rr831 - evaluation of tensioned and non-tensioned long ... · 8.2 suggested revision of dmciac...

Health and Safety Executive

Evaluation of tensioned and non-tensioned long tendon reinforcement in UK deep mining conditions

Prepared by Rock Mechanics Technology Limited for the Health and Safety Executive 2010

RR831 Research Report



David Bigby, PhD, BSc (Hons) Ken Hurt, PhD, BSc (Hons) Chris Reynolds, BSc (Hons) Robert Brown, BSc (Hons)

Rock Mechanics Technology Ltd Bretby Business Park Ashby Road Burton-on-Trent Staffordshire DE15 0QD

A research programme has been carried out by RMT in support of the revision of Part 2 of the British Standard for strata reinforcement components in coal mines, covering flexible systems for roof reinforcement. This continued work commenced under a previous HSE Project, ‘Testing and standards for reinforcement consumables’.

A particular focus was to compare tensionable and non-tensionable reinforcement systems, prompted by the introduction of tensionable systems to British coal mines. A review of previous research indicated conflicting claims for tensionable systems in terms of theoretical advantages and practical experience. The research included laboratory testing, underground measurement and analysis of underground monitoring data. Advice and draft Annexes were provided to the BS Committee and a revision of the DMCIAC guidance on the use of cable bolts to support roadways in coal mines drafted. The work highlighted practical problems concerning application of the tensionable systems in use in UK coal mines but did not exclude their future applicability provided they comply with the revised Standard.

This report and the work it describes were funded by the Health and Safety Executive and co-funded by the EU Research Fund for Coal and Steel. Aspects were also co-funded by UKCoal Ltd and various manufacturers through supply of materials for testing. The report’s contents, including any opinions and/ or conclusions expressed, are those of the authors alone and do not necessarily reflect HSE policy nor the opinions of any of the co-funding parties.

HSE Books

© Crown copyright 2010

First published 2010

You may reuse this information (not including logos) free of charge in any format or medium, under the terms of the Open Government Licence. To view the licence visit www.nationalarchives.gov.uk/doc/open-government-licence/, write to the Information Policy Team, The National Archives, Kew, London TW9 4DU, or email [email protected].

Some images and illustrations may not be owned by the Crown so cannot be reproduced without permission of the copyright owner. Enquiries should be sent to [email protected].

ii

www.nationalarchives.gov.uk/doc/open-government-licence

CONTENTS

EXECUTIVE SUMMARY vii

1. INTRODUCTION 1

2. THEORETICAL ASPECTS OF TENSIONED TENDONS 3 2.1 Rockbolting and Tensioning Practice 3 2.2 Analysis of the Pretensioning Effect 9 2.3 Evidence for the Pretensioning Effect 12 2.4 Summary 15 2.5 Bibliography 17

3. LABORATORY AND UNDERGROUND TESTING OF LONG TENDONS 21 3.1 Grout Encapsulation Testing 21 3.2 Variation of Bond Performance with Grout Strength 24 3.3 Field Sample Grout Testing 25

4. UNDERGROUND MONITORING OF PERFORMANCE AND EFFECTS OF TENSIONED LONG TENDONS 27 4.1 Historical Data Analysis 27 4.2 Underground Monitoring of Long Tendon Tensioning at Colliery A 30

5. IMPROVED MODELLING OF FLEXIBLE LONG TENDONS 37 5.1 Representation of Tensioned Tendons and Truss Systems 37 5.2 Application to Maingate Support on Drivage at Colliery C 39 5.3 Application to Maingate Support on Retreat at Colliery C 42 5.4 Conclusions on Modelling Flexible Long Tendons 44

6. ADVICE AND DRAFTS PROVIDED TO B S REVISION COMMITTEE 47 6.1 Advice and Recommendations 47 6.2 Draft Test Procedures / Annexes 48

7. LABORATORY TESTING OF ALTERNATIVE RIB REINFORCEMENT SYSTEMS 51

7.1 Rib Reinforcement Systems Tested 51 7.2 Test Procedures 52 7.3 Test Results and Discussion 53 7.4 Conclusions and Recommendations 60

8. SUGGESTED REVISION OF DMCIAC CABLEBOLTING GUIDANCE DOCUMENT 63

8.1 Introduction 63 8.2 Suggested Revision of DMCIAC Cablebolting Guidance 63 8.3 Revised Cablebolting Guidance – Appendix 1 70 8.4 Revised Cablebolting Guidance – Appendix 2 72 8.5 Revised Cablebolting Guidance – Appendix 3 74 8.6 Revised Cablebolting Guidance – Appendix 4 76

9. CONCLUSIONS AND RECOMMENDATIONS 79

iii

v

83 10. REFERENCES

FIGURES

Figure 2.1 Comparison of the load distribution of a normal anchor and a single bore multiple anchor (after Barley and Windsor, 2000) 85

Figure 2.2 The buckling beam concept for roof stability (after Strata Engineering, 2001) 86

Figure 2.3 External load applied to a pretensioned bolted joint 87 Figure 2.4 External load applied to a pretensioned lifting bolt 88

Figure 3.1 Laboratory short encapsulation pull test results for Megastrands and CBG grout at 1 and 3 days curing 89

Figure 3.2 Laboratory short encapsulation pull test results for Megastrands and CBG grout at 7 and 42 days curing 90

Figure 3.3 Variation of maximum load, bond strength and system stiffness with grout unconfined compressive strength (cube samples) 91

Figure 3.4 Variation of unconfined compressive strength with density for bottle samples of CBG grout obtained from UK coal mines 92

Figure 3.5 Variation of unconfined compressive strength with density for bottle samples of HPRG grout obtained from UK coal mines 93

Figure 4.1 Schematic of 10’s maingate, colliery C 94 Figure 4.2 Monitoring results for station 4, 10’s main gate, colliery C 95 Figure 4.3 Monitoring results for station 3, 10’s main gate, colliery C 95 Figure 4.4 Monitoring results for station 2, 10’s main gate, colliery C 96 Figure 4.5 Monitoring results from type ‘B’ telltales, 10’s main gate, colliery C 97 Figure 4.6 Monitoring results from cablebolt type ‘A’ telltales,

10’s main gate, colliery C 98 Figure 4.7 Monitoring results from cablebolt type ‘B’ telltales,

10’s main gate, colliery C 99 Figure 4.8 Monitoring results from type ‘A’ telltales, 22’s main gate, colliery C 100 Figure 4.9 Combined displacement from type ‘A’ and ‘B’ telltales,

570, 590 and 610 MM, 19’s tail gate, colliery C 101

Figure 4.13 Approximate position of installed instruments on 21 March 2007 105 Figure 4.14 Approximate position of installed Megastrands and tensioning on

Figure 4.10 Schematic of T18’s, face line, colliery B 102 Figure 4.11 Planned support pattern in widened face line 103 Figure 4.12 Section of widened face line support and instrumentation pattern 104

27 March 2007 106 Figure 4.15 Proposed Megastrand monitoring instrumentation, 302’s face line

March 2007 107 Figure 4.16 Time trend for cablebolt telltale number 13 at 196 m 108 Figure 4.17 Colliery C face line station 6 at 192 m

Figure 4.20 Sonic extensometer 196 m, displacement & strain against

109 Figure 4.18 Colliery C 302’ face 197 m CL 297 RREXTO1 110 Figure 4.19 Colliery C 302’ face 197 m CL 297 RREXTO1 – cont. 111

distance into strata 112 Figure 4.21 Sonic extensometer 196 m, displacement against time 113 Figure 4.22 Sonic extensometer 196 m, displacement against time during

Megastrand tensioning period 114 Figure 4.23 Strain gauged rockbolt 1. Mean microstrain and microstrain

difference against distance along bolt 115

iv

Figure 4.24 Strain gauged rockbolt 2. Mean microstrain and microstrain difference against distance along bolt 116





Figure 5.1 Modelled strata sequence 3.2 m roof mudstone 121

Figure 5.3 Modelled roof displacements with roof bolts failed and no

Figure 5.4 Modelled roof displacements with roof bolts failed and pretensioned

Figure 5.5 Modelled roof displacements with roof bolts failed and untensioned

Figure 5.6 Modelled roof displacements with roof bolts failed and tensioned

Figure 5.7 Shear strains and bolt loads with pretensioned flexible bolts

Figure 5.8 Shear strains and bolt loads with pretensioned truss system as

Figure 5.2a Coal strength properties 121 Figure 5.2b Siltstone strength properties 121 Figure 5.2c Mudstone strength properties 121

additional support 122

flexible bolts as additional support 122

flexible bolts as additional support 123

truss system as additional support 124

as additional support 125

additional support 125 Figure 5.9 Proposed bolt pattern at development face 126 Figure 5.10 Strata sequences 126 Figure 5.11 Roof condition with proposed support and sequence A 127 Figure 5.12 Roof condition with proposed support and sequence B 128 Figure 5.13 Alternative bolt patterns at development face 129 Figure 5.14 Roof movement with alternative support patterns (sequence A) 129 Figure 5.15 Bolt strains with alternative support patterns (sequence A) 130 Figure 5.16 Stress increase to represent face retreat 130 Figure 5.17 Modelled roof condition for retreat 131 Figure 5.18 Roof movements for face retreat 132 Figure 5.19 Bolt strains for face retreat 132 Figure 5.20 Fail bolts and cables 133

Figure 7.1 Mean residual loads at 50 mm displacement 134 Figure 7.2 Comparison of achieved mean maximum loads between different bolt 135

types and increasing hole size

Figure 8.1 Water diverting dual height tell-tale for cablebolting (white, blue, yellow 135 bands)

Figure 8.2 Water diverting triple height tell-tale for cablebolting (green yellow red 135 bands)

v

APPENDICES

APPENDIX 1 DRAFT TEST PROCEDURE – GROUT ENCAPSULATION TEST – ‘BOTTOM UP’ GROUTING 137 A1.1 Principle 137 A1.2 Apparatus 137 A1.3 Procedure 137 A1.3 Results 139

APPENDIX 2 DRAFT TEST PROCEDURE – GROUT ENCAPSULATION TEST – 141 ‘TOP DOWN’ GROUTING A2.1 Principle 141 A2.2 Apparatus 141 A2.3 Procedure 141 A2.3 Results 142

APPENDIX 3 DRAFT TEST PROCEDURE – DETERMINATION OF BOND 143 STRENGTH AND SYSTEM STIFFNESS – CEMENTITIOUS GROUT ANCHORED SYSTEMS A3.1 Principle 143 A3.2 Apparatus 143 A3.3 Procedure 144 A3.3 Results 146

APPENDIX 4 DRAFT TEST PROCEDURE – SHEAR TEST ON TENDON / GROUT 151 SYSTEM A4.1 Principle 151

A4.2 Apparatus 151 A4.3 Procedure 151 A4.3 Results 152

vi

EXECUTIVE SUMMARY

A three year research programme has been carried out by Rock Mechanics Technology Ltd for the Health and Safety Executive in further support of the revision of Part 2 of the British Standard for strata reinforcement components in coal mines which is intended to cover flexible systems for roof reinforcement. This programme continued work commenced under a previous HSE Project, “Testing and standards for reinforcement consumables”. Both Projects were cofunded by grants from the EU Research Fund for Coal and Steel and aspects were also funded by UKCoal Ltd and various manufacturers through supply of materials for testing.

A particular focus of this programme was to investigate the potential of tensionable reinforcement systems in comparison with non-tensionable systems, prompted by the application of tensionable systems by the British coal mining industry and the need to include them within the Standard. To this end a detailed review of previous research into tensionable systems was first carried out. This indicated conflicting claims for tensionable systems both in terms of theoretical advantages and practical experience, plus some potential problems.

Much of the Project work involved investigation of practical aspects of tensionable systems currently available to the UK industry, particularly aspects of grout mixing, pumping, encapsulation and strength development. This occupied a greater proportion of the available resources than originally envisaged, as the research encountered unexpected issues in this area. A laboratory programme of testing the ability of the systems to achieve full encapsulation was completed. This allowed development of encapsulation tests for both “top- down” and “bottom up” grouting systems to be incorporated into the draft revised Standard. The “top-down” testing found problems with achieving encapsulation with the currently used tensionable strand and thixotropic grout, which could encourage the addition of extra water to the mix, thus producing a weaker grout and jeopardising full encapsulation and system performance. This prompted a series of laboratory short encapsulation pull tests to investigate the relationship between compressive strength of grout and axial bond stress. It was concluded that it would be difficult to justify any reduction in the grout strength requirements contained in the current version of the Standard. A database was compiled of the compressive strength test results from samples of grout collected from underground over the last 14 years. This showed that the actual strengths achieved, particularly in more recent years, were generally lower than those required from laboratory tests as defined in the current Standard. However, where the grout achieved the manufacturer’s specified density, then the sample strengths were satisfactory. This highlighted the need for training and good practice underground.

Field experience and monitoring results from recent applications of tensionable tendons in UK coal mines were analysed, though data was scarce. This indicated that tensionable tendons appeared to have significant potential but that their use without timely post grouting had contributed to a severe roof fall. This fall was partly attributed to delaying tendon grouting until significant roof movement had occurred and the absence of further remedial action levels, exacerbated by the inherent difficulties in achieving full encapsulation with the particular tensionable system deployed. It was concluded that, where tensionable systems are used, they should be post-grouted as quickly as practical and at least within 24 hours of installation, earlier in rapidly deforming ground. Also, it is essential that secondary action levels and actions are incorporated into the managers’ monitoring scheme to ensure that appropriate action is taken if continued roof deformation occurs following a first level of remedial reinforcement.

An intensely instrumented field exercise, co-funded by UKCoal Ltd, was undertaken to investigate whether reported re-closure of bed separations during the tensioning of tendons

vii

could be reproduced and measured. In this case no compressive deformation of the rock was measured during the tensioning period, though some strain gauged rockbolts in the vicinity did experience compression. It was concluded that significant re-closure of bed separations was only likely to occur in the immediate roof and where large bed separation had previously occurred. This exercise demonstrated application of the new generation of geotechnical instrumentation which has been developed for design and safety monitoring of roadways reinforced with long tendons, including the remote reading extensometer system with local underground readout and strain gauged rockbolt readout with data logging facility.

A method of better representing tensionable tendon reinforcement and “truss” systems in FLAC3D numerical models of strata and support behaviour was developed. This was applied to hypothetical cases and real case studies. These studies did not show any advantage from tensioning fully encapsulated long tendons under the conditions modelled. However, the limitations of the modelling technique used must be borne in mind, whereby the models are not able to simulate bed separation and consequent loss of bedding shear strength, nor its reestablishment. This mechanism is often claimed as a significant feature of tensioned systems. The studies also showed that, where loading and deformation were symmetrical, deployment of reinforcement tendons towards the centre of a roadway was more efficient than towards the sides and that it was far more effective than placing tendons over the ribsides where they could act as “truss” anchors. However the modelling also showed that truss supports could be effective in preventing falls of ground albeit after significant rock failure and roof deformation had occurred. They were unable to prevent this rock failure and in typical UK conditions any pretension of trusses was rapidly lost due to horizontal stress induced roof shortening.

A significant amount of advice and recommendations were supplied to the relevant British Standards Committee over the period of the Project, including draft annexes and advice on how to categorise the various systems and system components. One key recommendation was that BS7861:Part 2 should only cover systems for roof reinforcement and not those applied to ribsides. This allowed exclusion of the lower capacity systems often used for ribside reinforcement, to avoid confusion. A new, Part 3 of the Standard, for ribside long tendon reinforcement, is recommended for development which can cover these lower capacity systems.

In preparation for such a Part 3 and for revision of the DMCIAC cable bolting guidance document, which includes rib reinforcement, a comprehensive laboratory short encapsulation pull testing (LSEPT) programme was undertaken on a range of potential rib reinforcement consumables, particularly considering their performance in coal under relatively high deformations. This work was also co-funded by UKCoal Ltd. This work showed that capsule PUR could be quite effective as a rib reinforcement encapsulant provided that the components could be properly mixed and that moisture induced foaming could be avoided. The limited comparisons between laboratory short encapsulation pull tests in coal and sandstone indicated quite similar results. The tests also revealed the differences in reinforcement performance depending upon confinement, which is of particular significance in broken coal ribs. In general this work confirmed that the current practice of employing AT capsule resin embedded rib reinforcement for application at the face of the heading and pumped cementitious grout encapsulated reinforcement for outbye remedial reinforcement was generally appropriate.

A draft revision of the DMCIAC document, “Guidance on the use of cablebolts to support roadways in coal mines” was prepared based upon the results of the research programme and developments in the technology since its publication in 1997. It pays particular attention to underground grouting operations and training. This is intended to be considered for development with a view to publication alongside the revised BS7861: Part 2. It was also noted that a minor revision of definitions within the DMCIAC supplementary guidance on the use of flexible bolts in reinforcement systems for coal mines is required.

viii

1 INTRODUCTION

Long tendon reinforcement of coal mine roof was introduced to UK coal mines around 1990 in response to increasing use of rockbolts as primary support, and the consequent need to reinforce above the bolted height, particularly during face retreat. Initially, birdcage cablebolts, introduced from Australia as part of a technology transfer process, were used with considerable success. The original British Standard for cablebolting consumables (reference 1) was written to formalise their use. Since the Standard was published (in 1997) however, a number of different long tendon designs have been introduced and adopted by the mining industry with varying degrees of success, and, around five years ago, systems designed to be tensioned before grout encapsulation, or “pretensioned”, were introduced. Pretensioned systems were researched extensively by Australian mining consultants, and used widely in Australian coal mines where both rockbolts and long tendons were optimised to accept, in some cases, very high pretension loads. More details on Australian applications and results will be found in later chapters, but suffice to say here that variants of Australian originated systems have been used in the UK. Currently there is considerable debate in academic circles concerning the relative merits and disadvantages of these systems, but little research has been undertaken to improve the level of understanding in UK mining conditions, which can differ from those experienced abroad, particularly regarding depth and rock strengths. Some workers argue that tension achieves a major improvement in ground control. Others suggest that any potential gains are outweighed by the reduced strain to failure and other potential disadvantages such as problematic installation. Yet others believe that there is no material benefit to strata control even when properly installed.

The widespread use of long tendons (other than birdcaged systems) prompted the need for a revision of the British Standard (reference 1), and a revision committee first convened in 2005. An earlier research programme (reference 2) carried out by Rock Mechanics Technology Ltd (RMT) provided the committee with suggested test methods for laboratory evaluation of long tendon performance, and an extensive test programme provided a performance comparison of current systems and produced performance bench marks for inclusion in the revised Standard. However, the project was completed at an early stage relative to revision of the Standard, and did not include research specific to pretensioned systems now being used. RMT therefore proposed to carry out a new project providing continuing assistance to the revision committee, and a comprehensive assessment of pretensioned long tendons. The project would run for three years from 1st October 2005 to 30th September 2008, with funding provided by the HSE, the European Research Fund for Coal and Steel (RFCS), and the mining industry.

Via the new project, RMT was to provide technical support to the revision committee, with particular regard to the provision of further test procedures for assessment of consumable performance. RMT would also undertake an appraisal of pretensioning and attempt to answer questions arising from current use, experience and the perceived level of understanding of this type of support. For example,

a) In what circumstances is their use appropriate? b) What level of pre-tension is acceptable? (pre-tensions approaching 70% of ultimate

tensile load have been used) c) What is the effect of pre-tension on the ability of a long tendon reinforcement system to

accept additional load and/or strain? d) What is the effect of pre-tension on the shear properties of the reinforcement system? e) Is there a genuine advantage in the use of pre-tension in certain circumstances and if so,

what circumstances?

1

f) What monitoring systems are appropriate for tensioned long tendons? g) What monitoring action levels should be employed to ensure that timely remedial action

is taken prior to system failure? h) What type of remedial action is appropriate in roadways reinforced by tensioned cables

and what response time is required? i) Do colliery ground control management systems need reviewing to take account of pre-

tensioned systems and if so, in what way? j) When should tension be applied and when should a tensioned system be grouted? k) How can it be ensured that a tensioned system, including any top down grouting system

employing thixotropic grout, is fit for purpose? l) Should roof support design methodologies such as numerical modelling be revised to

incorporate tensioned systems and if so in what way?

In order to provide satisfactory data in response to these questions, a framework of aims and deliverables was constructed. The aims of the Project were to achieve a better understanding of the comparative behaviour of tensioned and non-tensioned long tendon reinforcement in UK deep mining conditions, and to furnish this understanding to HSE and the industry. The individual deliverables were as follows:

1. Develop suitable laboratory tests for assessing tensioned long tendon system performance.

2. Produce a draft addendum/supplement to the DCMIAC document, “Guidance on the use of cablebolts to support roadways in coal mines”, covering the use of tensioned long tendon reinforcement.

3. Provide continued technical input to the Working Party and revision of BS7861:2– 1997.

4. Produce a final report (and interim reports) on the outcome of the Project, suitable for publication on the HSE website.

5. Provide a technical seminar to HSE and other interested parties on the findings of the Project.

The Project commenced on 1st October 2005, and delivered according to the requirements described above via regular progress statements. However, some two years in, it was found that research into and testing of thixotropic grouts required for the pretensioned system used in UK coal mines was absorbing unexpectedly high levels of effort, and problems encountered in the research, which reflected obvious difficulties being experienced in the field, needed to be resolved. After discussion with HSE supervising officials, it was agreed that the work programme would be altered to allow the completion of the grout studies. Some other aspects would be terminated or omitted, i.e. a cessation of work on a specific short encapsulation pull test for tensioned tendons (this being thought unnecessary), and abbreviation of work on analysis and comparison of tensioned and non-tensioned tendon monitoring data.

This document reports in full the outcomes of the Project in its final form and includes

• a review of theoretical treatments of pretensioning • a description of the laboratory and underground testing of long tendon systems • analysis of monitoring data from sites where pretensioned systems were used

(abbreviated) • theoretical modelling of flexible long tendons • recommendations and draft procedures (annexes) prepared for the revision committee • a review of laboratory short encapsulation testing of rib reinforcement systems, and • a proposed revision of the DMCIAC cablebolt guidance document.

2

2 THEORETICAL ASPECTS OF TENSIONED TENDONS

This review examines the available information worldwide on the tensioning of rockbolts and tendons, during or shortly after installation, and particularly on the recently adopted practice of applying higher pretension loads. The purpose of the review is to identify the potential benefits and drawbacks of high pretension loads.

2.1 ROCKBOLTING AND TENSIONING PRACTICE

2.1.1 What is pretensioning?

The term pretensioning is used in mining, mechanical and civil engineering to describe three somewhat different processes. It is therefore useful to compare these to avoid confusion and to see to what extent ideas and practices are generally applicable.

In mining, pretensioning is the practice of applying a predetermined load to rock reinforcement tendons during or shortly after installation. It was originally applied to mechanical point anchor bolts in order to lock the bolt head in position but with resin partially anchored types it is also necessary to tension the end plate to the roof to make the bolt effective as a support. Tensioning is also used with fully encapsulated rock bolts in conjunction with a two speed resin system. Tensioning loads as high as 50 tonnes are now used with higher capacity systems such as strand type bolts (Rataj, 2002). This practice has been claimed to increase tendon effectiveness as reinforcement.

Pretensioning in civil engineering is the tensioning of steel reinforcing elements passing through concrete beams before they are cast. Some of this tension is then locked in and generates compressive stress in the concrete when it sets so that no tension develops in the beam in service. Because concrete is strong in compression and weak in tension this allows beams to be lighter and to be used over longer spans. Post tensioning is a similar process except that the reinforcing tendons are placed in ducts passing through the beams after they have been cast and cured. The tendons are then tensioned against end plates or anchors to put the beams in compression.

This situation is fundamentally different from the way in which reinforcement is used around mine openings. In a mine, the tendon is installed approximately normal to the rock surface, rather than parallel to it, as a conventional reinforcing action requires. Consequently a pretensioned mining tendon compresses the rock under the end plate, but a corresponding tension develops close to the anchor position within the rock. This is in contrast to the concrete beam situation where the reinforcement induces compression in the concrete along the complete beam length.

In mechanical engineering, pretension loads are used in tightening bolted joints. A conventional mechanical bolt used to clamp components together is tensioned by the mechanical advantage which stretches the bolt. This tension is known as pretension because it exists before any other forces are applied to the joint. The pretension compresses the mating surfaces, preventing them from sliding or separating under load. Pretension loads are typically 70% of bolt tensile strength.

When an external tensile load is applied to the joint the effect is to decrease compression of the mating surfaces without significantly increasing tension in the bolt - typically 90% of applied load is absorbed by decompression of mating surfaces. This is because the compression of the

3

surfaces has a very high stiffness compared with the bolt stretch, so compressive load reduces much more quickly than the bolt tension increases as joint and bolt strain develop, up to the decompression point at which joint mating surfaces are no longer in contact.

At first sight this situation might be considered analogous to the mining one. However there are important differences. The main difference is that the material being “bolted together” in mining is rock rather than metal. The elastic modulus of rock is typically some ten to forty times lower than that of the steel tendon. Consequently we have the opposite situation to the mechanical engineering one - the “joint stiffness” could be much lower than the bolt stiffness so that external loads further increase tension in the bolt or tendon despite the pretension load. The other major difference relates to the situation geometry - the mechanical bolt clamps together two components across a single finite joint. In contrast the rock material is semi infinite in extent and multiple potential failure “joint” positions may be present both within the tendon length and beyond it.

In fact the mining pretension situation is analogous to gravity loading from a foundation as pointed out by Gray and Bates (1998). The elastic solution for a homogeneous medium (the Boussinesq equation) predicts a rapid dissipation of applied load with distance below the foundation. Using the solution for a circular footing as analogous to a tensioned tendon end plate suggests that, directly above the plate and at a distance equal to four times the plate radius, the maximum induced confining stress is just 10% of the applied stress from the pretension load. Consequently, if the rock behaves elastically, any strengthening effect from pretensioning should be limited to a small rock volume in the immediate vicinity of the end plate. Gray and Bates also point out that maximum confining stresses induced by pretensioning are at least an order of magnitude less than likely in situ stresses, and that the tensile stress generated just above the anchor horizon could reduce roof stability.

Fundamental questions therefore arise as follows:

• Does pretensioning really strengthen the rock mass? • Are pretensioned tendons more effective than non pretensioned ones as support? • Are there any disadvantages of pretensioning?

The answers to these could lead in turn to further questions:

• What is the optimum bolting system for use with pretensioning? • What is the optimum pretension load and how should it be applied? • What effect does tendon length have? Should pretensioned tendons be post grouted? Is

the failure strain of pretensioned tendons the same as non pretensioned ones? • Can pretensioned tendons be used for lifting or suspension of loads? • To try to answer these questions the available evidence from existing publications and

from consideration of engineering principles is considered in the review which follows.

2.1.2 Rockbolt systems

The large majority (about 85% according to Minova, 2006) of rockbolts used in coal mines worldwide are now full column bonded with the remainder being partially anchored with either a mechanical, resin or combination anchor. The use of mechanically anchored bolts in the USA has been in steep decline since 1985 (Tadolini and Mazzoni 2006), although combination resin/mechanical anchor types are used, particularly in difficult conditions (Mark 2000, Su and Poland, 2007).

4

All partially anchored types are tensioned in order to tighten the end plate against the rock. Without this the bolt would not be effective as support. Since the advent of fully encapsulated systems in the late 1970’s, a distinction has been made between “rockbolts” which are point anchored or partially encapsulated and tensioned and provide “active” support, and “dowels” which are fully encapsulated untensioned bolts and provide “passive” support. However this distinction is now misleading as most fully encapsulated systems outside the USA are also tensioned. The use of two resins of different gel times (“fast” and “slow” resins) with full column bolts has been common practice for some years, with the primary object of facilitating bolt installation. Tightening the nut after the fast resin sets, but before the slow resin gels, results in tension being imparted into a full column bolt. This is known as the torque tension system. The loads generated during normal installation (up to about 3 tonnes) are relatively small.

The majority of bolts used in the USA are forged head rebar bolts (Tadolini and Mazzoni 2006). Forged head bolts cannot be tensioned in the conventional way and normal practice has been to use the installation drilling mast to hold these in position until the resin cures. However the technique of “thrust bolting,” developed in the USA and patented in 1991, has emerged in which the drill crowd force is used to push the bolt end and compressible plate tight to the roof as the resin sets. This results in tension being developed in the bolt following the elastic expansion of the compressed rock and plate, once the thrust force is removed.

The advent of flexible bolts - strand type steel tendons installed with polyester resin - in the 1990’s has significantly increased the maximum length of rockbolt which is available to counter difficult support conditions. These bolts have an end plate secured using a nut or barrel and wedge anchor and can be tensioned. Tension loads of around 5 tonnes were originally used in Australia, but more recently proprietary types such as the Hi-Ten and Megastrand bolts, capable of being pretensioned to 25 tonnes or more, have been introduced (Rataj and Yearby, 1999). At the same time, additional tension loads of up to 10 tonnes have been applied to full column rockbolts. The use of higher pretension loads with bolts and cables in Australia is generally considered to improve roof control in more difficult conditions (Fuller 1999, Hebblewhite 2006).

2.1.3 Tensioning and tension loss with partially anchored bolts

The use of mechanically anchored bolts has now declined to less than 4% in US coal mines. The principal reason according to Tadolini and Mazzoni (2006) stems from the problem of tension loss with these systems during and following installation. This has resulted in additional labour cost as tension has to be checked regularly, and the bolts re-torqued if necessary. Costs then compare unfavourably with resin bolts.

The first problem is achieving the desired tension during installation. Generation of the preload by tightening is problematic. Tadolini (1991) found that torque wrenches over estimate preloads by more than 45% because of frictional losses as the bolt head is tightened against the plate. Lubricated thrust washers reduced the problem. Fuller et al (1981) produced design charts for estimating the tension achieved from the applied torque when tightening a threaded end nut against the collar plate. The tension developed depends on the torque applied and thread and collar plate friction. Lubrication again decreased friction. These problems occurred at modest loads. Consequently the development of very high pretension loads by tightening an end nut is likely to prove difficult in practice.

Mechanically anchored systems are also vulnerable to loss of tension following installation. This has been attributed to the high elastic modulus of the steel bolt in comparison with the surrounding rock (the ratio varies from 10 to 40 or more with a typical value of 15) (Unrug et

5

al). The explanation given is that only a relatively small anchorage slip, or decrease in bolt strain through local crushing, shrinkage, weathering or creep of the rock above the end plate, is therefore needed to allow this tension to dissipate. The use of wooden header boards was identified as a particular problem because of the tendency of the timber to creep and shrink, releasing the tension. However this ignores the elastic nature of the compressed rock close to the end plate which should compensate by expanding to maintain contact with the end plate. More likely explanations are that anchor slip with mechanically anchored bolts continues under load to significant levels, especially in weaker rocks, or that permanent deformation of failed material under the end plate limits the elastic compression taking place.

