Required technical specifications and standard testing methodology ...

SAFETY IN MINES RESEARCH ADVISORY COMMITTEE

Final Report

Required technical specifications and standard testing methodology for Thin Sprayed Linings

J.S. Kuijpers, E.J. Sellers, A.Z. Toper, T. Rangasamy, T. Ward, A.J. van Rensburg, H. Yilmaz and R.Stacey

Research agency

: CSIR Division of Mining Technology

Project number

: SIM 020206

Report number

: 2004-0404

Date

: November 2004

Executive summary Thin Sprayed Liners (TSLs), as well as shotcrete and mesh, are mainly used to provide a better support coverage in mining excavations. While it seems evident that increased support coverage leads to improved stability around mining excavations, it is not obvious how the actual support performance of surface-support liners should be quantified. The behaviour and performance of TSLs are addressed in this project. On the basis of underground and laboratory work, in combination with numerical modelling results, it was possible to formulate potential instability mechanisms that may be controlled by a TSL. The emphasis in this study is on the mechanical interaction between a TSL and a rock mass and the relevant and associated material properties. While it is acknowledged that a TSL may have other functions, such as the prevention of weathering through sealing, such other functions have not been directly addressed here. The main objective of this project is to formulate a standard testing methodology and to develop testing equipment in order to quantify relevant TSL parameters. During the course of this project, a series of international seminars was organised between the Australian Centre for Geomechanics, the University of the Witwatersrand, and the Université Laval in Canada. The objective of these seminars was also to formulate and reach consensus on internationally acceptable standard testing procedures for TSLs. By actively participating in these seminars, insights were shared between local and international participants. The testing methodology that is presented in this report is therefore not only relevant to South African applications, but is also broadly supported by the international mining community. The support performance of a TSL is largely determined by the substrate onto which it is attached. In addition, a TSL is typically part of a support system that contains other support components such as tendons. The requirements for a TSL are therefore strongly associated with the detailed morphology of the substrate as well as with the characteristics of the other support components. A simple mechanical model that quantifies the interaction between a TSL, the substrate, and the retaining support is described in this report. With the aid of this model, the required TSL parameters can, in principle, be quantified. One of the most revealing findings from this project was the strongly time-dependent behaviour of some of the TSL products. In essence, the strength of such products is directly associated with the duration of loading, and strength therefore needs to be carefully defined in those products. High strength and high stiffness are associated with rapid loading, while relatively low strength and stiffness are associated with slow loading. Loading rate is thus extremely relevant when the strength and stiffness of a particular material are being determined. The testing equipment that was developed for this project is capable of subjecting test specimens to variable loading rates. The proposed standard testing procedure requires the specification of loading rate in a tension test. Two standard tests were formulated, namely a tension test and an adhesion test. The test procedures have been optimised based on laboratory and underground testing and guidelines for the standard testing procedures are presented in the report. The adhesion test was designed for in situ evaluations as well and can be used to assess the underground performance in relation to potential performance under laboratory conditions. While it is not the intention here to prescribe detailed requirements for TSL products, it is apparent from this study that only tough and strong bonding materials are capable of providing effective support resistance when large deformations need to be accommodated. However, certain brittle and stiff materials may provide very effective support when they are not subjected to large deformations. In fact, the perception that large deformations need to be accommodated seems to be misplaced in many of the

2

observed cases. It is therefore important that any potential application is carefully assessed.

3

Acknowledgements The work presented here results from funding provided by SIMRAC. The co-operation, support and constructive criticism of members of the SIMGAP committee are appreciated and hereby acknowledged. Work done by M. Kiboka and S Saydam is also appreciated. The cooperation from Rock Engineering staff of the following mines was very helpful and is greatly appreciated: ?

Kloof

?

Placer Dome Western Areas’ JV

?

Savuka (two sites)

?

East Driefontein

Support and supply of TSL products is acknowledged from: ?

FOSROC, (Tekflex) (two products)

?

SA Mining and Engineering Supplies (Tunnelguard)

?

HYDROFLEX (Pty) Ltd

?

CHC Urethane Products

4

Page

Table of contents

Executive summary................................................................................................................2 Acknowledgements ................................................................................................................4 Table of contents ....................................................................................................................5 List of figures ..........................................................................................................................8 List of tables..........................................................................................................................11 1. 1.1 1.2 1.3 1.4 2. 2.1 2.2 2.3 2.4 2.4.1 2.4.2 2.4.3 2.4.4 2.4.5 3. 3.1 3.2 3.3 3.4 3.5 3.6 3.6.1 3.6.2 4. 4.1 4.2 4.2.1 4.2.2 4.2.3 4.2.4 4.2.5 4.2.6 4.3 4.3.1

Introduction ............................................................................................................13 General ..................................................................................................................13 Surface-support requirements ..............................................................................13 TSL requirements ..................................................................................................17 Aims and objectives...............................................................................................18 Literature review on laboratory testing of Thin Sprayed Liners .............................20 Introduction ............................................................................................................20 Development of TSL products...............................................................................21 Testing of TSLs .....................................................................................................22 Previous tests of TSLs ..........................................................................................24 Introduction ............................................................................................................24 Tensile strength and elongation tests....................................................................25 Adhesion tests .......................................................................................................28 Punch type of testing .............................................................................................36 Compression failure tests on coated samples .....................................................42 The functions of surface support...........................................................................44 Introduction ............................................................................................................44 Maintaining initial integrity.......................................................................................44 Single-block stability ..............................................................................................45 Composite beam ...................................................................................................51 Highly fragmented dilating rock..............................................................................52 Conceptual model..................................................................................................54 Introduction ............................................................................................................54 Description and examples .....................................................................................56 Proposed testing procedures ................................................................................61 Introduction ............................................................................................................61 Tensile strength test..............................................................................................61 Introduction ............................................................................................................61 Scope.....................................................................................................................63 Apparatus...............................................................................................................63 Procedure ..............................................................................................................64 Calculations ...........................................................................................................65 Reporting of results ...............................................................................................65 Proposed adhesion/bond strength test .................................................................66 Introduction ............................................................................................................66

