of roof strength index for various K values

0

1

2

3

4

5

6

7

0 1 2 3 4 5 6 7

Multiple linear regression value

Num

eric

al a

naly

sis

valu

e

Figure 90 Comparison between the results of multiple linear regression and the numerical data

Figure 91 Slight changes in failure zones near the excavation with no difference beyond 2m

2.2m high 2.8m high 3.3m high

98

8.2.2 SUPPORT DESIGN

Rock that has undergone compressive failure should not be assumed to have any self supporting capacity. The appropriate strategy is to suspend the failure mass from above. The design requires a prediction of the height and weight of the failed zone.

An equation for the height of failure based one specific roadway dimension has been determined and it would appear to be applicable for different roadway heights. It can used to estimate the length of the tendons and the required capacity.

The height of compressive failure is either,

Hf(development) = 1.25 + 6.12K ‐ 1.94RSI Equation 14

Hf (longwall) = 1.25 + 6.12(K*(Lf/Mv) ‐1.94RSI/Mv Equation 15

where Lf = the horizontal stress concentration at the maingate corner (Figure 52), and Mv = maingate vertical concentration (Figure 50, assume= 2?).

The shape of the strength factor contours is not readily amenable to curve fitting to allow the contained area to be determined by integration. It is assessed that a reasonable approximation to the area under the curve is 0.7 times Hf * roadway width.

An alternative approach is to determine the RSI value which defines a height of failure that potentially overrides the primary bolts. For example, allowing for the bolt tail and the anchorage, the active length of a 2.1m bolt may be 1.8m. Setting Hf to this value on development, the minimum RSI that would permit primary bolt only is:

Minimum RSI (2.1m bolt) = 3.15 K – 0.28. Equation 16

The following is an example of how the support design could progress:

Input: Depth of 250m, sonic velocity = 3240 m/sec, indicated UCS = 27, average density = 2.4 t/m3. L = major horizontal/vertical stress ratio = 1.6, M = minor horizontal/vertical stress = 1.4, roadway driven at 20o to the major principal horizontal stress. Secondary support is to use 45 tonne cables installed in 28mm holes.

Design values: Vertical stress = 250*2.4*0.0098 = 5.88MPa, RSI = 27/5.88 = 4.59, Mv = 2.0, Lf = 1.75, K = 1.42. Assume a friction angle of 30o.

Output: Height of compressive failure – development = 1.1m, mass of failed volume – development = 9 tonnes, height of compressive failure – maingate = 4.4m, mass of failure volume – maingate = 39 tonnes. Anchorage length = 1.31m.

Interpretation:

The height of compressive failure on development is 1.1m which is less than the length of the primary roof bolting that is to be used. The dead weight of 9 tonnes is very much less than the installed capacity of the primary bolting. No additional support on development is required. The height of compressive failure at the maingate is 4.4m which is higher than the length of the roof bolts. Long tendon support will be required prior to longwall retreat. A pattern of 2 by 50 tonne cables every 2.4m would be adequate to suspend the potential fall mass, and a pattern of 2 every 2m would give some

99

contingency. Long tendons would need to be 6m long if installed along the centreline or less if installed in pairs across the roadway.

Case Study #1 Kestrel Mine

The roof strength index has been used in a study of maingate failure in Kestrel Mine in Central Queensland (Gordon and Tembo, 2005). The trend at this mine has been for long tendon support to be required at shallow depths and not at greater depths where the overall stresses would tend to be higher. It was found that an RSI value of 3.5 was a good predictor of the need to install additional roof bolts on development and to install cables prior to longwall retreat. A value of less than 3.2 was a good predictor for the need to install cables on development. At Kestrel, the UCS was predicted from sonic velocity logs.

The minimum RSI value at the adjacent Gordonstone operation was about 2.5 and this operation was characterized with very difficult development roof conditions requiring long tendon support on development. Similar low strength roof units are also present in the adjacent Crinum Mine (payne, 2008)

In these operations the roof bolts were 2.1m long. The horizontal to vertical stress ratios was 1.2 and 1.6 and the gate roads were aligned sub‐parallel to the major principal horizontal stress.

The figure shows the relationship between the RSI and the height of brittle failure on the assumption that the longwall stress concentration was 1.2. Recognising the variations in roof strengths and the uncertainties in the stress information, it is considered that there is reasonable agreement between the mining conditions and a mode of when the bolts could be over‐ridden by a failure zone.

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.0

RSI

Hei

ght o

f com

pres

sive

failu

re (m

)

DevelopmentMaingate

Figure CS1‐1 Height of compressive failure as a function of the RSI for the stress assumptions at Kestrel

100

Case study # 2 ‐ Emerald Mine

Emerald Mine has been referred to in the body of this report in the context of measured stress directions. The paper also includes information on the stress magnitudes and the height of softening measured with extensometers. The UCS ranges between 24.4 MPa ‐36.4 MPa, the immediate roof is coal and then shales and clays. The roadway dimensions are 4.9m by 2.4m.

This stress data indicates an initial K value of 1.78 and then both the vertical and the horizontal stress increase equally.

The predicted height of brittle failure can be calculated as a function of the horizontal stress magnitude for different UCS values. It can be shown that the roof displacement data is consistent with the brittle failure model if the UCS is in the range of 30 to 35 MPa (which is the measured range).

1.5

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

7

7.5

8

10 11 12 13 14 15 16 17 18 19 20


Hei

ght o

f fai

lure

usi

ng R

SI m

odel

(m)

UCS = 20 MPaUCS = 25 MPaUCS = 30 MPaUCS = 35 MPaMark et al (2006), Figure 15

101

8.2.3 COMMENTARY ON DESIGN TO PREVENT COMPRESSIVE COLLAPSE

The design assumption is that the collapse zone can be suspended. This requires that the immediate roof line has some load‐carrying capacity so that the vertical loads can be transferred to the cables. The primary bolting and mesh is probably adequate for this role so long as the cables are located to minimise spans.

The three case studies from the literature suggest that the prediction of compressive failure rising above the bolt anchorage is in reasonable agreement with some mining conditions. It is noted that there is some uncertainty in the estimation of the UCS from the sonic velocity logs. This would suggest that any “factors of safety” should be close to unity.

Advocating compressive failure as the source of maingate instability, which is the implication of the RSI approach provides explanations to a number of key observations with respect to depth and stress:

• At the same mine, with the same roof types, there is no strong relationship between the need for secondary support and increasing depth.

• In the industry data base, there is a wide divergence in the use of long tendon support.

• Compressive failure higher in the roof is consistent with stress reductions within the “softened zone”.

Case study #3 Crinum Mine

At the adjacent Crinum Mine a move from a 6 bolt to 6:2 pattern was found to be necessary when the RSI approached 3.5 (Payne 2007). Once again, at Crinum the UCS was predicted from sonic velocity logs. Figure CS2 ‐ 1 shows the shape of the failure zone and the location of the 6 bolt pattern (thin lines) used at Crinum. The inner 4 bolts possibly do not have adequate anchorage. The thick lines show the location of the additional 2 bolts. It can be seen that these bolts have more of the anchorage outside the failure zone.

Figure CS3 – 1 Strength factor for RSI = 3.5 under development conditions (K=1.2)

102

• Compressive failure higher in the roof can induce stress guttering at the roof line as a result of vertical loading of the bolted roof.

• The magnitude of the bedding parallel shear stress that would be induced at the maingate if there was no relaxation in the softened zone exceeds the capacity of the bolting patterns used.