Whatever the reason, loosening or tension loss with partially encapsulated bolts is a common experience in mines (Mark 2000, Van de Merwe and Madden 2002). They also concluded that tensioned bolting in softer rocks such as mudstones was likely to be ineffective due to tension losses and a fully encapsulated system was needed in this case.

Resin assisted mechanical anchors are partially encapsulated with resin which flows around the mechanical anchor and provides additional bonding –these reportedly alleviate the problem of tension loss due to anchor slip. They currently comprise about 9% of the US Market although this appears to be declining (Tadolini and Mazzoni 2006). Su and Poland (2007) describe the use of resin assisted mechanical shell bolts to reinforce laminated roof by Consol in the USA. A longer resin column was found to be more effective in reducing the problem of tension loss. The recent introduction of low insertion force resin has made the installation of this bolt system easier, allowing four Pittsburgh Seam mines to adopt a full column of resin with 1.8 m long resin assisted mechanical shell bolts pretensioned to 9 tonnes. Full column bonding is claimed to lock in the pretension and eliminate the tension loss problem.

2.1.4 Tensioning and fully encapsulated bolts

Resin bonded rebar bolts comprise around 80% of US bolts installed. Although only 12% of these were reported as being tensioned in 2006, the proportion is growing (Su and Poland 2007, Tadolini and Mazzoni, 2006).

There is now a general recognition that fully encapsulated rebar bolts provide a high capacity stiff bolting system which reacts rapidly to rock dilation by transferring load via the bolt to other parts of the rock mass. Pretensioning is not a necessary part of this process, because of the high system stiffness, and untensioned rebar systems in the USA are generally considered to be better suited than partially anchored tensioned types to weak rock, high stress conditions (Tadolini and Mazzoni, 2006). Su and Poland (2007) however argue for the need for tensioning as well to counter immediate roof separation, which occurs particularly with place changing systems in which bolting is carried out up to 12 m from the face. This may provide the motive for the increasing adoption of tensioning with full column resin bonded systems in the USA.

The torque tension system using two speed resin cartridges dominates European and Australian rockbolting practice. However the primary motive for adopting the torque tension system in these countries was originally to facilitate rapid installation with pneumatic leg mounted drills and the relatively small tension loads (up to around 3 tonnes) were primarily to clamp the end plate firmly against the rock. The slow resin sets after any tensioning and this may result in the tension being “locked in” and not so easily dissipated as it is with partially encapsulated systems (Unrug and Thompson, 2002).

Thrust bolting is an alternative method of tensioning full column resin bolts. Unrug and Thompson (2002) describe in-situ tests using a forged head thrust tensioned rebar system using two speed resin. Strain gauged bolts were used to measure the levels of tension achieved, with

6

and without the use of a spring washer in addition to the plate. Tensions of 2.7 tonnes were successfully generated in both cases. Thrust bolting depends on elastic deformation of the rock, washer and end plate to generate the pretension load and the rock under the end plate must not fail and deform under the imposed loads. It is unclear from the literature to what extent thrust bolting is currently used in the USA as a deliberate technique to generate pretension loads.

The concept of increasing rockbolt pretension loads as a means of enhancing bolt performance originated in Australia in the mid 1990’s, (Rataj and Thomas 1997). A thrust bearing was originally used to reduce thread friction to achieve initial bolt tensions of around 7 tonnes. Later it was found that the use of high pressure lubricant plus modification to the nut could achieve the same result. A number of field trials were carried out and the installation of bolts with initial tensions of 10 tonnes or more using on board hydraulic bolters is now common in Australia (Hebblewhite, 2006).

2.1.5 Pretensioning of long tendons

Steel wire strand bolts encapsulated with resin (flexible bolts) also originated in Australia in the early 1990’s. As these generally have high load capacity they are obvious candidates for pretensioning. The maximum length which can be resin encapsulated is around 4 m, although development work is ongoing to increase the encapsulated length to 6 or even 8 m. Normal installation tension varied from 2 to 5 tonnes (Rataj, 2002). Tension loads of around 25 tonnes were used in initial pretension trials (Rataj and Thomas, 1997), but more recently proprietary types such as the Hi-Ten and Megabolt, capable of being pretensioned to 50 tonnes or more, have been introduced (Rataj and Yearby, 1999, McKenzie, 2001).

Post grouting of conventional strand bolts is difficult and is not usually undertaken. Most types can only be post grouted by installing into over-size holes (Fuller 1999, McKenzie 2001). Megabolts can be post grouted but according to Strata Engineering (2001) this is not necessary as the tendons are fully effective once installed near the roadway centre and pretensioned. The more recently introduced Megastrand bolts as used at colliery C in the UK (Adams and Rennison, 2003) were post grouted.

Techniques and equipment to allow the generation of pretension loads up to 60 tonnes have been developed. Rataj (2002) describes these developments for strand bolts. With barrel and wedge type end fittings special rimmed barrels have been developed to reduce tension bleed off as the tensioning device is removed. However 25% loss of applied tension is said to be typical for Hi-Ten bolts even with this modification and initial tension loss remains a serious problem for strand bolts. The application of these high loads in practice is also not without difficulty - Springvale Colliery in Australia reduced tendon tension from 20 tonnes to 12 tonnes in order to replace the 20 kg tensioning equipment with a lighter alternative (Anon, 2000, Bahr 2006).

Barczak et al (2004) describe the application of hydraulic prestressing units as an alternative means of tensioning tendons. These are inflatable metal bladders placed as a collar between the tendon end fitting and the rock. Pressurising with water generated 9 tonnes of pretension, following which the collars are left in place.

In excess of 80% of strand type bolts in Australia are reported to be pretensioned to around 20 tonnes or more (Rataj 2002) and the growth of this practice provides circumstantial evidence of effectiveness. The use of higher pretension loads with bolts and tendons in Australia is generally considered to improve roof control in more difficult conditions (Fuller 1999, Hebblewhite 2006). The main documented experience in the UK relates to the application of Megastrand bolts tensioned to 25 tonnes at colliery C. These were reported as being highly successful when early post grouting was used (Adams and Rennison 2003). Subsequently a roof failure was

7

experienced at another site at the same mine where these tendons were being applied without early post grouting.

2.1.6 Ground anchor installation and progressive debonding

There are obvious parallels between the installation of ground anchors in Civil Engineering and rock and cable bolting in mining - the only essential difference is one of scale as noted by Barley and Windsor (2000). Ground anchors are normally stressed to a design load equivalent to a maximum of 80% of the tendon strength after installation. Very large loads per anchor are often used compared with mining practice although bond stresses over extended anchor lengths are likely to be lower - typically around 1-3 MPa for stranded tendons (Barley and Windsor, 2000).

Prestressing of ground anchors has been cited as justification for introducing similar high levels of pretensioning for mining tendons. The purpose is to restrain the anchored structure in place against imposed loads, and prevent any strain and incipient failure from developing. The length of ground anchors, which usually includes a significant unbonded section, means that considerable displacement would have to take place to generate the design load if prestressing was not practised. Mining tendons are much shorter, frequently fully encapsulated and generate load very quickly in response to rock movement. Consequently there is not the same obvious requirement for prestressing mining tendons as there is for ground anchors.

Ground anchorages consist of a fixed anchor, a free tendon length and a substantial anchor head which transmits the tensile load in the tendon to the rock surface or structure. The free tendon length is deliberately decoupled using plastic sheathing and/or grease to limit the anchor length over which stress is transmitted to the ground (Littlejohn 1991). The limited length is used to ensure that stress is applied to the planned anchor zone in stable ground and also to prevent a phenomenon known as “progressive debonding” which is a perceived problem arising with long anchorage lengths.

Progressive debonding arises as a result of the non linear distribution of bond stress along extended anchorage lengths. Bond stress is concentrated at the proximal end and the distal end is not initially stressed. As the load increases, bond yield occurs at the proximal end and the load concentration zone progresses along the anchor length. Consequently the bond can fail despite the apparent average anchor bond stress being well below the bond failure stress. This has the consequence of limiting both the maximum load which can be applied to a single anchor and the maximum effective bond length. Typically a 6-7 m bond length is optimal according to Barley and Windsor (2000) - beyond that load capacity does not increase with bond length. The concept of single bore multiple anchors was devised by Barley to get round this problem. A number of tendons in one borehole are anchored in progression to distribute the load along the anchor length (Figure 2.1).

Progressive debonding has not been considered as a problem for mining tendons to date. It is a function of the anchor stiffness (controlled by the tendon) compared with the ground stiffness according to Barley and Windsor. The worst case applies with low stiffness tendons such as GRP installed in relatively stiff ground such as harder rock. Barley and Windsor give data relating to tendons installed in London Clay which indicate the maximum effective bond length could be as little as 2 m for GRP and 4 m for steel. This suggests that progressive debonding potentially could lead to bond failure developing at lower than expected loads for mining tendons installed into stiffer rock and then pretensioned to high loads. The potential for this problem to occur needs to be considered for tensioned tendons. This would be indicated by a non linear distribution of load along the tendon anchor length following tensioning, with load

8

not initially developing towards the distal end, and this could be checked with strain gauged tendons.

Progressive debonding can be suppressed by making the grouted length greater than the fixed anchor length, because the additional grout column along the decoupled length provides additional support to the proximal end of the fixed anchor length grout where the bond stress is applied (Barley and Windsor, 2000). In mining terms for pretensioned tendons this implies that decoupling the tendon over part of the encapsulated length could prevent progressive debonding from developing.

2.2 ANALYSIS OF THE PRETENSIONING EFFECT

2.2.1 Pretensioning and reinforcement mechanisms

The main mechanisms by which pretension improves roof support have been claimed to be as follows:

i. Compression of the roof above the end plate adds confinement and so strengthens the roof material (Zhang and Peng 2002)

ii. Clamping together of thinly laminated roof beds using pretension forms them into a thicker beam which better resists shear deformations (Stankus and Peng 1996)

iii. Pretension increases the support stiffness so that it provides immediate resistance to roof movement and so enhances support performance (Tadolini and Mazzoni 2006) (Seedsman 1998)

iv. The application of pretension reduces or eliminates bed separations, which develop before bolts are installed, to restore the structural integrity of the roof (Seedsman 1998, Zhang and Peng 2002, Hebblewhite, 2006)

Most advocates would consider that a combination of some or all of these mechanisms operates in practice. Perhaps the most prominent explanation of the pretensioning effect to date has been developed by Strata Engineering in Australia (Rataj M and Thomas R 1997, Strata Engineering 2001). Pretensioned tendons installed near the roadway centre prevent or limit buckling of immediate roof beds under end loading from horizontal stress. The mechanical advantage accruing from the tendon position means that a relatively small restraining tension load provides a large additional resistance to roof bed buckling. Pretensioning is effective because it prevents buckling from developing more effectively than “passive” tendons (Figure 2.2).

It is unclear if Strata Engineering consider that application of pretension to tendons can close up existing bed separations. This is put forward as a mechanism by Seedsman 1998 and others and has been reported in practice - for example by Adams and Rennison (2003).

Gray and Finlow Bates (1998) considered pretensioning loads as applied through the tendon end plate to an elastic rock mass and drew the analogy with application of foundation load to the ground. This is dissipated rapidly in three dimensions so any strengthening effect should only be local to the plate. In any case the additional confining pressures generated by pretension across potential shear planes are small in comparison with likely in situ stresses.

There are also disadvantages to pretensioning according to Gray and Finlow Bates. A tensile stress zone is generated above the anchor zone and, because rock is weak in tension, this could make failure just above the tendon more likely. The bolt stiffness is unchanged by pretensioning since it depends on the elastic modulus of steel which is constant. However, the remaining elastic stiffness range before yield is reduced - effectively a loss of capacity. Overall they consider that pretensioning tendons brings no significant advantage, and that longer bolts and tendons should be fully grouted for maximum effectiveness.

9

The presumption that the rock behaves elastically is however questionable - at higher in situ stresses shear deformation and possibly bed separation and buckling may be developing before reinforcement takes place. It is in this latter situation that pretension is usually claimed to be effective in preventing tensile bedding plane separation in the first metre or so, resisting bedding plane shear and increasing the residual strength of failed rock within the compression zone.

In contrast to Strata Engineering, Seedsman (1997) advocates, in laminated roof conditions, the use of pretension at the edges of intersections to resist shear movements, with unpretensioned types to maximise capacity in the centre. He also considers that the outward inclination of bolts away from the roof centre line could be as effective as pretensioning as a means of increasing shear resistance.

2.2.2 Numerical modelling of bolt pretension

Very few modelling studies to examine rockbolt pretension mechanisms have been reported to date. Some work has been done in the USA and Canada looking at partially anchored and tensioned bolts within larger roadway models. Stankus and Guo (1997) described the development of a two dimensional elastic finite element model with gap elements to represent bedding plane partings and one dimensional bar elements as bolts. This approach is too simplistic to give reliable predictions as noted by Mark (2000).

Bouteldja (2000) used modelling to predict load distributions in pretensioned and non pretensioned cable bolts, again using two dimensional finite elements. A spring connection between the cable and ground was used to represent the grout column and the end plate was fixed to the rock. The load distribution along tensioned post grouted cable bolts and rock bolts was found to depend on the type of grout and the type of anchor used. Pretension directly added to loads imposed by ground movement.

Zhang and Peng (2002) describe a more sophisticated three dimensional modelling approach using the ABAQUS finite element program. In this case element failure was detected using the Mohr Coulomb failure criterion. Strata layers and two bedding planes were simulated and the bolts were modelled as three dimensional beams with the anchor length tied to the rock and the end plate in contact with the rock surface. A 10 tonne pretension load was applied to the bolts. Bolt installation 8 m from the face was simulated by which time significant displacement could develop depending on loading conditions.

The model suggested that bed separation develops as the immediate roof bed deforms under horizontal stress loading. Pretensioned bolting was effective within the lower zone of compression up to 0.9 m into the roof in closing the separation and increasing the residual strength of yielded roof material. The compressive stress magnitude in the upper compressive zone was very small compared with the vertical stress at this point and therefore had no effect on model behaviour. They conclude that pretension closes cracks and separations and gives a high resistance to subsequent movement.

In subsequent papers, Zhang and Peng modelled three way and four way intersections supported by tensioned bolts and verified the model by comparing predictions with an instrumented case study from 1991. Agreement was reported as good. They simulated bedding planes 0.6 m and 1.5 m into the roof. These opened when mining was simulated. With a three way intersection 10 tonnes pretension closed both bedding planes. For a four way intersection, 12 tonnes pretension closed the lower bedding plane but not the upper. They conclude that the benefits of pretensioning in terms of increased shear resistance and closing of bed separations was normally limited to the first 0.9 m of roof.

10

Although recent modelling work demonstrates improved sophistication, more development is needed to give a fully convincing simulation of rockbolt pretension mechanisms. In particular the simulation of the bolt system needs to be improved to include the resin annulus and associated interfaces, together with their engineering properties. Morsey et al (2004) describe a more realistic 3D simulation of a fully grouted bolt using beam elements for the rebar embedded in brick elements for the grout. Friction properties for the interfaces can be input. A pull test was simulated in this case, but potentially similar models would be of value in examining pretension in a more realistic way.

2.2.3 Pretension and the lifting or suspension of loads

The use of rockbolts to suspend pipe ranges or other services, or as anchor points for lifting equipment is normal practice in mines. UK guidance limits the load which can be imposed on standard support bolts to less than 1 tonne. For greater loads dedicated lifting bolts should be installed. These are normally partially anchored and unplated to prevent load build up on the bolt end resulting from rock deformation. Adams and Rennison however report the use of pretensioned Megastrands as lifting bolts at colliery C. In this case a 25 tonne pretension load was first imposed. This raises the question of the remaining bolt capacity for lifting-does the pretension load reduce the available lifting capacity by the same amount, and does the load being lifted then add its full weight to the load in the pretensioned bolt?

The weight of evidence from the above review suggests that the answer in both cases is yes, principally because the steel bolt is stiffer than the host rock.

There is general agreement that the imposed pretension loads the bolt between the anchor and the end plate, and therefore directly reduces the remaining bolt capacity. Any subsequent strata loading is likely to directly add to this load.

Figure 2.3 illustrates the mechanical engineering situation described in section 2.1.1 in which a bolt fastening together two metal components is pretensioned. A subsequent service load applied to the joint or bolt end, decreases compression of the mating surfaces without significantly increasing tension in the bolt because the compression of the surfaces has a very high stiffness compared with the bolt stretch, so compressive load reduces much more quickly than the bolt tension increases as joint and bolt strain develop. So in the mechanical engineering situation an extra tensile load imposed on the bolt end does not significantly increase the load along the bolt, providing it is not so large that the joint mating surfaces are no longer in contact.

The mining lifting bolt situation is different because of the relatively low stiffness of the surrounding rock compared with metal components (Figure 2.4). A tensile load imposed on the end of a lifting bolt will stretch the bolt, but because the compressed zone of rock has a lower stiffness, the corresponding elastic expansion of the rock will not be large enough to significantly reduce the pretension load. Consequently most of the imposed end load will be seen along the bolt length as additional loading. In this case we would expect to break the bolt before the plate loses contact with the rock surface.

The technique of thrust bolting (section 2.1.4) relies on this stiffness contrast between the bolt and rock to generate bolt pretension. Unrug and Thompson (2002) showed that even without a spring washer in the system, tension loads could be generated by elastic expansion of the rock when the bolt was released.

11

It follows that it is inadvisable to use pretensioned bolts for lifting loads as the load capacity is reduced both by the pretension load and by any subsequent strata loading. The latter load is unknown unless instrumented bolts are used.

2.3 EVIDENCE FOR THE PRETENSIONING EFFECT

2.3.1 Field trials

Mark et al (2000) summarise instrumented field measurements of roofbolt performance in the USA which investigated the effect of a range of parameters. Three studies compared tensioned and non tensioned bolts. The results were inconclusive because of changes in conditions and mining procedures which so often compromise field trials. At one site bolt loads in non tensioned fully bonded bolts equalled or exceeded those of tensioned partially anchored bolts after a few days, suggesting that in actively deforming roof, high pretension loads may not be as important as the use of a stiff bolting system.

The wider introduction of pretensioning in Australia was preceded by field trials with rockbolts and with pretensioned flexible bolts (Rataj and Thomas, 1997), (Strata Engineering, 2001). Pretensioned roofbolts were reported to reduce roof displacement at Teralba Colliery, and combinations of roofbolts and flexible bolts gave similar results at West Wallsend and Newstan Collieries.

As part of ACARP project C8019, completed in 2001, tensioning equipment was developed to allow 50-60 tonnes of preload with strand tendons. Project trials took place in a maingate travel road at Crinum mine, longwall installation roads at Oaky North, and a maingate belt road at Wyee Colliery.

The Megabolts at Wyee were installed after development. Increase of bolt pretension from 20 tonnes to 40 tonnes in the trial area reportedly slowed roof movement and improved visible conditions in front of the retreating face.

At Oaky North, Megabolts were being installed as secondary support in longwall installation roads. Road widening tended to cause instability at this site. In this case the Megabolts were post grouted after tensioning to loads between 20 and 50 tonnes. However conditions during the trial were better than expected and the results are not particularly conclusive. Five bolt-end load cells all registered slow loss of pretension load by around 20% during the course of the trial. This presumably reflects slight bedding in or creep at the bolt ends and suggests the absence of dynamic strata loading in the lower part of the grouted length.

At Crinum Mine, Megabolts were again being used as secondary support on development in weak roof conditions. Where the tension load was increased from 20 to 48 tonnes the roof movement after tensioning was significantly reduced. The 48 tonne tensioned Megabolts were not post grouted although the others were. Load cells showed static end load for the ungrouted higher pretension load bolts and slowly increasing loads for the others. The effects of face retreat are not reported in this study.

Although not a field trial as such, the introduction of Megabolts at colliery C in the UK (Adams C and Rennison G, 2003) is described in some detail, including reports of the roof being pulled back up into position as 25 tonnes pretension was applied, together with success in limiting movement and improving conditions when early post grouting was being used. Greater pretension is said to have had more effect. However a roof failure subsequently occurred at a nearby site at the mine where early post grouting was not applied.

12

Mining field trials are extremely difficult to conduct because of the number of uncontrolled variables - such as rock conditions and installation practices - which can influence the results. Although the field trial data described here suggests that pretensioning may bring benefits, it is not sufficiently comprehensive to be fully conclusive. The wider adoption of high pretension loads in Australia as reported by Hebblewhite (2006) however provides further corroborative evidence of this.

2.3.2 Laboratory studies

It is difficult to replicate the geometry associated with pretensioned bolt systems in the laboratory. A single bolt or tendon can be installed across a joint plane in rock or concrete, and tensioned for example, but, unless the sample blocks are very large, this does not realistically reproduce the in-situ geometry - especially the dissipation of compression from the end plate. It is difficult to obtain intact samples of weaker rocks and, in order to simulate an in-situ stress field, the rock sample would also need to be installed into some sort of confinement system which is only practical for small samples. Consequently there are few laboratory studies reported, and these involve compromises in terms of test geometry, loading conditions and the test medium.

A number of studies have looked at the distribution of load along a bonded tendon. These were reviewed by Hagan (2003) who noted that applying load to a grouted bolt end generates a tension which usually dissipates exponentially along the bolt, whilst loads generated by bed separations within the bolted length are associated with linear load reduction on either side. Laboratory testing of a strain gauged bolt with a bonded length of 175 mm, installed into a simulated rock sample in a biaxial cell was undertaken to investigate this effect. Based on limited results, the method of loading appeared to be the cause. The explanation put forward is as follows: in the pull test a jack is used to stress the tendon and this bears against and compresses the rock surface around it. The end plate maintains this compression after tensioning and constrains the rock free surface around the bolt. In contrast when bed separation occurs at some point along the bolt there is no corresponding surface restraint or compressed zone and therefore rock confinement conditions differ between the two cases. The rate of force reduction (a measure of efficiency) is greater in the former case suggesting that the bond strength and load transfer rate in the initial bonded length are increased. This suggests that pretension loads may not be distributed linearly through the resin bonded length although it should be noted that post grouting would further change the situation.

The use of a 150 mm internal diameter biaxial cell to confine the rock specimen during tendon testing was introduced by Fabjanczyk et al (1998). Previously steel tubes were commonly used. Apart from the absence of initial stresses, these also impose an unrealistic boundary condition because of their high stiffness compared with the rock test medium. Fabjanczyk et al showed that in the biaxial cell bond strength results were lower than in steel tubes. A 500mm long strain gauged rockbolt installed into sandstone in the cell with 10 MPa confinement was pretensioned to 15 tonnes. The resulting load distribution was highest towards the proximal end as expected, and initially non linear. After four days the distribution had changed towards linear, with a drop at the proximal end and an increase towards the distal end. This suggests that either creep or progressive debonding (see section 2.1.6) may have taken place, and the difference between the two loading cases discussed by Hagan may be more complex than assumed. The results suggest that in a static situation, high pretension loads could dissipate through resin creep, or could damage the resin bond, and further investigation of these effects is needed.

Clifford et al (2001) report a laboratory pull test with a 2.4 m long Megabolt installed into sandstone core contained in steel tubes. The bolt was anchored using resin, tensioned to 20 tonnes and post grouted. The tension load increased beyond the pretension value as additional

13

strain was applied to the bolt end. This would be expected if the compressed rock is more elastic than the bonded tendon and so tends to confirm the conclusion from the discussion in section 2.2.3.

Mahoney et al (2005) describe the development of a laboratory test facility for shear loading of reinforcement tendons. Shear loading along a single shear plane can be applied to fully installed tendons. In this mode the tendon clamps the surfaces together and applies a normal force across the shear plane to maximise the frictional resistance to movement. Concrete blocks of 1 m total length were used as the test medium, with a single smooth joint surface formed at 250 mm as the parting. Only a few test results with rockbolts were reported. Although initial pretension increased early shear stiffness, it had no effect once stiffness reduced with further shear strain. Axial collar loads of 16 tonnes developed in both pretensioned and non pretensioned bolts prior to bolt shear failure.

The same conclusion - that pretension did not change the maximum shear resistance, only the shear stiffness - was reached by Japanese experimenters in 1981 according to Jalalifar et al (2006). Ferrero (1995) conducted laboratory shear tests with encapsulated steel bars installed across a sawn joint in rock or concrete, and again concluded that pretension had no effect on the final shear resistance, only on the initial stiffness.

McHugh and Signer (1999) studied the shearing behaviour of strain gauged resin grouted bolts installed in high strength concrete blocks. The 600 mm long fully threaded bolts were tensioned using end plates and nuts prior to the resin setting. Pretension loads up to 10 tonnes were used. They found that pretension load had little effect on shear resistance.

Jalalifar et al (2006) reported a laboratory study of bolt double shear in concrete containing two preformed joint planes. Total test length was 600 mm with joints at 150 mm from each end. The fully grouted test bolt was tensioned axially with nuts and load cells to apply confinement across the joint planes. Increased bolt tension again increased the early shear stiffness and tended to increase the shear resistance at bolt yield. The tests were not continued to bolt failure but the trends again suggest that pretension does not have a strong influence on maximum shear resistance.

In summary these studies suggest that, for encapsulated bolts installed across a shear plane, the only measurable effect of pretension is to increase the initial shear stiffness up to bolt yield. At first sight this finding is surprising, but it presumably reflects the high stiffness of the bonded bolts, such that high axial loads are generated by dilation during shearing, irrespective of the initial tension load.

It should also be noted that these laboratory tests do not fully simulate real conditions in terms of the stress field, the artificial nature of the joints and the bolt/rock installation and geometry. The latter effect could be especially important as the laboratory set up maximises pretension induced confinement across the joint plane and the pretensioned bolt axial stiffness. In the real situation both parameters could be significantly reduced compared with the test set-up and this would be likely to further reduce any benefit from pretensioning.

In summary it appears that, even under favourable test conditions, the laboratory studies undertaken show little or no benefit in terms of shear resistance. For encapsulated bolts installed across a shear plane, the only measurable effect of pretension is to increase the initial shear stiffness up to bolt yield. As a significant increase in roof shear strength is the main anticipated benefit usually claimed by advocates, this is surprising.

14

2.4 SUMMARY

This review has attempted to answer the questions posed in section 2.1.1. The most basic of these is simply does pretensioning work? In other words does the application of high pretension loads to untensioned fully encapsulated tendons, or lightly tensioned fully or partially encapsulated types, really strengthen the rock mass, and are highly pretensioned tendons more effective as support?

There does not currently appear to be any definitive proof or demonstration that high pretension loads enhance support performance or significantly strengthen the rock mass. There is some evidence from field trials of improved support performance and the wider adoption of increased pretension loads provides further circumstantial and anecdotal evidence. Laboratory tests to date have not however demonstrated any significant advantage.

A mechanism by which higher pretension loads improve support effectiveness has not been confirmed in practice. The most widely asserted explanation - that higher pretension increases shear resistance across joints and bedding planes - would seem to be ruled out by laboratory studies. For encapsulated bolts crossing a shear plane these have consistently shown that an increased initial shear stiffness up to bolt yield is the only measurable effect of pretension. Even this effect would be limited to the zone of compression in the vicinity of the end plate.

An alternative explanation is that applying high pretension loads closes bed separations which develop prior to bolting, and therefore re-establishes frictional contact between beds in the immediate roof. This would provide a significant enhancement to the roof shear strength. Some modelling work suggests that bed separations do develop at weak interfaces and that the installation of tensioned bolts can close separations within the compressed zone-the higher the pretension, the higher into the roof that separations can be closed.

The Strata Engineering concept of buckling immediate roof beds, which can be stabilised by centrally placed pretensioned tendons, also involves the bed separation idea. In this case roof stability is achieved from the mechanical advantage obtained by centrally supporting a buckling beam, so it does not rely on closing the bed separation. The higher the pretension, the higher the resistance to compressional forces acting on the ends of the beam due to horizontal stress. Consequently horizontal stress is retained in the immediate roof and stability enhanced.

Acknowledging the practical difficulties associated with underground investigations, it should still be possible to use field measurements to check the validity of these ideas by monitoring roof deformations at multiple points as tendons are installed and tensioned, but this does not appear to have been carried out to date. The identification of an underlying mechanism would be the key to confirming that high pretension loads improve roof support.

The question of disadvantages associated with pretensioning was also raised in the introduction. The main one cited is reduction of available tendon capacity. Additional strain resulting from rock deformation will add directly to that imposed by the pretension load, so the tendon yield strain will be reached with less additional strain. Tendon failure could in theory also be expected at a slightly lower level of roof movement. However by far the major proportion of total tendon strain to failure occurs post yield, and this is unaffected by pretensioning. So in deforming roof conditions in which tendons are strained beyond yield this disadvantage is unlikely to be important.

It follows from this that it is inadvisable to use pretensioned bolts for lifting loads as the load capacity is reduced both by the pretension load and by any subsequent strata loading. The latter load is unknown unless instrumented bolts are used.

15

There are other possible drawbacks. Progressive debonding (section 2.1.6) is one. There is at least a theoretical possibility of failing the bond through this phenomenon, by application of higher pretension loads, or subsequent additional loading, and this needs to be considered.

A reduction in rock confining stress above the anchor position is another potential drawback. The pretension load generates a tensile reaction at the top of the tendon anchor length (Gray and Bates 1998). In practice this should only slightly reduce the compressive stresses acting at this position due to the in-situ stress field, but it could initiate instability in a marginal case.

It seems evident that tensioned tendons should be post grouted. One publication does question the need for it (Strata Engineering, 2001), but the consensus is strongly in favour, with the following reasons put forward:

i. It locks in the pretension load, preventing subsequent loss. ii. It ensures the tendon acts as an effective stiff support even without any benefit from the

pretension itself-the “belt and braces” principle. iii. It reduces moisture access to the rock through the annular space-this can cause long

term deterioration in weaker mudstones (Unrug et al 2004). iv. It reduces potential corrosion of the tendon from moisture and salts emanating from the

rock. v. It reduces the risk of injury to personnel through violent failure of the tendon or end

fittings.

Post grouting could also provide a precaution against progressive debonding, should this prove necessary.