5

4.3.2 4.3.3 4.3.4 4.3.5 4.3.6 4.4 5. 5.1 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.2.5 5.2.6 5.3 5.3.1 5.3.2 5.3.3 5.3.4 5.3.5 5.3.6 6. 6.1 6.1.1 6.1.2 6.1.3 6.2 6.3 6.3.1 6.3.2 6.3.3 6.3.4 6.4 6.4.1 6.4.2 6.4.3 6.4.4 6.5 6.5.1 6.5.2 6.5.3 6.5.4 6.6

Scope.....................................................................................................................67 Apparatus...............................................................................................................67 Procedure ..............................................................................................................67 Calculations ...........................................................................................................68 Reporting of results ...............................................................................................68 Conclusions ...........................................................................................................69 Results from laboratory testing .............................................................................71 Introduction ............................................................................................................71 Tensile tests ..........................................................................................................71 Introduction ............................................................................................................71 Product A...............................................................................................................72 Product B...............................................................................................................72 Product C...............................................................................................................73 Product D...............................................................................................................75 Product E...............................................................................................................76 Adhesion tests .......................................................................................................77 Introduction ............................................................................................................77 Product A...............................................................................................................78 Product B...............................................................................................................78 Product C...............................................................................................................80 Product D...............................................................................................................81 Product E...............................................................................................................82 Underground Monitoring.........................................................................................84 Introduction ............................................................................................................84 Disk installation procedure ....................................................................................84 Pull test procedure.................................................................................................87 Quantity of testing..................................................................................................89 Description of the test sites ...................................................................................89 Savuka pipe raise ..................................................................................................90 Site location ...........................................................................................................90 History of TSL at the test site ................................................................................92 Installed support.....................................................................................................92 Major natural and induced discontinuities..............................................................92 Savuka strike gully .................................................................................................93 Site location ...........................................................................................................93 TSL history.............................................................................................................94 Installed support.....................................................................................................95 Major natural and induced discontinuities..............................................................95 Placer Dome Western Areas Joint Venture (PDWAJV) .......................................95 Site location ...........................................................................................................95 TSL history.............................................................................................................96 Installed support.....................................................................................................97 Major natural and induced discontinuities..............................................................97 Kloof 8 Shaft ..........................................................................................................97

6

6.6.1 6.6.2 6.6.3 6.6.4 6.7 6.7.1 6.7.2 6.7.3 6.7.4 6.8 6.9 6.9.1 6.9.2 6.9.3 6.9.4 6.10 7. 8. 9.

Site location ...........................................................................................................97 TSL history.............................................................................................................98 Installed support.....................................................................................................99 Major natural and induced discontinuities..............................................................99 Driefontein 1 East................................................................................................100 Site location .........................................................................................................100 TSL history...........................................................................................................101 Installed support...................................................................................................102 Major natural and induced discontinuities............................................................102 Geotechnical characteristics of the test sites .....................................................102 Test results ..........................................................................................................104 Monitored data .....................................................................................................104 Product types.......................................................................................................106 Interpretation of results ........................................................................................107 Perceived effectiveness of TSLs.........................................................................109 Conclusions .........................................................................................................111 Discussion and recommendations .....................................................................112 References ..........................................................................................................114 Appendix A...........................................................................................................118

7

Page

List of figures Figure 1.2.1

Load transfer in a typical support system ...................................................... 14

Figure 1.2.2

Energy absorption through bulging (top) and into tendons (bottom).......................................................................................................... 16

Figure 1.2.3

Failure of the support system due to lack of adequate coverage.................. 17

Figure 2.4.1

Schematic describing a classification of existing tests to determine properties and characteristics of TSLs (Potvin, 2002)................. 25

Figure 2.4.2

Die cut tensile strength test specimens and tensile strength testing apparatus (Archibald, 2001) ............................................................... 26

Figure 2.4.3

Shape and dimensions of a Type 1 test specimen using ASTM D638............................................................................................................... 26

Figure 2.4.4

Test set-up for a tensile test on a "dog-bone"-shaped specimen a) Tannant et al. (1999), b) Spearing and Gelson (2002) .............................. 27

Figure 2.4.5

De-bonding mechanisms............................................................................... 28

Figure 2.4.6

Core adhesion test (Spearing, 2001)............................................................. 29

Figure 2.4.7

(a) Generic view of a test dolly and (b) Typical test set-up in the laboratory (Tannant et al. 1999) ..................................................................... 30

Figure 2.4.8

Adhesion pull test assembly schematic and test equipment (Archibald, 2001) ............................................................................................ 31

Figure 2.4.9

Specimen assembly prior to over coring and at finish of adhesion bond strength test (Archibald, 2001).............................................................. 32