Coal measures are transversely anisotropic, especially for the finer grained units. Transverse anisotropy can increase the roof stresses such that the failure heights increase (Figure 92), so ignoring them is non‐conservative to the roof support design. Because of this, there may be a need to apply higher factors of safety in highly bedded material if using the simple equation and prior to field verification. Alternatively, a separate analysis using Examine2D or more sophisticated codes could be considered. As discussed earlier in this report, there is little guidance on the choice of the independent shear modulus.

(a) Isotropic elastic

(b) Transverse anistropic elastic (G=200 MPa)

Figure 92 Different isotropy and failure conditions

103

8.3 TENSILE FAILURE

The important to point is that a rock mass in compression may behave as a stable continuum. In a destressed state, small imposed or gravitational loads can cause large displacements of component rock units. (Brady and Brown, 1985)

A rock mass must be assumed to have zero tensile strength (by virtue of its discontinuities) so the development of a zone of elastic tensile stress must be assumed to be a zone of failure and associated redistribution of stresses. Whether a roof collapse develops depends on the nature of the joints sets – closely spaced joints, particularly if oriented parallel to the roadway axes will promote a collapse.

Since the collapse mode requires shear along vertical joints, it is appropriate to examine the elastic horizontal stresses that are developed in the immediate roof (Figure 93). As discussed in an earlier chapter, the stresses will become negative (tensile) when the K ratio is less than 0.65. Note how the stresses are still compressive at the roof/rib corner and that the maximum height is at the centre of the roadway. The height of the tensile zone (Ht) increases as the K value decreases (Figure 94) but does not change with increasing magnitudes of the stresses. Slightly different relationships exist for different roadway aspect ratios.

An approximate relationship between the K value and height of the tensile zone is:

Ht =2.3K2 ‐4.74K +2.12 (for K between 0.1 and 0.62). Equation 17

Figure 93 Negative horizontal stress (K=0.2) for a 5m by 3m roadway

104

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

K value

Maxim

um heigh

t of zon

e of negative

horizontal stress (m

)

5m by 3.5m

5m by 3m

5m by 2.5m

5m by 2m

Figure 94 Height of negative horizontal stress (5.5m by 3.3m roadway)

K values of less than 0.65 may be more common than expected in coal mines. Certainly, the stress fields measured within coal seams in advance of the mining face have such K values (Table 15). Other locations where the mining geometry results in an increase in vertical stress and a reduction in horizontal stresses include close proximity to large excavation “voids” where non‐linear behaviour of the overburden results in the development of stress arches higher in the roof and lower into the floor. Possible locations include tailgates where there is an adjacent goaf and under pillar goaf edges in multiple seam operations. There is a particular hazard if pillars are designed to yield.

8.3.1 COLLAPSE MODE

The collapse mode is simply the gravity‐driven collapse of joint bounded blocks (Figure 95) or, in the case of a tailgate, may be the gravity collapse of blocks defined by the onset of brittle failure at the maingate. In the case of joints, whether the collapse occurs may be related to the kinematic acceptability of the blocks (Figure 96) whereby block interactions may result in a metastable roof that has adequate serviceability. It is apparent that roadway orientation has the potential to be a major control on collapse.

Figure 95 Collapse mode for tensile stress regime

105

Figure 96 Roadways oblique to joint sets will produce better conditions in both headings and cutthroughs

8.3.2 SUPPORT

Roofs exposed to tensile stresses are particularly difficult to stabilize. The quote at the beginning of this section highlights the fact that such roofs can appear to be stable and then collapse with little if any warning. Visual inspections and TARPS cannot be relied on as there may be no slow deterioration and no noise.

The support strategy should be to provide a tensile member in the roof (strap or mesh) and then to suspend from the zone of compressive stresses above the roadway which also has an adequate spanning capability. In terms of dead weight loading, the bolting density would be very low, but there is a need to consider bolt spacings and the loading of the straps or mesh. With the roadway geometries under considerations, it is not practical to attempt to reinforce the vertical joints themselves. The loading on the bolts will not be high, the major practical consideration is the loading on the strap or panel.

Figure 97 Loading on mesh panels

8.3.3 COMMENTARY ON DESIGN TO PREVENT TENSILE COLLAPSE

Supporting roofs exposed to tensile stresses is relatively difficult as it is easy to overload the mesh or strap. The use of standing support would remove a number of uncertainties – but is it compatible with the roadway use?

106

If examining this collapse mode in numerical models, you should examine plots of horizontal stress to assess possibility of movement on joints, plots of negative σ1 and mean stress, and the distribution of failure zones provided that the tensile strength is set to zero (in the Mohr‐Coulomb criterion).

In a zone of tensile roof stresses, it is possible for cantilevers to form which can then fail (Figure 98). This failure mode can be deceptive as the onset of collapse, as viewed from the roadway, may involve compression. Given the hazards intrinsic in tensile failures, it is important that the full geometry of a roof fall is examined before a mechanism is decided upon.

Figure 98 Generation of compressive stresses and failure in a regime of no imposed horizontal stress

8.4 DELAMINATION FAILURE

In this section we analyse the impact of the presence of bedding partings. Partings define jointed rock beams. These beams may be able to span across the coal mine roadway or, if they too thin, there is a need to use roof bolts to create a composite beam. This composite beam can be created if shear along the bedding partings is prevented.

Unless it can be demostrated otherwise, geological variability is such that it should always be assumed that close to extremely‐close spaced bedding partings will be present just ahead of mining. Extended‐cut mining can provide such a demonstration. For immediate face bolting, where bolting is conducted within about 2.7m of the face, this cannot be demonstrated and, furthermore, there are still changes in the stresses acting in the roof at the position at which the bolting is conducted.

As the roadway is formed, shear can develop along bedding partings, the bedding may open, and individual beams of rock develop (Figure 99). If the beams are too thin, they may collapse.

So there are two parts to the design exercise – identifying the characteristics of a spanning unit and determining how to create such a unit if there is insufficient confidence that one is present.

107

Figure 99 Slip and separation in a layered roof rock (Brady and Brown, 1985)


The mechanics of the jointed rock beam is shown in Figure 99 – failure can be by shear along joints, compressive failure through the rock substance, or by snap‐through (buckling). Voussoir beam or linear arch analysis can used to analyse the mechanics: a number of methods are available ‐ Brady and Brown (1985), Cpillar (www.rocscience.com), or Sofianos and Kapensis (1998). The latter is amenable to spreadsheet procedures and has been used in this report. A listing of the spreadsheet is included in an appendix.

Figures 100 and 101 present some typical relationships. As expected, the critical thickness (factor of safety = 1.0) increases with span and decreases with rock strength. For a 5.5m span, and a typical strength of 40 MPa, the necessary thickness without a surcharge is about 0.1m, and this increases to 0.3m for a 2m surcharge or a 10m span and a 1m surcharge. The associated deflections are about 20mm. The authors’ experience is that these numbers are reasonably consistent with interpreted field behaviour.

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

1 2 3 4 5 6 7 8 9 10

Span (m)

Crit

ical

thic

knes

s (m

)

UCS = 10 MPaUCS = 20 MPaUCS = 40 MPaUCS = 60 MPa

Figure 100 Critical thickness and deflection of voussoir beams as a function of span and rock strength (E/UCS=250)

108

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 0.5 1 1.5 2 2.5 3 3.5Surcharge (m)

Crit

ical

thic

knes

ss (m

)


Figure 101 Critical thickness and deflection of a 5.5 m span voussoir beam (1m surcharge)

Based on Figures 100 and 101, some simple rules of thumb can be derived – the required thickness is about 2.5% to 4% of the span, and the resulting deflection is about 0.5% of the span.