It seems to be generally assumed that post grouting locks in the pretension load, preventing subsequent tension loss, although this does not appear to have been verified by test work, either in the laboratory or in-situ. Some laboratory work suggests that the initial load distribution may change with time. However, pretensioned bolts are typically used in actively deforming roof where additional loading rapidly develops, so longer term load change or loss is usually not an issue.

It follows that the optimum bolting system for this application combines the facility to pretension and post grout, with a high bond strength and stiffness in both the anchor and post grouted lengths. There is however insufficient information to reach definitive conclusions on the optimum pretension load and tendon length.

In order to fully resolve the considerable uncertainties which still surround the practice of pretensioning, the following investigation programme is recommended:

i. The question of the mechanism(s) by which higher pretension loads may improve support effectiveness needs to be resolved. Field measurement, supplemented by numerical modelling, appears to be the appropriate method in this case. Field studies should seek to identify in detail the pattern of deformation of roadway roof and, in particular, confirm the development of bed separations (as opposed to bed dilation) prior to support, and establish if these are reduced or closed by the installation of pretensioned support. Numerical modelling should be used to simulate the field situation and add parametric studies. It is likely that a number of field sites will be needed to obtain a representative range of conditions.

16

ii. The detailed mechanics of tendon load transfer do not seem to be fully known, and this adds to the difficulty in resolving the pretension question. Laboratory studies of rock deformation and tendon behaviour under more realistic loading conditions should be carried out to investigate load distributions in bonded tendons. These should make use of findings regarding in-situ rock deformation processes obtained in the field studies. In particular the difference between load distributions associated with tensioning the bolt and those induced by rock deformation within the bonded length needs to be explained, and if necessary allowed for in the experimental procedures. Factors to be investigated should include pre and post yield interface shear strengths and stiffnesses, the progressive debonding process, tension losses and the effect of post grouting.

iii. The use of 3D modelling adequately simulating the tendon, encapsulant and rock, is considered essential to complete the study of both the load transfer and pretensioning processes. The shear behaviour of the resin/rock and tendon/resin interfaces, especially at large strains, is an important factor and laboratory measurements should be used to obtain the necessary data to allow simulation of this behaviour.

2.5 BIBLIOGRAPHY

Tadolini SC 1991 The Effects of torque tension relationships on roofbolt systems CIM Bulletin July vol 84 no 951

PG Fuller GW Cadby 1981 Pre-tensioning rockbolts Division of Applied Geomechanics, CSIRO technical report 112, March

Tadolini and Mazzoni 2006 Twenty four conferences, 170 papers and understanding roofbolt selection and design still remains priceless 25th International Conference on Ground Control in Mining, Morgantown

Unrug K et al 2004 Tensioned vs non tensioned systems 23rd International Conference on Ground Control in Mining, Morgantown pp258-261

P Gray PF Bates 1998 The pretensioning placebo: Australia’s longwalls March pp78-81

Minova 2006 The Minova guide to resin-grouted rockbolts, Minova International Ltd, Chipping Norton, Oxfordshire, UK.

Su D and Poland R 2007 Fully grouted high strength mechanical shell tensioned bolt improves Pittsburgh Seam primary Roof Support 26th International Conference on Ground Control in Mining, Morgantown pp235-241

M Rataj and M Yearby 1999 The Development of roofbolting in Australian coal mining, in Rock Support and Reinforcement practice in Mining, Villaecusa, Windsor and Thomson, editors Balkema Rotterdam pp425-435

P G Fuller 1999. Roof strata reinforcement - achievements and challenges Keynote Lecture in Rock Support and Reinforcement practice in Mining, Villaecusa, Windsor and Thomson, editors Balkema Rotterdam pp405-415

Tadolini SC 1991 The effects of torque tension relationships on roofbolt systems CIM Bulletin July vol 84 no 951

17

PG Fuller GW Cadby 1981Pre-tensioning rockbolts Division of Applied Geomechanics, CSIRO technical report 112, March

Van der Merwe JN and Madden B 2002 Rock Engineering for Underground Mining SAIMM Special Series Publication no 7.

Mark C, Dolinar D, Mucho T 2000 Summary of field measurements of roofbolt performance. Proceedings of New Technology for Coal Mine Roof Support NIOSH IC9453 pp81-86

Mark C. 2000 Design of roof bolt systems. Proceedings of New Technology for Coal Mine Roof Support NIOSH IC9453 pp111-131

Hebblewhite B 2006 25 Years of ground control developments, practices, and issues in Australia 25th International Conference on Ground Control in Mining, Morgantown pp111-117

Su D and Poland R 2007 Fully grouted high strength mechanical shell tensioned bolt improves Pittsburgh Seam primary Roof Support 26th International Conference on Ground Control in Mining, Morgantown pp235-241

Rataj M and Thomas R 1997 New methods and technologies of roof bolting in Australia coal mines 16th Conf on Ground Control in Mining, Morgantown pp149-157

Strata Engineering (Australia) Pty Ltd. 2001. Application of 50 to 60 tonne cable pre-loads to roof control in difficult ground conditions. End of Grant Report, ACARP Project C8019. report no 97-079-ACR March

McKenzie R 2001 Cablebolts - a market review. International Longwall News, March 1st, Aspermont Limited.

Anon 2000. Domestic market focus alters Springvale plan. International Longwall News, March 31st, Aspermont Limited.

Bahr A 2006. Springvale: solution to tough conditions. International Longwall News, October 19th, Aspermont Limited.

Adams C and Rennison G 2003 The application of pretensioned grouted tendons at colliery C UK 22nd Int. Conf on Ground Control in Mining, Morgantown pp249-255

Rataj M 2002 Improvement in pretensioning strand bolts in Australian coal mines. 21st Int. Conf on Ground Control in Mining, Morgantown pp145-149

Littlejohn S 1993 Overview of rock anchorages Comprehensive Rock Engineering vol 4 pp413-449 Pergamon Press

Barley AD and Windsor CR 2000 Recent advances in ground anchors and ground reinforcement technology with reference to the development of the art GEO 2000 Int. Conf on Geotechnical Engineering, Melbourne 19-24 Nov

Unrug K and Thomson E 2002 Field testing of fully grouted thrust tensioned bolts 21st

International Conference on Ground Control in Mining, Morgantown pp141-144

18

T M Barczak S C Tadolini P McKelvey 2004 Hydraulic prestressing units: an innovation in roof support technology 23rd International Conference on Ground Control in Mining, Morgantown pp286-294

Strata Engineering (Australia) Pty Ltd. 2001 Application of 50 to 60 tonne cable pre-loads to roof control in difficult ground conditions. End of Grant Report, ACARP Project C8019. report no 97-079-ACR March

Seedsman R 1998Less steel in the roof, more brass in your pocket Australian Longwalls March p83

Seedsman R 1997 Geotechnical design of roof support and reinforcement ACARP SUMMARY OF PROJECT c3027

Guo S, Stankus JC 1997 Control Mechanism of a tensioned bolt system in the laminated roof with large horizontal stress 16th Conf on Ground Control in Mining, Morgantown WV

Zhang Y and Peng S 2002 Design considerations for tensioned bolts 21st International Conference on Ground Control in Mining, Morgantown pp131-144

Peng SS Zhang Y 2003 Numerical model for tensioned bolting design: a case study Trans Soc Min Metall. Explor. vol 314 sect 3 pp59-65

Zhang Y and Peng S 2003 Intersection stability and tensioned bolting 22nd International Conference on Ground Control in Mining, Morgantown pp208-217

Morsey K, Yassien A, Han J, Khair Aw and Peng S 2004 3D FEM simulation for Fully Grouted Bolts 23rd International Conference on Ground Control in Mining, Morgantown pp273-277

Bouteldja M 2000 Design Of Cable Bolts Using Numerical Modelling PhD thesis, Department of Mining and Metallurgical Engineering, McGill University, Montréal, Canada, April

Fabjanczyk M, Hurt K and Hindmarsh D 1998 Optimisation of Roof Bolt Performance Proceedings International Conf on Ground Control in Mining, Wollongong

Clifford B, Kent L, Altounyan P, Bigby D 2001 Systems used in Coal Mining: Long Tendon Developments. 20th International Conference on Ground Control in Mining, Morgantown

Mahoney L Hagan P Hebblewhite B Hartman W 2005 Development of a laboratory facility for testing shear performance of installed rock reinforcement tendons 24th International Conference on Ground Control in Mining, Morgantown pp357-365

Jalalifar H, Aziz N, Hadi M 2006 The effect of surface profile, rock strength and pretension load on bending behaviour of fully grouted bolts. Geotechnical and geological engineering 24 pp1203-1227

Ferrero A M 1995 The shear strength of reinforced Rock Joints Int. Jnl. of Rock Mechanics and Mining Sciences 32(6) pp595-605

Hagan P 2003 Observations on the differences in load transfer of a fully encapsulated bolt. Proceedings of 1st Australasian Ground Control in Mining Conference UNSW Sydney, November.

19

Frith R 2000 The use of cribless tailgates in longwall extraction. 19th International Conference on Ground Control in Mining, Morgantown pp84-92

20

3 LABORATORY AND UNDERGROUND TESTING OF LONG TENDON SYSTEMS

Long tendon system testing was carried out to determine how effectively long tendons could be encapsulated using candidate grouts and available mixing and pumping systems, and, in the process, establish testing procedures for inclusion in the revised cablebolting Standard. Work was also done to determine the effect on system bond performance of varying grout strength, and the long term trends in grout strength were established via monitoring of comprehensive field sample data. This work is described below.

3.1 GROUT ENCAPSULATION TESTING

This work was concentrated on development of laboratory tests for grouting and encapsulation of flexible reinforcement systems. These should be suitable for:

• checking that a particular grout is suited for use with a range of flexible reinforcement systems,

• checking that any particular flexible reinforcement system is suitable for use with grouts which comply with other aspects of the Standard.

The key criterion is that full encapsulation of the strand can be achieved.

3.1.1 ‘Bottom – up’ testing

Grouting from the ‘bottom-up’ is where a feed tube is located at the mouth of the installation hole, and encapsulation progresses from the bottom to the top of the cablebolt. It is the most common form of grouting. A test based upon that for double birdcaged cablebolts in the current British Standard (reference 1) was devised, and tests on double mini-cage, double nut cage and tensionable Reflex cablebolts carried out with Pozament CBG grout, the only grout currently used for this type of installation in UK mines.

A 6 m length of strand was selected as the test datum - most cable bolts are 6 m long. The strand was housed in a clear pvc (or similar) tube so that the encapsulation could be observed. For a realistic test, the internal diameter of the tube selected was as near as possible to that of the underground installation hole. The double mini and nut cable samples were prepared by first attaching a breather tube to the strand – the purpose of which was to bleed air from the installation during grouting. Next, the grout tube was secured to the lower section of the strand, the top of the tube being located approximately 0.5 m above the cable end. The tube was placed on the opposite side of the assembly to the breather. Overall length of the grout tube was approximately 2.5 m. The assembly was then located in the 45 mm diameter tube and slid into the tube until the top of the cable was around 100 mm below the tube cap. To complete the assembly, a seal around the mouth of the tube was required and this was achieved initially using an expansive foam sealant. However, this proved only partially reliable and a fabric sock filled with pre-mixed grout was also used. Unlike the mini and nut cage cables, one version of the Reflex strand, is designed such that it can be pre-tensioned prior to grouting – by installing into polyester resin capsules located at the top of the borehole, and then tensioning against a washer plate fitted at the borehole mouth. The provision of a seal at the mouth of the hole prior to grouting is at least partially to be achieved by the tensioning-induced tight fit of the plate against the roof strata. The strand supplied was factory fitted with a steel breather tube located in the lay of the strand. The Reflex strand is

21

designed to be grouted using a feed tube and a special arrangement simulating a counter bored hole at the mouth of the installation was used to accommodate this.

The tube assembly was placed alongside an aluminium scaffolding pole which was to act as a support providing straightness and rigidity when erected to the vertical. The tube was taped to the pole at 1 m intervals (to assist as a marker for grout flow) and ‘jubilee’ clips were also used to secure the assembly. The assembly was lifted and placed into a purpose made swivel bracket mounted approximately 2.5 m above ground level. The bracket acted as a support and swivelled to assist in placing the assembly in a vertical position. Once vertical, the assembly was clamped in position. The breather tube (pvc in the case of mini and nut cage, steel for the Reflex bolt) was immersed in a clear container of water. The flow (and cessation) of air down the breather tube would be detected as bubbles in the container of water.

Mixing and pumping were carried out using a high-shear-mix system comprising tank, mixing paddle with air motor, and Whyte-Hall Model GB7 pump unit. Separate air lines were connected from the compressor to the mixer motor and pump to minimise pressure losses. The pump was connected to the test piece via an 11 m long hose with pneumatic usage rating. Prior to grouting, the grouting hose was flushed clear with compressed air to remove any standing water and avoid dilution of the mix. The hose was then connected between the pump and the grout tube in the test assembly. Grout pumping was then started and progress monitored first by the show of air bubbles at the water container and then flow in the test piece, rotation of the mixing paddle being continued throughout the process. Progress of grout encapsulation was videoed. Following encapsulation, the equipment was flushed and cleaned.

The tests on double mini cage and nut cage cables, using 13 mm bore grouting tube and 10 mm breather, were successful, and showed the method used produced consistent results. Average encapsulation time was slightly shorter for the nut cage cables but this was almost certainly due to variation in operator control. The tests showed that successful encapsulation can be indicated reliably by a show of grout at the mouth of the breather tube, and this could be incorporated both into codes and rules for underground installation and the acceptance procedure for the revision of the British Standard (reference 1) - at least for systems which incorporate a substantial breather tube. For the Reflex bolt tests, a show of grout at the breather exit was not in evidence - due to the necessarily small bore of the tube. In this case full encapsulation was coincidental with a cessation of bubbles at the breather exit. Provided this feature is shown to be consistent, it could also be incorporated into procedures for this type of design.

Sealing at the mouth of the hole is probably the most unreliable part of the operation. The most effective seal was formed by using a ‘tubi-grip’ sock filled with grout and surrounding the cable assembly. The Reflex bolt tests showed that sealing around the end plate was problematic but, once accomplished, full encapsulation could be achieved. However, due to the sealing problems, only one fully successful test was carried out; another test achieved full encapsulation but leaked.

Procedure for mixing and grouting for the mini and nut cage tests involved connecting the grout hose to the assembly grout tube before any pumping took place. This was subsequently changed for the Reflex bolt tests following experience with Megastrand trials. Here, the grout hose was first flushed clear with compressed air, and then the outlet returned to tank to allow a flow to be established prior to grouting. This was shown to have advantages in allowing the flow to be observed and regarded as satisfactory (or otherwise) prior to the grouting operation. This part of the procedure should be incorporated into any future tests and the recommended acceptance procedure for revision of the Standard.

22

The testing programme showed that the mounting arrangement with support pole and swivel bracket was a practical proposition. However, the steel support pole used initially was relatively heavy, and use of an aluminium pole made the procedure much easier and probably safer.

The testing procedure evolved from this test series was formally written up as a suggested Annex to the revised Standard and this is included in a later chapter.

3.1.2 ‘Top -down’ testing

Long tendons which are encapsulated from near to the top of the tendon to the hole mouth are less common than ‘bottom-up types, and at the time of writing, only the Megastrand system uses this type of encapsulation, in the UK. The Megastrand comprises a grouping of steel wires around a central hollow steel tube. The distal end of the assembly is welded together to form a termination, and the proximal end has a more sophisticated arrangement which locks the individual wires around a fitting which is threaded. The fitting is hollow, designed to accommodate a grouting lance, and is also fitted with a nut. An end plate can be placed over this end, and a hydraulic jack can also be connected to facilitate loading the installed strand (or pretensioning). The Megastrand is designed to be encapsulated with grout injected into the central tube which has an outlet at least 2 m below the distal end of the assembly. A systematic and steady flow of grout down the assembly to the mouth of the hole requires the use of a thixotropic grout. However, for pretensioning, the assembly is first spun through a resin capsule placed at the top of the installation hole, and anchored prior to loading and subsequent grouting. Resin anchored length is usually around 2 m.

Testing equipment and procedure were similar to that described above. A 6 m Megastrand was located in a clear pvc tube and a grout lance attached to the feed connection in the end fitting of the strand. The mouth of the hole/tube was not sealed, and a successful test would be where grout would fully encapsulate the strand from top to bottom and start to issue from the hole mouth. The grout selected was Pozament HPRG thixotropic grout which is designed to have a water-to- solids ratio (WSR) of 0.3. This is the only thixotropic grout currently accepted for use in UK mines.

The tests used the standard pump / mixing system currently used in UK mines - Whyte-Hall GB7 pump and mixing assembly sourced from Australia, powered by a mobile compressor delivering 0.69 MPa (100 psig). The tests identified problems with the mixing and pumpability of the HPRG grout at the recommended water:solids ratio (WSR) of 0.3. It was found that it was necessary to add further water to achieve a pumpable mix, even though it was later found that the grout was supplied in bags low in weight, resulting in an already higher WSR than expected when mixing the required quantity of water with a bag of grout. [All future tests were carried out with accurate weighing of grout and water prior to mixing]. Tests on grout cubes indicated relatively low strength due to low densities derived from excessive water quantities. The results suggested the possibility that the system may not be being used properly underground, with additional water added to the grout to achieve pumpability. It was found that it was possible to achieve full encapsulation with a higher WSR of 0.33, the thixotropic nature of the grout being maintained, but the additional water consequently reduced the strength of the grout. This problem was referred to the manufacturers (grout and strand) for resolution and a revised thixotropic grout formulation (AGH10) with modified properties and a higher recommended WSR was developed.

Three laboratory pump tests were undertaken with the new grout at the recommended WSR (0.33) and these were all successful. Tests on cube samples cured for 28 days, taken from two of the mixes, produced mean strengths of 70 and 77.5 MPa respectively, compared with the current British Standard requirement of 80 MPa. This result contrasted to the manufacturer’s UCS cube

23

test results which indicated compliance with the Standard. The probable explanation for this discrepancy is the difference between the laboratory mixing and curing regimes specified in the Standard in comparison with those achieved with the mixing system and regimes used underground (and in the laboratory encapsulation tests), which could include air entrapment and curing temperature variation.

An underground field placement trial of the HPRG and AGH10 thixotropic grouts for “top down” tensionable reinforcement tendons was carried out with Megastrands at colliery B but this was inconclusive due to on-site limitations, in particular very low pneumatic pressure. The results were at odds with the laboratory experience in that the standard grout performed (in encapsulation terms) at least as well as the variant. On-site air and water temperatures were very high (in excess of 34 deg C), and it is possible that this could have affected the thixotropic nature / viscosity of the grout. Grout samples were taken during the trials and laboratory tested after curing for 28 days. Results for UCS and density were low compared to the manufacturers’ specification, for both the standard and variant grout, despite careful attention to mixing ratios. This, however, is fairly typical of field trial results and (to some extent) results from laboratory trials.

These trials and underground experience at other mines suggested that the grout pumping equipment currently being used in the UK may not provide the performance required for this application. Therefore an alternative pump / mixer unit (Blue Heeler) was imported from Australia by the Megastrand supplier and a further set of two laboratory pump tests was undertaken using the AGH10 grout (with a WSR of 0.33) and Megastrand combination, again 6 m of encapsulation being sought. However these tests were problematic with one failing to achieve grout flow into the cable and the other achieving two-thirds encapsulation before the grout delivery hose burst. The issue of whether these tendons can be installed satisfactorily with an appropriate thixotropic grout has still to be resolved. Nevertheless the “top down” pump test has been fully documented and recommended for inclusion in the revised Standard.

3.2 VARIATION OF BOND PERFORMANCE WITH GROUT STRENGTH

The problems which arose during the encapsulation trials of Megastrand led to some debate in the revision committee as to the requirement for a 28 day grout strength of 80 MPa. The figure was carried forward to the revised draft from the original Standard, and there was some discussion that if this could be lowered, a more easily pumpable grout could be used, leading to more reliable encapsulation of the Megastrand tendon. The original strength criterion was decided upon so that system bond strength, as measured in a laboratory short encapsulation pull test (LSEP), could be optimised, and it follows that any reduction in strength might lead to reduction in bond performance.

In order to resolve the issue, a laboratory test programme was set up to determine the variation of bond performance with grout strength. The test was based on the LSEP procedure used for determination of long tendon performance for research work carried out for an earlier project, and fully described in reference 2. Samples of Megastrand 1 m long were encapsulated to a depth of 325 mm in sandstone cores using Pozament CBG grout. The cores were housed in a biaxial cell pressurised to simulate typical underground stresses. CBG was chosen as the encapsulant as this grout was easier to mix using bench equipment, more pourable for easy sample assembly, and its properties with respect to curing time very well known (when compared with thixotropic grouts). At the same time as the samples were encapsulated, grout from the same batch was poured into sample bottles and cube moulds. The intention was to pull test groups of two Megastrand samples and to crush test corresponding cube and bottle samples at intervals as the grout cured, so producing measured bond performance over a range of grout strengths. Eight samples were prepared with grout mixed to the standard WSR, and two more

24

with a weak mix – to be tested within 24 hours at a very low grout strength. The samples with a standard grout mix were tested with 1 day, 3 days, 7 days and 42 days curing time.

The bond performance results are shown graphically in Figures 3.1 and 3.2 as variation of load with bond displacement for samples cured from 1 to 3 days, and 7 to 42 days respectively. The graphs were used to determine maximum load, bond strength, and system bond strength in the load range 150 – 300 kN. Bond strength is defined as load at which the slope of the load / displacement characteristic falls below 20 kN/mm, and system stiffness is the slope of the load / displacement characteristic over the load range given above. The values determined were then plotted against corresponding grout strength from cube crushing tests, and the results are shown Figure 3.3. A logarithmic curve fit was applied to each set of data points and these are also shown in Figure 3.3.

It will be seen from Figure 3.3 that all fitted curves show increasing bond performance with grout strength, as would be expected, although the bond strength individual values tend to plateau above a grout UCS of 60 MPa. Although the existing Standard (reference 1) calls for a minimum UCS of 80 MPa after 28 days, it was known that CBG, and the associated thixotropic grout, would continue to increase in strength beyond this time, and 42 day results confirm this, with an average cube strength of 94 MPa. However, it should be noted that grout strengths determined from field samples, although often measured well in excess of 28 days curing, rarely achieve the 28 day requirement, and are often well below this figure. This issue is outlined fully in the next section. Since bond performance has been shown to continue to increase with grout strength, and field measurements indicate that the existing requirement for strength is not being met (see below), it is difficult to justify a relaxation in the requirement.

3.3 FIELD SAMPLE GROUT TESTING

UK coal mines have used only one type of cementitious grout for ‘bottom-up’ long tendon encapsulation, and some other applications, for at least the last 12 years. Pozament CBG grout is subject to routine sampling at the mines, and the bottle samples are always weighed – for density – but usually also crushed at a specialist laboratory to determine strength. Crushing after 28 days curing is recommended. Applications requiring a thixotropic grout utilise Pozament HPRG grout which is also required to meet strength criteria specified in the Standard, and also subjected to routine testing as described above.

A database of bottle sample test results for CBG was established in 2001. In 2006 the database was brought up to date, and results for HPRG included. The database was further revised in 2007 to include CBG and HPRG samples from colliery A. The variation of grout strength with density for bottle samples of CBG grout up to an including 2007 data is shown in Figure 3.4. The results indicate that the CBG grout had consistently met the current British Standard, provided that it had achieved the density (2090 kg/m3) previously furnished by the manufacturer as being consistent with their recommended WSR. However, on many occasions, this density had not been achieved in the field. This is particularly true of 2007 samples supplied by colliery A. Only one of 36 samples collected during 2007 from colliery A achieved the 28 day UCS (80 MPa) and none reached the 2090 kg/m3 density. This prompted a mixing test, witnessed by the manufacturers, which was conducted on the surface at the colliery, using the Whyte Hall mixer and pump, and where cube samples were taken. Tests on these samples showed a problem with the CBG grout strengths. This was referred to the manufacturer who reformulated the grout in order to resolve the issue. With reference to achievable density, the manufacturer, at one stage, proposed to lower the requirement to 2.045 kg/m3. However, this was an error arising from quoting the requirement for density at 24 hours curing. The 28 day figure quoted above is correct, and still valid.

25

The results from HPRG bottle samples collected at colliery A are given in Figure 3.5. None of the samples achieved the required 28 day strength and density values of 80 MPa and 2090 kg/m3 respectively.

26

4 UNDERGROUND MONITORING OF PERFORMANCE AND EFFECTS OF TENSIONED LONG TENDONS

The effectiveness of tensioned long tendon systems in UK mining conditions was to be determined in two ways:

a) analysis of the historical and monitoring data available at UK Coal collieries and Headquarters in order to compare the actual in-situ behaviour of tensioned and untensioned systems. In the event, as described in chapter 1 above, this work was abbreviated so that issues arising from the use of thixotropic grouts could be more fully explored and resolved.

b) underground monitoring to be undertaken on the effects of installing and tensioning long tendons on the surrounding strata and accompanying support systems. This work was carried out during installation and tensioning of Megastrands at a face line drivage at colliery A.

4.1 HISTORICAL DATA ANALYSIS

4.1.1 Introduction

Tensioned long tendon systems were first used in UK mines in the late 1990’s but came to more prominence in 2002 when the Megastrand system was used for remedial support in a main gate at colliery C. This project was regarded as a success and use of the Megastrand became more widespread. Although installation of the Megastrand was intended to be a staged process where initial installation and pretensioning would be quickly followed by grouting, development in the UK became oriented toward delaying grouting due to operational constraints. This development led ultimately to a failure at a UK mine, where inspection of a fall of ground clearly showed that Megastrands had not been previously grouted and so would not have been as effective as properly installed (fully grouted) long tendons.

In the UK the Megastrand has been widely used at colliery C and at some sites at collieries A, B and D.

We have tried to establish the success or otherwise of use of the Megastrand in terms of its contribution to the stability of a site during development and face retreat but time constraints and availability of analysable information have limited the study to two sites at colliery C, and to a lesser extent, a face heading at colliery B.

4.1.2 Site studies – 10’s main gate, colliery C

10’s main gate at colliery C was driven 1638 m inbye to the face start position. Primary support was via 2.4 m full column roofbolts. Cablebolting was carried out systematically over at least 1000 m of the drivage using non-tensioned mini-cage cables, 8 m long, in various configurations, but well behind the development face and in response to action levels indicated by monitoring.

A diagram of the drivage layout and snapshots of development monitoring are given in Figure 4.1. At the development stage, a section of the maingate at around the 500 to 600MM suffered significant roof deformation and stability was re-established by installation of minicage cable bolts in response to telltale action levels. On face reteat, a section of the maingate roof inbye of

27

the 500 metre mark (MM) area, which had been stabilised by cable bolting on development, began to move again well ahead of the retreating face resulting in severe problems close to the face end. Eventually the face was halted with face end machinery jammed. Following salvage work, Megastrands were installed ahead of the retreating face together with steel supports. The 6 m Megastrands were installed in pairs with 1.2 m spacing along the gate, tensioned to 25 tonnes and then grouted with thixotropic grout. Conditions no longer deteriorated ahead of the face and the face was able to resume and continue to complete extraction of the panel. Conditions were reported to be superior to most recent experience in the same seam.

Figures 4.2, 4.3 and 4.4 show monitoring results for monitoring stations at 607, 423 and 248 metre marks respectively. The results span several months and indicate increasing movement at the inbye stations with the face retreating to within 100 m or so of these stations. In particular Figure 4.2 shows the activation of previously stabilised roof at a horizon just above the rock bolts at 607MM.

Results from tell tale measurements - both those installed for roofbolts and after cable bolts were installed – are shown in Figures 4.5, 4.6 and 4.7. These give histograms of ground behaviour along the main gate from the face positioned at the 1000m mark through to completion. Figure 4.6 and 4.7 both show the ‘B’ indicator cable bolt tell tale results but from differing perspectives. They indicate a particularly large roof dilation at the 565 MM. but the dilation here did not progress significantly with face retreat and the effects of front abutment. Figure 4.7 shows how the roof inbye of 565MM remained stabilised for some time after cable bolt installation but then began to deteriorate significantly as the face approached.

It is probably fair to conclude from the above that the significant roof dilation experienced on development where support was via roofbolts and passive mini-cage cablebolts, was prevented from progressing significantly during face retreat by the installation of the additional tensioned support.

4.1.3 Site studies – 22’s main gate, colliery C

Following the introduction of tensioned tendons in 10’s main gate, their mode of application was changed for 22’s development in that 8 m Megastrands were installed behind the development face as secondary support in place of the previously used passive mini-cage cables. The 8 m Megastrands were installed in pairs, 1.2 m apart, tensioned to 25 tonnes and grouted using thixotropic grout. This was successful and conditions in 22’s main gate were significantly better than those experienced during 10’s drivage and retreat. Figure 4.8 is a histogram of the complete development / retreat history of 22’s main gate via type ‘A’ tell tale results. It will be seen that roof dilation was moderate along the gate during drivage and face retreat did not produce any significant additional roof movement.

4.1.4 Site studies – 19’s tail gate, colliery C

The successes experienced with 10’s recovery and the retreat of 22’s panel encouraged the mine to modify the deployment of tensioned systems still further, to the extent that Megastrands were installed at the development face in pairs, 6 m long and tensioned to 20 tonnes. However grouting was carried out outbye of the face and behind the development machinery in order to speed up or at least not impede development rate, and grouting was only undertaken in response to the action levels set for the ‘A’ type tell tales (25 mm of indicated dilation).

It was on this gate road that a major fall of ground occurred. The fall occurred during face retreat between the 588 and 600 MM with the inbye end bounded by cribs, and the outbye end bounded by wood props. Wood props had been removed from the area in which the fall

28

occurred to facilitate crib building. Subsequent investigation showed that high levels of roof dilation had taken place prior to the fall. Figure 4.9 shows plots of the combined indicated movement from ‘A’ and ‘B’ tell tales (in the vicinity of the fall) at the 570, 590 and 610 MM. Total movement measured by these tell tales prior to the fall was 360, 322 and 207 mm respectively. This level of movement would indicate that failure of rockbolts and long tendons was highly likely, and examination of the installed roof support after the fall showed that this was indeed the case. Examination of the (failed) Megastrands in the fall area also indicated that grouting may not have been effective with very little grout evident around the strand assemblies.