Figure 2.4.10 Typical views showing adhesion bond surfaces after completion of failure tests (Archibald, 2001) .................................................................... 33 Figure 2.4.11 Underground adhesion testing a) Apparatus b) Test site (Espley et al., 2001)..................................................................................................... 33 Figure 2.4.12 Schematic diagram of pull-test apparatus (Erck 1994)................................. 34 Figure 2.4.13 Baggage Load Test Frame & Set-Up (Swan and Henderson, 1999) .............................................................................................................. 35 Figure 2.4.14 Liner supports the load from loose rock ........................................................ 36 Figure 2.4.15 Final stages of a plate pull test showing tensile rupture in the TSL and adhesion loss with the rock (Tannant et al., 1999) ................................. 37 Figure 2.4.16 Pull strength test assembly schematic and test equipment (Archibald, 2001) ............................................................................................ 38 Figure 2.4.17 Test assembly before and after the pull strength test (Archibald, 2001) .............................................................................................................. 38 Figure 2.4.18 Typical views showing pull test failure conditions at the completion of failure tests (Archibald, 2001).................................................. 40 Figure 2.4.19 Test set-up and typical test results from a large-scale pull test on Mineguard coated concrete blocks (Espley et al., 1999)............................... 41 Figure 2.4.20 Punch testing set-up (Spearing et al., 2001) ................................................. 42 Figure 2.4.21 Coated and uncoated cylinders are tested to failure (Espley et al., 1999; Archibald and DeGagne, 2000)............................................................ 42 8

Figure 3.3.1

Maximum block height in relationship to base dimensions (dead weight only) .................................................................................................... 45

Figure 3.3.2

Cross section of the axi-symmetric model (half symmetry) ......................... 46

Figure 3.3.3

No de-bonding (perfectly plastic material) ..................................................... 47

Figure 3.3.4

Partial de-bonding (perfectly plastic material)................................................ 47

Figure 3.3.5

Punch resistance with varying membrane properties (soft membrane)..................................................................................................... 48

Figure 3.3.6

Punch resistance with varying membrane properties (stiff membrane)..................................................................................................... 48

Figure 3.3.7

Bond stress distribution affected by stiffness................................................ 49

Figure 3.3.8

Effect of membrane and interface stiffness on punch resistance (yield).............................................................................................................. 49

Figure 3.3.9

Effect of membrane and interface properties on critical bond strength .......................................................................................................... 50

Figure 3.3.10 Results of laboratory punch tests: A) soft, B) stiff ......................................... 51 Figure 3.4.1

Stress distribution through a section of a coated layer.................................. 51

Figure 3.5.1

Development of vertical stress into a pillar abutment (after Barron, 1982) ................................................................................................. 53

Figure 3.6.1

Distribution of inertia forces in the unstable rock mass skin ......................... 56

Figure 4.2.1

Proposed tensile strength test specimen dimensions for testing of TSL material............................................................................................... 62

Figure 4.2.2

Tensile loading apparatus and recording equipment..................................... 66

Figure 4.2.3

Recommended dimensions for the tensile testing specimen ....................... 66

Figure 4.3.1

Recommended dimensions for (axi-symmetric) steel dolly.......................... 69

Figure 4.3.2

Example of pull testing equipment connected to glued dolly ......................... 69

Figure 5.1.1

Test rig for dog-bone specimens ................................................................... 71

Figure 5.2.1

Variation in tensile strength for different curing times (product A)................. 72

Figure 5.2.2

Variation in tensile strength for different curing times (product B)................. 73

Figure 5.2.3

Variation in tensile strength for different curing times .................................... 74

Figure 5.2.4

Variation in tensile strength with respect to loading rate................................ 74

Figure 5.2.5


Figure 5.2.6


Figure 5.2.7


Figure 5.2.8


Figure 5.3.1

Pulling equipment used for adhesion tests in the laboratory......................... 77

Figure 5.3.2

Adhesion results for product A....................................................................... 78

Figure 5.3.3

Adhesive test results for product B................................................................ 79

Figure 5.3.4

Adhesive test results for product C................................................................ 81

Figure 5.3.5

Adhesive test results for product D................................................................ 82

9

Figure 5.3.6

Adhesion test results for product E................................................................ 83

Figure 6.1.1

Square 50 mm by 50 mm decoupling slot..................................................... 85

Figure 6.1.2

Removed slotted disk along pre-cut lines ...................................................... 85

Figure 6.1.3

Disc pasted onto liner and left for a 24-hour curing period............................ 86

Figure 6.1.4

Disk positioning along the sidewall of a tunnel............................................... 87

Figure 6.1.5

Marking of positions for disks above the grade line ....................................... 87

Figure 6.1.6

Frame of pull tester being held in position by mining technician ................... 88

Figure 6.1.7

Calliper vernier used to measure slot size and thickness of liner ................. 88

Figure 6.3.1

Regional layout of the Savuka pipe raise site showing the line of stoped areas and position of strike-stabilising pillars .................................... 91

Figure 6.3.2

Local layout of the pipe raise test site indicating test site positions .............. 91

Figure 6.4.1

Regional layout of the gully test site at Savuka gold mine ............................. 93

Figure 6.4.2

Local setting of the Savuka gully site with positions of tests indicated ......................................................................................................... 94

Figure 6.5.1

Local site layout for the PDWAJV test programme....................................... 96

Figure 6.6.1

Regional setting of the 14L Footwall Drive East at Kloof 8#.......................... 98

Figure 6.7.1

Test site at the 36-22 FW Drive East at Driefontein No.1 East................... 100