The critical thickness quoted above relates to a factor of safety of 1.0, and in most cases against compressive failure. Greater thicknesses result in a rapid increase in the factor of safety (Figure 102). A 30% ‐ 40% increase in thickness results in change from 1.0 to 2.0 in the factor of safety.

1

2

3

4

5

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Thickness (m)

Fact

or o

f saf

ety

(com

pres

sion

)


Figure 102 Stability of a voussoir beam increases with increasing thickness (1m equivalent surcharge)

8.4.2 REINFORCEMENT DESIGN

If a spanning unit has not or cannot be determined to be present, there is a need to install reinforcement across the bedding surfaces so that an adequately spanning beam can be formed. Determining the stresses to be applied to the reinforcement element is complex involving considerations of the full three dimensional stress field in a discontinuous material which is known to

109

behave non‐linearly. In the following analysis, a 3 dimensional elastic continuum code will be utilized. This approach is assessed to be conservative (i.e. will over estimate the bedding parallel shear stress) as it does not account for the established stress redistributions to higher in the roof and relaxation near the roof line once the roof deflects. It will be assumed that at the development mining face the roof beam is exposed to the full elastic horizontal stress regime.

Section 6.2.2 discussed bedding‐parallel excess shear stresses (BPXS) and how they develop away from the face. Figure 64 highlights the fact that about 60‐70% of the BPXS have developed by the time the roadway has advanced 2.7m. If there are bedding partings present in the zone of positive BPXS, shear movements will have occurred. It follows that bolts are installed in a rock mass where movements have developed. The bolts will only be exposed to subsequent potential movements and the associated stresses.

8.4.2.1 DRIVING FORCES

A series of analyses have been conducted using Examine3D (www.rocscience.com.au) with a grid of prediction points near the face and a plan located 37.5m distant (Figure 103). The roadway was 5.0m wide and 3m high, and a Poisson’s ratio of 0.25 was used (more details in Appendix B). The Suu and Sue stress tensors were imported into an Excel spreadsheet to be manipulated to generate data files of positive BPXS for a range of friction angles. Simple incremental linear assumptions were used to integrate the area under the curves. Because this is a linear elastic code, the various stress components can be normalized – in this case to the magnitude of the vertical stress.

Figure 103 Examine 3D geometry

The approach can result in a variety of plots. Figure 104 shows the distribution of BPXS that will develop after the face has advanced 2.31m for 3 different heights into the roof. The peak values are about 0.5m from the rib line, which is closer to the rib line when compared to the simple beam models (Stimpson, 1987). The data can be further manipulated to show the progressive increase in BPXS with distance from the rib line (Figure 105)

Plane strain

Mining face

110

2.31M FROM FACE LINE

0

0.1

0.2

0.3

0.4

0.5

0.6

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Distance across roadway (m)

BPX

S @

35

degr

ee fr

ictio

n an

gle

(MN

) 0.2M INTO ROOF

0.4M INTO ROOF

0.6M INTO ROOF

Figure 104 BPXS at 2.31 m from the face

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.5 1 1.5 2 2.5DISTANCE FROM RIB LINE (m)

INC

REM

EN

TAL

BPX

S (%

)

Figure 105 Cumulative increase in BPXS with distance from the rib line

The average BPXS decreases as an inverse function of the distance from the face raised to the fourth power (Figure 106). For typical bolting locations (about 2.3m to 2.7m from the face), the BPXS represent about 20‐25% of the plane strain value. Figure 107 shows how the BPXS varies with both height into the roof and timing of the bolting. It can be seen that there is a “ridge” of high values about 0.5m ‐ 1.0m into the roof for about the first 2m of the driveage. At typical bolting locations, the ridge is very much less prominent.

111

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 2 4 6 8 10 12Distance from the face line (m)

Ave

rage

BPX

S at

35

degr

ee fr

ictio

n (M

Nm

)

0.2

0.4

0.6

0.8

1

1.2

Height into roof (m)

Figure 106 Average BPXS across the roof line for different heights into the roof and different locations of bolts (area under curves in Figure 107 expressed as uniformly distributed load

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6Bolting location with respect to face (m)

0

0.5

1

1.5

Hei

ght i

nto

roof

(m)

Figure 107 BPXS as a function of the location and height

In the following analysis, the BPXS developed on two bedding horizons are discussed – 0.2m and 0.4m above the roof line (dataset in an appendix). The data was manipulated in the context of the following relationship:

BPXS (kN/m) = XS * vertical stress (MPa) * Ffriction *Fbolt timing, Equation 18

where XS is a function of the height into roof and the K values acting normal and parallel to the roadway, Ffriction is a correction for different friction angles, and Fbolt timing is a correction for bolt location.

The XS values are presented in Figure 108 (0.2m into roof) and Figure 109 (0.4m into roof) for the case of a friction angle of 35o and 2.3m from the face.

1

0 0.5 1 1.5 2 2.5 3Normal/vertical

0

0.5

1

1.5

2

2.5

3

Par

alle

l/ver

tical

Figure 108 XS factor for 0.2m into the roof

0 0.5 1 1.5 2 2.5 3

Normal/vertical

0

0.5

1

1.5

2

2.5

3

Par

alle

l/ver

tical

Figure 109 XS factor for 0.4m into the roof

The Ffriction parameter is itself a function of the K values such that Ffriction = 1+(G *(35‐x)/25) where G is given in Figure 110 and x is the design friction angle. The Fbolt timing parameter (Figure 111) does vary with the K ratios, but at this stage it has not been further analysed.

0 0.5 1 1.5 2 2.5 3Normal/vertical

0

0.5

1

1.5

2

2.5

3

Para

llel/v

ertic

al

Figure 110 G parameter to account for different friction angles

0.1

1

10

0 1 2 3 4 5 6 7 8Distance from the face line (m)

Bolti

ng lo

catio

n fa

ctor

2.5-1.5-11.5-1-12.0-1.33-1

Normal-parallel-vertical

Figure 111 The bolt timing factor

8.4.2.2 RESISTING FORCES

The deformations that develop close to the roof mean that it should be assumed that the bedding surfaces will be open. Pragmatically there is a need to close the open bedding surface, otherwise only the relatively inefficient dowel resistance will be available.

Based on observations of the roof lifting and adjacent bolts losing tension made during the introduction of pre‐tensioned roof bolts in the Australian industry, it is assessed that correct installation of tensioned bolts can close bedding surfaces. It is not necessary to apply a preload against, not necessarily to apply significant preload across the closed surface and the shear stiffness is such that full capacity is mobilized within a few milimetres of movement (Figure 112). Shear resistance is given in Figure 79.

111

Figure 112 Fully grouted bolts are very stiff during initial shear loading (Stjern, 1995)

8.4.3 DENSITIES AND PATTERNS

The number of ideally placed bolts is calculated by dividing the BPXS value by the frictional resistance provided by the bolts. The number of bolts will depend on the strength of the bolts, the depth of cover, the K ratios, and the friction angle of the bedding. Expressed in terms of installed capacity (Section 4.2.1.2), there is good agreement between this BPXS model and current mining practice (Figure 113) especially if the mines with low RSI values are excluded. Friction angles of between 30o and 35o are indicated and this is consistent with the rock types and joint roughness.