4.1.5 Site studies – T18’s face line, colliery B

Drivage of T18’s face line experienced support difficulties soon after commencement. Width of the roadway was initially 6.4 m and primary support was via 2.1 m fully encapsulated roofbolts. Monitoring results showed instability above the bolted height and, at 44 MM, cablebolting was commenced using 8 m double nutcaged cables. Cables were installed behind the face, and a programme of cable bolting back toward the heading entry was also commenced. At the same time, drivage width was reduced to 4.8 m, with ‘cheeking out’ to full width being carried after initial drivage. At 62 MM, ‘cheeking out’ was stopped.

Installation of Megastrands commenced at the 105 MM and continued for the rest of the drivage. Results were reported to be favourable with anecdotal evidence of the roof being ‘driven back up’ on pretensioning – i.e. strata beds separated on excavation were recompressed by loading the Megastrands prior to grouting.

A diagram of the site is given in Figure 4.10. As the Megastrands were installed in a section of heading driven at a substantially smaller width than the initial nutcaged section, and it was not cheeked out during drivage, it would not be appropriate to attempt to compare the monitoring results from the Megastrand and nutcage sections.

4.1.6 Discussion and conclusions

Initial use of tensioned long tendon reinforcement in UK mines gave encouraging results with previously problematic longwall districts showing improved face end conditions on retreat, less requirement for additional support around the face end, and improved retreat rates.

Tensioned systems were first used as secondary or remedial support well behind the development face and additional to ‘passive’ non-tensioned cablebolts, and then replacing the non-tensioned support system.

Problems arose when the tensioned systems were installed at the face with subsequent grouting delayed until some time later – in response to tell tale movement reaching action level status. This procedure can be expected to have had three major consequences. Firstly, the initial high stiffness of a tensioned system that might have been effective in controlling early roof movement, had the tendons been grouted on installation, was not available.

Secondly, and probably more importantly, the roof movement which occurred prior to grouting probably restricted the tendons’ central and external annuli and prevented proper subsequent grouting, The laboratory testing of thixotropic grout mixing and pumping, described elsewhere in this report, has shown proper grouting of Megastrands to be problematic even in controlled conditions. These problems would be exacerbated considerably in-situ especially where immediate roof shear had been allowed to distort the bottom end of the tendon prior to grouting.

29

Thirdly, there is considerable anecdotal evidence that grouting teams in the field, have difficulty grouting newly installed tensioned cables with thixotropic grout, and in this event, will resort to ‘watering down’ to achieve full grouting of an installation. The temptation to water down the grout in this case will have been exacerbated by the additional grouting difficulties experienced due to roof deformation. The operating principle behind grouting the Megastrand, and other “top down” groutable tendons, is that the thixotropic nature of the grout will cause it to flow down the strand from the ejection point in the central feed tube, encapsulating the strand until reaching the mouth of the hole, where a show of grout will indicate successful grouting. If the grout is watered down, compromising the thixotropic nature of the grout, then clearly the show of grout at the mouth of the hole cannot be guaranteed to indicate successful encapsulation. Furthermore, watered down grout will be weaker and slower to develop strength than properly mixed material.

Less comparison of monitoring data from tensioned and untensioned long tendon sites was possible during the project than was originally envisaged for two reasons. Firstly, considerably more project time was required to investigate the grout mixing, pumping and strength issues described above as these were only discovered during the course of the Project. Secondly, such comparison was impeded by a scarcity of directly comparable examples of roadway behaviour with and without the use of tensioned tendons.

However, the work carried out indicates that, at least initially, the deployment of tensioned systems in gate roads where non tensioned support had failed to provide workable retreat conditions, was successful, with colliery C achieving satisfactory retreat rates where the previous main gate exhibited prohibitive roof failure around the face end before tensioned systems were deployed. Use of tensioned systems replacing non-tensioned cablebolts as additional support at 22’s main gate, colliery C, was also successful, and available monitoring information indicates good roof stability in the main gate during face retreat. However, installation of tensioned systems at the development face with subsequent grouting dependent on tell tale action levels appears to have been a step too far, particularly in the light of subsequent investigation revealing the difficulties associated with grouting these systems with available equipment.

4.2 UNDERGROUND MONITORING OF LONG TENDON TENSIONING AT COLLIERY A

This monitoring exercise was undertaken in order to measure prior to grouting the effects of tensioning to approximately 10 tonnes, 8 m long Megastrand reinforcement tendons in a roof which had already been subject to a moderate level of dilation and had been previously reinforced with a combination of 3 m long KT rockbolts and tensioned 8 m long Megastrands.

4.2.1 Drivage and support aspects

A face line at colliery A was driven from the Tailgate (left hand gate looking inbye) to Coal gate using a Joy BM ED15 bolter miner equipped with 4 hydraulic roofbolting rigs and two rib bolting rigs. The drivage was commenced with 8.2 m width with the roof supported using a reinforcement row spacing of 0.8 m, each row comprising 9 x 3 m KT rockbolts, 6 x 4 m flexible “Reflex” bolts, both installed through AT polyester resin capsules, and 3 x 8 m Megastrand tensioned tendons. Two additional 1.8 m KT rockbolts were placed behind the bolting machine to fill an unreinforced triangular area created by the bolt installation angles of the bolter miner drill rigs. The Megastrand tendons were installed with a point anchor using a 28 x 1200 mm medium set AT resin capsule, then tensioned to a nominal 10 tonnes within several hours of installation and finally grouted using Pozament HPRG thixotropic cementitious grout

30

as a batch at the end of each shift. The Megastrand is fitted with a central grout tube and is designed to be grouted from the top down using a suitable thixotropic grout.

The face heading commenced at the Tailgate junction at a roof horizon approximately 1.45 m above the top of the Two Yard seam and then, leaving a lip at the 30 m mark, moved forward beneath a 0.8 m thick coal top. The face heading experienced considerable difficulties in maintaining roof control at full width, requiring considerable remedial support and making poor progress. It was therefore decided to switch to a two stage drivage system and from the 80 m mark onwards, drivage continued at the narrower width of 5.2 m with roof support provided, again at 0.8 m row spacing, by 10 x 3 m KT rockbolts and four x 8 m Megastrands at 0.8 m spacing plus one central 1.8 m KT bolt.

Following thirling with the coalgate and formation of the junction, the maingate ABM25 bolter miner commenced driving back along the face line widening it on the face side to a total width of 8.2 m. The support pattern is shown in Figures 4.11 and 4.12. This shows that an additional 4 x 3.0 m KT bolts plus 2 x 3.85 m flexible “Reflex” bolts were installed across the additional 3.0 m width at a 0.8 m row spacing, with a further alternating pattern comprising rows of six and then four 8 m Megastrands installed across the full width of the face heading at the same row spacing.

4.2.2 Installation layout and reading of instruments

The instrumentation installed specifically for this exercise comprised 3 rows of instruments installed between straps and centred around the cable bolting telltale (13) at 196 m (nominally 194 m). For convenience these rows have been labelled as 195 m, 196 m and 197 m, though the strap spacing is actually 0.8 m. The relative positions of the instruments and the excavation on 21 March 07, when installed, and 27 March 07, when the neighbouring Megastrands were tensioned, are shown on Figures 4.13 and 4.14.

The instruments were installed towards the face side of the narrow section of the face, six days prior to the machine cutting past during the widening operation. Each row of instrumentation included 2 x 3 m strain gauged KT rock bolts and a roof extensometer. The strain gauged bolts were additional to the existing rockbolting pattern previously installed. Each was fitted with 9 horizontally opposed pairs of 120 ohm resistance strain gauges spaced at equal intervals along its length and mounted in longitudinal machined slots, to measure axial and bending load distribution These were read with both a Soil Instruments SG1041 Strain Meter and an RMT SM01 SGMeter data logger, the latter being used for continuous readings every 10 seconds on each bolt whilst a nearby Megastrand was being tensioned.

The extensometer in the central row of instruments was a Magnesonic type with 19 readable anchors installed at intervals in a 43 mm diameter borehole up to a maximum height of 7.0 m. The extensometers on either side were RMT RME04, 4 height remote reading extensometers which could be read using a RMT RRT-1442-PR ATEX approved portable readout with a resolution of better than 0.1 mm. The top anchors were installed at 10 m into the roof.

Figures 4.12 and 4.15 show the originally intended positions of the instruments with respect to the existing supports. The main difference between planned and actual was that, due to practical constraints, strain gauged bolts 2, 4 and 6 were situated on the opposite (goaf side) of the already installed Megastrand and thus further from the positions of the nearest Megastrands to be installed and tensioned than originally intended.

31

4.2.3 Results – telltale and wire extensometer data

The weekly reading history for cable bolting telltale 13, nominally at 194 m (actually 196 m), from installation on 10 November 2006 until its replacement due to going over-scale (+75 mm on A) on 17th May 2007, is shown in Figure 4.16. It was anchored at 7 m (A) and 9 m (B). This shows that the roof had stabilised at a displacement of approximately 11 mm in the zone below 7 m, in the Megastrand reinforced height, with an additional 7 mm between 7 m and 9 m above the reinforced height, prior to the widening process. By the time that the machine had just cut past on 27th March the roof had begun to move again in the reinforced height with the readings increased to 27 mm (A) and 6 mm (B). The roof continued to dilate in the reinforced height at a relatively high but gradually slowing rate over subsequent weeks and had not re-stabilised by the time the telltale went overscale on 17 May, 8 weeks later.

A 5 wire extensometer was located at 192 m, anchored to a maximum height of 7 m and installed on 14th November 2006 near the face of the heading. This had registered 24 mm of total roof lowering by the 19 March, just prior to installation of the new instruments, 23 mm of which had occurred below a height of 2.9 m (first anchor height). This is shown in Figure 4.17.

4.2.4 Results – remote reading extensometer data

No data was obtained from RRExto 2. It appears that, when installed, the wires were not properly crimped to the instrument and so no subsequent changes were registered.

RRexto 1 was anchored at 10 m, 7.5 m, 3.7 m and 2.1 m with readings taken relative to roof level. The total roof dilation measured during the 15 days between installation on 21/3/07 and the final reading on 5/4/07, to a height of 10 m, was 32 mm of which 9 mm occurred below 2.1 m, 18 mm occurred between 2.1 m and 3.7 m and 2.5 mm occurred between 3.7 m and 7.5 m (see Figure 4.18). This compares with 39 mm total roof displacement measured on the nearby telltale between 19/3/07 and 2/4/07.

The relative distribution of dilation between anchor bays remained similar throughout the measurement period. The rate of roof lowering was approximately 3 mm per day over the weekend period prior to cutting past the exto, 7 mm per day between the 26th and 27th March as the machine cut past the station and up until the Megastrands were tensioned, 3 mm per day for the 2 days following Megastrand tensioning and 1 mm per day over the next 7 days until readings ceased. This final rate of 1 mm per day was the same as that measured by telltale 13 for the 7 weeks following cutting past.

No significant change in roof dilation rate or lowering rate was measured during the tensioning period itself (see Figure 4.19). Certainly no closure of dilation or roof lifting was measured.

4.2.5 Results – sonic extensometer data

The sonic extensometer results are shown in Figures 4.20, 4.21 and 4.22. Figure 4.20 shows 2 main strain zones forming in the roof, at 1.2 m near the top of the coal roof (26 mm/m) and 3.2 m (30 mm/m). No movement was measured above a height of 4.1 m. The sonic extensometer data agrees generally with the Remote Reading Extensometer data, showing about 6 mm more total roof lowering over the period to 29th March. The sonic extensometer was lost to roof shear prior to the 5th April when the final RRE readings were taken. The large number of anchors used with the sonic extensometer allowed the strain zone at 3.2 m to be identified as a newly developed zone of dilation associated with the cut past, being above any zone of significant strain previously detected by the wire extensometer at 192 m. This significant strain

32

zone was located just above the top of the installed 3 m rockbolts within the zone reinforced by Megastrands alone.

4.2.6 Results – strain gauged rock bolt data

The strain gauge rock bolt data is set out as mean microstrain and microstrain difference graphs in Figures 4.23 – 4.28. During the period from first reading on 21st March until the day prior to cutting past and Megastrand tensioning (26th March) there was considerable increase in bolt loading, with most bolts reaching at least 20 tonnes towards their centre and bolt 3 definitely going well into yield. During this period gauges were lost in the central sections of bolts 3 and 4, probably due to high strains.

Table 4.1 Sequence of SG bolt readings, Megastrand tensioning and bolt load development on 27th March 2007

SG bolt Reading Number Reading Comments on bolt load/strain changes during number times of and Mega corresponding periods

during readings tensioning tension- at 10s sequence ing intervals

SG1 10:44 2 Initial readings

10:54 7 Mega 1

11:14 6 Mega 3

2 Final readings


11:04 7 Mega 2

11:50 2 Final readings



Compression from previous day except at 1m & 2.4m No change No change during tensioning Compression 1.5 – 2.0m No change during tensioning Compression 1.5 – 2.0m & below 1m Reloading back to earlier tension levels in subsequent days

Compression from previous day around 0.8 - 1.1m

Further compression 1.5 – 2.7m No change during tensioning Very little change Considerable additional tensional loading 1.1 -1.5m on subsequent days

Considerable compression from previous day

Some small changes (tension and compression) during overall tensioning period. Considerable additional tensional loading on subsequent days

SG4 10:49 4 Initial Considerable tensional load development from readings previous day, greatest towards centre of bolt.

33



11:22 6 Mega 7



11:18 8 Mega 6


No change during overall tensioning period. Considerable additional tensional loading on subsequent days

Considerable tensional load development from previous day at 1.1m taking bolt to yield. No change No change No change Additional tensional loading on subsequent days at 1.1m and eventual gauge failure at 1.1m

Tensional load development towards centre of bolt from previous day reaching yield. Some further tensional load development in bottom half of bolt No change No change Considerable additional tensional strain on subsequent days in centre of bolt leading to eventual gauge failure at >13000 microstrain

The original plan had been for each bolt to be monitored using the data logger whilst the nearest Megastrand was tensioned. The actual sequence of monitoring and tensioning differed somewhat from the ideal and is set out in Table 4.1 above. The Megastrand and bolt numbers are defined on the plan shown in Figure 4.14.

During the period when the machine cut past (26th to 27th March), different bolts behaved in different ways. Bolts 1, 2 and 3 all moved towards axial compression along some sections of their length, though retaining a similar shape of overall axial load distribution along their lengths. This resulted in the tops and bottoms of the bolts going into compression with the central sections remaining in tension. Bolts 4, 5 and 6 all continued to develop tensional loading along their lengths. The period from installation to cut past was accompanied by very high levels of bending in most bolts, particularly bolts 2, 4 and 6, towards the centre of the faceline.

During the intensive periods of monitoring whilst nearby Megastrands were tensioned, no changes in bolt strain/load were measured though changes did occur on both bolts 1 and 2 before, between and/or after successive intensive periods of monitoring. In the case of Bolt 1, the successive periods were 20 minutes apart, occurring during the period of other Megastrand tensioning and when some other bolts were moving into tension.

Following the machine cut past and Megastrand installation, tensioning and grouting, all bolts continued to develop load, with some gauges being lost due to high strain and each bolt probably going beyond yield at some location along its length. Some gauges reached 18000 microstrain (1.8%) prior to failure. This should be considered in the context of the steel characteristics, with yield occurring at approximately 0.3% and strain at maximum force being above 8% according to the BS7861-Part 1:1996 (reference 3).

34

4.2.7 Results – grouting

Two bottle samples (samples 7 & 8) were taken from the mix of Pozament HPRG thixotropic grout which was used to encapsulate the newly installed Megastrands on 27th March 2007. These were tested at an accredited test house at an age of 38 days. Sample 7 had a density of 1.950 g/cc and UCS of 62.8 MPa and Sample 8 had a density of 1.925 and UCS of 60.7 MPa. Sample 8 was slightly misshapen due to storage and two small air pockets had formed on the sample surface.

BS7861 Part 2:1997 (reference 1) calls for a 28 day cube strength of at least 80 MPa, for samples cast in the laboratory. The manufacturer’s data sheet states that a density of 2.2 g/cc will be achieved with the correct water solids ratio of 0.3 and proper mixing without air entrapment. However the measured bottle sample strengths and densities are comparable or higher than those typically obtained for HPRG from other underground sites in the UK, though higher strengths can be achieved for a pumpable mix in laboratory conditions using the same equipment. It is probable that the densities and strengths achieved on site were attributable to a combination of the following factors:

• problematic pumpability of HPRG at the recommended water to solids ratio, • probable air entrapment during the mixing process using the Whyte Hall pump and

mixer system, • strength reduction due to higher water temperatures than those specified for laboratory

tests.

4.2.8 Discussion and conclusions

The Megastrand tensioning process to approximately 10 tonnes at this site did not produce any re-closure of previous roof dilation measurable with the highly sensitive extensometers used. This was the case even though the immediate roof had dilated by approximately 20 mm during the preceding 5 days. It can therefore be concluded that, previous reports of the roof being lifted during Megastrand tensioning most probably refer to relatively exceptional circumstances where very large displacements have already occurred. Also when roof lifting does occur it is likely only to apply to the very bottom roof strata where discrete bed separations and bolt debonding or bolt failure may have already occurred.

The strain gauged rock bolt results are difficult to interpret. Although no change in strain was measured on any of the six bolts specifically during the one or two minutes of tensioning of a nearby Megastrand, there was definitely a trend towards bolt compression on bolts 1, 2 and 3 over the period when Megastrands were being tensioned in the vicinity and in some cases the amount of compression was significant. The compressive strain experienced along sections of these bolts was 2000 to 4000 microstrain, equivalent to localised compressive loading of up to 25 tonnes or more. However this did not appear to prevent these bolts developing further tensional loading over subsequent days as the widening process progressed away from the measurement site. Whilst it is possible that this bolt compression was associated with the Megastrand tensioning operations, it is also possible, that it was associated with the general redistribution of loading within the roof which would have occurred during the widening operations, with cantilever and destressing effects occurring in the previously exposed section of roof, where these bolts were located, as the face heading width increased.

It can be concluded that the Megastrand tensioning operation in itself did not damage the roof and there is no evidence that it impaired the operation of the previously installed reinforcement.

35

5 IMPROVED MODELLING OF FLEXIBLE LONG TENDONS

As described in Chapter 2, very few modelling studies of tensioned tendons in mining have been reported. The work undertaken during this Project has concentrated on incorporation of tensioned long tendons into geotechnical numerical models using the FLAC finite difference continuum modelling package supplied by ITASCA Inc. FLAC is an acronym for Fast Lagrangian Analysis of Continua. The work has involved the following;

a) developing a methodology for incorporating pre-tensioning of fully grouted tendons into the FLAC models generally used for assessing and comparing reinforcement designs for underground coal mining applications,

b) developing a methodology for examining tensioned “truss” type supports incorporating flexible long tendons using FLAC – this was prompted by a suggestion from senior personnel within the UK mining sector that application of such systems, as used in the USA, could be highly cost effective if adopted in the UK

c) use of FLAC models for comparison of the effectiveness of the two systems described above (grouted tensioned long tendon reinforcement and tensioned “truss” systems) with conventional non tensioned flexible reinforcement.

d) application of the techniques developed to specific support design problems at a UK colliery

A further area of flexible long tendon modelling work under the Project has been to compare alternative support strategies for both hypothetical and specific sites where the use of a less dense primary rock bolt support pattern at the face of the heading combined with a secondary support system installed outbye incorporating flexible long tendons has been considered as an alternative to a higher density of reinforcement applied close to the face of the heading. The incentive for such strategies comes from the potential increase in drivage rates that may be achievable and it is important to identify whether this is a safe option.

The work has therefore concentrated on developing an approach for representing tensioned and untensioned long tendons in FLAC and applying the resulting models to real mining problems, rather than fundamental studies of tendon / grout / rock interaction.

All the modelling described has used FLAC3D to represent a short “slice” of mine roadway and the surrounding strata one support cycle thick. This has allowed the spacing of reinforcement along the roadway to be more realistically represented than in a purely two dimensional model, but avoids the complexities and long run times associated with a truly three dimensional model of the complete tunnel including the heading face

5.1 REPRESENTATION OF TENSIONED TENDONS AND TRUSS SYSTEMS

Both FLAC and FLAC3D allow a pre-tension to be specified for a bolt or cable as it is installed. However, when using this feature the forces in the tendon will not be in equilibrium with those elsewhere in the model. As a result a substantial proportion of the specified pretension may be lost as the model deforms to achieve equilibrium. The alternative is to model pre-tensioning as it is actually achieved, by applying load to the bolt or cable. The procedure adopted was as follows:

• Specify bolt geometry, dimensions & steel properties • Set bond properties for anchored length at top of bolt (spin into resin)

37

• Apply load to base of bolt and reaction force to roof (pressurise jack) • Step model to equilibrium

o Tension spreads along free length of bolt o Balancing loads develop along anchored length

• Fix base of bolt to rock in roof (tighten against face plate) • Remove applied loads at base of bolt (remove jack) • Set bond properties for rest of bolt (if fully grouted)

Pre-tensioned truss systems are sometimes employed as an alternative or in addition to bolts to provide confinement and support to mine roadway roofs [Oldsen et al 1997, Pile et al 2004]. They consist of angled bolts or cables installed each side of the roadway connected by a truss bar or cable which is tensioned during installation. The reaction forces where the truss bears against the rock generates horizontal in addition to vertical confining stresses in the roof. The structural elements and commands provided in FLAC allow a truss system to be simulated either by continuous cables or separate components with appropriate connections. Pre-tensioning of the truss was achieved in a similar manner to that for pre-tensioning bolts by applying the tensioning force to each end of the truss, stepping the model to reach equilibrium, connecting the truss ends and then removing the tensioning forces.

The effectiveness of these systems as remedial support was compared by modelling a non specific roadway (5.2 m wide x 3.0 m high) with weak roof that softened on excavation. This comprised mudstone up to 3.2 m into the roof with a seam thickness of 2.4 m as shown in Figure 5.1 (strata sequence and model grid). The material behaviour adopted for the rock incorporated the following features:

• Increasing strength with confinement • Bi-linear strength envelope with higher friction at low confinement • Reduced strength parallel with stratification • Post-failure softening or strength loss

The property values assigned to the different rock types were typical of those found applicable in UK coal measures from previous work. The strengths used were reduced from those typically measured in the laboratory to represent the in-situ or rock-mass strengths. Figure 5.2 shows the strength envelopes assigned for the three rock types used in the model.

The initial vertical stress in the models was set to 20 MPa at the level of the roadway, corresponding to a depth of 800 m. The lateral stress acting across the roadway was set at 16 MPa in rock, with a lower value of 12 MPa in coal.

The models incorporated 7 x 22 mm x 2.4 m fully encapsulated high strength steel rockbolts per row in the roof with a yield strength of 280 kN, bond strength of 100 kN/m and bond stiffness of 200 MN/m/m. Each rib was reinforced by 3 x 24 mm x 1.8 m GRP rockbolts per row with a yield strength of 300 kN, bond strength of 100 kN/m and bond stiffness of 20 MN/m/m. Model runs were undertaken at various bolt densities to identify sensitivity. These showed that increasing the bolt density beyond 1.9 bolts per m2 (0.7 m spacing) did not improve roof control significantly. Also increased bolt length from 2.4 to 3.0 m had little positive effect. Subsequent models examining the effects of flexible long tendons therefore used a row spacing of 0.7 m. The flexible tendons were 2 x 4 m long flexible bolts per row with a yield strength of 450 kN, bond strength 430 kN/m and bond stiffness of 200 MN/m/m. Encapsulation length was 2 m.

The model sequence comprised, installing the remedial support systems, then failing the original roof-bolts and following the model response. The model roof displacements with no

38

secondary support installed are plotted in Figure 5.3 showing the vertical roof movement or lowering and the horizontal convergence of two points either side of the roof (termed roof shortening). On excavation of the roadway, both roof lowering and shortening increased but then stabilised. When the roof-bolts were failed, the roof displacements increased again. The roof lowering continued to increase with no evidence of stabilising, in effect indicating a roof fall.

The response with tensioned cables installed as remedial support is plotted in Figure 5.4 (a). In this case the roof lowering started to increase again after the bolts were failed but soon restabilised. Figure 5.4(b) plots the tension in the cables showing the initial application of pretension to 200 kN. On failure of the roof-bolts the cable tension increased to 450 kN, the yield strength of the cabled specified in the model.

The response with non tensioned cables installed as remedial support is plotted in Figure 5.5 (a). The roof lowering again stabilised after the primary bolts were failed and was slightly lower than with pre-tension. The final tension developed (Figure 5.5 (b)) was significantly less than with pre-tension and the flexible bolts did not reach yield in this case. This is a significant consideration with regard to the advisability of using tensioned reinforcement systems.

The response with a tensioned truss system as remedial support is plotted in Figure 5.6 (a). The roof lowering again stabilised but after substantially more movement than for the tensioned cables. The response of the truss tension in Figure 5.6(b) is of particular interest. Once again it shows the initial application of pre-tension. However, when the roof bolts were failed the tension in the truss first started to reduce as a result of the roof shortening before finally increasing again as the roof lowering became more significant.

Figures 5.7 and 5.8 illustrate the final state of the models showing strain contours and cable tensions for the two cases. The results showed both systems stabilising the roof. The truss system allowed much larger roof displacements before the roof stabilised, the tension generated in the truss was less than in the tensioned cables. The initial reduction in the truss tension illustrates what may be a general point, that tensioning of roof truss systems is unlikely to be effective in mining environments where the roof is shortening.

5.2 APPLICATION TO MAINGATE SUPPORT ON DRIVAGE AT COLLIERY C

Colliery C was working longwall retreat panels at a depth of approximately 1000 m. Conditions experienced in the Maingate of the previous panel during retreat had been poor resulting in production delays. For the next face being developed, a dense pattern of reinforcement was planned as follows:

• installed at face during development at 1.5 m cycles o 7/8 x 2.4 m x 22 mm roof bolts o 2 x 8 m tensioned tendons o 3/2 x 1.8 m x 22 mm rib bolts

• installed approximately 70 m behind the face o 1 x 8 m tensioned tendons o 2 x 8 m tensioned tendons, angled above ribs with provision to attach truss/sling

• Installed before retreat o 2 x 3.8 m rib tendons o Truss/sling between angled tensioned tendons

39

The mine operator was concerned that this support would result in slow drivage rates and RMT were requested to examine alternatives to enable faster drivage rates. The support system had to be adequate for development and additional support could be installed for retreat if required. The alternative support patterns examined concentrated on a reduced density of 2.4 m roof bolts but with a higher than usual density of longer tendons, in the form of un-tensioned flexible bolts.

The comparison between different reinforcement patterns was made using computer modelling using the techniques described in section 5.1. An initial comparison was made for expected conditions during development of the gate, with the reinforcement represented in the model being that intended to be installed at the face during development. Additional support or reinforcement to be installed back from the face or prior to retreat were not included at this stage, nor were stress increases that might be experienced during face retreat allowed for.

The FLAC3D models described in section 5.1 were modified to reflect the geological sequence, rock properties and stresses expected at the site. Rock properties were available from tests carried out for previous modelling exercises for the mine.

5.2.1. Site description

For the modelled panel the depth varied between 970 m-1030 m and the seam thickness was 1.8 m - 2.2 m. In-situ stress measurements were available from two sites at the mine, the results are summarised in Table 5.1.

Table 5.1 Measured in situ stress

Site Method Depth from Elastic Maximum Minimum Azimuth surface modulus of horizontal horizontal of

(m) strata stress stress maximum (GPa) (MPa) (MPa) horizontal

stress Underground Overcore 960 25 24 13 334o

Surface borehole Hydrofrac 916-1001 - 26-29 14-15 149o-154o

Surface borehole Breakout - - - - 142o

The gates of the previous and planned panels were oriented with azimuths of 150o/330o placing them approximately in-line with the expected maximum horizontal stress direction.

5.2.2. Model description

The proposed nominal width for the Maingate was 4.7 m. The model roadway was represented as 4.8 m wide and 3.2 m high with the roof horizon formed at the top of the seam. The planned reinforcement system as represented in the computer model is illustrated in Figure 5.9. In all cases the lower end of the bolts or cables was set to be rigidly connected to the model grid, representing a strong plate to retain bolt load as well as the grout bond between bolt and rock. The bolt properties used are summarised in the table below:

40

Table 5.2 Modelled support parameters

Roof Solid rib Face rib Roof Roof Bolt type KT KT GRP Tensioned tendon Flexible Bolt yield strength (kN) 280 280 400 600 500 Bond strength (kN/m) 400 50 50 430 400 Pre-tension (kN) 0 0 0 100 0

Model runs were conducted comparing alternative reinforcement patterns with the two geological sequences shown in Figure 5.10. The main difference between the sequences is that sequence B includes a band of weaker mudstone in the roof at 4 m above the seam not present in sequence A. The properties assigned to the different strata types were obtained from previous modelling work.

Results obtained with the planned reinforcement system are illustrated in Figures 5.11 and 5.12 showing shear strain contours, bolt and cable loads and simulated roof extensometer displacements. The range of roof displacements indicated by the simulated extensometers is representative of those that can be experienced during gate development at the mine. Except for the lowest of the stresses plotted, the results showed both bolts and cables (in this case tensioned tendons) being loaded up to their yield strength.

The difference between the two geological sequences examined was relatively small. The mudstone at 4m in sequence B allowed additional movement at this horizon for the highest value of lateral stress used. This value for the lateral stress range was higher than the expected range. In effect this would be similar to the rock mass strength being lower than represented in the model.

5.2.3 Comparison of alternative reinforcement patterns on drivage

Model runs were carried comparing the performance of alternative reinforcement patterns. The basic form of the alternatives examined is shown in Figure 5.13. The tensioned tendon bolts were omitted from the original pattern and 7 of the 2.4 m KT bolts replaced by flexible bolts in a staggered 4/3 pattern.

Results obtained using geological sequence A are plotted in Figure 5.14 (showing roof extensometer movement at roof level and at 2 m into the roof) and in Figure 5.15 (showing maximum bolt and cable strains). With the tensioned tendon cables omitted but no flexible bolts installed the roof displacements and KT bolt strains increased.

Replacing 7 KT bolts with 4 m flexible bolts reduced the roof displacements, but the total movement remained more than for the original pattern. The bolt strains reduced below those for the original pattern. The strains developed in the flexible bolts were similar to those for the tensioned tendons. Increasing the flexible bolt length from 4 m to 6 m did not result in a major difference in the results.

A final run was conducted to check the potential effect if full encapsulation could not be achieved with 6 m flexible bolts. The upper 4 m of the bolts was assigned the same bond strength as previously, the lower 2 m was not bonded. The results obtained indicated considerably poorer roof control with increased roof displacements and higher strains in the KT

41

bolts. This result demonstrates the importance of obtaining full encapsulation and of maintaining a high density of effective reinforcement in the bolted horizon.