Figure 6.7.2

Schematic illustrating the test positions at the 36-22 FW Drive East at Driefontein No.1 East....................................................................... 101

Figure 6.9.1

Distribution of thickness for 1-mm-range intervals ...................................... 107

Figure 6.9.2

Percentage sample distributions of pulloff pressures for specified pressure-range intervals .............................................................................. 108

Figure 6.9.3

Cumulative maximum pull out resistance for all test sites .......................... 109

10

Page

List of tables Table 2.1

Existing TSL products..................................................................................... 22

Table 2.2:

Relevant standards for testing of TSLs .......................................................... 24

Table 2.3

Tensile properties of TSL materials according to ASTM D638 (Archibald, 2001) ............................................................................................ 28

Table 2.4

Adhesive strengths of various products and their curing time ....................... 34

Table 2.5

Pull strength properties of TSL materials (Archibald, 2001)........................... 39

Table 3.1

Induced stresses associated with “beam” behaviour..................................... 58

Table 3.2

Induced stresses associated with “bag” behaviour........................................ 58

Table 3.3

Induced stresses in composite beam and contributing rock mass ............... 59

Table 5.1

Adhesion results for product A........................................................................ 78

Table 5.2

Adhesion results for product B ....................................................................... 79

Table 5.3

Adhesion results for product C ....................................................................... 80

Table 5.4

Adhesion results for product D ....................................................................... 81

Table 5.5

Adhesion results for product E ....................................................................... 83

Table 6.1

Mix ratios for the adhesive .............................................................................. 84

Table 6.2

Quantity of adhesion tests conducted per site ............................................... 89

Table 6.3

Brief outlines of the geotechnical characteristics of the five mines hosting the test sites ...................................................................................... 89

Table 6.4

History of TSL application ............................................................................... 92

Table 6.5

Support systems in place at the site apart from TSL..................................... 92

Table 6.6

Dominant natural and induced discontinuities ................................................ 92

Table 6.7


Table 6.8

Support systems in place at the site apart from TSL..................................... 95

Table 6.9


Table 6.10


Table 6.11

Support systems in place apart from the TSL................................................ 97

Table 6.12


Table 6.13


Table 6.14

Support systems in place apart from the TSL................................................ 99

Table 6.15


Table 6.16

History of TSL application at the Driefontein site.......................................... 101

Table 6.17

Support systems in place apart from the TSL.............................................. 102

Table 6.18

Dominant natural and induced discontinuities .............................................. 102

Table 6.19

Quality classification of the rock mass based on q-values .......................... 103

Table 6.20

Listing of rock mass rating values for the test sites ..................................... 103

Table 6.21

Adhesion test results for Savuka pipe raise ................................................. 104

11

Table 6.22

Results for Savuka gully ............................................................................... 104

Table 6.23

Results for Placer Dome Westonaria Joint Venture 2B11SR South sidewall .............................................................................................. 105

Table 6.24

Results for East Driefontein 1E 36-22FW drive east................................... 105

Table 6.25

Results for Kloof 8# 14level FW drive east .................................................. 106

Table 6.26

Type and providers of TSL placement for the test sites............................... 106

Table 6.27

Matrix of holistic effectiveness of TSL .......................................................... 109

12

1. Introduction 1.1 General Lack of support coverage can be singled out as the most obvious cause of rock falls in mining excavations (Daenke et al., 1998; Klokow, 1999; Daenke et al., 1999). In addition, it is becoming increasingly evident that (support) coverage can reduce weathering and unravelling of exposed rock surface (Ortlepp, 1983; Wojno et al.; 1986; Applegate, 1987). Currently, the design of stope and tunnel support systems is based on tributary area theory according to which primary support units are designed to retain a particular volume of rock, based on instability thickness and support spacing (Wagner, 1983 ; Daenke et al., 1998). The containment support is often just regarded as an additional safety measure to prevent smaller rocks from falling into the excavation. However, recent work (Haile, 1999; Kuijpers et al., 2002) suggests that relatively large rock mass instabilities are often caused by failure of the containment support, even when the primary support units remain intact. Such occurrences of failure are typically associated with a highly fragmented and unravelled rock mass that is not capable of transmitting sufficient shear loads to the primary support units. Under these conditions, the uncovered rock mass needs to be reinforced and/or supported, so that sufficient competency is available. Reinforcement could be supplied by surface support such as shotcrete and TSLs, as well as by mesh and lacing. Alternatively, additional secondary tendons could also provide such reinforcement, for instance. However, as the objective of this study is the performance of TSLs, only the behaviour of this type of surface support will be discussed in more detail.

1.2 Surface-support requirements In order to appreciate the contribution of surface support to the stability of an underground excavation, it is necessary to view such surface support as part of a support system. This support system not only comprises the retaining and containing support components, but also the rock mass itself. In fact, the rock mass is the most essential part of the system. Figure 1.2.1 depicts the load transfer in a typical support system in which the rock mass is also incorporated. This conceptual model is described in more detail in Chapter 3. With respect to the rock mass characteristics, it is typically assumed that the rock mass in between individual retaining units can be regarded as a single continuous plate or shell with a certain thickness. The function of the primary retaining support is to pin this competent skin to the solid rock under design loading conditions. If the solid continuous skin is sufficiently thick and adequate confining conditions prevail, it is not necessary to provide additional, secondary surface support, as the resistance, or competency, of the skin itself is sufficient. This is in fact the basis for the design of primary retaining support. However, this assumption is in many cases not valid, as the rock mass skin either may be subject to deterioration in time, or may have insufficient resistance because of its morphology in association with the loading conditions. Weathering and unravelling of an initially competent rock mass may be prevented by surface support either by sealing, or by providing minimal support to the immediate surface. These functions are not necessarily associated with excess loading conditions, but with maintaining the initial rock mass integrity under general conditions. If the rock mass integrity is insufficient, reinforcement and/or support are required. The demand on the surface support is determined by the loading conditions in combination

13

with the detailed structure of the rock mass. In a worst-case scenario, the contribution of the rock mass to resisting load and deformations can be neglected.