0

10

20

30

40

50

60

70

80

0 100 200 300 400 500 600 700Depth (m)

Prim

ary

Supp

ort C

apci

ty (t

onne

s/m

2 )

Stone roof253035Linear (Stone roof)

Low RSI values

Figure 113 Comparison between ideal BPXS support capacity at 0.4m into the roof (Kn=2.0, Kp=1.5, G=1.3) and the Australian database

Ideally the bolts should be placed so that they are equally loaded and Figure 105 can be used for guidance. According to this analysis, there is no advantage with centreline bolts as the locations are strongly biased towards the rib sides (Table 23) and this can result in a large “unsupported” span in the centre of the roadway that is bigger than the cut out distance. If the roof is thinly bedded this

112

span may not be adequately stable (Figure 100) indicating the need for mesh panels. The possibility of adverse loading on the panels may require additional bolts or the relocation of the “ideal” towards the roadway centerline. The rib side bias has been examined with physical models and these certainly show the advantage of a bias towards to the ribs (Figure 114).

Table 23 Bolt locations to resist BPXS

Pattern Distance from rib line (m) Centreline span (m) Compromise locations (m) 2 0.55 4.1 1.35 4 0.3,0.8 3.6 0.6,1.6 6 0.2,0.55,1.05 3.1 0.4,0.9,1.6 8 0.15,0.4,0.7,1.15 2.9 0.2, 0.5, 1.0,1.7

Figure 114 Optimum bolt patterns from physical models (Stimpson, 1987)

The following is an example of how the support design could progress:

Input: Depth of 250m, average density = 2.4 t/m3. L=1.6, M=1.4, roadway driven at 20o to the major principal horizontal stress. Thinly bedded mudstone roof, JRC = 4. Bolting conducted at 2.3m from face. Pre‐tensioned 22mm diameter X grade bolts.

Design values: Vertical stress = 250*2.4*0.0098 = 5.88MPa, Kn = 1.42, Kp=1.52. Assume a base friction angle of 25o, and a JRC contribution of 8o. Required beam thickness is at least 0.3m to carry 2m of surcharge.

Output: XS =122 (use 0.4m height chart), G=1.48, Ffriction= 1.11, Fbolt timing = 1.0, BPXS=796kN, Boltfriction = 340 *tan 32 = 212 kN, Anchor length =0.9m.

Interpretation:

The ideal bolt density is 3.75 bolts per metre of roadway advance. A 4 bolt pattern is recommended. The thinly bedded mudstone could mean large loads on the mesh panel, so less than ideal bolt locations (in terms of BPXS) is recommended. Bolt locations of 0.6m and 1.6m from each ribs is recommended. The centre span is 2.0m and some loading on the mesh is anticipated. An outbye centre bolt could be considered if there is concern about mesh loading. A minimum bolt length of 1.4m is possible, but precedent practice is the use of 1.8m bolts. 1.8m bolts can be installed in a single pass operation.

113

8.4.4 COMMENTARY ON DESIGN TO PREVENT DELAMINATION

Based on the authors’ experiences, the voussoir beam tool is appropriate for estimating spanning thickness. The tool can also be used to identify the requirements for a spanning unit higher in the roof if a suspension strategy is to be adopted (e.g. if destressed coal is to be suspended from a stone roof). A key aspect is the estimation of the surcharge, which should include the anchorage length of the bolt and any suspended load. At this stage, it is considered that a factor of safety of 1.5‐2.0 should be applied to the required thickness – this means a thickness of 0.4m may be appropriate.

It is noted that you could include the magnitude of the horizontal stress in the beam calculation if you wish – the input UCS value is the available compressive capacity (UCS – σh).

One of the implications of the BPXS analysis is that bolting densities reduce if bolting can be deferred until the roadway is advanced. The voussoir beam analysis can be used to assess the risk of loss of roof skin with span in this case being from face to last line of bolts. If risk is acceptable, the bolting densities are substantially reduced. This is probably why the USA industries uses generally lower bolt densities as that industry uses extended cut mining to a much greater level than Australia.

The design tool for bedding plane shear assumes that the roof beam is exposed to the elastic stress field in a continuum with no relaxation due to roof deflection. While this appears to be a major simplification, the validation to current mining practice indicates that it is reasonably valid at least for a height of 0.4m. It appears that factors of safety close to unity may be appropriate.

The XS value varies by no more than 20 % within the range of K values encountered in the Australian industry. This suggests that there is no real advantage in having different support densities in headings compared to cut‐throughs.

The use of ultimate strength of the bolts is acceptable, if you can rely on their post yield and post failure capacity. The greater the tolerable deformations in the system, the more likely that the immediate roof stresses will relax.

The BPXS tool should only be applied for initial roadway development and not at the maingate corner because there is little doubt that the relaxation of roof stresses has developed. If there was no such relaxation, the maingate BPXS magnitudes would be similar to those shown in Figure 110 with the bolt timing being set at 0m – say 5 times greater than those used in the design for the development face. If this was the case, one implication would be that maingate roofs without cable support would fail – and this is demonstrably not the case in Australia or other countries. It is noted that Seedsman (2000) had not identified this relaxation mechanism.

Deterioration of the maingate conditions can be explained increase in height of stress arch and the resulting increase in the vertical surcharge to the beam when the RSI is low. This vertical surcharge induces compressive stresses in the roof beam that has been created in the “softened zone” by the roof bolts so that the observer may see additional guttering.

It is possible that the reports of stress notching in intersections are a reflection of changes in the height of the stress arch and not increased horizontal stresses at the roof line.

It is noted that the roadway width used in the BPXS analysis was 5.0m. The results can be reasonably applied to roadways with widths between 4.5m and 5.5m. They should not be directly extrapolated

114

to wider roadways because the impact both in terms of the wider roadway and the bolt location factor has not been examined.

115

9 PREVENTATION OF RIB COLLAPSE

A similar logical framework has been developed for ribs (Figure 115). For ribs, there are 3 steps in the framework. Different from roof, ribs undergo a simple step‐wise increase in the applied stress field from development to the maingate and then the tailgate. At one level, it is appropriate to consider the hazards of coal ribs in the context of a vertical pitwall. This analogy is not sufficient so it is then necessary to consider the possibility of the onset of additional failure modes related to the additional vertical stresses that ribs carry compared to rock faces.

The steps are:

1. Check the height of the roadway. Rib support is concerned with falls from height, with the greatest concern with falls from above shoulder level. Assuming that roadway width is such that reasonable access is available around the various pieces of mining machinery and hence a relatively clear workspace, a roadway height of 1.8m is considered to be a reasonable limit for pattern rib support.

2. Check for alignment with respect to coal joints and cleat.

3. Check for the possible onset of mining‐induced failure. Note that this may not develop if the alignment is such that joints and cleats dominate behaviour. Also note that this failure may be time and mining system dependent.

Consider realignment ofroadway Possibiity of

large wedge and planarslides Bolt length based

on slide geometry

More than 20degrees to

cleatN

Y

Strength/verticalgreater than 5.0 N

Y

Mining-inducedfracturing will developimmeidately or over

time

Secureintersections

Extractionheight greater

than 1.8mN

Y

Pattern supportnot required

Pattern supportnot required

Figure 115 Logical framework for support of ribs

116

9.1 STRUCTURE CONTROL

In one aspect, ribs can be considered as vertical pit walls (Figure 116). They are therefore exposed to the same hazards of joint‐controlled failure with planar slides, wedges and topples. Highly broken coal, for example in a fault zone, could also generate a circular slip mode. Numerous papers and books discuss planar slides and wedges and there are software programs available (such as Rocplane and Swedge, www.rocscience.com).