The shear strain contours, bolt and flexible bolt loads obtained form the run with 4 m flexible bolts indicated that the two outer flexible bolts were not fully loaded and were less effective than the others. This is likely to be the case where the roadway behaviour is symmetrical (as in the model) with displacements extending higher into the roof above the centre of the roadway. In these circumstances, longer tendon reinforcement is likely to be more effective placed towards the centre of the roadway than at the sides.

Results obtained using geological sequence B were very similar to those obtained with sequence A. There was slightly more benefit to longer flexible bolts in this case.

5.2.4 Conclusions from modelling drivage support design

The following conclusions can be drawn from the modelling of alternative support on drivage for the particular set of conditions modelled.

• The results obtained indicated that to maintain roof conditions during gate development in these conditions it is important to maintain a high bolt density.

• If flexible bolts are adopted to replace ordinary (KT) roof bolts as part of the pattern then it is important that they are fully encapsulated.

• The roof condition was not strongly sensitive to varying the length of flexible bolt between 4 m and 6 m.

• The alternative patterns examined used flexible bolts in place of KT bolts. These gave roof conditions comparable to the original pattern but no clear improvement.

• Using flexible bolts in addition to rather than in place of KT bolts may produce improved roof control and condition.

5.3 APPLICATION TO MAINGATE SUPPORT ON RETREAT AT COLLIERY C

As described above, the next panel to be developed at Colliery C was planning to use a combination of tensioned tendons and a trussing system as additional support to be installed at various stages following drivage and during face retreat.

The work described in section 5.2 for development included examination of different geological sequences and initial stresses. For the modelling of face retreat, these variations were not included. The geological sequence used was Sequence A (see Figure 5.10). The vertical stress was taken as 25 MPa and the lateral stress acting across the gate was set as 16 MPa.

5.3.1 Simulation of face retreat

When the face is retreated, the gate road ahead of the face will be subjected to altered, and probably increased, stresses. The effect of this was simulated by starting with the existing model for development of the roadway and increasing the stresses acting at the outer boundary in several stages. This procedure has been used effectively to provide guidance on expected conditions during retreat at many sites.

The stress redistribution around a longwall face and the resultant stress changes for the gate are complex and will be dependant on a large number of factors including:

• Caving geometry

42

• Waste consolidation • Rock properties • In-situ stress regime

The stress increases applied to the model are shown in Figure 5.16. Although model runs were conducted using both the stress-notched and non-stress-notched options shown in Figure 5.16, only the results obtained for the stress-notched option are included in this report.

5.3.2 Support system

The support system planned for the gate was listed in section 5.2 and involved support components being installed at different stages during the life of the gate. For the modelling of secondary support described here, there was no differentiation in the model between support installed 70 m behind the face and that to be installed prior to retreat, both were added to the model at the same stage.

Model runs were conducted both with and without the additional support planned for retreat. For the runs with the additional support, this was added to the model after the development phase had been completed and prior to the stress increases being applied.

The bond properties for the tensioned long tendons were also set at the same stage, implying that they were grouted immediately after installation. The truss/sling was assigned a tensile yield strength of 600 kN and tensioned to 100 kN when installed.

5.3.3 Results

Model results obtained with the planned additional support installed are illustrated in Figure 5.17, showing shear strain contours, bolt and cable loads and simulated roof extensometer displacements. As the stresses were increased the roof extensometer displacements also increased, but the height of softening did not extend further into the roof. By the final stage, all the roof bolts and cables were loaded up to yield.

The model results showed a substantial increase in roadway convergence, primarily due to floor heave and rib movement, as the stress increases were applied. The additional support had a relatively minor effect in reducing this convergence.

The roof displacements generated in the model are plotted in Figure 5.18. Despite the high support/reinforcement density installed, the roof displacements were still large, although they were much less than those in the floor or ribs.

The (maximum) axial strains generated in the roof bolts and cables are plotted in Figure 5.19. With the bolts at yield, higher strains imply a greater risk of broken bolts. The additional support planned to be installed for retreat was effective in reducing the strains generated in roof and long tendons.

It was found that although the truss/sling was tensioned to 100 kN when installed, the tension in this support element actually reduced as the model was run and the stresses increased. This was due to the separation between the end-points of the truss (where it was anchored to the cables) reducing as the stresses were increased and the roadway profile reduced.

Final model runs were conducted with the roof bolts and/or long tendons deliberately failed. With this done the roof displacements (which had stabilised) increased rapidly, Figure 5.20.

43

With no additional support installed and all the original bolts and cables failed, the roof displacements continued to increase with no sign of stabilising; this effectively indicates the roof falling out. With additional support installed and the original bolts and cables failed, the displacements stabilised. With the additional vertical cables failed but the truss system left intact, the roof again stabilised although after a larger amount of movement. In this instance it was the truss system providing support; in doing so the tension in it had increased again although it remained well below yield, the loads in the angled tendons to which the truss was attached did reach yield.

5.3.4 Conclusions from modelling face retreat effects on support system

• As the stresses were increased to simulate the effects of face retreat, the model results showed large deformations, particularly in the ribs and floor resulting in severe convergence of the gate.

• The roof displacements generated in the model also increased substantially, but the height of softening remained below the height of the long tendons.

• The increasing roof displacements generated increasing strains in the roof and long tendons, increasing the risk of tendon failure. The additional support planned to be installed prior to face retreat reduced the strains developed in the roof bolts and long tendons.

• With the rest of the support system failed, the truss system was effective in stabilising the roof although it allowed substantial displacements before doing so. The loads within the truss remained below yield, although those in the angled long tendons to which the truss was attached did reach yield.

5.4 CONCLUSIONS ON MODELLING FLEXIBLE LONG TENDONS

The work described in this Chapter has shown that a method has been developed for modelling and comparing the behaviour of tensioned and untensioned flexible long tendon reinforcement when used as part of a coal mine support system in realistic geological and geotechnical conditions. The method developed simulated the process of tendon installation and tensioning so as to ensure that the tension was retained in the tendon following grout encapsulation. The modelling code used was FLAC3D which allowed reinforcement row spacing to be included in the analysis. The work was extended to examine the effectiveness and applicability of tensioned cable truss systems.

For the particular examples modelled, both hypothetical and real, no benefit was seen from applying pre-tension to the flexible tendons. Indeed, the effect of pre-tensioning appeared to be to take the tendons beyond their yield load at an earlier stage in the loading process resulting in marginally higher roof displacements before stability was established. However the limitations of the modelling should be noted. In particular the models used could not realistically simulate loss of shear resistance in the bedded strata due to bed relaxation prior to tendon installation and so could not simulate any benefit which may be derived from pre-tension in remobilisation of such shear resistance, a claimed benefit of tensioned tendons. Considerable further work is required to investigate this claimed benefit through modelling.

The modelling exercise investigated the effect of not achieving full grout encapsulation of a non tensioned long tendon. In the particular conditions modelled, typical of UK coal mining, full encapsulation was shown to be critical for achieving stability, as would be expected due to the major loss of reinforcement stiffness resulting from non encapsulation.

44

It was also found that, in the symmetrical geotechnical conditions modelled, flexible long tendon reinforcement was more beneficial when biased towards the centre of the roadway. This may not be the case in asymmetric conditions such as those resulting from a lateral stress “notch” or due to geological and structural variations.

The investigations into the applicability of trussing systems employing flexible long tendons indicated that, where horizontal stresses are relatively high, as found in most UK underground coal mining situations, any tension applied to the truss was rapidly lost during roof shortening following installation. Trusses did not supply any useful control or support of the roof until there had been very high levels of roof failure and roof lowering, at which point they were capable of acting as a “last resort” sling support provided the density was sufficient to avoid failure of the anchoring cables. The cables used for anchoring the trusses would have been much more effective if used as a reinforcement system towards the centre of the roof span and had little effectiveness as reinforcement when placed over the ribsides as required for truss anchorage.

45

6 ADVICE AND DRAFTS PROVIDEDTO BS REVISION COMMITTEE

The need to revise the British Standard for cablebolting consumables (reference 1) was recognised following the widespread introduction of long tendon designs not covered by the original standard, in particular, tensioned tendons. Advice and material input to the revision committee was enabled by an earlier project reported in reference 2, and continued via the project reported here.

Advice and recommendations were given at the regular monthly meetings of the committee, and draft test procedures (suggested for inclusion as annexes to the revised Standard) were submitted when appropriate. A summary of advice given, plus draft test procedures submitted during the period covered by this project are reported below.

6.1 ADVICE AND RECOMMENDATIONS

6.1.1 Scope of the revised Standard

The existing Standard covers only birdcaged cablebolts of ‘single’ and ‘double’ strand configuration, with corresponding capacities of 300 and 600 kN. Since publication, usage of long tendons has evolved so that tendons with 300 kN are used only for ribside reinforcement. It was recommended that this should be formally recognised in the revised Standard, in that the scope should include roof reinforcement only, and so exclude single strand systems. This would allow minimum strength criteria to be based on:

• the Reflex bolt for resin encapsulated (or non-post grouted) systems, and • the double minicage cablebolt (or similar) for cementitious post grouted systems.

The single strand configuration would continue to be usable for ribside reinforcement, and would probably be included in a future addition to the Standard – BS7861:Part 3 – covering flexible tendons for reinforcement of ribsides. Some initial work on this is described in Chapter 7.

6.1.2 Organisation of the revised Standard

Advice to the revision committee regarding the way the revised Standard would categorise long tendons was that categorisation should be based on whether or not a system was designed to be post grouted, i.e. whether grout encapsulation would take place after or during installation of the long tendon.

Post grouted systems include birdcaged (rarely used now), minicage, and nutcaged cablebolts, and the Reflex bolt when equipped for post grouting. A cementitious grout would normally be used, and, as this would require many days to develop full strength, a post grouted system would not be regarded as providing reinforcement on installation. Post grouted systems could also be classified as cementitious grouted systems.

Tensionable strands should be regarded as a subset of post grouted systems, since they would normally be post grouted after tensioning. Again, since the cementitious grout will require many days to develop full strength, a tensionable system would not be regarded as providing full

47

reinforcement capability upon installation. Since these tensionable systems utilise resin as well as grout encapsulants, they could also be classified as composite systems.

Non-post grouted systems currently comprise only the Reflex bolt, and only when installed in, and fully encapsulated with, resin. Since resin takes less than one hour to achieve near-to full strength, these systems would be regarded as providing full reinforcement upon installation. Non-post grouted systems could also be classified as resin grouted systems.

6.1.3 Tensionable system issues

It was advised and agreed that performance tests carried out on tensionable systems would be carried out in the ‘untensioned state’, as this was the worst case scenario. This would enable tests developed for non-tensionable systems to be applied to tensionable ones as well.

A recommendation was given as to an adequate anchorage length for initial installation of a tensionable system, so that the required tensioning load could be applied safely. The recommended criterion was that the anchorage length provided at initial installation should be sufficient to withstand a tensile load of at least the minimum ultimate strength of the tendon. This anchorage length (in mm) would be equivalent numerically to 3 x the ultimate tendon strength (in kN). This is based on a requirement to be included in the revised Standard which specifies that a resin bonded length of 450 mm shall have a minimum system bond strength of 300 kN and a 2:1 in-situ factor of safety (FOS).

6.1.4 Encapsulant properties

It was recommended that encapsulating resin should have the same mechanical performance and characteristics as those required by BS7861: Part 1: 2007 (reference 4).

Cementitious grout should have mechanical properties, as specified in the existing BS7861: Part 2: 1997 (reference 1), with additional requirements regarding water quality.

6.2 DRAFT TEST PROCEDURES / ANNEXES

Procedures for tests to be carried out on long tendons as part of an acceptance programme, and suggested for inclusion in the revised Standard, were developed during the project period and are described below.

6.2.1 Grout encapsulation test – ‘bottom up’ grouting

A series of tests intended to fully encapsulate long tendons from the proximal end upwards (bottom up) were carried out in the laboratory. The objectives of the tests were to • fully encapsulate commonly used long tendons using the only currently approved grout,

and • establish a standardised method for inclusion as an annex to the revised Standard.

The test programme is described in Chapter 3. The method used was based on that given in the present Standard, but altered to; • reflect use of new long tendon types • allow new grouts to be included, provided they meet other approval criteria, and • enable encapsulation of the length of strand specified by the supplier for use in the field.

48

The procedure should be applied with a caveat that this must be a system test, i.e. a candidate long tendon must be tested with the grout and hole size specified by the supplier(s) and the result, and any approval statement would apply only to that system. A change to the tendon / hole size / grout combination would require new tests.

The draft procedure is given, in full, in Appendix 1.

6.2.2 Grout encapsulation test – ‘top down’ grouting

A method for grouting from the distal end of a tendon downwards (top down) was developed from a testing programme, as mentioned in section 6.2.1 above and reported in Chapter 3. Although the test programme was not fully successful in terms of encapsulation of the Megastrands tested, the test is regarded as valid for this type of system.

The comments regarding the need for a system test, noted in 6.2.1, also apply here.

The draft procedure is given, in full, in Appendix 2.

6.2.3 Determination of bond strength and system stiffness – cementitious grout anchored system

A test procedure for determination of bond performance of resin encapsulated long tendons was supplied to the revision committee in the timescale of a previous project (reference 2). A further procedure, for cementitious grouted systems, has been finalised during the current project, and, as with the encapsulation tests given above, it is intended that this should be a system test of the long tendon, hole size and the grout intended for use with it. The test results obtained from it would apply only to the test combination, and should it be intended to use a different combination, further tests would be required since the performance of the new combination would be an unknown quantity.

The pass criteria for this performance test were determined by the revision committee, following provision of results from a comprehensive series of tests undertaken in the previous HSE project. The acceptance criteria are as follows:

• the minimum system bond strength (load at which the slope of the load / displacement characteristic falls below 20 kN/mm) shall be 400 kN, and

• the minimum system stiffness (slope of the load / displacement characteristic) shall be 95 kN/mm, measured between loads of 150 kN and 300 kN.

The draft test procedure is given, in full, in Appendix 3.

6.2.4 Shear test on tendon / grout system

A test procedure for testing long tendons encapsulated with resin or cementitious grout under shear load conditions was supplied to the committee during the project period together with proposed acceptance criteria derived from tests undertaken in the previous HSE project. The acceptance criteria are as follows:

• the mean of three test results should exceed 325 kN for resin encapsulated tendons, and • the mean of three test results should exceed 380 kN for grout encapsulated tendons.

The draft test procedure is given, in full, in Appendix 4.

49

6.2.5 Composite systems

It was recommended that the components of a composite system should meet the requirements of the tests described in 6.2.3 and 6.2.4 above for the resin and grout encapsulated lengths respectively.

50

7 LABORATORY TESTING OF ALTERNATIVE RIB REINFORCEMENT SYSTEMS

The revised Standard for cablebolting materials is not expected to include within its scope, systems intended for use in reinforcement of the sides of underground roadways (ribsides). But, very difficult conditions experienced with longwall roadways at colliery A, due to very large displacements of ribsides, and culminating in a fatality attributed to a fall of ground from ribs, have highlighted the importance of rib reinforcement. An extensive testing programme has been conducted on existing and candidate materials for ribside use, over an eleven month period. The results of this work are expected to provide two major benefits:

• optimisation of ribside materials – bolts, long tendons and encapsulants – for UK coal mines, thus facilitating a safer working environment, and

• provision of key data for inclusion in a possible new section (Part 3) of the British Standard 7861 – specifying ribside consumables.

The analytical objectives of the test programme were to:

• compare a range of potential rib tendon systems, it being necessary to compare both flexible and non flexible systems. Modelling had shown the need for at least 100 kN/m of encapsulation bond strength to have a significant effect on rib behaviour at colliery A,

• determine whether this could be achieved with polyurethane which was perceived to have possible advantages in ribside use,

• develop a set of tests and possible acceptance criteria for ribs which would be different to those for roof – in particular, residual load after 50 mm displacement and displacement to maximum load, rather than look for high system stiffness and bond strength, and

• determine whether a test could be developed in coal core, and how results in coal compared with those in sandstone as specified for current and proposed British Standard tests.

The results of the test programme are summarised below.

7.1 RIB REINFORCEMENT SYSTEMS TESTED

Rib reinforcement in colliery A longwall face roadways has previously been via:

• on the cuttable side, FT500 (Weldgrip) solid GRP dowels and FT500 (Weldgrip) hollow dowels installed using cementitious grout injected through the centre hole.

• on the solid side, KT type steel rockbolts and, (where necessary) grout injected FT500 hollow dowels.

• Solid KT and FT500 bolts have both been encapsulated with Minova Lokset ‘AT’ capsule resin. Other systems have also been used, including Osborn ‘Reflex’ steel flexible bolts, again with ‘AT’ resin.

For this laboratory testing exercise, it was agreed that a comprehensive series of tests would be carried out with the FT500 dowel and performance of this product would be taken as the benchmark against which other products/ systems could be assessed. The initial programme of

51

tests set out below in coal/CBG cylinders (except where stated) and comprising 3 tests for each combination of parameters in each was agreed with colliery A:

Weldgrip FT500 GRP bar / CBG grout at 10MPa confinement (in sandstone) Weldgrip FT500 GRP bar / CBG grout at 3 confinements (1 MPa, 5 MPa and 10 MPa) Weldgrip FT500 GRP bar / AT resin (43 mm hole) at 2 confinements to be decided Weldgrip FT500 GRP bar / AT resin (27 mm hole) at 2 confinements

Weldgrip 28 mm GRP bar / CBG grout at 2 confinements Weldgrip 28 mm GRP bar / AT resin at 2 confinements Weldgrip 28 mm GRP bar / PUR resin at 2 confinements

Big (28 mm) steel bolt / CBG grout at 2 confinements Big (28 mm) steel bolt / AT resin at 2 confinements

KT steel bolt / CBG grout at 2 confinements KT steel bolt / AT resin at 2 confinements KT steel bolt / PUR resin at 2 confinements

Osborn Reflex bolt / CBG grout at 2 confinements Osborn Reflex bolt / AT resin at 2 confinements Osborn Reflex bolt / PUR resin at 2 confinements

7.2 TEST PROCEDURES

The laboratory short encapsulation pull test (LSEP test) is now a well established tool in deriving and comparing the performance of reinforcement systems, and is incorporated into the current British Standard for rockbolting consumables used in coal mines (reference 4). The method documented in the British Standard uses a homogeneous rock core (sandstone) of known properties housed in a cell which provides a bi-axial confining pressure. For comparison purposes, some of the initial tests were conducted in sandstone, but in order to try to determine more effectively the performance of rib reinforcement systems in the coal seam, it was decided to manufacture cores with an axially central section composed of coal lumps obtained from colliery A as described below.

Lumps of coal were machined and glued together to form a piece approximately 200 mm long and of sufficient diameter to fit inside a cylindrical plastic mould of 145 mm diameter. The coal piece was set in sufficient CBG grout to form a 145 mm diameter core that was long enough to fit in a biaxial cell. After sufficient curing time, the grout / coal cores were held in a biaxial cell on a lathe, under a nominal light load, and a hole was drilled into the coal to a depth of 160 mm, using a drill bit of appropriate diameter for the bolt size and resin / grout type that were to be installed in that core. The internal diameters of the drilled holes were measured and, in due course, the bolts were set in the cores using whichever resin or grout was pertinent to the test. After an appropriate length of time, at least 24 hours for polyester and polyurethane resins and 14 days for CBG grout, the core assemblies were pull-tested in the biaxial cell. The LSEPT test requires a confining pressure simulating stresses exerted on reinforcement systems underground. A nominal 10 MPa is used in the British Standard Test (reference 4). For these tests in coal it was intended to use a range of pressures simulating the effect of increasing depth into the rib, from 1 MPa to 10 MPa. It was found however, during the initial testing programme, that the “coal cores” could not reliably withstand 10 MPa and it was agreed to carry out tests at 1 MPa and 5 MPa.

52

Bond length is obviously critical to performance and, initially, a length of 320 mm was trialled in line with that used for assessment of long tendons. However the initial test programme results, including those from tests in sandstone cores, showed that it would not be possible to determine the peak and residual load characteristics at this embedment length. Results showed bond strengths exceeding yield strength of the tendons used. It was decided, therefore, to reduce the bond length to 160 mm in line with that used for assessment of rock bolts in the British Standard (reference 4).

7.3 TEST RESULTS AND DISCUSSION

The results of the pull tests on the various bolt types with the different confining pressures and resin/grout types are best summarised as a series of tables. The mean figures take into account all available data for tests under a particular set of parameters. This will usually be a mean of 3 tests. A figure in brackets indicates the number of tests averaged if other than 3.

7.3.1 Weldgrip FT500 24 mm GRP bolts

The initial tests were carried out with 320 mm embedment. These are not tabulated below for reasons previously mentioned (see above). The results of the tests conducted on the Weldgrip FT500 24 mm GRP bolt at 160 mm embedment are summarised in Tables 7.1 and 7.2, below:

Table 7.1 LSEP test results for Weldgrip FT500 GRP: 160 mm embedment in coal

Grout/resin type

Test nos. Confining pressure (MPa)

Mean bond

strength (kN)

Mean system

stiffness (kN/mm between

40-80kN)

Mean maximum

load (MPa)

Mean residual load (kN) at

50mm displacement

AT 1, 2, 3 5 96.7 70.3 116.2 32.7 (2)

AT 4, 5, 6 1 51.9 n/a 58.8 22

CBG 8, 16, 17 1 76.2 41.2 (1) 79.0 9.5

CBG 7, 18, 19 5 106.9 110.8 113.0 19.9

CBG 20 10 138.2 (1) 153.8 (1) 144.9 (1) n/a

PUR 15P, 16P, 17P 5 24.1 n/a 25.0 9.1

PUR 18P, 19P 1 14.3(2) n/a 15.2(1) 3.9

In AT resin, the pull tests were characterised by multiple failures where the load built up and was then rapidly released as stick/slip behaviour. Higher performance parameters were achieved at the higher confining pressure and the consistency between the tests under each confining pressure was good. All samples failed at the coal / resin interface.

53

Table 7.2 LSEP test results for Weldgrip FT500 GRP: 160 mm embedment in sandstone

Grout/resin type


Mean bond

strength (kN)

Mean system


Mean maximum

load (MPa)


50mm displacement

40-80kN)

CBG 13, 14, 15 10 138.8 149.5 139.8 23

In CBG grout the consistency between tests under similar conditions was also good. Again, the higher the confining pressure, the higher the performance parameters. Test No.20 was conducted at 10 MPa confining pressure, but the core sample fractured during the test. The tests were generally characterised by stick / slip behaviour. All samples failed at the bolt / grout interface although sample No.8 also failed at the coal/grout interface. The CBG tests in coal produced higher mean bond strengths, mean maximum loads and mean bond stiffnesses than the equivalent tests in AT resin, but lower residual loads. Three CBG tests carried out in sandstone at 10 MPa confining pressure showed similar results. The tests were characterised by stick / slip behaviour and all failed at the bolt / grout interface. Interestingly the mean performance parameters for these tests were similar to those achieved in coal at 10 MPa confinement (Test 20).

The tests carried out in PUR resin gave significantly lower performance parameters than those in AT resin and CBG grout. After achieving maximum load at a similar displacement to the other tests, the loads rapidly levelled to values similar to the residual loads. Tests carried out at 1 MPa confining pressure showed markedly lower results than those at 5 MPa. All failed at the coal/resin interface, although sample No. 19P also failed partially at the bolt / resin interface.

7.3.2 Weldgrip 28 mm GRP bolts

The results of the tests conducted on the Weldgrip 28 mm GRP bolt are summarised in Table 7.3, below.

In AT resin at 5 MPa confining pressure, the tests were characterised by a rapid increase to maximum or near maximum load. This was followed by stick / slip behaviour, each slip being followed by a steady increase in load before the next slip. Successive peak loads were lower than the previous peak and all three tests showed broadly similar results. All samples failed at the coal / resin interface.

In AT resin at 1 MPa confining pressure, after an initial increase in load, bolt load was maintained or developed steadily for 15-25 mm of displacement. Abrupt failures did not occur on 2 of the tests until peak load had been well exceeded and one test did not show any sudden failure at all. All samples failed at the coal / resin interface. Again all mean performance parameters were higher for 5 MPa confinement than for 1 MPa confinement.

54

Table 7.3 LSEP test results for Weldgrip 28 mm GRP: 160 mm embedment in coal

Grout/resin type


Mean bond

strength (kN)

Mean system


40-80kN)

Mean maximum

load (MPa)


50mm displacement

AT 24, 25, 26 5 136.5 103.2 190.6 68

AT 44, 45, 46 1 41.8 8.13 (1) 81.3 39.5 (2)

CBG 9, 9A, 10, 11 5 143.8 134.9 (3) 266.3 16.7 (2)

CBG 12, 13, 14 1 68.8 14.9 89.3 17.2

PUR 27, 28, 29 5 13.7 n/a 15.4 6.4

PUR 30 ,31 1 12.1 (2) n/a 13.4 (2) n/a

In CBG grout at 5 MPa confining pressure, the tests were characterised by extreme stick / slip behaviour with rapid increase in load to a high maximum followed by alternate cycles of rapid failure and increase in load, often to near maximum. The tests were concluded with a rapid decline in loading to a low residual load. Two samples failed at the grout / coal interface and two failed at the bolt / grout interface. The general behaviour in these tests was extremely unusual and could warrant repeat testing to examine whether it is truly representative. The unusual graphs may have been generated by a combination of high stress build up in the flexible bolt followed by sudden release leading to shock loading of the pressure transducer and rebound of the displacement transducer during the extreme stick / slip behaviour.

In CBG grout at 1 MPa confining pressure, again there was a rapid increase in load with little displacement after which minor stick / slip behaviour occurred. In two of the tests, loading gradually increased after these failures and maximum load was achieved after 27 mm and 37 mm displacement, respectively. The loading rapidly decreased to residual level after maximum load was achieved. All samples failed at the bolt / grout interface. Again the performance parameters at 5 MPa were higher than those at 1 MPa. Also, as for the FT500 GRP bolts, the CBG tests produced higher bond strength, maximum load and system stiffness results than the AT resin, but significantly lower residual strengths.

In PUR resin, at both confining pressures, a low maximum loading was rapidly achieved and the loads then gradually decreased to residual level. All samples failed at the coal / resin interface.

7.3.3 28 mm steel ‘big’ bolts

The results of the tests conducted on the 28 mm steel bolt are summarised in Table 7.4, below.

In AT resin at 5 MPa confining pressure, loading rapidly increased with very low bolt displacement followed by a gradual increase in displacement as maximum load was reached. Test Nos. 42 and 43 gave very similar results in that after maximum load was passed there was

55

stick / slip behaviour as mean loading decreased with bolt displacement to a consistent residual load. Test No. 41 followed a similar shaped loading curve to a much higher maximum load, but with only one significant sudden failure after maximum load was passed. This test failed at the bolt / resin interface, whilst the other two failed at the coal / resin interface.

Table 7.4 LSEP test results for 28 mm steel: 160 mm embedment in coal

Grout/resin type


Mean bond

strength (kN)

Mean system


40-80kN)

Mean maximum

load (MPa)


50mm displacement

AT 41, 42, 43 5 164.2 451.1 199.8 49.7

AT 38, 39, 40 1 55.9 5.8 (2) 75.4 43.6

CBG 47, 48, 49 5 194.3 622.2 196.1 41.1

CBG 50, 51, 52 1 67.9 50.1 (2) 78.6 45.9

In AT resin at 1 MPa confining pressure the three tests showed similar load / displacement graphs. A rapid initial increase in load with very low displacement was followed by a gradual climb to maximum load with several minor failures after which the loading gradually decreased to a steady residual load. All tests failed at the coal / resin interface.

As with the GRP tests, all performance parameters were lower at the lower confinement although residual load was only a little lower.

In CBG grout at 5 MPa confining pressure, loading increased with very low bolt displacement and a high maximum load was rapidly achieved. Thereafter, the load decreased rapidly with displacement. All three tests had a similar residual load. Test No. 47 had a lower maximum load than the other two and failed at the bolt / grout interface. The other two tests failed at the coal / grout interface.

In CBG grout at 1 MPa confining pressure, the initial load increased to an initial failure. After that, the load increased slowly to a maximum before a gradual reduction to a consistent residual loading. All three tests failed at the coal / grout interface.

In contrast to the other tests commented upon so far, the mean residual load achieved at 1 MPa confinement was a little higher than that at 5 MPa. The other mean performance parameters were lower at 1 MPa than at 5 MPa confinement. Comparison of the results of the tests on Big steel bolts with those on GRP bolts shows similarity in that mean bond strengths and system stiffnesses were greater for CBG than for AT resin at each confinement, but a difference in that the mean maximum loads and residual loads were similar for the CBG and AT for each confinement.

56

7.3.4 KT 22 mm steel bolts

The results of the tests conducted on the KT 22 mm steel bolt are summarised in Table 7.5, below:

Table 7.5 LSEP test results for KT 22 mm steel: 160 mm embedment in coal

Grout/resin type


Mean bond

strength (kN)

Mean system


40-80kN)

Mean maximum

load (MPa)


50mm displacement

AT 32, 33, 34 5 110.9 165.2 115.5 32.7

AT 35, 36, 37 1 54.2 n/a 64.7 34.7

CBG 66, 67 5 153.6 (2) 533.9 (2) 153.6 (2) 26.5 (2)

CBG 68, 69 1 97.6 (2) 259.5 (2) 97.6 (2) 34.5 (2)

PUR 71, 75, 76 5 93.7 1454.6 (2)

94.7 37

PUR 72, 77, 78 1 85.4 90.91 (1) 85.8 21.7

In AT resin at 5 MPa confining pressure, all three tests showed a rapidly increasing load to a maximum. After maximum load was achieved, mean bond load gradually reduced with displacement often accompanied by moderate stick / slip behaviour to a consistent residual load. All three samples failed at the coal / resin interface.

In AT resin at 1 MPa confining pressure, loading increased rapidly initially. In two tests, the maximum load occurred at a considerable displacement beyond the bond strength determination point. After maximum load, all three samples gradually declined to a similar residual load.

In CBG grout at 5 MPa confining pressure, maximum load was achieved very quickly. There was no significant stick / slip behaviour and both load / displacement graphs have a high degree of similarity.

In CBG grout at 1 MPa confining pressure maximum load was achieved very quickly. One test maintained a high consistent load resulting in a high residual load, whilst in test No.69 the loading reduced rapidly to a much lower residual load.