1

1

2

2

3

3 Tendons

Fabric, mesh, lacing

1) Directly through the rock mass 2) Into the tendons 3) Via the fabric into the tendons

Figure 1.2.1

Load transfer in a typical support system

Figure 1.2.1 illustrates the interaction between various support components in a support system, represented as a reinforced plate (beam) loaded in an orthogonal direction. This model is, however, not limited to plates and orthogonal loading directions and is applicable to other conditions as well. The conceptual model is used as a means to demonstrate the contributions of the various components in a support system. The surface support can maintain the rock mass integrity, reinforce the rock mass, and support the rock mass. Any combination of these three functions is possible, but a clear distinction needs to be made between each of them, as different design requirement are associated with each function. These support functions are described in more detail in Chapter 3. In case of a competent excavation wall (skin), no forces need to be transmitted via the fabric, and component three is associated with zero values. If the potentially unstable skin is completely detached from the “base” rock, component one will have a value of zero, as all the unbalancing forces will have to be transmitted to the retaining support (component two). This simple scenario is in fact the basis of current support design; a competent skin and full detachment from the surrounding rock mass. The surface, or fabric, support does not play a direct role in this type of support design, as it is assumed that the skin is internally stable. However, this assumption is often not valid, as the skin has a limited strength and this strength may be exceeded under certain loading conditions. However, in order to design the surface support, the reinforcing effect of the surface support has to be understood and quantified. In the case of a discontinuous and fragmented skin, the first function of the surface support is to prevent unravelling and loosening of the fragments. In doing so, the initial 14

competence and resistance of the skin are maintained. Referring to Figure 1.2.1, this implies that a minimum amount of load needs to be transferred via the surface support into the tendons. If unravelling is not prevented, the resistance of the skin will gradually decrease and the demand on surface support will inversely increase. The competency, or strength, of a skin can be quantified by clearly defining the loading and boundary conditions. In addition, the morphology of the skin and the contact conditions of the fragments have to be specified. For any realistic situation, this becomes an impossible task and the following is therefore only an attempt to illustrate the concept. The reinforcing effect of a surface liner is quantified for a few simple test cases consisting of single plates. Assuming the plate is discontinuous and that there is no bonding (cohesion) between the discontinuities, the shear resistance of the plate is controlled by the surface support. The shear load that can be supported is directly proportional to the thickness of the liner and its shear strength. For instance, a liner with a thickness of five millimetres and a shear strength of three mega Pascals could, in principle, resist a shear force of 15 KN/m. Assuming plane strain, symmetric conditions and a support span of one metre, a dead weight in excess of one metre of rock could, in theory, be accommodated by this shear resistance. This value would be higher in a practical three-dimensional configuration, as the out-of-plane resistance needs to be considered as well in that case. The bending resistance is directly related to the thickness of the plate, so that the reinforcing effect is proportionally stronger in thicker plates. A liner with a thickness of five millimetres and a tensile strength of three Map is capable of developing a tensile force of 15 KN/m. This translates into a maximum bending moment of 15 KN.h/m, whereby h is the plate thickness. For a span of one metre, the maximum distributed load that can be resisted is 120 KN.h/m 2. Thus, a plate with a thickness of 0,1 m could carry 12 KN/m 2 for instance. This translates into a dead weight of around 0,5 m of rock. The above examples ignore the contribution of additional layers underneath the skin and consider only the first layer at the surface. However, the mere presence of the first layer can have substantial reinforcing effect on the next layer, etc. By keeping the first layer in place, the subsequent layers will be able to contribute to the total resistance, and the requirements for the surface support will remain limited. While it appears to be impossible to quantify the reinforcing effect of surface-support liners in a more complex, realistic situation, a more practical solution is suggested. On the basis of the simple model shown in Figure 1.2.1, it can be argued that the load transmitted internally within the skin is controlled by the strength, or competence of the skin. Any load in excess of this strength will have to be transmitted through the surface support. The strength of the skin may or may not be enhanced by the surface support, and the strength of the skin is negatively affected by deformations that lead to the loosening of fragments. The strength of the skin is further controlled by the fragment size, shape, orientation, distribution, etc. and some classification scheme needs to be developed in order to relate these parameters to the resistance of the skin. The stronger the rock mass skin, the smaller the requirements for surface support. Energy can be absorbed in various ways. In the context of tunnel (and stope) support systems, energy absorption is often associated with yielding tendons (and props). However, this mode of energy absorption is only relevant in a relatively competent rock mass. As soon as relatively large rock mass deformations between the support units can take place, energy is absorbed locally by the fabric support and within the unravelling rock mass. In such cases, a stiff fabric support attracts large forces and small deformations, while a softer fabric will be subject to large deformations and smaller forces. The stiff fabric may therefore be more prone to failure, but is able to prevent rock mass unravelling, as long as it does not fail. The softer fabrics have a much smaller capacity to prevent rock 15

mass unravelling, but they do have a relatively large (elastic) energy absorption capacity. It is therefore important and relevant for any support-system design to decide what the prevailing rock mass conditions are and how potential energy impulses are to be absorbed in the various components of such a system. Figure 1.2.1 demonstrates the concept of load transfer and Figure 1.2.2 illustrates the two extremes in terms of energy absorption.