Figure 116 Failure modes for rock slopes that can be observed in coal ribs (Hoek and Bray, 1981).

In the rock slope engineering, there is a preference to avoid “noses” in pit walls as these protuberances are unstable due to the lack of lateral restraint. In coal mine roadways, all intersections are “noses” (Figure 117) and may need special attention.

117

Figure 117 All intersections are noses

9.1.1 SLIDES

The key variables in a planar slide analysis are the dip and friction angle of the discontinuity. Recognizing the low heights, the stresses are low and JRC can become a significant contributor to friction angle. The JRC for coal surfaces ranges between 4 and 14, averaging 8 (Table 9). Thus, for a typical rib geometry, the friction angle increment related to JRC may be up to 20o.

Figure 118 shows the installed bolt capacity to achieve a factor of safety of 1.5 using with passive bolts. The installed capacities are only in the order of 1 tonne/m advance to 2.5 tonnes/m advance for a 3m high rib. For different rib heights, the loads can be reduced according to the square of the height.

0

5

10

15

20

25

30

40 45 50 55 60 65 70 75 80 85

Dip of plane with friction angle of 45o

Bol

t for

ce (k

N/m

)

0

1

2

3

4

5

6

7

8

9

10

Fact

or o

f saf

ety

with

out b

olts

or f

all m

ass

(tonn

es/m

)Bolt force at 3.0m highFactor of safety without passive boltsMass at 3m

Figure 118 Planar geometry and required face support for a 3m high rib

118

9.1.2 WEDGES

At more than 20o to the strike of a discontinuity, there is still the possibility of a wedge falling from the rib line. The maximum dimensions of the wedge are in the order of 5m long and about 1.5m into the face (Figure 119). The required bolt capacity is in the order of 0.5 tonnes/metre of roadway advance for a 3m high rib.

0123456789

10

10 15 20 25 30 35 40 45

Angle between face and strike of one of the conjugate joint sets

Leng

th a

nd d

epth

of r

ib

impa

cted

(m) o

r wei

ght o

f w

edge

(ton

nes)

012345678910

Inst

alle

d pa

ssiv

e su

ppor

t to

give

FoS

= 1

.5 (k

N/m

)

Length (m)Depth (m)Mass (tonnes)kN/metre

Figure 119 Wedge geometry and required face support

9.1.3 TOPPLES

Sagaseta et al (2001) provides a general solution for the anchor force to limit toppling failures. They argue that the anchor force (Ft) is given by the following relationship:

Ft = 0.5*γkTH2 Equation 19

Where γ = density, H is face height, and kT is a dimensionless factor which is provided in a series of design charts. For the geometries of interest, the range of kT value is 0.2 to 0.3.

The anchor force to prevent toppling is therefore in the order of 1 tonne/m to 2 tonnes/m (Figure 120).

0

5

10

15

20

25

30

1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4

Rib height (m)

Inst

alle

d ca

paci

ty (k

N/m

)

0.20.3

Figure 120 Anchor force to prevent toppling

119

9.2 STRESS INDUCED RIB COLLAPSE

9.2.1 MINING INDUCED FRACTURES

O’Bierne et al (1986) discussed mining induced fractures in coal ribs. Brittle rock behaviour provides the theoretical background to these observations.

As discussed in Section 5.3.4, there are 3 components to the brittle rock failure criterion – the onset of failure related to the loss of cohesive strength, a spalling limit, and then the standard Hoek Brown criterion in the “far field”. The far field in this context may be only a few metres away from the excavation. Readily available software cannot yet handle this composite criterion, so there is a need to apply it stepwise.

Figure 121 shows the distribution of the strength factor with respect to the brittle parameter for the case of 20 MPa coal with vertical stresses of 5 MPa and 10 MPa. This simple analysis would suggest that rib failure would progressively extend well into the rib with increasing depth. It is known that this is not the case as observations of rib spall at depth suggest that it is still only a skin effect. Colwell (2004) examined rib data over a depth range of 125m to 525m and reported that average rib bolt length was 1.2m and the maximum length was 1.8m.

(a) 5 MPa vertical stress (b) 10 MPa vertical stress

Figure 121 Distribution of the cohesive loss component about roadways

The explanation for this discrepancy is the need to consider the spalling limit. The spalling limit is the ratio of the magnitude of the major and minor principal stress and hence is independent of any strength parameter. The distribution of the spalling limit with depth at mid‐height is a function of the K ratio and the aspect ratio of the opening, and in general is reaches a value of 10 somewhere between 0.6m and 1.1m into the rib (Figure 122).

Similar to the RSI concept, it is possible to manipulate the data in the context of a coal strength index (CSI) to produce a design chart for the depth of brittle failure (Figure 123). The shape of the failure zones can be obtained by inspection of Figure 121a. Pattern rib support may only be required for failure depths in excess of about 0.1m, or a CSI of about 7.5.

120

0

2

4

6

8

10

12

14

16

18

20

0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5

Depth into rib (m)

Spallin

g ratio

K=0.1, 5m by 3mK=0.2, 5m by 3mK=0.4, 5m by 3mK=0.1, 5m by 2mK=0.2, 5m by 2mK=0.4, 5m by 2m

Figure 122 Value of spalling ratio along the spring‐line of the rib for different K ratios and roadway aspect ratios

0

0.2

0.4

0.6

0.8

1

1.2

3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9Coal strength index

Dep

th o

f brit

tle fa

ilure

(m)

0.1, 5m by 3m0.2, 5m by 3m0.3, 5m by 3m0.4, 5m by 3m0.1, 5m by 2m0.2, 5m by 2m0.3, 5m by 2m0.4, 5m by 2m

maximum depth set by spalling limit = 10

Figure 123 Estimation of maximum depth of brittle failure

9.2.2 BUCKLING

For roadway orientations that are within say 20o of the trend of a coal joint set, there is a possibility of buckling of coal beams if the joints are closely spaced. Euler buckling concepts offer some insight into the mechanisms involved, with the limitation that there is a large degree of uncertainty regarding the imposed loading.

The axial loading of slabs was discussed by Hoek and Brown (1980) who provided the following relationship for the axial stress (σα) at which the plate will buckle:

121

σα = Π 2 E/(12q2(l/t)2 Equation 20

where E=modulus of elasticity,

l = height of the slab,

t = thickness of slab, and

q = constant with values of 1 for both ends pin‐jointed, and 0.5 for both ends clamped.

The axial stress applied to a joint bound slab at the side of a roadway is difficult to accurately determine due to the development of yielding in the coal and any low strength bands. A simple 2 dimensional elastic analysis suggests that the vertical stress at the rib line can increase by about 220% (Table 18, Figure 57). Using this result, Figure 124 quantifies the relationship between slab thickness, extracted coal thickness and depth. At shallow depths, slabs less than about 200mm thick may buckle, and this increases with depth to 400mm at depths in excess of 400m.

0

0.1

0.2

0.3

0.4

0.5

0.6

1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4

Roadway Height (m)

Buc

klin

g th

ickn

ess

(m)

200

300

400

500

Simple beam, pin jointedE= 1500 MPaVertical stress related to depth of cover and density of 2.5 t/m3 , and an abutment peak at 2.16 the average stress

DEPTH

Figure 124 Thickness of slabs that may undergo buckling.