Similarly to the Big steel bolts and again in contrast to the GRP bolts, the mean residual loads achieved at 1 MPa confinement for both CBG and AT resin were higher than those achieved at 5 MPa confinement. Other performance parameters were higher at 5 MPa than at 1 MPa confinement. The general trend of performance parameters between CBG and AT resin for the KT bolts is similar to that found for the GRP bolts in that mean bond strengths, maximum loads and system stiffnesses were greater for CBG at each confinement but mean residual loads were lower.

57

Problems were experienced in mixing the components of the PUR resin in some earlier tests with KT bolts and the results of these tests have been discounted. On pulling the bolts, it was seen that some of the resin had not set properly. Further investigations indicated that, whichever component of the PUR capsule was poured first, the components always settled into two layers of immiscible fluid and rotating the bolts in the hole in the sample cores was not sufficient to mix them properly. Therefore, an alternative approach to mixing was adopted whereby the components from the PUR capsules were extracted, separately weighed and then mixed in the correct proportions and poured into the hole and the bolt installed before gelling had taken place.

In these (reported) tests for PUR resin at 5 MPa confining pressure, maximum load (and bond strength) with high system stiffness was achieved rapidly. After maximum load, the loading dropped but relatively high residual load was achieved. Results of tests for PUR resin at 1 MPa confining pressure were very similar in shape to those conducted at 5 MPa, but with lower mean performance indicators.

These results with PUR are significantly better than those achieved in earlier reported tests on GRP bolts and the performance parameters are much closer to those achieved with CBG and AT resin than was the case for the GRP bolt tests. In fact some of the performance parameters achieved for PUR were better than those achieved in the equivalent tests with CBG and AT resin. This suggests that comparable results may be achievable with PUR provided the components can be properly mixed on site in the correct proportions. It also suggests that retesting of PUR with GRP bolts may be warranted.

7.3.5 Reflex flexible steel bolts

The results of the tests conducted on the Reflex flexible steel bolt are summarised in Table 7.6, below:

Table 7.6 LSEP test results for Reflex flexible steel: 160 mm embedment in coal

Grout/resin type


Mean bond

strength (kN)

Mean system


40-80kN)

Mean maximum

load (MPa)


50mm displacement

AT 56, 57, 59 5 83.0 65.2 99.9 55.9

AT 60, 61, 62 1 53.2 n/a 60.6 25.5

CBG 1, 2, 3 5 81.4 70.8 (2) 119.1 65.0 (2)

CBG 53, 54, 55 1 44.4 n/a 64.5 22.2

PUR 73, 79, 80 5 85.5 366.7 (2) 96.6 55.9

PUR 74, 81, 82 1 51.0 n/a 54.3 33.2

58

The shapes of the load displacement curves for AT resin at 5 MPa confinement were similar to those achieved with steel bolts, particularly the KT bolt, with steep initial profiles followed by a plateau of variable length with gradual load shedding to a relatively high residual load. During load shedding two bolts displayed moderate stick / slip behaviour and one did not, this one achieving a significantly higher residual load than the others.

Again, for 1 MPa confinement, the shapes were similar to steel bolts at the same confinement with good consistency between residual loads and only minor stick / slip behaviour. All performance parameters were lower at 1 MPa than at 5 MPa confinement.

For CBG embedment at 5 MPa, the load displacement curves obtained were significantly different to those for steel bolts. Rather than a high peak load being achieved at very low displacement with rapid load shedding, load development was more gradual and a relatively long load plateau was achieved before bond failure and the onset of stick / slip behaviour. Mean performance parameters for Reflex bolts in CBG grout at 5 MPa confinement, were lower than for all other steel bolts, except for residual load which was higher than for all other steel bolts in CBG at 5 MPa confinement.

At 1 MPa confinement, the curves were more similar in shape to those obtained with other steel bolts in CBG grout, with relatively low bond strengths and peak loads but gradual load shedding with increased displacement.

For both CBG and AT resin, all performance parameters, including residual strengths, were lower at 1 MPa confinement than those achieved at 5 MPa confinement.

The shapes of the curves for two of the PUR embedded Reflex bolts at 5 MPa were very similar to the shapes of the curves for KT bolts but with some post failure stick / slip. One curve showed different behaviour with a long gradually rising curve, following initial steep load development, with peak load occurring at between 30 and 35 mm displacement.

The performance parameters for PUR at 5 MPa confinement were of a similar order or better than those obtained with CBG and AT resin.

In contrast the curves for PUR embedded Reflex bolts at 1 MPa confinement were a little different to those obtained for the KT steel bolt. The bond strengths and peak loads were considerably lower but the curves had well developed post yield plateau and reached comparable or better residual loads. As for 5 MPa confinement, the performance parameters for PUR at 1 MPa confinement were of a similar order to those obtained with CBG and AT resin.

7.3.6 Discussion of residual loads

One of the most important performance parameters for coal rib reinforcement is probably the residual load achieved, as relatively large displacements of coal rib can be expected, particularly at colliery A. For this series of tests, this parameter was quantified at a displacement of 50 mm. Figure 7.1 shows a graph of the mean residual loads at 50 mm displacement plotted for each bolt type, embedment material and confinement.

It can be seen that, at low confinements, as are likely to be encountered in the yielded rib up to a depth of several metres at colliery A, for GRP bolts the AT resin gave considerably higher residual loads than the other embedment materials, CBG grout gave 40-45% of the AT resin residual loads and PUR gave very low residual loads. The bigger diameter bolts gave the higher residual loads.

59

For higher confinements (5 MPa) with GRP bolts, as are likely to be encountered deeper into the yielded coal, the residual loads were generally higher and, again, the AT resin gave higher residual loads than the CBG grout with poor results for the PUR.

For the steel bolts, the differences in the residual loads achieved between embedment materials and confinements were much less marked and they were generally higher than for the equivalent size GRP bolt in the same embedment material with the exception of the large diameter GRP bolt in AT resin.

The residual loads achieved by the Reflex bolts were somewhat lower than for the 22 mm KT steel bolts at 1 MPa confinement but considerably better at 5 MPa confinement. Again the embedment material was much less influential and high residual loads were achieved with the PUR resin

7.3.7 Discussion of maximum loads

Figure 7.2 shows a graph of the mean maximum loads achieved with the different tendons, in the various grouts and at the two confining pressures with respect to increasing (nominal, i.e. bit) hole size. Increasing hole size means, of course, increasing bolt diameter except in the case of the use of CBG grout. In these cases all holes were drilled with a 43mm bit as the installation of CBG grouted tendons requires an additional grout tube in the hole. The unreliable results with PU grout and GRP bolts have been omitted from the analysis.

The graph shows that, at 1MPa confining pressure, there is little difference in the mean maximum loads achieved, irrespective of bolt type, grout type or hole size. The maximum loads for all five bolt types are only marginally higher in the 43mm diameter hole. At 5MPa confining pressure, however, the higher maximum loads are readily discernible. Both the 28mm steel bolt and the 28mm GRP bolt achieved higher maximum loads, over twice, at 5MPa confining pressure than at 1MPa confining pressure. The KT bolt recorded a higher maximum load in CBG grout (and therefore the largest diameter hole) than in other hole sizes, but otherwise the KT bolt, the Reflex bolt and the FT500 GRP bolt did not show much variation in maximum load with increasing hole size. Higher mean maximum loads were achieved with larger diameter bolts in larger diameter holes.

7.4 CONCLUSIONS AND RECOMMENDATIONS

The earlier performed tests with PUR capsule resin and GRP bolts gave relatively poor results. Later tests with steel bolts, where the resin components were carefully weighed and mixed to the manufacturer’s recommended proportions by weight outside the hole, gave much improved results. This may warrant repetition of the PUR tests with GRP bolts.

However, the experience of mixing capsule PUR in the laboratory indicated that it was likely to be problematic to achieve satisfactory mixing underground. The fact that the proportion of components, vigour of mixing, application of confinement and exclusion of water were all critical to the good results achieved with PUR and steel bolts mitigates against PUR being a practical reinforcement encapsulant. The presence of water causes the material to foam with consequently very poor reinforcement performance. Loss of relatively small amounts of either one of the components appears to result in failure of the PUR to gel properly and absence of confinement, even without the presence of water, was found to result in foaming on some occasions.

Generally the GRP bolts produced significantly more pronounced post failure stick / slip behaviour than the steel bolts. It is not clear whether this is an undesirable behaviour.

60

GRP bolts generally achieved lower or similar residual loads and other performance parameters when compared with the equivalent steel bolts, encapsulant and confinement, a significant exception being the 28 mm GRP bolt in AT resin at 5 MPa confinement. As GRP bolts are known to be very weak in bending, it is recommended that steel bolts are used for rib reinforcement wherever possible.

Where GRP bolts are used then encapsulation in AT resin is to be preferred, where this can be achieved to good standards and installation is at the face of the heading. Elsewhere, loss of AT resin into rib breaks may mean that good installation with this encapsulant is not possible.

However, the achievement of better than 10 tonnes residual load per metre of encapsulation (as indicated necessary by previous modelling) for the 28 mm GRP bolts with CBG grout at both confinements, indicates that they can provide significant reinforcement to the coal ribs at colliery A. This was only achieved by the 24 mm GRP bolts at the higher confinement. It should be noted however that the CBG grout was tested after 14 days compared with 1 day for the AT resin. The reinforcement provided by CBG will be highly time dependent over the first 14 days.

The Reflex bolt performance was mixed when compared with the equivalent diameter solid steel bolt, with some performance parameters being better and some worse. Its residual loads were worse at lower confinement and better at higher confinement. However in relation to the 28 mm Big steel bolts, the Reflex bolt performed generally less well on most measures, though residual loads at 5 MPa confinement were still a little higher. On balance, as Reflex bolts have lower elongation to failure than steel bolts and are more problematic to install, it would seem advisable to use Big steel bolts in the ribside where practicable.

The similar performances achieved by both steel bolts and Reflex bolts in different encapsulants indicates that whichever encapsulant is likely to produce the best quality embedment from a practical viewpoint is to be preferred. The low rate of strength development with time achieved by CBG should also be considered. All tests with CBG grout were undertaken in 43 mm diameter holes in order to provide sufficient annulus for a grout tube.

This work confirms that AT resin is suitable for application at the face of the heading where immediate reinforcement is required and that CBG is more likely to be appropriate for application outbye as remedial or secondary support where the lower rate of strength increase can be accommodated by additional temporary support measures where necessary and loss of AT resin into breaks during installation is likely to be more severe.

61

8 SUGGESTED REVISION OF DMCIAC CABLEBOLTING GUIDANCE DOCUMENT

8.1 INTRODUCTION

Guidance on the use of cablebolts in coal mine roadways was issued by HM Inspectorate in 1996 (reference 5). Since then, as has been amply illustrated earlier in this document, cablebolting technology and methodology has moved on, for example with respect to long tendon system design, and use of pretensioning.

A comprehensive revision of the DMCIAC document has been drafted under this Project – incorporating latest practices and systems. The draft revised guidance includes installation of long tendons in ribsides as well as roof, but does not include flexible bolts which, although included in the revision of the British Standard for cablebolting materials, are the subject of separate DMCIAC guidance on use.

The revised guidance document is intended for introduction following official adoption of the revised BS7861 Part2 currently being prepared.

A full text of the suggested revision of the guidance document is given below.

8.2 SUGGESTED REVISION OF DMCIAC CABLEBOLTING GUIDANCE

Guidance on the use of cablebolts to support roadways in coal mines

Introduction

1 The introduction of new designs of cablebolt, largely replacing previously used birdcaged types, has prompted a review of the guidance on use of cablebolts to support roadways in coal mines originally published in 1996. This revised guidance is appropriate to current use of cablebolts, and includes a section on use of tensionable tendons which are now being used in some roadways, and which require particular attention in use, especially with regard to grouting using thixotropic grouts.

Application

2 This guidance applies to situations where cablebolts are installed as additional support when excessive strata movement is experienced or expected in places principally supported by rockbolts.

3 Operational necessity usually requires that cablebolts are installed outbye of the face, often weeks or months after roadway excavation. Recent developments in cablebolt design, however, have allowed installation at or near the face. Even in these circumstances, the cablebolts must be regarded as additional support, since they cannot achieve optimum performance immediately.

Definitions

4 The following definitions apply throughout this guidance: Cablebolt: a flexible tendon comprising one or more steel strands, or a group of steel wires, installed using a pumped fully encapsulating grout, to provide reinforcement of a mine roadway roof or side. For the purposes of this document a flexible bolt is not a cablebolt as it is installed

63

in a single operation through encapsulating resin capsules and provides immediate reinforcement in a similar manner to a rockbolt. Flexible bolts are covered by a separate supplementary document to the guidance on use of rockbolts to support roadways in coal mines (NB definition of flexible bolt in flexible bolt guidance needs revising to bring into line)

Tensionable cablebolt: a cablebolt, as above, designed to be installed with a short upper anchorage with rapid strength development so that a tensioning load can be applied prior to full encapsulation with a pumped grout.

Rockbolt: a bar inserted into the roof or side of a roadway which is used in conjunction with fully encapsulating resin or some other appropriate substance to provide reinforcement of the roof and sides of a roadway or working place in a mine.

Rockbolted heading/roadway: a heading/roadway in which rockbolts provide the principal means of support.

Site Investigation and cablebolting scheme design

5 The Approved Code of Practice – The control of ground movement in mines (reference 1), hereinafter referred to as ‘ACOP’, requires a detailed technical analysis of ground conditions and the preparation of a design document, where rockbolting or any other support system is being considered for principal support in coal mines. The ACOP lists the factors to be taken into account, and the contents of the design document. The work required can be summarised as an investigation of site conditions and an assessment of geotechnical factors which affect support system design for that site. The ACOP also requires that the assessment of ground conditions and design document be reviewed if there is a reason to suspect a material change in the matters to which they relate. For rockbolted roadways this includes response to the results of the manager’s monitoring scheme

6 In accordance with the ACOP, the manager would carry out these functions, if suitably qualified, or appoint a competent person or persons to carry out these and other functions defined by this guidance document.

7 Where the site investigation, geotechnical assessment and/or monitoring indicate that the strata require the use of cablebolts, a support design needs to be prepared. Where an existing design has already been proven, reference to it may be made for other roadways of similar dimensions, in the same seam, provided that suitable and sufficient steps are taken to show that the geological conditions, rock properties, stress fields and monitoring results at both sites are substantially similar.

8 Where cablebolts are to form part of the systematic support, the initial design of the cablebolting system should be prepared on the basis of the results of the site investigation. As a minimum, the design needs to take account of the following: • the profile of the heading; • the length and type of cablebolts and any associated equipment to be used in the roof

and ribs; • the density and pattern of cablebolts in the roof; • the distance of installation from the face of the heading.

9 When the initial design has been completed, documentation needs to be prepared detailing: . the use of the roadway/junction; . the free-standing supplementary support if applicable;

64

• the layout and dimensions of the cablebolting pattern; • the specification of cablebolting materials to be used; • the method of work; • the design verification monitoring system.

10 The design documentation will be incorporated into the Design Document for rockbolted support required by the ACOP.

11 Where cablebolts are used as a remedial support system, in response to monitoring or where previous experience has shown that roof movement is occurring or likely to occur above the rockbolted height, their length, position and density should be determined by the mine manager or a competent person and incorporated into a revised design document. The design should take into account location, mechanism and degree of roof movement.

Cablebolt specification

12 Cablebolts used in coal mines for reinforcement of the roof must comply with the requirements of the British Standard for strata reinforcement support system components, part two, (ref 2).

13 Cablebolts used in coal mines for reinforcement of the roof should normally have a minimum length of 8 m when used in roadways less than 5 m wide, and 10 m when used in other applications, unless monitoring and geotechnical information indicates otherwise.

14 Holes drilled for cablebolts will be formed in such a way as to optimise system bond strength. Suppliers will advise on the most suitable hole diameter and drill bits to be used.

15 Cablebolts will not be considered fit for purpose unless and until full column grouted.

16 Where cementitious grouts are used, the supplier will specify the water to solids ratio of the mixed grout. The mine manager or competent person should ensure that means are available at the workplace to allow accurate measurement of water volume, and that operators understand the importance of ensuring that the correct water to solids ratio is used, and that ‘watering down’ of the grout is not acceptable.

17 The mine manager or competent person should put in place a sampling procedure to monitor the properties of the mixed grout, and should specify sampling frequency. See Appendix 1. Failure to achieve the required properties may necessitate additional cablebolts being installed.

18 Where possible, sufficient grout to fully encapsulate a cablebolt installation should be mixed in one operation.

19 Where thixotropic grouts are used, grouting must be carried out in one continuous operation, and care should be taken to ensure sufficient grout is mixed to allow this. Experience shows that thixotropic grouts can be difficult to place, and the greatest vigilance is required to ensure that the required grout properties are achieved.

20 All cablebolts installed in the roof should be installed as near to vertical as is practicable, unless the design specifies otherwise.

21 All cablebolts shall be fitted with an end termination capable of transferring load to the surrounding rock surface, unless they are to be connected to other cablebolts. End fittings, and

65

couplings used for cable connection, must comply with the requirements of the British Standard for strata reinforcement support system components, part two, (ref 2).

Tensionable Cablebolts

22 Tensionable cablebolts are installed using an initial point anchorage with rapid strength development to allow a tensioning load to be applied. A tensioning load is applied using a tensioning jack or other means of tensioning approved by the tendon supplier, prior to full grout encapsulation of the remaining cable length. The cable supplier shall provide technical data on suitable tensioning equipment, safe tensioning procedures, and the maximum tensioning load which may safely be applied.

23 Tensionable cablebolts used in coal mines for reinforcement of the roof must comply with requirements of the British Standard for strata reinforcement support system components, part two, (ref 2).

24 Tensionable cables will normally be installed via a machine which will thrust the cable to the back of the hole through an encapsulating material such as resin. Particular care should be taken to ensure that the correct hole depth is drilled in order that

a) the required encapsulation length is achieved, and b) the proximal end of the cable is properly located to allow safe and efficient tensioning.

25 Where encapsulated resin is used for the initial point anchorage of a tensionable cablebolt used for reinforcement of the roof, this must comply with the performance requirements specified in the British Standard for strata reinforcement support system components, part two, (ref 2). The gel and setting times should be appropriate to the installation time required for the cablebolt used.

26 Resin capsule requirement should be determined from a calculation of the safe cablebolt encapsulation length. .The bond length requirement is given by the formula: Minimum bond length (mm) = numerical equivalent of [3 x cable ultimate tensile strength (kN)]

27 Tensionable cables should be fully grouted as soon as possible after installation, and within a maximum period of 24 hours.

Monitoring

28 The ACOP makes reference to the necessity for a monitoring scheme based on a comprehensive system of instrumentation and trend analysis, where rockbolts are used as principal support. Where cablebolts are to be installed systematically as part of the principal rockbolting support system, the manager or competent person should prepare a monitoring scheme which takes account of this. Where cablebolts are to be installed as remedial support, an addendum to the scheme will be required.

29 The scheme or addendum should set out the manager’s requirements for: • the procedures for the auditing of routine monitoring devices where cablebolts • form part of the support system; • the equipment to be used; • the duties of individuals; • plans, schedules and reports;

66

• the maximum levels of movement allowable on the monitoring devices before action is required;

• the action to be taken and the person responsible for taking the action.

30 The scheme or addendum needs to recognise: • the need to monitor and report physical changes affecting the security of the support

system; • the need to take remedial measures.

Routine monitoring devices

31 Cablebolt dual height tell-tales are normally used for the routine monitoring of the roofs of roadways where cablebolts form part of the support system. Where cablebolts are to be installed systematically, and close to the face, it may be appropriate to combine rockbolt and cablebolt tell-tales using triple height units as described in the guidance for use of flexible bolts document (ref 3). The construction, installation procedures and method of reading of cablebolt tell-tales are shown in Appendices 2 and 3 below.

32 Tell-tales need to be installed: • to at least the height of the cablebolt length + 1 m; • at intervals not greater than 20 metres; • as near vertical as practicable and sited as close to the centre of the roadway, and • as soon after cablebolting as practicable, or as directed by a competent person.

33 Tell-tales can also be set at increased frequencies by persons working on site or through supervisory or managerial instruction.

34 All other arrangements for monitoring should be the same as for rockbolts. Alternative arrangements for monitoring roadway sides reinforced by cable bolts may be appropriate.

Training (See Appendix 4)

35 All personnel involved with the installation of cablebolts should have received appropriate operational and safety training, and be duly authorised.

36 Management and officials/supervisors should have general training on the action of cablebolts, correct installation techniques, monitoring arrangements and testing procedures. They should be aware of the need for routine grout sampling, and the need to implement and maintain such a procedure. Maintenance of drilling and mixing/pumping equipment is of great importance and training to appropriate personnel must be given.

37 Operators should receive training to ensure that they are competent in the use of the machinery and materials to be used and the procedures to be followed when installing cablebolts. Emphasis needs to be given to maintaining satisfactory standards at all times, particularly with regard to correct grout mixing and full column encapsulation.

Cablebolting materials

38 All cablebolting reinforcement materials forming part of the roadway support system must be suitable for the purpose for which they are to be used.

39 A material intended for cablebolt roof reinforcement shall be regarded as suitable if:

67

a) it has previously received an acceptance number under the British Coal Corporation’s procedures for the Acceptance of Strata Reinforcement Materials and Equipment between 1 April 1992 and 30 June 1995; or

b) it can be shown, by means of independently conducted and assessed type testing, to meet the criteria set out in the British Standard for strata reinforcement support system components used in coal mines, Part 2 (ref 2).

40 A material intended for cablebolt reinforcement of the side of a roadway shall be regarded as suitable if:

a) it has previously received an acceptance number under the British Coal Corporation’s procedures for the Acceptance of Strata Reinforcement Materials and Equipment between 1 April 1992 and 30 June 1995; or

b) it can be shown, by means of independently conducted and assessed type testing, to meet the criteria set out in the British Standard for strata reinforcement support system components used in coal mines, Part 2 (ref 2) or,

c) if it does not meet the above criteria, it is: i) subjected to an independently conducted and assessed laboratory test programme, designed to simulate as accurately as possible, conditions of use, and ii) independently risk assessed with respect to safety and/or health, with due regard to conditions of intended use, and iii) the results of the laboratory test programme and the risk assessment are found to be satisfactory

41 All material should be subjected to carefully controlled field trial before being accepted for general use.

Installation equipment

42 Drill rigs, grout mixers, grout pumps, tensioning equipment, and associated cablebolt installation equipment must be suitable for purpose, and independently assessed in terms of performance, and safety in use. Where appropriate, testing should be carried out according to procedures given in the British Standard for strata reinforcement support system components used in coal mines, Part 2 (ref 2).

43 Equipment suppliers shall provide evidence via documentation of such assessment, full instructions as to use, maintenance and health and safety, and be prepared to provide training to operators.

44 Mine management should ensure: a) where compressed air operated equipment is used, that the compressed air supply is adequate (with respect to pressure and volume flow rate) for its correct operation and suitably filtered, b) that mine water supplies are adequate, and of suitable quality, and c) that maintenance procedures are in place to ensure that equipment is maintained correctly and at suitable intervals, in accordance with manufacturers’ instructions.

References

68

1 Approved code of practice and guidance. The control of ground movement in mines, the mines (control of ground movement) regulations 1999. HSE Books.

2 British Standard BS7861:rev. Strata reinforcement support system components used in coal mines. Part 2. Specification for cablebolting. British Standards Institution, 389 Chiswick High Road, London W4 4AL.

3 Supplementary guidance on the use of flexible bolts in reinforcement systems for coal mines . Mines04 pdf. HSE website. May 2007.

4 British Standard BS EN ISO 7500-1: 2004 Metallic materials. Verification of static uniaxial testing machines. Tension/compression testing machines. Verification and calibration of the force-measuring system. British Standards Institution, 389 Chiswick High Road, London W4 4AL.

5 British Standard BS 6319-2: 1983 Testing of resin compositions for use in construction. Method for measurement of compressive strength. British Standards Institution, 389 Chiswick High Road, London W4 4AL.

69

8.3 REVISED CABLEBOLTING GUIDANCE – APPENDIX 1

Appendix 1 Sampling and testing of grout mixed underground

Principle A1.1 Cementitious grout is prepared underground by mixing a known weight of powder with a specified volume of water in an approved mixing unit. With a cablebolt installed in a pre-drilled hole, the mixed grout is pumped into the hole to fully encapsulate the cablebolt.

A1.2 It is vital to achieve the correct water to solids ratio if optimum grout strength development is to be achieved. Experience has shown that, frequently, grout quality is poor due to over –watering, usually because difficulties in grout pumping are encountered. It is very important therefore that the grout quality is monitored by sampling and testing on a regular basis according to the Manager’s sampling procedure.

Preparation of test specimens A1.3 The samples should be collected in plastic bottles having nominal internal dimensions of 57 mm diameter x 100 mm deep (250 ml capacity).

A1.4 After mixing the grout to the manufacturer’s instructions, pump a sample into a plastic bottle of the size specified, sufficient to completely fill the bottle. At least three specimens should be prepared in this way.

A1.5 The samples should be clearly labelled with the date sampled, name of the mine, district/heading location, and type of grout.

A1.6 Store the samples upright in a secure place underground for 24 hours.

Testing A1.7 The manager or an appointed person will decide whether the samples should be tested at the mine (i.e. weighed for density determination) or sent to a laboratory for density and strength measurement.

A1.8 Frequency of sampling should be carried out according to the Manager’s instructions.

A1.9 The grout supplier will provide the following information to allow verification of grout quality:

a) a chart showing the variation of filled bottle weight with density, and indicating the range of acceptable weight / density, at 2, 3, 5 and 7 days curing, and

b) a table indicating the acceptable range of density and unconfined compressive strength at 28 days curing for samples collected in bottles.

Procedure 1 - Testing at the mine (density measurement) A1.10 Remove the samples from the mine after 24 hours and test as soon as possible.

A1.11 Weigh the bottle, cap and grout (filled to the brim). Check this weight against the chart provided by the manufacturer. If this comes within the range indicated on the chart, the grout is being mixed correctly.

A1.12 If the bottle is not completely full of grout, the following procedure should be used:

i) weigh the bottle, cap and grout, and record the weight.

70

ii) top up the bottle with water, replace the cap and record the weight. iii) take the difference between the two weights, double it and add to the original

weight of bottle, cap and grout. iv) consult the manufacturer’s chart, as in A.1.11 above.

An example calculation is as follows: Weight of cap, bottle and grout 600 grams Weight of cap, bottle, grout and water 610 grams Weight of water 10 grams Water weight corrected to grout weight = (10 grams) x 2 20 grams Corrected weight of full bottle = 600 + 20 = 620 grams

Compare the result with the manufacturers’ chart as in A.11 above.

Procedure 2 – Laboratory testing (density and strength measurement) A1.13 Samples should be tested after curing for 28 days. Samples received at the laboratory prior to this time should remain sealed in the bottle and stored at a temperature of 20 deg C +/1 deg C until required for testing.

A1.14 Remove the bottle from around the sample, and prepare the ends to give a cylinder 90 mm long with parallel faces.

A1.15 Weigh and measure the sample length, and diameter. Calculate and record density (i.e. weight of sample (gm) / volume of sample (cm3)) to a resolution of 0.1 gm/cm3.

A1.16 Use a testing machine calibrated to BS EN ISO 7500-1:2004 (ref 4) with a suitable capacity and load rate capability. When spacing blocks are used between the platens and test specimen, the requirements of BS 6319-2:1983 (ref 5) apply.

A1.17 Wipe clean the bearing surfaces of the testing machine and of any auxiliary platens. Remove any loose grit or other material from the surfaces of the test specimen that are to be in contact with the compression platens. Place the test sample on the lower machine platen and carefully centre. Load should be applied to the prepared parallel faces. Do not use packing at any of the interfaces between the test specimen, auxiliary platens, spacing blocks and machine.

A1.18 Apply load (without shock) and increase it continuously at a nominal rate of 45 N/mm2/min until no greater load can be sustained. Record the maximum load applied to the specimen.

A1.19 The uniaxial compressive strength (UCS) of each laboratory sample is calculated as follows: • UCS (N/mm2) = Maximum load (N) /Original cross sectional area (mm2) - equation 1 • Calculate values of UCS to the nearest 0.1 N/mm2. • If the height/diameter ratio (h/d) of the prepared sample is other than 2:1, then a

correction should be applied as indicated below: o Corrected UCS N/mm2 = UCS (equation 1) / (0.304 x d/h + 0.848)

where d = diameter of the sample (mm) h = height of the sample (mm)

A1.20 Compare the results with the manufacturers’ supplied data.

71


Appendix 2 Cablebolt dual height tell-tale

Introduction A2.1 This device is designed to be installed in vertical holes as soon as practicable following the installation of long tendon reinforcement, by purpose trained personnel. The general assembly of a water diverting type, which is designed to reduce corrosion of the tell-tale indicators, is shown in Figure 8.1.

Installation (water diverting type) A2.2 Drill hole using recommended bit to 1 m plus above reinforcement height or 6 m, whichever is the greater. e.g. 11 m for 10 m cablebolts

6 m " 4 m " "

A2.3 Insert top anchor, attached to smallest indicator “B” to top of hole. Use graduated purpose-designed insertion rods to confirm anchor position. Ensure the water seal sleeves are not pushed up the hole. Tug wire to seat anchor and check for firm anchorage.

A2.4 Position lower anchor attached to larger indicator “A”, at 1 m below the top of the reinforcement height using graduated purpose designed insertion rods, but not beyond.

e.g. 9 m for 10 m cablebolts 3 m “ 4 m “ “

A2.5 Ensure the water seal sleeves are fully located to the top of the reference tube.

A2.6 Keeping the suspension cables under tension, the reference tube can now be inserted into the tell-tale hole. The reference tube should be pushed fully into the hole to ensure that the stabilising fins locate against the hole mouth.

A2.7 Position indicator “A”, zero line (top of white band) to be level with bottom of reference tube. Align to scale. Crimp ferrule in position using crimping pliers.

A2.8 Position indicator “B”, zero line (top of white band) to be level with bottom of indicator “A”. Align to scale. Crimp ferrule in position using crimping pliers.

A2.9 Record details. At all tell-tale sites, a sign must be placed bearing a unique reference code for reporting and identification purposes giving the type of tell-tale, it’s position, date and time of installation, and anchor heights. This information should be passed to relevant officials.