Disintegration:

Typical

Stiff tendon; Soft containment

Elastic response: Preferable

Figure 1.2.2

Soft tendon; Stiff containment

Energy absorption through bulging (top) and into tendons (bottom)

A situation as shown at the top of Figure 1.2.2 reflects a disintegrated rock mass that has lost its resistance against deformations. Without the presence of effective containment support, the rock mass would unravel as shown in Figure 1.2.3. The containment support needs to accommodate deformations while the rock mass itself is losing its integrity. This reflects the “worst-case” scenario in the sense that any imposed loading has to be transmitted via the containment support. Any imposed impact energy would be absorbed by the containment and the dilating rock mass; the tendons hardly contribute in this scenario. The design requirements for any containment support would obviously be extremely demanding under such conditions and failure as shown in Figure 1.2.3 can typically be expected. A response as sketched at the bottom of Figure 1.2.2 may be obtained with a relatively stiff and/or active containment support. In this case, the initial integrity of the rock mass is maintained and the rock mass will contribute to resisting imposed deformations. In principle, this should limit the requirements for the containment support. However, depending on the shape of the excavation, the associated deformations may be restricted, because only uniform deformations can be accommodated, as differential deformations within the rock mass are prevented. If that is the case, energy cannot be absorbed through large deformations and the system may thus attract large forces and be prone to violent failure. This issue requires further research, as it is extremely important to have a realistic understanding of energy absorption potential in any practical scenario.

16

Figure 1.2.3

Failure of the support system due to lack of adequate coverage

As can be appreciated from Figure 1.2.3, the weakest link in a highly fragmented rock mass is the surface, or containment, support. The rock mass is able to unravel around the primary support units, which are rendered ineffective. As will be further explained in Chapter 3, there is typically a lack of compatibility between passive surface support (mesh and lacing) and the primary support, both in terms of strength and in terms of stiffness. Active support, such as shotcrete and TSLs, can in principle maintain the integrity of the fragmented rock mass and thus provide a more viable support alternative.

1.3 TSL requirements As may be deduced from the above, TSLs are a subclass of surface support and their main function is similar to that of any other surface support, namely the provision of local stability by area coverage. While there are important differences between mesh and lacing, shotcrete and TSLs, it is important to acknowledge that common design criteria are applicable. A major distinction can be made between active surface support, such as shotcrete and TSLs, and passive surface support, such as mesh and lacing. While maintaining initial rock mass integrity is a major function of the active surface support, this function can hardly be fulfilled by passive support that only becomes activated in response to relatively large deformations and associated rock mass disintegration. The distinction between shotcrete and TSLs is more subtle and there is, in fact, an overlap between (fibre) reinforced shotcrete and thin liners with a basis of cement. The obvious difference is in the thickness of the application, as shotcrete is applied with a layer of thickness in the order of tens of millimetres and TSLs in the order of millimetres. From a simplistic point of 17

view, a TSL therefore needs to be more competent per unit area than shotcrete, in order to provide the same support capacity. It is, however, important to consider the actual application and the associated load and deformation mechanisms as described earlier. The rock mass (fragmentation, integrity) determines largely the effectiveness of the surface support. For instance, in a highly fragmented and disintegrated rock mass, a thin liner will act as a “bag”, while a thicker coating should act as a shell or a plate. The design requirements for both cases are substantially different. A large portion of this report deals with the testing of TSLs, which is integral to the quantification of their potential as surface support. The tests especially allow a comparison between the various products so that the selection process may be less biased. However, while the relevant properties of any TSL may be quantified in this way, the selection of appropriate surface support also requires a proper insight into the local rock mass conditions and the loading and deformation conditions that may be expected. Some form of rock mass rating that takes cognisance of the rock mass integrity is required, in order that the contribution of the rock mass to the resistance of deformations can be quantified. This rock mass rating system still needs to be developed and the only likely way is by empirical means. It is therefore important that cases of surface-support failure are thoroughly investigated in order to provide insight in (local) rock mass and loading conditions in relation to surface-support capacity.

1.4 Aims and objectives The main objective of the project is to provide realistic guidelines for the design and testing of Thin Sprayed Liners as rock surface support. In order to achieve these goals, a worldwide literature survey, as well as an industry-application survey was conducted. Workshops with key players, such as manufacturers, consultants, users and researchers in the South African industry were organised. International workshops were attended in Canada and South Africa, as well as international seminars, in which papers from various countries were presented. In addition, underground monitoring and testing, laboratory testing, and numerical modelling provided data that was incorporated. On the basis of this information and a classification according to mechanical behaviour, four areas of potential liner application were identified, namely: ?

Maintaining rock mass integrity through the prevention of unravelling

?

Single-block stability

?

Stability of a blocky beam/plate or shell

?

Highly fragmented, dilating rock

The interaction between a liner and the rock mass substrate was analysed for these cases and is presented in Chapter 3. Even in the relatively simple case of individual-block stability, the mechanics are quite complex and will be difficult to calibrate. None of these cases can therefore be considered as suitable for a standard testing procedure, which needs to be: ?

Simple

?

Cost effective

18

?

Repeatable

?