9.3 COMMENTARY ON RIB SUPPORT DESIGN

Hoek and Bray (1981) suggests that the kinematic acceptability of planar slides is reduced if the rock face in an open excavation is aligned at more that 20o from the strike of the joint. This recommendation should be considered to be a minimum in the underground coal environment as there may still be concerns with the practicalities of the installation of rib support at greater angles. The authors have found that an angle of about 35o would be preferred.

The discussion about coal stresses (Section 5.5.2) speculated that there may be a reduction in vertical stress to below that related to depth of cover at the mining face and that this increases outbye. If this

122

is the case, there may be an apparent time dependent deterioration of the ribs related to brittle failure as the CSI reduces as the vertical stress increases.

The Euler analogy needs to be applied with care as it requires that the loading system is of low stiffness such that there is no stress relaxation if the rib compresses to any degree. If the rib slabs are formed by a brittle failure mechanism, the development of more intense failure zones at the roof and rib corner would not be expected to transfer the same levels of vertical loading. The mechanics being invoked are the same as those that drive the relaxation of the softened zone in the roof.

123

10 REFERENCES

ACIRL. (1986). Evaluation of coal pillar stability and extraction methods. In ACA End of Grant Report.

AusIMM, (1995). Field Geologist Manual (3rd e.d.). (p.390). The Australasian Institute of Mining and Metallurgy. Carlton, Australia.

Barton, N.R. (1973). Review of a new shear strength criterion for rock joints. Engineering Geology 8, 287‐322.

Barton, N.R. (1980). Evaluation of shear strength in rockfill and between rockfills and rock foundations. NGI report.

Barton, N.R., & Choubey, V. (1977). The shear strength of rock joints in theory and practice. Rock Mechanics 1/2:1‐54.

Barton, N., Lien R., & Lunde, J. (1974). Engineering classification of rock masses for the design of tunnel support. Rock Mechanics 6 (4), 163‐177.

Bieniawski, Z.T. (1976). (Eds. Bieniawski, Z.T.) Rock mass classifications in engineering practice. In Exploration for rock engineering. 1, 97‐106. Cape Town : AA Balkema.

Bieniawski, Z.T. (1993). (Eds. Hudson, J., Brown, E.T., Fairhurst, C., & Hoek, E.). Design methodology for rock engineering: Principles and practice. Comprehensive Rock Engineering, 2:779‐93. Oxford: Pergamon.

Bjurstrom, S. (1974). Shear strength of hard rock joints reinforced by grouted untensioned bolts. Proceedings 3rd ISRM Congress, Denver, 1194‐1999.

Brady, B.H.G. & Brown, E.T. (1985). Rock mechanics for underground mining. London, UK: George Allen & Unwin.

Brown, E.T. (2008). Estimating the Mechanical Properties of Rock Masses. SHIRMS 2008 (Ed Potvin,Y. Carter, J., Dyskin,D. and Jeffrey,R), Perth, 3‐17.

Canmet. (1977). Pit Slope Manual, Chapter 6, Mechanical Support. Ottawa: Canmet.

Coates, D.F. (1970). Rock mechanics principles. Mines Branch, Department of Energy, Mines and Resources. Monograph 874.

Colwell, M.G. (1998). Chain pillar design: Calibration of ALPS. (p.67). Final Report: ACARP Project C6036.

Colwell, M.G. (2004). Analysis and design of rib support (ADRS): A rib support design methodology for Australian collieries. (p. 325). Final report: ACARP Project C11027.

Colwell, M.G., & Frith, R. (2006). Why uniaxial compressive strength and Youngs modulus are commonly poor indicators of roadway roof stability – except in the tailgate. Paper presented in Coal Operators Conference at University of Wollongong, Illawarra branch, AusIMM.

Cram, K. (September, 2008). Production from Australian longwall mines. Australian Longwall Magazine.

Diederichs, M.S., & Kaiser, P.K. (1999). Tensile strength and abutment relaxation as failure control mechanisms in underground excavations. In International Journal of Rock Mechanics and Mining Sciences: 36, 69‐96.

124

Diederichs, M.S., Kaiser, P.K., & Eberhardt, E. (2004). Damage initiation and propagation in hard rock during tunneling and the influence of near‐face stress rotations. In International Journal of Rock Mechanics and Mining Sciences: 41, 785‐812.

Donaldson, A.C. (1979). (Eds. Donaldson, A.C., Presley, M.K., & Renton, J.J.) Origin of coal seam discontinuities. (p.102‐132). In Carboniferous Coal Guidebook. WV Geological Survey, Vol B‐37‐1.

Enever, J.R. & Doyle, R. (November, 1996). (Eds. NcNally, G.H., & Ward, C.R.). The role of horizontal stress in longwall design: A case history at Dartbrook Colliery, NSW. (p.143‐151). Symposium on Geology in Longwall Mining, 12‐13.

Enever, J.R., Jeffrey, R.G. & Casey, D.A. (2000). (Eds. Girard, J., Liebbman, M., Breeds, C. & Doe,T.). The relationship between stress in coal and rock. (p.409‐414). Pacific Rocks 2000. Rotterdam: Balkema.

Fookes, P.G., Baynes, F.J., & Hutchinson, J.N. (2000). Total geological history: A model approach to the anticipation, observation, and understanding of site conditions. Paper presented in GeoEng 2000, An international conference on geotechnical and geological engineering in Melbourne. Lancaster, USA: Technomic Publishing Company.

Gale, W. J. (2008). Stress issues in underground coal mines and design approach. Paper presented in “Stress Measurements, Monitoring and Modelling Techniques and their Design Applications”, Wollongong: Eastern Australia Ground Control Group.

Gale, W. J. & Matthews, S. M. (1992) Stress control methods for optimised development and extraction operations. Report to National Energy Research Development, and Demonstration (NERD&D) Program, Project 1301.

Gordon, N. & Tembo, E. (2005). The roof strength index: A simple index to one possible mode of roof collapse. Paper presented in Bowen Basin Symposium. (Eds. Beeston, J.W.) Yeppoon, Queensland.

Haile, A. (2004). A reporting framework for geotechnical classification of mining projects. Bulletin of the AusIMM. September, p.30‐37.

Hanna, T.H. (1982). Foundations in tension: ground anchors. Trans Tech Publication Series on Rock and Soil Mechanics, Volume 6.

Hasenfus, G.J. & Su, D.W.H. (2006). Horizontal stress and coal mines: Twenty‐five years experience and perspective. (p.256‐267). Paper presented in 25th International Conference on Ground Control in Mining.

Herget G. (1988). Stress in rock. Balkema.

Hillis, R.R., Enever, J.R. & Reynolds, S.D. (1999). In situ stress field of eastern Australia. Australian Journal of Earth Sciences. 46, 813‐825.

Hobbs, D.W. (1967). The formation of tension joints in sedimentary rocks: An explanation. Geology Magazine, 104 (6), 550‐556.

Hobst, L. & Zajic, J. (1983). Anchoring in rock and soil. (p.570). Developments in Geotechnical Engineering, 33. Elsevier, Amsterdam.

Hoek, E. & Bray, J.W. (1981). Rock slope engineering. 3rd Edition. London: Instn Min. Metall.: London.

Hoek, E & Brown E.T. (198). Underground Excavations in Rock. Instn Min Metall.: London .

Hoek, E & Brown E.T. (1997). Practical estimates of rock mass strength. International Journal of Rock Mechanics and Mining Sciences, 34 (8), 1165‐1186

125

Hoek, E., Marinos, P.G. & Marinos, V.P. (2005). Characterisation and engineering properties of tectonically undisturbed but lithologically varied sedimentary rock masses. International Journal of Rock Mechanics Mining and Sciences, 42 (2), 277‐286.