Reading methods A2.10 By colour Report whole and part bands visible, for example: ‘A’ : white, blue, yellow ‘B’ : 3/4 white, blue, yellow See Figure 8.1

A2.11 By scale Report measurement, in millimetres, lining up with reference mark for each anchor. Reference for “A” is bottom of reference tube. Reference for “B” is bottom of indicator “A”.

72

Scale has millimetre divisions, with centimetre marks. For example: “A” : = 12 mm; “B” : = 31 mm; total = “A” + “B” = 43 mm

Interpretation A2.12 1.Movement of “A” relative to its reference (bottom of plastic tube) represents the strata expansion within the reinforced height.

A2.13 Movement of “B” relative to its reference (bottom of “A” represents the strata expansion above the reinforced height (assuming no movement above top anchor).

A2.14 The total strata expansion is “A” plus “B”.

A2.15 Expansion of strata above the top anchor is not detected.

73


Appendix 3 Cablebolt triple height tell-tale

Introduction A3.1 This device is designed to be installed where cablebolts are installed systematically as part of the rockbolting support system. It should be installed in vertical holes as close to the face as possible immediately after grouting of the cablebolts, and by purpose trained personnel. The general assembly of a water diverting type, which is designed to reduce corrosion of the tell-tale indicators, is shown in Figure 8.2.

Installation (water diverting type) A3.2 Drill hole using the recommended drill bit to 1 m plus above longest reinforcement height or 6 m, whichever is the greater.

e.g. 7 m for 6 m cablebolt 6 m “ 4 m “ “

A3.3 Insert top anchor, attached to smallest indicator “C” to top of hole. Use purpose-designed, graduated, insertion rods to confirm anchor position. Ensure the water seal sleeves are not pushed up the hole. Tug wire to seat anchor and check for firm anchorage.

A3.4 Position mid anchor attached to middle indicator “B” 1m below the top of the longest reinforcement height using purpose graduated insertion rods, but not beyond. Tug wire to seat anchor and check for firm anchorage.

A3.5 Position lower anchor attached to thickest indicator “A” at 0.3 m below the top of the rockbolts e.g. 2.1 m for 2.4 m rockbolts, 1.5 m for 1.8 m rockbolts using purpose graduated insertion rods, but not beyond. Tug wire to seat anchor and check for firm anchorage.

A3.6 Ensure the cable guide sleeves are fully located to the top of the reference tube.

A3.7 Keeping the suspension cables under tension, the reference tube can now be inserted into the tell-tale hole. The reference tube should be pushed fully into the hole to ensure that the stabilising fins locate against the hole mouth.

A3.8 Position indicator “A”, zero line (top of green band) to be level with bottom of reference tube. Align to scale. Crimp ferrule in position using crimping pliers.

A3.9 Position indicator “B”, zero line (top of green band) to be level with bottom of indicator “A”. Align to scale. Crimp ferrule in position using crimping pliers.

A3.10 Position indicator “C”, zero line (top of green band) to be level with bottom of indicator “B”. Align to scale. Crimp ferrule in position using crimping pliers.

A3.11 Record details. At all tell-tale sites, a sign must be placed bearing a unique reference code for reporting and identification purposes giving the type of tell-tale, it’s position, date and time of installation, and anchor heights. This information should be passed to relevant officials.

Reading methods A3.12 By colour

Report whole and part bands visible, for example: “A” : green, yellow, red

74

“B” : ¾ green, yellow, red “C” : yellow, red

See Figure 8.2.

A3.13 By scale Report measurement, in millimetres, lining up with reference mark for each anchor. Reference for “A” is bottom of reference tube. Reference for “B” is bottom of indicator “A”. Reference for “C” is bottom of indicator “B”. Scale has millimetre divisions, with centimetre marks. For example: “A” : = 12 mm “B” : = 31 mm “C” : = 40 mm Total = “A” + “B” + “C” = 83 mm

Interpretation A3.14 Movement of “A” relative to its reference (bottom of plastic tube) represents the strata expansion within the rockbolt bolted height.

A3.15 Movement of “B” relative to it’s reference (bottom of “A”) represents strata expansion within the section of roof reinforced by the cablebolt only.

A3.16 Movement of “C” relative to its reference (bottom of “B”) represents the strata expansion above the reinforced height (assuming no movement above top anchor).

A3.17 The total strata expansion is “A” plus “B” plus “C”.

A3.18 Expansion of strata above the top anchor is not detected.

75


Appendix 4 Training

A4.1 Training courses must be provided for all personnel involved in or responsible for decisions as to whether cablebolting is to be carried out, and in implementation and monitoring of such activity. Content of these courses must include the following topics.

For managers A4.2 Provide an understanding of the forces present in the rock and the redistribution of these as a consequence of mining operations. A4.3 Illustrate the differences between passive support, rockbolting and cablebolting. A4.4 Explain the action of cablebolts in limiting roof movement and roadway deformation. A4.5 Highlight the adverse effects of poor installation standards and describe the Managers grout sampling scheme. A4.6 Provide an appreciation of monitoring techniques and the information obtained, together with details of the installation and inspection procedures for the telltale monitoring system and also the setting of action levels and associated actions. A4.7 Give guidance on the construction and implementation of the manager’s ‘Scheme for the routine monitoring of rockbolted roadways’. A4.8 Give instruction on the inspection of bolted roof and ribs for signs of excessive bolt loading or deterioration, and the action to be taken if these are discovered.

For rockbolting co-ordinators / engineers As for managers (above) plus A4.9 An introduction to rock mechanics principles as applied to cablebolting including such topics as stress, the strata, design of reinforcement systems, underground engineering, consumables and several detailed case histories, including site visits where possible. A4.10 Management of the manager’s ‘Scheme for the routine monitoring of rockbolted roadways’. A4.11 Installation, replacement and reading of all routine monitoring devices used at the mine. A4.12 Familiarisation and use of the appropriate computer software. A4.13 Setting of appropriate routine monitoring action levels for each area of the mine. A4.14 Setting of appropriate corresponding remedial action for action levels for each area of the mine. A4.15 Determination of appropriate measuring frequencies for routine monitoring devices within the mine. A4.16 Follow up on remedial actions to secure stability. A4.17 Formulation and updating of the ‘Schedule of measurement zones and measuring frequency’ and related measuring timetables. A4.18 Production of tell-tale checklists for officials.

For officials / supervisors A4.19 An appreciation of basic rock mechanics as applied in cablebolting. A4.20 Aspects of the mine monitoring system. A4.21 Action levels. A.4.22 Action, duties and responsibilities. A4.23 Remedial measures. A4.24 Follow up actions. A4.25 Communication links. A4.26 All aspects of the tell-tale checklist system at the mine. A4.27 Appreciation of tell-tales.

76

A4.28 The correct installation of tell-tales. A4.29 The replacement of tell-tales. A4.30 The identification of tell-tales. A4.31 Reading of tell-tales and appropriate action levels and the associated action to be taken. A4.32 Instruction on the inspection of bolted roof and ribs for signs of excessive bolt loading or deterioration and the actions to be taken if these are discovered. A4.33 Appropriate types of extra support to secure the roof in adverse conditions. A4.34 Highlight the adverse effects of poor installation standards and describe the Managers grout sampling scheme.

For Operators A4.35 The action of cablebolts and typical cablebolt patterns, highlighting the importance of good installation practice. A4.36 Correct installation of cablebolts including adequate practical on-site training. A4.37 The sequence of operations and the time at which cablebolting is carried out. A4.38 Maintenance of the cablebolting equipment (drilling machines etc) to ensure that performance is maintained at designed levels. (Particular attention needs to be directed to ensuring provision of a sufficient supply of either hydraulic fluid, or compressed air (as appropriate) to allow the drilling equipment to operate within design parameters). A4.39 Provision of the correct length of drill-rods in an undamaged condition and arrangements to ensure that the correct depth of hole is drilled. A4.40 The type of grout together with the importance of the recommended mixing and pumping procedures. Highlight the adverse effects of poor installation standards and describe the Managers grout sampling scheme. A4.41 An appreciation of the manager’s ‘Scheme for the routine monitoring of rockbolted roadways’, the information indicated by means of the tell-tale monitoring system and action, where appropriate. A4.42 An instruction that operators, in the event of difficulty in the application of cablebolting and monitoring, need to bring those matters to the attention of those having statutory responsibility for the supervision of operations. A4.43 Personal protective equipment.

A4.44 Experience has shown that particular attention needs to be paid by all personnel involved in cablebolting to the following, and any training course should stress their particular importance: • that grout mixers and pumps are approved for use, well maintained and unmodified

from the manufacturers’ specification • that air supplies (in the case of compressed air equipment) are adequate in terms of

delivery pressure, flow rate and are filtered • that hosing is in good condition, of adequate bore and of minimal length conducive to

carrying out the task safely • that operators are aware of the grout bag weight and the necessary water volume

required for that weight • that means is available to accurately apportion the required water volume • that the correct mixing technique is used, and mixing time is as required • that grouting of a cablebolt is carried out in one continuous operation (implying that

sufficient grout should be mixed to achieve this) • that the mixed grout is sampled in accordance with requirements (Manager’s scheme),

and the samples stored properly • that, during grouting, the operators ensure that a flow of air bubbles followed by

cessation and show of grout (if possible) occur in the case of bottom-up grouting, and that a flow of grout at the bottom of the hole occurs in the case of top-down grouting.

77

A4.44 The above factors are important where any cablebolting operations take place, but become even more acute when thixotropic grouts are used - because the viscosity of the grout causes extra demands to be made on equipment and people.

78

9 CONCLUSIONS AND RECOMMENDATIONS

There is some evidence in the literature of field trials of tensioned reinforcement tendon systems indicating enhanced performance over untensioned systems, though conclusive proof has not been identified. The wider adoption of increased pretension loads provides further circumstantial evidence of advantages from tensioned reinforcement. Laboratory tests to date do not appear to have demonstrated any significant advantages, though effective laboratory simulation of insitu conditions is very difficult to achieve.

A mechanism by which higher pretension loads may improve support effectiveness has still to be confirmed. The most widely asserted explanation - that higher pretension increases shear resistance across joints and bedding planes - would not appear to have been demonstrated by laboratory studies reported in the literature. For encapsulated bolts crossing a shear plane these have consistently shown that an increased initial shear stiffness up to bolt yield is the only measurable effect of pretension. Even this effect would be limited to the zone of compression in the vicinity of the end plate.

An alternative explanation is that applying high pretension loads closes bed separations which develop prior to bolting, and therefore re-establishes frictional contact between beds in the immediate roof. This would provide a significant enhancement to the roof shear strength. Some modelling work suggests that bed separations do develop at weak interfaces and that the installation of tensioned bolts can close separations within the compressed zone-the higher the pretension, the higher into the roof that separations can be closed. The concept of buckling immediate roof beds, which can be stabilised by centrally placed pretensioned tendons, also involves the bed separation idea. Acknowledging the practical difficulties associated with underground investigations, it should still be possible to use field measurements to check the validity of these ideas by monitoring roof deformations at multiple points as tendons are installed and tensioned, but this does not appear to have been carried out to date. A field measurement exercise to measure this phenomenon carried out under this Project did not show any measureable re-closure at the site concerned. The identification of an underlying mechanism would be key to confirming that high pretension loads improve roof support.

The question of disadvantages associated with pretensioning has been assessed. The main one cited in available research work is reduction of available tendon capacity. Additional strain resulting from rock deformation will add directly to that imposed by the pretension load, so the tendon yield strain will be reached with less additional strain. Tendon failure could in theory also be expected at a slightly lower level of roof movement. However by far the major proportion of total tendon strain to failure occurs post yield, and this is unaffected by pretensioning. So in deforming roof conditions in which tendons are strained beyond yield this disadvantage is unlikely to be important.

It follows from this that it is inadvisable to use pretensioned bolts for lifting loads as the load capacity is reduced both by the pretension load and by any subsequent strata loading. The latter load is unknown unless instrumented bolts are used.

There are other possible drawbacks. Progressive debonding is one. There is at least a theoretical possibility of failing the bond through this phenomenon, by application of higher pretension loads, or subsequent additional loading, and this needs to be considered.

A reduction in rock confining stress above the anchor position is another potential drawback. The pretension load generates a tensile reaction at the top of the tendon anchor length. In

79

practice this should only slightly reduce the compressive stresses acting at this position due to the in-situ stress field, but it could initiate instability in a marginal case.

It can be concluded that tensioned tendons should be post grouted, for the following reasons: i. It locks in the pretension load, minimising subsequent loss.

ii. It ensures the tendon acts as an effective stiff support even without any benefit from the pretension itself - the “belt and braces” principle.

iii. It reduces moisture access to the rock through the annular space-this can cause long term deterioration in weaker mudstones (Unrug et al 2004).

iv. It reduces potential corrosion of the tendon from moisture and salts emanating from the rock.

v. It reduces the risk of injury to personnel through violent failure of the tendon or end fittings.

Post grouting could also provide a precaution against progressive debonding, should this prove necessary.

It seems to be generally assumed that post grouting locks in the pretension load, preventing subsequent tension loss, although this does not appear to have been verified by test work, either in the laboratory or in-situ. Some laboratory work suggests that the initial load distribution may change with time. However, pretensioned bolts are typically used in actively deforming roof where additional loading rapidly develops, so longer term load change or loss is usually not an issue.

It follows that the optimum tensioned bolting system for this application combines the facility to pretension and post grout, with a high bond strength and stiffness in both the anchor and post grouted lengths. There is however insufficient information to reach definitive conclusions on the optimum pretension load and tendon length.

In order to fully resolve the considerable uncertainties which still surround the practice of pretensioning, the following investigation programme is recommended:

i. The question of the mechanism(s) by which higher pretension loads may improve support effectiveness needs to be resolved. Field measurement, supplemented by numerical modelling, appears to be the appropriate method in this case. Field studies should seek to identify in detail the pattern of deformation of roadway roof and, in particular, confirm the development of bed separations (as opposed to bed dilation) prior to support, and establish if these are reduced or closed by the installation of pretensioned support. Field measurements carried out at colliery A, and presented in this report indicated no significant re-closure of roof dilation during tensioning to 100kN, but this needs further confirmation. Numerical modelling should be used to simulate the field situation and add parametric studies. It is likely that a number of field sites will be needed to obtain a representative range of conditions.

ii. The detailed mechanics of tendon load transfer do not seem to be fully known, and this adds to the difficulty in resolving the pretension question. Laboratory studies of rock deformation and tendon behaviour under more realistic loading conditions should be carried out to investigate load distributions in bonded tendons. These should make use of findings regarding in-situ rock deformation processes obtained in the field studies. In particular the difference between load distributions associated with tensioning the bolt and those induced by rock deformation within the bonded length needs to be explained, and if necessary allowed for in the experimental

80

procedures. Factors to be investigated should include pre and post yield interface shear strengths and stiffnesses, the progressive debonding process, tension losses and the effect of post grouting.

iii. The use of 3D modelling adequately simulating the tendon, encapsulant and rock, is considered essential to complete the study of both the load transfer and pretensioning processes. The shear behaviour of the resin/rock and tendon/resin interfaces, especially at large strains, is an important factor and laboratory measurements should be used to obtain the necessary data to allow simulation of this behaviour.

NOTE: Some of the work recommended above was expected to be carried out during the lifetime of this project. However, the need to concentrate on practical aspects of safely and effectively designing, installing and using tensioned tendons has resulted in a lack of time available to address the complex laboratory testing and numerical modelling required to fully understand the mechanisms involved.

It was found that a considerable amount of project time needed to be allocated to the issue of post grouting of Megastrands, and many laboratory tests were carried out. It must be concluded from this work that, in the form tested during this Project, the Megastrand cannot be reliably post grouted using thixotropic grouts and mixing / pumping equipment currently approved for the UK coal mining industry. This is probably a function of the small internal diameter (10mm) of the central grout tube in the tested Megastrands. Operators using this system, and available grout and equipment, are almost certain to ‘water down’ the grout to enable encapsulation, resulting in a weakened grout, and therefore compromised system performance.

Studies of field samples of cementitious grouts, both thixotropic and non-thixotropic, have shown that, in the main, grout pumped into tendon installations does not meet the required density, and, it follows, strength. It is recommended that guidance on use of cablebolts is updated, partly with a view to stressing and reinforcing the need to mix grout to the required water to solids ratio. A suggested, updated guidance document is included in this report.

A laboratory study of the effect on tendon bond performance of varying grout strength showed increasing bond performance with grout strength, as would be expected, although the bond strength individual values tend to plateau above a grout UCS of 60 MPa. In view of this and the fact that field measurements indicate that the existing requirement for strength is not being met, it is difficult to justify a relaxation in the grout strength requirement specified in the existing Standard for cablebolting materials.

A study of available monitoring data from sites where tensioned tendons were installed in UK mines, suggests that use of these systems was at least partially successful. However, installation of tensioned systems at the development face with subsequent grouting dependent on tell tale action levels appears to have been a step too far, particularly in the light of subsequent investigation revealing the difficulties associated with grouting these systems with available equipment. It is essential that post grouted tendons are grouted as early as is practical after installation and certainly no longer than 24 hours after installation (earlier in rapidly deforming ground).

The newly developed ability to underground data log strain gauged rockbolts, and the newly developed intrinsically safe remote reading extensometer system with a portable readout, allowed accurate and detailed measurements of the effects on roof behaviour of in-situ tensioning of a flexible tendon. Results from the highly sensitive extensometers used at colliery A showed that the Megastrand tensioning process to approximately 10 tonnes at this site did not

81

produce any re-closure of previous roof dilation measurable with the equipment used. This was the case even though the immediate roof had dilated by approximately 20 mm during the preceding 5 days. It can therefore be concluded that, previous reports of the roof being lifted during Megastrand tensioning most probably refer to relatively exceptional circumstances where very large displacements had already occurred. Also when roof lifting does occur it is likely only to apply to the very bottom roof strata where discrete bed separations and bolt debonding or bolt failure may have already occurred.

A method has been developed for modelling and comparing the behaviour of tensioned and untensioned flexible long tendon reinforcement when used as part of a coal mine support system in realistic geological and geotechnical conditions. For the particular examples modelled, both hypothetical and real, no benefit was seen from applying pre-tension to the flexible tendons. Indeed, the effect of pre-tensioning appeared to be to take the tendons beyond their yield load at an earlier stage in the loading process resulting in marginally higher roof displacements before stability was established. The modelling exercise also investigated the effect of not achieving full grout encapsulation of a non tensioned long tendon. In the particular conditions modelled, typical of UK coal mining, full encapsulation was shown to be critical for achieving stability, as would be expected due to the major loss of reinforcement stiffness resulting from non encapsulation. However, the limitations of the modelling should be noted. In particular the models used could not realistically simulate loss of shear resistance in the bedded strata due to bed relaxation prior to tendon installation and so could not simulate any benefit which may be derived from pre-tension in remobilisation of such shear resistance, a claimed benefit of tensioned tendons. Considerable further work is required to investigate this claimed benefit through modelling.

Considerable effort was directed to developing test procedures for establishing the performance, and installation characteristics of tensionable systems in the laboratory, and from this work, the project facilitated drafting of a revised British Standard, provided information valuable in development of a new section directed to ribside reinforcement materials, and provided a draft Guidance Document revision.

82

10 REFERENCES

1. British Standards Institute. Strata reinforcement support system components used in coal mines. Part 2. Specification for cable bolts. BS 7861-2:1997.

2. Rock Mechanics Technology Ltd. Testing and standards for rock reinforcement consumables. Final report, HSE project D5017. September 2005

3. British Standards Institute. Strata reinforcement support system components used in coal mines. Part 1. Specification for rockbolting. BS 7861-1:1996.

4. British Standards Institute. Strata reinforcement support system components used in coal mines. Part 1. Specification for rockbolting. BS 7861-1:2007.

5. Health and Safety Executive. Guidance on the use of cablebolts to support mine roadways. July 2007.

83

Anchor Load Distribution for Normal Anchor Tendon Bond Length

Bond Stress

OLoading

Initial Loading Ultimate Loading

Load Distribution along Fixed Anchor Fixed Anchor Length Generally 10m Max

Unit Anchor Tendons

Single Bore Multiple Anchor Load Distribution

Bond Stress

O

Load Distribution along Fixed Anchor

Figure 2.1 Comparison of the load distribution of a normal anchor and a single bore multiple anchor (after Barley and Windsor, 2000)

85

Roof Displacement

• Tensile bending • Roof measures sub-divide into discrete units • Buckling roof-high displacements

PP

F

Mechanical Advantage (MA)= Uv / Uh

Beam Equilibrium Condition : F = P/MA (ignoring the load bearing capacity of the beam itself)

At small values of Uv : F<<P for equilibrium

Figure 2.2 The buckling beam concept for roof stability (after Strata Engineering, 2001)

86

0.5Uh0.5Uh Uv

a) Bolt Pretensioned

T

C

Joint

Joint compression force C = Bolt pretension force T

C, T

Joint Compression Bolt Extension

Joint Stiffness >> Bolt Stiffness

b) External force Fe applied

C΄

Joint Compression Bolt Extension

Fe = C΄ + T΄ Bolt Tension = T + T΄ Joint Compression Force = C – C΄

T΄

Fe

Most of the external load is absorbed by reduction in joint compression force

Figure 2.3 External load applied to a pretensioned bolted joint

87

a) Bolt Stiffness >> Rock Stiffness

C, T Force

Rock compression Bolt extension

b) External force Fe applied

Fe = C΄ + T΄ Bolt Tension = T + T΄ Joint Compression Force = C – C΄

C, T

C΄

T΄Fe

Most of the external load is added to the bolt pretension

Figure 2.4 External load applied to a pretensioned lifting bolt

88

Figure 3.1 Laboratory short encapsulation pull test results for megastrands and CBG grout at 1 and 3 days curing.

89

Figure 3.2 Laboratory short encapsulation pull test results for megastrands and CBG grout at 7 and 42 days curing.

90

91

Figu

re 3

.3

Var

iatio

n of

max

imum

load

, bon

d st

reng

th a

nd s

yste

m s

tiffn

ess

with

grou

t unc

onfin

ed c

ompr

essi

ve s

treng

th (c

ube

sam

ples

).

92

Figu

re 3

.4 V

aria

tion

of u

ncon

fined

com

pres

sive

stre

ngth

with

den

sity

for b

ottle

sam

ples

of C

BG

gro

ut o

btai

ned

from

UK

coa

l min

es

93

Figu

re 3

.5 V

aria

tion

of u

ncon

fined

com

pres

sive

stre

ngth

with

den

sity

for b

ottle

sam

ples

of H

PR

G g

rout

obt

aine

d fro

m U

K c

oal m

ines

Figu

re 4

.1 S

chem

atic

of 1

0’s

mai

n ga

te –

col

liery

C

94

DS10’s new main gate station 4 @ 607m mark

Figure 4.2 Monitoring results for station 4, 10’s main gate, colliery C

Station 3 @ 423m ( 5 – wire roof exto) 10’s main gate


95

Station 2 @ 248m


96

Figu

re 4

.5 M

onito

ring

resu

lts fr

om ty

pe ‘B

’ tel

ltale

s, 1

0’s

mai

n ga

te, c

ollie

ry C

97

Figu

re 4

.6 M

onito

ring

resu

lts fr

om c

able

bolt

type

‘A’ t

ellta

les,

10’

s m

ain

gate

, col

liery

C

98

99

Reading (mm)

160

227

265

305

345

385

425

465

505

545

585

618

645

685

725

765

825

865

905

945

19 N

ov 0

1

03-D

ec-0

1

22-D

ec-0

1

14-J

an-0

2

11-F

eb-0

2

13-M

ar-0

2 25

Mar

02

050100

150

200

250

300

350

400

450

500

Har

wor

th 1

0's

Mai

n G

ate

Cab

le B

olt T

ype

'A' T

ell T

ales

19

Nov

01

27-N

ov-0

1

03-D

ec-0

1

17-D

ec-0

1

22-D

ec-0

1

05-J

an-0

2

14-J

an-0

2

23-J

an-0

2

11-F

eb-0

2

25-F

eb-0

2

13-M

ar-0

2

17-M

ar-0

2

25 M

ar 0

2

02-A

pr-0

2

Roa

dway

Pos

ition

(MM

)

Figu

re 4

.7 M

onito

ring

resu

lts fr

om c

able

bolt

type

‘A’ t

ellta

les,

10’

s m

ain

gate

, col

liery

C (a

ltern

ativ

e vi

ew)

Figu

re 4

.8 M

onito

ring

resu

lts fr

om t

ype

‘A’ t

ellta

les,

22’

s m

ain

gate

, col

liery

C

100

Figu

re 4

.9 C

ombi

ned

disp

lace

men

t fro

m ty

pe ‘a

’ and

‘B’ t

ellta

les,

570

, 590

and

610

mm

, 19’

s ta

il ga

te, c

ollie

ry C

101

Figure 4.10 Schematic of t18’s face line, colliery B

102

Figure 4.11 Planned support pattern in widened face line

103

Exte

nsom

eter

SG Bolt

SG Bolt

Figu

re 4

.12

Sec

tion

of w

iden

ed fa

ce li

ne s

uppo

rt an

d in

stru

men

tatio

n pa

ttern

104

TT 207MM approximately 3.5m from cut/dint

Cut Dint

TT 196MM

X6 R2 X5

X4 +s X3

X2 R1 X1

Cut and Dint

KEY

Pre Installed Mega Strands +s Sonic Extensometer

X1 Positions of Installed Strain Gauged Bolts R1 Remote Reading Telltale

Figure 4.13 Approximate position of installed instruments on 21 March 2007

105

Tail Gate

Machine approximately 4-5m beyond TT position.

X2

X6

X1

+s

R1

R2

2

4

3

87

5

1.2m 6

X4 X31

X5 TT 196MM

KEY

Installed Mega Strand and order of Tension +s Sonic Extensometer

Pre Installed Mega Strands R1 Remote Reading Telltale

X1 Positions of installed Strain Gauged Bolts

Figure 4.14 Approximate position of installed megastrands and tensioning on 27 March 2007

106

R

S

R

196mm (Telltale)

Planned Sonic Extensometer position Planned Megastrand positions S

Planned Remote Reading Exto positions Planned SG Bolt positions R

Figure 4.15 Proposed megastrand monitoring instrumentation 302’s faceline march 07

107

Figu

re 4

.16

Tim

e tre

nd fo

r cab

le b

olt t

ellta

le n

umbe

r 13

at 1

96m

108

Figu

re 4

.17

Col

liery

C 3

02’s

face

line

stat

ion

6 at

192

m

109

Figu

re 4

.18

Col

liery

C 3

02’s

face

197

m c

l 297

rrex

to 1

110

Figu

re 4

.19

Col

liery

C 3

02’s

face

197

m c

l 297

rrex

to 1

111

Figure 4.20 Sonic extensometer 196m, displacement & strain against distance in strata

112

Figu

re 4

.21

Son

ic e

xten

som

eter

196

m, d

ispl

acem

ent a

gain

st ti

me

113

114

Figu

re 4

.22

Son

ic e

xten

som

eter

196

m, d

ispl

acem

ent a

gain

st ti

me

durin

g m

egas

trand

tens

ioni

ng p

erio

d

Figure 4.23 Strain gauged rockbolt 1. mean microstrain and microstrain difference against

distance along bolt

115


distance along bolt

116


distance along bolt

117


distance along bolt

118


distance along bolt

119


distance along bolt

120

Siltstone

Coal

Mudstone

Mudstone

Figure 5.1 Modelled strata sequence 3.2 m roof mudstone

Figure 5.2a Coal strength properties

COAL

0

10

20

30

40

50

60

70

80

90

100

-5 0 5 10 15

Confining stress, MPa

Axi

al s

tress

, MP

a

Hoek-Brown

bilinear fit

Figure 5.2b Siltstone strength properties

SILTSTONE

0

10

20

30

40

50

60

70

80

90

100

-5 0 5 10 15


Axi

al s

tress

, MP

a

Hoek-Brown

bilinear fit

Figure 5.2c Mudstone strength properties

MUDSTONE

0

10

20

30

40

50

60

70

80

90

100

-5 0 5 10 15


Axi

al s

tress

, MP

a

Hoek-Brown

bilinear fit

121

Roof sag/bulking

Roof shortening

Fail bolts Excavate roadway

Figure 5.3 Modelled roof displacements with roof bolts failed and no additional support

Roof sag/bulking

Roof shortening

(a. Displacements)

Excavate roadway

Install flexible bolt tension to

200kN

Fail bolts

Flexible bolt tension

(b. Tension)

Figure 5.4 Modelled roof displacements with roof bolts failed and pre-tensioned flexible bolts

as additional support

122

Excavate roadway

Install flexible bolts

Fail bolts

Roof sag/bulking

Roof shortening

Flexible bolt tension

(b. Tension)

(a. Displacements)

Figure 5.5 Modelled roof displacements with roof bolts failed and untensioned flexible bolts as additional support

123

Excavate roadway

Install truss -Tension to 200kN

Fail bolts

Roof sag/bulking

Roof shortening

Truss tension

(b. Tension)

(a. Displacements)

Figure 5.6 Modelled roof displacements with roof bolts failed and tensioned truss system as additional support

124

Figure 5.7 Shear strains and bolt loads with pre-tensioned flexible bolts as additional support

Figure 5.8 Shear strains and bolt loads with pre-tensioned truss system as additional support