Practical

?

Representative of relevant properties and behaviour

For this reason, it was decided to propose two testing procedures. The first one is a modified tensile test, based on ASTM D638, which deals with the tensile properties of plastics. The second one is an adhesion test based on ASTM 4541, which deals with the pull-off strength of thin coatings (200

94.72

22.35

1.21-1.83

Rockguard

11.36

8.78

86.69

61.28

28.48

1.43-1.65

RockWeb*

13.05

9.00

>200

61.56

41.38

2.06-2.82

Masterseal

2.50

------

65.91

------

3.60

4.84-7.90

* - All specimens exhibited 50 mm limit of extension with no break of test sample occurring

2.4.3 Adhesion tests 2.4.3.1 Introduction In the past, many adhesion test procedures were applied by mining researchers to assess the characteristic bond strength of TSL products. In such an adhesion test, the adhesive or bonding strength between a TSL and the (rock) substrate is measured. Both the adhesive strength and the tensile strength are important physical properties controlling a liner's ability to adequately support loose rock between rock bolts. Two types of bond strength need to be considered: tensile and shear. Tensile bond strength is a measure of the ability of TSL to remain in contact with the rock when a tensile stress is applied normal to the rock-TSL interface. Shear bond strength is concerned with the ability to resist stresses that act parallel to the rock-TSL interface (see Figure 2.4.5). In practice, some combination of these stresses may be acting on the TSLrock interface.

Shear

Rock

Shear

Tension

TSL Tension

Figure 2.4.5

De-bonding mechanisms

Failure may occur as a result of inadequate adhesion between the TSL and the rock surface. Adhesive strength of different rock types and the factors influencing the adhesion are important test considerations. Therefore, investigations are carried out to determine 28

the effect of the following factors on bond strength: surface roughness of rock, tensile strength of rock, curing time of the TSL, and environmental conditions. The TSL’s bond to the ground normally increases with time, provided that the ground is firm, clean and dry. Adhesion to smooth, wet, and soft ground is generally poor (Tannant et al., 1999).

2.4.3.2 Core adhesion test This simple test was verbally described at the Perth workshop (2001) and found general acceptance among the researchers. The direct adhesion test consists of two pieces of core bonded together by a TSL as shown in Figure 2.4.6. Top and bottom halves are subjected to a unit-axial pull until failure takes place at the TSL rock interface. No further references or information are available on how this test is conducted; however, because of its simplicity it has the potential to become a testing method in determining the bonding strength of TSL. Sample preparation is an important issue in that both halves should lie along the same axis and in the direction of pull, to prevent eccentric loading and premature failure. The test has one obvious shortcoming in that it does not allow a natural curing of the product. The liner starts to cure from the edges of the core and the curing process progresses towards the centre of the assembly. This may not only affect the curing time but, more importantly, it may also affect the chemical processes as a result of limited exposure.

Figure 2.4.6

Core adhesion test (Spearing, 2001)

2.4.3.3 Embedded dolly Spearing and Gelson (2002) describe a wet bond strength test that emulates a wet substrate in laboratory conditions. In this test, a concrete block is cut into 25-mm-thick tiles of 150 mm2. These are then fully submerged in water for 24 hours. A block is removed from the water and wiped dry. The liner is then applied with a dolly to the cut concrete surface. The dolly is pulled off after the required time, using a standard Alcometer. Typically, three tests are carried out and the average is calculated. Considering the importance of the sensitivity of a TSL material to environmental conditions Tannant et al. (1999) developed testing procedures to measure adhesive and tensile 29

strengths of TSLs in the laboratory. They showed that high humidity or wet rock surfaces might significantly degrade the adhesive bond between the TSL and the rock. Tannant et al.’s (1999) adhesion test consists of pulling a test dolly embedded underneath the TSL. Test dollies of varying diameters (25 cm, 12,5 cm, 5,9 cm) and thickness (6,0 mm, 3,5 mm) were used. These dollies consisted of perforated steel discs that are welded to a central bolt onto which the load cell could be attached (Figure 2.4.7).

Figure 2.4.7

(a) Generic view of a test dolly and (b) Typical test set-up in the laboratory (Tannant et al. 1999)

For spray-on products, applied to irregular rock surfaces, adhesion bond variation has been shown to result from differences in substrate strength, surface roughness, porosity and degree of alteration characteristics (Archibald, 1992). Archibald (2001) modified Tannant et al.’s (1999) test in order to achieve more consistent bond strength values. To avoid problems associated with variable surface conditions when using natural rock slabs, a paving stone product was selected and used as the standard test material for all adhesion tests. The paving stone product was found to exhibit uniform strength and surface properties. Pull plates of average size relative to those utilised by previous researchers were also applied in this test procedure. Archibald (2001) recommended the following points for the test: (see Figure 2.4.8 and Figure 2.4.9): ?

A test product is sprayed onto a flat surface of a cut circular concrete paving stone or rock (30,5 cm diameter by 3,8 cm thick).

?

A test dolly (circular perforated steel plate with 10 cm diameter size and 3,2 mm thickness) is immediately placed on the fresh uncured coating. The TSL is still in its initial liquid state and permitted to seep through the numerous perforation holes of the test dolly.

?

Immediately following the initial curing, the TSL forms an adhesion bond with the test surface and produces an embedment bond about the pull plate.

?

A second coating is sprayed over the test dolly to fully embed it within the TSL.

?