Hoek, E & Diederichs,M. (2006) . Empirical estimates of rock mass modulus. International Journal of Rock Mechanics and Mining Sciences, 43, 203‐215.

ISRM (International Society for Rock Mechanics) Commission on Standardization of Laboratory and Field Tests. (1978). Suggested Methods for the quantitative description of discontinuites in rock masses. International Journal of Rock Mechanics and Mining Sciences, 5, 316‐368.

Jaeger, J.C., & Cook, N.G. (1979). Fundamentals of Rock Mechanics. London: Chapman and Hall.

Jemeric, M.L. (1985). Strata mechanics in coal mining. Rotterdam: A.A.Balkema.

Ji, S., Zhu, Z., & Wang, Z. (1998). Relationship between joint spacing and bed thickness in sedimentary rocks: Effects of interbed slip. Geology Magazine, 135 (5), 637‐655.

Kaiser, P.K & Kim, B‐H. (2008). Rock Mechanics Challenges in Underground Construction and Mining. SHIRMS 2008 (Ed Potvin,Y. Carter, J., Dyskin,D. and Jeffrey,R), Perth, 23‐38.

Lambe, T.W. (1973). Predictions in soil engineering. Geotechnique, 23 (2) 149‐202.

Littlejohn, S. (1993). Overview of rock anchorages. Comprehensive Rock Engineering, 4:413‐450. Oxford: Pergamon.

Mark, C. (1990). Pillar Design Methods for Longwall Mining. United States Bureau of Mines Information Circular 9247, 53p.

Mark, C., & Molinda, G.M. (2005). The coal mine roof rating (CMRR): A decade of experience. Coal Geology, 64, 85‐103.

Mark, C., Gale, W., Oyler, D., & Chen, J. (2007). Case history of the response of a longwall entry subjected to concentrated horizontal stress. International Journal of Rock Mechanics and Mining Sciences, 44, 210‐221.

Martin, C.D., Kaiser, P.K., & McCreath, D.R. (1991). Hoek‐Brown parameters for predicting the depth of brittle failure around tunnels. Canadian Geotechnical Journal, 36: 136‐151.

McCabe, K.W., & Pascoe, W. (1978). Sandstone channels: Their influence on roof control in coal mines. (p.24). Mine Safety and Health Administration, IR 1096.

Medhurst, T.P. (1999). Highwall mining: Practical estimates of coal seam strength and the design of slender pillars. Trans Instn Min Metall (Sect A Min Industry), 108, A161‐A171.

Medhurst, T.P., & Brown, E.T. (1998). A study of the mechanical behaviour of coal for pillar design. International Journal of Rock Mechanics and Mining Sciences, 35, 1087‐1105

Molinda, G.M. (2003). Geological hazards and roof stability in coal mines. US Department of Health and Human Services, NIOSH, Pittsburgh Research Laboratory, Information Circular IC 9466, Publication number 2003‐152.

Molinda, G.M., & Mark, C. (1994). The coal mine roof rating (CMRR): A practical rock mass classification for coal mines. United States Bureau of Mines IC 9387.

Morgenstern, N.R. (2000). Performance in geotechnical practice. Paper presented at the Inaugural Lumb Lecture, 10 May, Hong Kong.

Muirwood, A.M. (1972). Tunnels for roads and motorways. Quart J Eng Geol, 5, 111‐126.

Nemcik, J.A., Gale, W.J, & Fabjanczyk, M.W. (2006). Methods of interpreting ground stress based on underground stress measurements and numerical modelling. Paper presented in Coal 2006, 7th Underground Coal Operators conference. 104‐112.

126

O’Beirne, T, Shepherd, J, Rixon, K and Napper, A, 1986. Coal rib stabilisation – a new perspective. The Coal Journal, pp7‐11.

Palmstrom, A., & Broch, E. (2006). Use and misuse of rock mass classification systems with particular reference to the Q system. Tunnelling and Underground Space Technology, 21, 575‐593.

Pattison, C.I, Fielding, C.R, McWatters, R.H. & Hamilton, L.H. (1995). Nature and Origin of Fractures in Permian Coals from the Bowen Basin and the Relationship with in situ reservoir permeability. Bowen Basin Symposium, Mackay. 217‐234.

Payne, D. (2007). Crinum Mine – 15 longwalls in weak roof. What did we learn?. Paper presented in Coal 2006, 7th Underground Coal Operators conference. 22‐41.

Peck, R.B. (1969). Advantages and disadvantages of the observational method in applied soil mechanics. Geotechnique, 19, 171‐187.

Pells, P.J.N. (2002). Developments in the design of tunnels and caverns in the Triassic rocks of the Sydney region. International Journal of Rock Mechanics and Mining Sciences, 39, 5, 569‐588

Pells, P.J.N., & Bertuzzi, R. (2007). Limitations of rock mass classification systems. Tunnels and Tunnelling International, April, 33‐37.

Peng, S.S. (1986). Coal Mine Ground Control. (p.491). John Wiley and Sons Inc.

Potvin, Y. & Nedin, P. (2003). Management of rockfall risks in underground metaliferous mines: A reference manual. Minerals Council of Australia

Sagaseta, C., Sanchez, J.J., & Canizal J. (2001). A general analytical solution for the required anchor force in rock slopes with toppling failure. International Journal of Rock Mechanics and Mining Sciences, 38 (3), 421‐435.

Schach, R., Garshol, K., & Heltzen, A.M. (1979). Rock bolting: A practical handbook. (p.84). Oxford: Pergamon Press.

Seedsman, R.W. (2000). Progress in the development of a roof bolt design methodology based on resisting shear. Paper presented in 19th International Conference on Ground Control in Mining. 272‐277.

Seedsman, R.W. (2003). Roof support in gateroads in multiple seam longwalls: Lessons from ultra‐close mining at Gibsons Colliery. Paper presented in Coal 2003 4th Australasian Coal Operator’s Conference, Wollongong.

Seedsman, R.W. (2004). (Eds. Villaescua, E., & Potvin, Y.). Failure modes and support of coal roof. Paper presented in Ground Support in Mining and Underground Construction. Proceedings of the 5th International Symposium on Ground Support, Perth. AA Balkema.

Shen, B., Guo, H., King, A. &Wood, M. (2006). An Integrated Real‐time Roof Monitoring System for Underground Coal Mines. Paper presented in Coal 2006 7th Australasian Coal Operator’s Conference, Wollongong.

Shepherd, J. & Gale, W. (1982). Colliery roof stability and the role of geology. The Australian Journal of Coal Mining Technology and Research, No 1, 47‐67.

Sofianos, A. & Kapenis, A.P. (1998). Numerical evaluation of the response in bending of an underground hard rock voussoir beam roof. . International Journal of Rock Mechanics and Mining Sciences, 35(8), 1071‐1086.

Stimpson, B. (1987). Optimising the reinforcement effect of full‐column, untensioned, grouted bolts in bedded mine roof. International Journal of Mining and Geological Engineering, 5, 285‐298.

127

Stjern, G. (1995). Practical Performance of Rock Bolts. Doktor Ingenior Thesis, Instituttt for Geologi Bergteknikk, norges Tekniske Hogskole, Universtitetet i Trandheim.

Tarrant, G. (2006). Skew roof deformations mechanism in longwall gateroads: Concepts and consequences. Paper presented in COAL 2005, 6th Australian Coal Operators Conference. Brisbane, AusIMM.