125

x x

x x

x x

Density Length

KT bolts 2.1 2.4m

Tensioned tendons x 0.28 6.0m

Flexible bolts + 0 -

Figure 5.9 Proposed bolt pattern at development face

Sequence - a Sequence - b 9

8

7 Silty mudstone (UCS 40MPa)

6

5 Mudstone (UCS 30MPa)

4

Silty mudstone (UCS 40MPa) Silty mudstone (UCS 40MPa) 3

2 Siltstone (UCS 50MPa) Siltstone (UCS 50MPa)

1

Silty mudstone (UCS 40MPa) Silty mudstone (UCS 40MPa) 0

-1 Coal seam

-2

-3

Coal seam

Figure 5.10 Strata sequences

126

0

1

2

3

4

5

6

7

Hei

ght a

bove

roof

, mm

12 16 20

Lateral stress MPa

ROOF EXTENSOMETER

0 20 40 60 80 100

Displacement, mm

Figure 5.11 Roof condition with proposed support and sequence A

127

ROOF EXTENSOMETER

0

1

2

3

4

5

6

7

12

16

20

Lateral stress MPa

Hei

ght a

bove

roof

, mm

0 20 40 60 80 100

Displacement, mm

Figure 5.12 Roof condition with proposed support and sequence B

128

140

Density Length

KT bolts 1.1 2.4m

Tensioned tendons x 0 -

Flexible bolts + 1.0 4/5/6m

+ + + +

+ + +

+ + + +

+ + + Figure 5.13 Alternative bolt + + + + patterns at development face

+ + +

+ + + +

TOTAL ROOF EXTENSOMETER MOVEMENT

10 12 14 16 18 20 22

Lateral stress, MPa

Mov

emen

t, m

m120

100

80

60

4+4, 2.4m KT @1.5m 4+3, 6m FB 40

20 4+4, 2.4m KT @1.5m 6m FB 2m free length

0

Figure 5.14 Roof movement with alternative support patterns

MOVEMENT ABOVE 2m (sequence A) 35

7+8, 2.4mKT @ 1.5m 2, 6m Megastrand

7+8, 2.4mKT @ 1.5m no long tendons

4+4, 2.4m KT @1.5m 4+3, 4m FB

10 12 14 16 18 20 22

Lateral stress, MPa

Mov

emen

t, m

m30

25

20

15

10 4+4, 2.4m KT @1.5m 4+3, 6m FB

5 4+4, 2.4m KT @1.5m 6m FB 2m free length

0

129


7+8, 2.4mKT @ 1.5m No long tendons

4+4, 2.4m KT @1.5m 4+3, 4m FB

BOLT STRAIN



4+4, 2.4m KT @1.5m 4+3, 4m FB

4+4, 2.4m KT @1.5m 4+3, 6m FB

4+4, 2.4m KT @1.5m 6m FB 2m free length

10 12 14 16 18 20 22

Lateral stress, MPa

CABLE STRAIN



4+4, 2.4m KT @1.5m 4+3, 4m FB

4+4, 2.4m KT @1.5m 4+3, 6m FB

4+4, 2.4m KT @1.5m 6m FB 2m free length

10 12 14 16 18 20 22

Lateral stress, MPa

Figure 5.15 Bolt strains with alternative support patterns (sequence A)

STRESS INCREASE

Vertical stress

Lateral stress – no stress notch

Lateral stress – slight Stress notch

0 1 2 3 4 5 6 7

STAGE

Lateral stress for development = 16MPa

Stage Face position

0 Development

1 120 – 160 m

2 50 – 100 m

3 30 – 50 m

4 20 – 30m

5 10 – 20m

6 <10m

7 higher than expected

Figure 5.16 Stress increase to represent face retreat

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

4.5%

5.0% Sd

trai

n

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

4.5%

5.0%

Stra

in

0%

50%

100%

150%

200%

130

ROOF EXTENSOMETER

0

1

2

3

4

5

6

7

8

0 20 40 60 80 100

Displacem ent, m m

Hei

ght a

bove

roof

, mm

0

1

2

3

4

5

6

7

Stage

Figure 5.17 Modelled roof condition for retreat

131

Stra

inM

ovem

ent,

mm

Stra

inM

ovem

ent,

mm

200

180 160 140 120 100 80 60 40 20 0

45

40

35

30

25

20

15

10

5

0

6%

5%

4%

3%

2%

1%

0%

6%

5%

4%

3%

2%

1%

0%

ROOF EXTENSOMETER MOVEMENT

0 1 2 3 4 5 6 7

Stage Figure 5.18 Roof movements for face retreat

MOVEMENT ABOVE 2m

0 1 2 3 4 5 6 7

Stage

BOLT STRAIN

0 1 2 3 4 5 6 7

Stage

CABLE STRAIN

0 1 2 3 4 5 6 7

Stage

Additional support for retreat

No additional support





Figure 5.19 Bolt strains for face retreat



132

ROOF LO W ERING

0

200

400

600

800

1000

1200

1400

0 100000 200000 300000 400000 500000

S olution ste ps

Mov

emen

t, m

m

No addit ional s upport ins talled Fail roof bolts and c ablesA ddit ional s upport ins talled - Fail roof bolts and c ablesA ddit ional s upport ins talled - Fail roof bolts and old c ables

Additional support installed – fail bolts and cables

-

ROOF LOWERING

0

200

400

600

800

1000

1200

1400

1600

0 100000 200000 300000 400000 500000

Solution steps

Mov

emen

t, m

m

No additional support installed - Fail roof bolts and cables Additional support installed - Fail roof bolts and old cables Additional support installed - Fail roof bolts and cables

Initial drivage

Increase stresses for face retreat

Fail bolts and cables

Unstable roof

Roof movements stabilise

Figure 5.20 Fail bolts and cables

133

80 70

68

AT re

sin

1 M

Pa60

56

56

50

C

BG g

rout

at 1

MPa

4850

46

26

9.5

17

22

22

33

17

27

PU re

sin

at 1

MPa

44

41

40

40

37

35 3

5 33

33

AT

resi

n at

5 M

Pa30

C

BG g

rout

at 5

MPa

22

20

20

PU re

sin

at 5

MPa

10

64

4 #N

/A

#N/A

#N

/A

0 W

eldg

rip G

RP

Wel

dgrip

GR

PKT

Ste

el 2

2mm

Bi

g Bo

lt St

eel

Osb

orn

Ref

lex

FT50

0 24

mm

Bi

g Bo

lt 28

mm

28

mm

St

eel S

trand

22

mm

Bol

t typ

e

Figu

re 7

.1 M

ean

resi

dual

load

s at

50

mm

dis

plac

emen

t

134

Mean residual load (kN)

050100

150

200

250

300

27 m

mho

le A

Tre

s in

27 m

mho

le P

Ure

s in

33 m

mho

le A

Tre

s in

43 m

mho

le C

B G

gro u

t

27 m

mho

le A

Tre

s i n

27 m

mho

le P

Ure

s i n

33m

mho

le A

Tre

s in

43 m

mho

le C

B G

gro u

t

1 M P

a C

o n

fin in

g P

re s s

u re

5 M

P a

C o

n fin

in g

P re

s s u

re

Me a n Ma x i m u m Loa d ( k N )

W e

ldg r

ip G

R P

F T 5

00 2

4 m m

KT

Ste

e l 2

2 mm

O s b

o rn

R e f

lex

S te

el S

tra nd

22 m

m

W e l

d g rip

G R

P B

ig B

o lt 2

8 mm

Big

Bo l

t Ste

e l 2

8 m m

135

Figu

re 7

.2 C

ompa

rison

of a

chie

ved

mea

n m

axim

um lo

ads

betw

een

diffe

rent

bol

t typ

es a

nd in

crea

sing

hol

e si

ze

Figure 8.1 Water diverting dual height tell-tale for cablebolting&(white, blue, yellow bands)&

Figure 8.2 Water diverting triple height tell-tale for cablebolting&(green yellow red bands)&

136)

APPENDIX 1 DRAFT TEST PROCEDURE GROUT ENCAPSULATION TEST – ‘BOTTOM UP’ GROUTING

A1.1 PRINCIPLE

This test is used to verify that a ‘bottom-up’ flexible reinforcement system, where grouting takes place from the proximal end of the tendon toward the distal end, can consistently achieve full encapsulation.

A1.2 APPARATUS

A suitable test arrangement is shown in Figure A1.1.

The overall length of the test tendon should be equal to the manufacturer’s maximum recommended length for vertical installation using standard pumping equipment. It should be installed in a clear rigid tube with an internal diameter equal to the manufacturer’s recommended installation hole diameter +/- 1mm. The tube should have a minimum wall thickness of 2mm. The tube may be a continuous length or be made up from several pieces provided the joints are made with made-for-purpose sockets and bonded with a suitable proprietary cement. One end of the tube will be capped with a made-for-purpose end fitting – again bonded with a suitable proprietary cement. The joints of the assembled tube must be air tight. It is important that the wall thickness, construction of the tube and all ancillary components are selected taking full account of the pressures that could be generated during the test.

Flexible breather and grouting tubes may be required and, if used, should conform to the recommendations of the tendon manufacturer.

The assembled test tendon and tube will need to be supported in a vertical position, and this may be via a structure attached to the side of a building, a scaffolding tower, a support pole erected into the vertical position via a pivoting bracket, or other suitable arrangement.

The mixing and pumping equipment should be selected from the range typically used and accepted for underground use in coal mines. If the manufacturer recommends a specific pumping and mixing system then this should be used. If pneumatically operated, an adequate supply of compressed air (consult the manufacturer’s literature) should be made available, and connected with suitable low-restriction hosing and couplings. A 25mm bore hose should be used for grouting. The hose should be no less than 10m long and fitted with a suitable connector at one end for connection to the pump, and a means of connection to the test tendon grouting tube at the other. The apparatus should include a T-piece arrangement at the pump outlet with associated valves, couplings and an additional hose to provide a means of relieving the pressure at the pump after grouting, and delivery of excess grout or flushings to a suitable container.

Sufficient grout for efficient mixing in the apparatus to be used, to fully encapsulate the test tendon, and to fill grout sample moulds, should be available.

137)

A1.3 PROCEDURE

A1.3.1 Preparation of the test tendon. Assemble the tube components and bond together. Fit the breather tube and grouting tube (if required) to the tendon and secure using pvc or other suitable tape. Ensure sufficient tubing extends beyond the proximal end of the tendon for immersion in water (breather) and connection to the grout hose. The test tendon can be anchored either at the top of the tube using, for example, mixed encapsulated resin, and/or at the bottom using a combination of a seal and clips/tape. The test assembly must be sealed at the proximal end and this can be affected using a surgical sock filled with pre-mixed grout – as is typically practiced underground – or some other means.

Erect the test assembly to the vertical using a means of support as described above. Ensure that the test tendon is straight to within 100mm of the notional vertical axis of the test assembly. The proximal end of the test tendon should be approximately 2.5m above ground level.

A1.3.2 Preparation of the grout.

The grout should be stored at a temperature of 20 +/-1 deg C. Water for mixing should also be at this temperature as a result of storage or blending. Weigh the components (water and grout) in the proportions required with reference to the manufacturer’s recommended water-to-solids ratio, using a digital scale calibrated using equipment with calibration traceable to National Standards. Sufficient material should be available for efficient operation of the mixing equipment, encapsulation of the grout, and for cube samples.

A1.3.3 Test Method.

Ensure that the test tendon is securely in position. If a grout seal has been used, this should be allowed to cure for at least six hours. Immerse the breather tube outlet in a container of clear water.

Pour the water for grout mixing into the mixing tank, having first established that the tank is clean and empty. Add the grout in increments while using a suitable rotational speed for the mixing paddle. Once all the grout has been added, continue to mix for the period recommended by the grout manufacturer prior to grouting. Continue mixing throughout the grouting operation. Having established that the grout hose is clear of water and contaminants (by, for example, blowing through with compressed air), connect the hose to the pump outlet, direct the other end to the mixing tank, and commence pumping to establish a return flow to tank. Observe the flow to establish that it is continuous, free of air voids and that the flow appears to be consistent with the pump manufacturer’s quoted flow rates. Note: flow can be checked by filling a known volume and timing the event.

Stop the pump, and connect the grout hose securely to the test bolt grouting tube.

Start the pumping operation and record events as follows: show of grout at the proximal end of the test tendon (time required) full encapsulation of the test tendon (time required) cessation of bubbles and /or show of grout at the breather outlet (time required).

Terminate the test when full encapsulation has been achieved, indicated by either a) a show of grout at the breather tube outlet if a breather tube with a bore of 7mm or

more is used, or b) a cessation of bubbles at the breather tube outlet

or, there is an event, such as pump stall or leakage, which prevents full encapsulation being achieved.

138)

man

ufac

ture

r’s m

axim

um

reco

mm

ende

d pu

mpe

d co

lum

n le

ngth

End Cap

Clear Tube

Flexible Tendon

Grout Tube 0.5m into Clear

Tube

Seal

Breather Immersed in

Water Container

Mixing Tank

Mixer Motor

Grout Pump

2.5

m

At least 10mAt least 10m

Figure A1.1 Example of an arrangement for grout encapsulation test – bottom-up grouting

Disconnect the hose from the test assembly having first de-pressurised the circuit by opening the secondary line at the pump outlet. Use the grout hose to fill three moulds with the remaining grout for density and UCS measurements in accordance with annex C.

A1.4 RESULTS

Leave the test tendon in position for at least 12 hours to allow the grout to cure. Remove the tendon to the ground and remove any supports. Photograph the tendon in approximately 1m long segments in order to have a record of the encapsulation. Inspect the test tendon and confirm that full encapsulation has been achieved or otherwise. Inspect for any suspected voids in the encapsulation and note their position. Section the tendon at points where possible voids have been noted, or at four points equi-spaced along the tendon. Photograph the sections.

Five tests should be carried out. For a candidate system to be acceptable, all tests should be completed satisfactorily as described above.

139)

APPENDIX 2 DRAFT TEST PROCEDURE GROUT ENCAPSULATION TEST – ‘TOP DOWN’ GROUTING'

A2.1 PRINCIPLE

This test is used to verify that a candidate ‘top-down’ flexible reinforcement system, where grouting takes place from or near the distal end of the tendon toward the proximal end, can consistently achieve full encapsulation.

A2.2 APPARATUS

A suitable test arrangement is shown in Figure A2.1.

A breather tube may not be required, and the grouting tube may be either integrated into the tendon structure (internal) or attached along the length of the tendon (external). Determine the exact positioning of an external grouting tube using the manufacturer’s specification.

For other aspects of the apparatus, see Appendix 1.

A2.3 PROCEDURE

A2.3.1 Preparation of the test tendon.

Assemble the tube components and bond together. Fit the grouting tube (if required) to the tendon and secure using pvc or other suitable tape. Ensure that the grouting tube arrangement complies with the manufacturer’s recommendation and it is not blocked during assembly and anchorage. Ensure sufficient tubing extends beyond the proximal end of the tendon for connection to the grout hose. The tendon can be anchored either at the top of the tube using, for example, mixed encapsulated resin, and/or at the bottom using a combination of a seal and clips/tape. The test tendon must be sealed at the proximal end. This arrangement will depend upon whether an end plate is used to complete the installation. If not, a tape seal around the joint between tube and tendon will probably be effective.

Erect the test tendon to the vertical using a means of support as described in Appendix 1.

Ensure that the test tendon is straight to within 100mm of the notional vertical axis of the test assembly. The proximal end of the test tendon should be approximately 2.5m above ground level. A means of egress of air at the proximal end will be required: this may be furnished via an end plate, but if not, drill an 6mm hole in the tube just above the seal.

A2.3.2 Preparation of the grout.

See Appendix 1.

A2.3.3 Test method.

Ensure that the test tendon is securely in position. Pour the water for grout mixing into the mixing tank, having first established that the tank is clean and empty. Add the grout in increments while using a suitable rotational speed for the mixing paddle. Once all the grout has been added mix for the period recommended by the grout manufacturer. Continue mixing

141)

throughout the grouting operation. Having established that the grout hose is clear of water and contaminants (by, for example, blowing through with compressed air), connect the hose to the pump outlet, direct the other end to the mixing tank, and commence pumping to establish a return flow to tank. Observe the flow to establish that it is continuous, free of air voids and that the flow appears to be consistent with the pump manufacturer’s quoted flow rates. Note: flow can be checked by filling a known volume and timing the event. Stop the pump, and attach the grout hose securely to the test tendon using a proprietary lance, grouting tube or other suitable fitting.

Start the pumping operation and record events as follows: � show of grout at the distal end of the test tendon (time required), and � full encapsulation of the test tendon (time required).

Terminate the test when grout issues from the breather hole, or there is an event, such as pump stall, which prevents full encapsulation being achieved.

Depressurise the circuit by opening the secondary line at the pump outlet and then disconnect the hose from the test assembly. Use the grout hose to fill three moulds with the remaining grout for density and UCS measurements in accordance with appropriate annex.

A2.4 RESULTS

See Appendix 1.

End Cap

Clear Tube

Flexible Tendon

Grout Tube

Seal

Mixing Tank

Mixer Motor

Grout Pump

2.5

m

Breather Hole

man

ufac

ture

r’s m

axim

umre

com

men

ded

pum

ped

colu

mn

leng

th

At least 10mAt least 10m

Figure A.2.1 Example of an arrangement for grout encapsulation testing – top-down arrangement

142)

APPENDIX 3 DRAFT TEST PROCEDURE DETERMINATION OF BOND STRENGTH AND SYSTEM

STIFFNESS – CEMENTITIOUS GROUT ANCHORED SYSTEM

A3.1 PRINCIPLE

The bond performance of a flexible tendon and an associated cementitious grout is determined from a laboratory short encapsulation pull test where the tendon sample is installed in a confined, rock core using cementitious grout which complies with Part 5 of this standard, and, after grout curing, is pull tested under controlled conditions. Bond performance is assessed in terms of load and slope values obtained from the load / bond displacement characteristic.

A3.2 APPARATUS

A3.2.1 The installation apparatus comprises a machine tool lathe, such as that shown in Figure A3.1, a hydraulic biaxial cell, a water feed system and drill assembly. Use pull test equipment, as shown in Figure A3.2, to carry out the test. An autographical recording facility or other means of producing a load / extension graph for recording the test data during pull testing are required, as indicated in Figure A3.2, and as described below.

A3.2.2 Machine Tool Lathe

A machine tool lathe with a sufficient bed length to allow the drilling operations to be carried out in a single pass is required. The lathe should be capable of a traverse of at least 320mm, a rotational speed of 440 rpm, and offer a minimum torque of 200Nm. An automated feed rate of 1.25mm/revolution is also desirable, but not essential.

A3.2.3 Biaxial Cell

A hydraulic biaxial pressure cell is shown in Figure A3.3. The cell should have a nominal internal diameter of at least 145mm and a minimum confining membrane length of 500mm. The cell should be capable of applying a confining pressure of at least 10MPa.

A3.2.4 Water Feed

The system should allow flushing water to be delivered effectively through a rotating drill rod, fixed in the chuck of the lathe, to the tip of the drill bit during drilling operations.

A3.2.5 Drilling Consumables

A drill bit as typically used for the underground installation of the candidate tendon should be used where possible. Most underground installations exhibit rifling of the hole wall, and this should be duplicated in the laboratory test through choice of drilling equipment (bit and rod) and positioning of the drill rod, as indicated in the test procedure for resin grouted systems. However, hole wall rifling is not as critical to bond performance when testing cementitious grout bonded tendons, principally because the larger hole diameter used provides a high level of shear strength at the rock /grout interface (relative to that achieved with smaller diameter holes used for resin bonded systems) without relying on geometrical discontinuity.

143)

A3.2.6 Pull Test Equipment

Pull test equipment must be suitably calibrated, and will comprise a hydraulic hollow ram jack, a pressure bearing plate or stressing stool, hydraulic hose, pressure gauge and/or load cell, and hydraulic pump fitted with a non-return valve, as shown in Figure A3.2. The centre hole of the hollow ram should be sufficiently large to allow assembly onto a steel embedment tube. Experience shows that a 95 tonne capacity ram is required to provide a sufficiently large centre hole.

A3.2.7 Recording Apparatus

Use a Linear Variable Differential Transformer (LVDT) or dial indicator to record displacement and an in-line pressure gauge, preferably with an electronic sensor and / or suitable capacity load cell, to record the applied pressure and hence load. See Figure A3.2.

A3.2.8 Rock Test Specimens

The rock test specimens should consist of sandstone rock cores with an external diameter to match the internal diameter of the biaxial cell used, and of sufficient length to extend approximately 10mm beyond the membrane at both ends of the cell. The rock test specimens should have suitable properties and meet the performance criteria specified in A3.5. The core should comprise poorly cemented, medium grained, homogeneous sandstone with rounded, well-sorted grains. When tested according to ISRM Suggested Methods, the uniaxial compressive strength should lie between 21 and 31 MPa and the Young’s Modulus should lie between 7 and 10 GPa. An example of a suitable rock is “Hollington Stone”, extracted from Hollington Quarry, Hollington in Staffordshire.

A3.2.9 Embedment Tube

The tendon should be installed in the core and a steel embedment tube mounted on the core. The tube is used to facilitate loading of the section of tendon embedded in the rock core, while maintaining the shape of the tendon where it exits the core. The tube should be 450mm long, and have an internal diameter nominally the same as the diameter of the drill bit used to drill the core. The internal surface should be machined with a groove 1mm deep and 2mm in pitch along its full length. Tube thickness should be a minimum 10mm and the material should have a minimum yield stress of 400 MPa. One end of the tube should be threaded on the outer surface to accept a reaction plate.

A3.3 PROCEDURE

A3.3.1 Rock Core Preparation

Core specimens with major irregularities, bedding or discontinuities should be discarded. Any minor irregularities or depressions found in the outer surface of the rock core must be removed or filled with a suitable self hardening filler compound to avoid localised deformation of the cell membrane under pressure.

A3.3.2 Installing Rock Core in Biaxial Cell

The biaxial cell should be securely mounted on the lathe stock such that the axis of the rock core is in alignment with the axis of the lathe chuck. The rock core should be located inside the biaxial cell ensuring that the cell membrane has full circumferential and axial contact with the

144)

rock core. Approximately 10mm of rock core should overlap the membrane at both ends of the cell. A confining pressure of 10MPa should then be applied to the rock core.

A3.3.3 Drilling

Sharp undamaged drill bits of the correct type and dimensions (see A3.2.5) should be used. Drill rods in good condition, which are clear of debris and with full flushing functionality, should be used. Mark the drill rod 320mm from the bit end. Mount the drill rod in the lathe chuck such that it is concentric, that the required length of drill rod extends beyond the face of the chuck, and the water feed is attached. Advance the lathe stock until the face of the rock core is close to the drill bit. Operate the lathe at the correct rotational speed (approximately 440rpm), apply flushing water, then, manually advance the lathe stock to initiate drilling. Once the drill bit has begun to penetrate the rock core, ensure that rock penetration continues at the appropriate rate (approximately 1.25mm/revolution), preferably by engaging an autofeed mechanism. When the rock core has been drilled to the correct depth (320mm), disengage the feed mechanism and withdraw the stock slowly, maintaining lathe rotation and flushing water pressure. Ensure the hole is free of debris and has a depth of 320mm.

A3.3.4 Tendon Installation and Pull Testing

Measure and record the internal diameter of the drilled hole using a calibrated borehole micrometer, recording the diameter for at least six positions evenly distributed along the length of the borehole. From these readings determine the average borehole diameter. Ensure the flexible tendon to be tested is clean and free from contaminants. The tendon, in its design geometrical form (complete with cages, bulbs etc) should be of sufficient length to allow installation into the borehole and a 450mm long embedment tube placed on top of the rock core. Additionally the tendon should have a ‘tail’ section of straight strand(s) which will protrude beyond the embedment tube and be of sufficient length to allow fitting of spacer plates and barrel/wedge assemblies or other end termination. Overall length of the test tendon is likely to be around 1 metre. The end of the tendon to be inserted into the rock core should have a fully developed profile (not tail section) and should be cut normal to the axis of the tendon. Depending on the application, a tendon to be used underground would be fitted with a breather or grout tube along most of its length. Complete the assembly of the test tendon by securing a 320mm length of breather or grout tube, as specified by the tendon supplier, to the distal end. A breather tube of less than 10mm bore should be sealed at both ends. A larger breather tube, or grout tube, should be left open to admit grout.

Remove the biaxial cell, complete with confined rock core, from the lathe and place it upright on the laboratory floor. Place a sealing membrane around the periphery of the mating face of the steel tube and place the tube on top of the core, taking care to centralise the core and tube holes. Carefully insert the tendon into the assembly and locate to the back of the core hole. Prepare a sufficient quantity of grout to fully encapsulate the tendon, taking care to adhere to manufacturers’ requirements for water-to-solids ratio, mixing time etc. Pump the grout slowly into the assembly until full. Top up as necessary as settling occurs. Fill a 100mm cube mould from every mix prepared, for UCS testing of the grout. Allow the grout to cure for 14 days before testing.

With the core installed in the biaxial cell and pressurised to 10 MPA, install the pull test equipment on the test assembly, as shown in Figure 3.2. Install a reaction plate onto the embedment tube, fit an end termination onto the tendon, and locate to bear on the reaction plate. Fit a dial indicator or LVDT to the end fitting in order to record bond displacement.

145)

Operate the hydraulic pump at a slow and smooth rate (approximately 1 kN/sec) applying increasing pressure to the hydraulic jack. Record load and bond displacement incrementally so that a minimum of twenty data points have been recorded when displacement exceeds 10mm in total, or at 90% of the load at which the yield strength of the tendon would be reached. Cease pump operation at this point. After pull testing, relieve the pressure from the pull test jack and then relieve the pressure in the biaxial cell.

A3.3.5 Core Examination

After testing, withdraw the rock core from the biaxial cell and split the core in the axial plane in order to inspect the quality of installation and mode of bond failure. Examine the grout / tendon and the grout / rock interfaces and note the location of any shear failure.

A3.4 RESULTS

Plot load to a base of measured displacement and determine a) bond strength which is the load at which the slope of the load / displacement characteristic falls below 20 kN / mm b) system stiffness which is the slope of the load / displacement characteristic in the load range 150 - 300 kN.

A3.5 ROCK CORE PERFORMANCE CRITERIA

A standard test is required to determine the performance characteristics of the rock core under pull test conditions to ensure consistent rock material is used. When tested in accordance with the procedure described in Annex E of Part One of this Standard, and using the standard consumables and criteria listed in Table A3.1, the rock core should provide test results which lie within the performance envelope shown in Figure A3.4. Bond failure must be at the rock/resin interface.

Table A3.1 Standard Consumables and Criteria for Rock Core Performance Testing

Hole diameter 28.5mm+/- 0.5mm Bond length 160mm Confining pressure 10MPa (biaxial cell) Drill rod type 19mm AF, hollow, hexagonal Bar type M24 High tensile continuously threaded

steel bar, grade 10.9 steel (Yield Strength 312kN, UTS 346kN)

Grout type Resin grout complying with Part 1 of this Standard.

146)

Figure A3. 1 Lathe test&

Figure A3. 2 Pull test equipment&

147)

Tendon End Fitting and

Bearing Plate

LVDT'Attached to

Tendon'

Figure A3.3 Tensile testing machine&

Upper M/C

Lower M/C

Bearing Plate

Bi-axial Cell with Test Sample Installed

Figure A3.4 Biaxial cell

148)

F ve&

149)

APPENDIX 4 DRAFT TEST PROCEDURE SHEAR TEST ON TENDON / GROUT SYSTEM

4.1 PRINCIPLE

The ultimate shear strength of a flexible reinforcement system is determined in the laboratory using a single shear frame in conjunction with a flexible tendon / grout double embedment assembly.

4.2 APPARATUS

4.2.1. The design of a single (guillotine) shear frame suitable for the testing of a flexible tendon / grout double embedment assembly is shown in figure A4.1. Use a test machine calibrated to BS EN ISO 7500 Part 1 1999, having an autographic recording facility or other means of producing a force / displacement graph.

4.2.2. The arrangement of the test assembly is shown in figure A4.1. It consists of two thick-walled hollow steel tubes, each 450mm long, with an internal diameter nominally the same as the diameter of the drill bit recommended for tendon installation. The tubes should have a wall thickness of at least 10mm, and have a 1.0mm deep by 2.0mm pitch thread machined onto their internal surface in order to provide a standard surface finish intended to inhibit failure between this surface and the grout.

4.2.3. Use a displacement transducer, for example the stroke measurement device on the testing machine, to record the separation of the two tubes.

4.3 PROCEDURE

4.3.1. Sample size

Prepare three test assemblies.

4.3.2. Preparation of test assemblies – resin bonded systems

Blank off one end of each tube with strong adhesive tape. Prepare sufficient slow set resin, mixed in accordance with the manufacturer’s instructions, to fully encapsulate one tube / tendon assembly, and pour into one tube. Push the tendon, which should be 900mm long, into the resin by hand, whilst at the same time slowly rotating the tendon and ensuring as far as is possible that the tendon is centrally positioned within the tube assembly. Remove any excess resin from the tube outer surface and face. Allow to cure for one hour. Prepare another batch of resin, as described above, and fill the second tube. Encapsulate the remaining section of tendon in the second tube, taking care to centralise the tendon within the tube and that the tube faces are fully butted together. Allow the assembly to cure for at least 24 hours at a room temperature of 20+/- 2 deg C.

4.3.3. Preparation of test assemblies – cementitious grout bonded systems

Blank off the end of one of the tubes with strong adhesive tape, and butt together the open ends, securing the joint temporarily with strong adhesive tape. Prepare sufficient grout, mixed in accordance with the manufacturer’s instructions, to fully encapsulate the tube / tendon assembly

151)

1

whilst at the same time slowly rotating the tendon, and ensuring, as far as is possible, that the

tendon is centrally positioned within the tube assembly. Allow the assembly to cure at a room

temperature of 20+/- 2 deg C for 14 days.

4.3.4. Method

Place a test assembly in the test machine and apply load at a rate not exceeding 2 kN / sec until

such time as the maximum force is achieved. Record load and displacement (platen

displacement or piston stroke are sufficient) at intervals of not more than 2 seconds during the

test.

4.4 RESULTS

Plot a load vs displacement characteristic and note maximum load. The ultimate shear strength

of the grouted flexible tendon is determined from the mean of the three test results.

Figure A4. 1 Sectional diagram of shear frame

Published by the Health and Safety Executive 02/11



A research programme has been carried out by RMT in support of the revision of Part 2 of the British Standard for strata reinforcement components in coal mines, covering flexible systems for roof reinforcement. This continued work commenced under a previous HSE Project, ‘Testing and standards for reinforcement consumables’.

A particular focus was to compare tensionable and non-tensionable reinforcement systems, prompted by the introduction of tensionable systems to British coal mines. A review of previous research indicated conflicting claims for tensionable systems in terms of theoretical advantages and practical experience. The research included laboratory testing, underground measurement and analysis of underground monitoring data. Advice and draft Annexes were provided to the BS Committee and a revision of the DMCIAC guidance on the use of cable bolts to support roadways in coal mines drafted. The work highlighted practical problems concerning application of the tensionable systems in use in UK coal mines but did not exclude their future applicability provided they comply with the revised Standard.

This report and the work it describes were funded by the Health and Safety Executive and co-funded by the EU Research Fund for Coal and Steel. Aspects were also co-funded by UKCoal Ltd and various manufacturers through supply of materials for testing. The report’s contents, including any opinions and/ or conclusions expressed, are those of the authors alone and do not necessarily reflect HSE policy nor the opinions of any of the co-funding parties.

RR831

www.hse.gov.uk

www.hse.gov.uk

rr831 - evaluation of tensioned and non-tensioned long ... · 8.2 suggested revision of dmciac...

Documents