Thompson, A.G. (2004). Rock support action of mesh quantified by testing and analysis. Paper presented in Surface support in Mining. (Eds. Potvin, Y., Stacey, D., & Hadjigeorgiou, J.) Australian Centre for Geomechanics, pp 391‐398.

Zipf, R.K. (2005). Failure mechanisms of multiple seam mining interactions. Paper presented in 24th International Conference on Ground Control in Mining. 93‐106.

128

APPENDIX A VOUSSOIR BEAM SPREADSHEET

By implementing the spreadsheet as detailed below, you acknowledge that you appreciate the limitations of the method as outlined in Sofianos, A. & Kapenis, A.P. (1998). Numerical evaluation of the response in bending of an underground hard rock voussoir beam roof. International Journal of Rock Mechanics and Mining Sciences, 35(8), 1071‐1086.

INPUT m Span 70 m Beam thickness 10 Gpa Modulus 7.5 kn/m3 Density 25 MPa UCS 30 m Surcharge thickness 5 kN/m3 Surcharge density 24

Output Crushing factor =C18/C23 Buckling factor =C19/C23 Deflection mm =C30*C27*1000

calculations n check 0.1-0.3 =0.3-0.14*C16*POWER(C23,0.333) Sn =C2/C3 m Equivalent surcharge =C7*C8/C5 equation 9 =4*C14*C6/C4/1000/C26*SQRT((1-

0.5*C14*C6/C4/1000*C24*C26*C26)) buckling =3/C26/C26/C26 Q =(C3+C17)/C3*C5*C3 Qn =(C17+C3)/C3*C5*C2/C4/1000000 lamda =0.2+0.06*SQRT(C16)+0.0005/SQRT(C23) Zon =1-2*C14/3 Sz =C2/C27 Zo =C25*C3 delzo =C23*C26*C26*C26/16 omega =1/3*ATAN(SQRT(1/27-C28*C28)/C28) delz =1-COS(C29)/SQRT(3)-SIN(C29)

thickness (C3)

Surcharge thickness and density (C7,C8)

span (C2)

UCS, modulus, density (C6,C4,C5)

129

APPENDIX B EXAMINE 3D MODELLING

This appendix provides more details on some of the Examine3d modelling.

Figure B1 shows the difference between 5m, 5.5m, and 8m wide roadways. The plots support the use of the BPXS plots for roadways between 4.5m and 5.5m in width, but not for roadways that are wider.

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50

0.5

1

1.5

2

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50

0.5

1

1.5

2

8M WIDE

5M WIDE

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50

0.5

1

1.5

2

5.5M WIDE

Figure B1 Results for different roadway widths, assuming 3m height, σ1 = 12 MPa, σ2 = 9 MPa, and σ3 = σv = 6 MPa, roadway parallel to σ1 , and PR = 0.25

130

Figure B2 shows the effect of Poissons ratio. For Bottling at 2.31m and a 35º friction angle, the value of the BPXS at 0.4m into the roof is 0.906, 0.863, and 0.838 for Poissons Ratio values of 0.35, 0.25, 0.15 respectively.

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50

0.5

1

1.5

2

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50

0.5

1

1.5

2

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5Bolt location (m)

0

0.5

1

1.5

2

Hei

ght i

nto

roof

(m)

0.15

0.25

0.35

Figure B2 Results for different Poissons ratios, assuming 3m height and 5m wide, σ1 = 12 MPa, σ2 = 9 MPa, and σ3 = σv = 6 MPa, roadway parallel to σ1 .

131

Table B1 Summary of BPXS data as refered to in the body of the report.

Normal to roadway

Parallel to roadway

Vertical BPXS (kN)

Height into roof (m0 0.2 0.4 0.4 0.4 0.4 0.4 0.4 Friction angle 35 35 30 25 20 15 10

1 1 6 277 387 440 497 554 615 673 1 3 6 345 483 538 588 644 699 751 1 6 6 421 599 659 705 759 810 856 1 12 6 560 820 894 932 988 1029 1066 3 1 6 312 410 466 526 593 658 720 3 3 6 370 490 551 606 672 734 791 3 6 6 452 607 673 724 788 846 897 3 9 6 529 724 793 842 903 956 1002 3 12 6 603 833 913 955 1019 1066 1108 3 24 6 879 1240 1356 1394 1466 1499 1524 6 1 6 347 422 490 559 635 711 782 6 3 6 401 497 571 635 710 784 850 6 6 6 490 620 700 758 830 900 959 6 9 6 576 738 820 877 947 1011 1065 6 9 6 576 738 820 877 947 1011 1065 6 12 6 656 853 940 993 1063 1121 1171 6 12 6 656 853 940 993 1063 1121 1171 6 12 12 904 1214 1347 1449 1575 1692 1794 6 15 6 730 962 1060 1106 1179 1230 1277 6 18 6 804 1071 1177 1219 1294 1340 1382 6 24 6 951 1278 1402 1444 1516 1558 1589 8 12 6 683 862 956 1017 1090 1156 1211 9 3 6 437 511 597 671 754 838 913 9 6 6 523 628 722 789 870 951 1019 9 9 6 608 746 842 908 986 1062 1125 9 12 6 693 863 962 1026 1102 1172 1230 9 12 6 693 863 962 1026 1102 1172 1230 9 15 6 775 976 1082 1141 1217 1281 1336 9 18 6 851 1085 1201 1254 1333 1442 1442 9 24 6 999 1304 1434 1480 1563 1609 1652

12 1 6 413 440 533 623 718 814 903 12 3 6 469 519 617 703 795 889 973 12 6 6 554 637 744 825 912 1001 1079 12 8 6 612 715 825 907 989 1076 1150 12 9 6 639 753 863 947 1027 1113 1184 12 12 6 723 870 983 1069 1143 1222 1289 12 12 12 981 1240 1399 1515 1660 1799 1918 12 15 6 808 987 1103 1191 1258 1332 1395 12 18 6 890 1099 1222 1309 1374 1441 1501 12 24 6 1045 1317 1461 1539 1605 1660 1711 15 6 6 588 649 767 859 955 1053 1139 15 9 6 671 764 886 980 1070 1164 1244 15 12 6 756 881 1006 1102 1185 1274 1350 18 1 6 476 461 574 685 802 915 1022 18 3 6 534 541 660 767 881 991 1093 18 6 6 601 636 765 872 981 1090 1188

132

Normal to roadway

Parallel to roadway

Vertical BPXS (kN)

Height into roof (m0 0.2 0.4 0.4 0.4 0.4 0.4 0.4 Friction angle 35 35 30 25 20 15 10

18 9 6 703 776 908 1013 1112 1215 1304 18 12 6 788 893 1028 1125 1228 1325 1409 18 18 6 957 1127 1267 1380 1459 1544 1620 18 24 6 1125 1352 1506 1618 1690 1763 1831 24 1 6 509 448 583 722 864 998 1126 24 3 6 528 566 667 804 941 1072 1196 24 6 6 655 650 798 930 1060 1186 1303 24 6 12 938 1037 1234 1404 1590 1777 1945 24 9 6 737 765 921 1050 1174 1297 1407 24 12 6 814 873 1034 1166 1284 1403 1509 24 18 6 1021 1151 1312 1446 1544 1647 1740 24 24 6 1190 1385 1551 1690 1775 1866 1951

analytical tools for managing rock fall hazards in australian coal mine roadways

Documents

general information

qualified person

organisation person

analytical tools

seedsman geotechnics

gordon geotechniques

n gordon n aziz civil

reli liability