After the test product has cured, the embedded test dolly and coating are overcored with a specially designed hole-cutting saw to isolate the test area from the 30

rest of the TSL. The hole cutting saw is pilot mounted to the 19 mm diameter threaded pulling collar centrally attached (Figure 2.4.9). (Over-coring of the pull plate is conducted to insure that only the adhesion bond associated with the area immediately beneath the pull plate is actually measured during pull testing.) ?

An eye-nut is threaded onto the test dolly and a load cell and a displacement transducer are attached.

?

The load cell and pulling device are centred over the test dolly.

?

After two days of curing for the last layer of the TSL the test dolly is pulled normal to the concrete surface at a displacement rate of about one mm/minute for a maximum distance of 25 mm while loads and displacements are logged.

?

This procedure is continued until full release or loss of adhesion contact between the pull plate assembly and the substrate occurs.

?

Load-deformation data is recorded and the stress from the measured load divided by the over-cored area is calculated. The adhesive strength is determined from the peak stress.

Figure 2.4.8

Adhesion pull test assembly schematic and test equipment (Archibald, 2001)

31

Figure 2.4.9

Specimen assembly prior to over coring and at finish of adhesion bond strength test (Archibald, 2001)

Archibald also recommends the following points for the test. ?

A minimum initial coating layer of a thickness approximating one mm and a second coating layer of a thickness in the region of two mm to prevent any pullthrough of the plate from the material, should excessively high adhesion bonds form at the substrate surface.

?

Curing of the TSL is permitted for a minimum of two days.

?

Ideally, the adhesive failure should occur within 90 seconds of stress application.

?

Standard surface material and a minimum population size of ten test samples per material should be used in order that a reliable measure of the adhesive strength can be obtained.

For most adhesion tests conducted, peak bond strengths were generated after 1 to 3 mm of plate displacement. Following the development of peak adhesion strength, for all sample materials, a stepwise loss of strength took place thorough extension and progressive release of localised bond attachments between the pull plate and concrete test material surfaces. The manner in which material adhesion loss occurs was shown to vary between materials assessed, as shown in Figure 2.4.10. The following failure modes are observed by Archibald (2001): 1 Partial failure to adhering layer beneath the pull plate, with the material remaining attached to the concrete slab after failure of the top encapsulating layer. This effect occurred because the top cover was placed in too thin a layer, and could not be effectively anchored over top the pull plate. 2 Full detachment of the layers from the concrete slabs at the limit of extensional failure. 3 Shear failure of the liner product at the interface between the perforated pull plate and the bottom layer of material, indicating that the material tensile strength was exceeded by the adhesive strength. 4 Failure of the lining material accompanied by localised tensile failure of the concrete surface, resulting in attachment of concrete fragments to the TSL layer after ultimate bond failure. This would indicate that the bond strength of the TSL material exceeds the concrete tensile strength at localised contact sites.

32

Figure 2.4.10

Typical views showing adhesion bond completion of failure tests (Archibald, 2001)

surfaces

after

Archibald’s test classifies the adhesive strength of a TSL, after its placement onto rock. The test is designed to permit the determination of the interface bond strength between the TSL and the (concrete) substrate. The test does not assess typical conditions that may be encountered by spraying onto rock surfaces in situ. Espley et al. (2001) performed underground adhesion testing of one particular liner on rock and shotcrete, with a range of curing times and for various moisture levels. The surface substrates were cleaned prior to liner application. Pull plates (dollies) were embedded in the liner for the adhesion testing. After the liner had cured, each test dolly was over-cored and pulled while the loads were measured. The test apparatus and an underground pull site are shown in Figure 2.4.11. The results indicate a correlation between surface moisture and adhesive strength – that is, adhesion decreases with increasing surface moisture.

Figure 2.4.11

Underground adhesion testing a) Apparatus b) Test site (Espley et al., 2001)

Archibald’s (2001) test allows for the quantification of the adhesive properties of a TSL. It offers several advantages, as it is fairly easy and practical. It was also shown that the technique could easily be used at underground sites. One of the shortcomings of this method is, however, that the dollies need to be installed before the liner is sprayed, which excludes testing of any existing application. In addition, it is questionable to what extent the properties of the liner are affected by the presence of the perforated plate, especially if the holes are relatively small. Another consideration is that Espley at al.’s (2001) test cannot be used on an existing liner and it is debatable to what extent the bond strength plays a role, as the liner mainly fails in shear and tension. This aspect of testing is further discussed in Section 2.4.4. 33

Lacerda and Rispin (2002) also tried adhesion tests at different curing times with GSM CS1251. Details of their testing are not available. Table 2.4 is a summary of adhesive strengths as a function of curing time, according to various authors. Temperature and humidity readings at the time of the tests are not included, although they play a significant role in adhesive strength.

Table 2.4

Adhesive strengths of various products and their curing time Curing Time, Hrs 1/4

8

Product

24

72

Adhesive Strength, MPa

Mineguard*

-

-

0.56

-

Rockguard*

-

-

0.43

-

RockWeb*

-

-

0.40

-

Masterseal*

-

-

0.50

-

TekFlex

-

>0.16

>0.51

>0.65

GSM CS1251‡

1.0

-

-

-

†

*Archibald (2001), †Swan and Henderson (1999), ‡Lacerda and Rispin (2002)

2.4.3.4 Glued dolly In the coatings industry, a common adhesion test is performed with a pull test machine as illustrated in Figure 2.4.12. Tapered pins are bonded to the coating with an epoxy adhesive that is cured for 30 minutes at 15 degrees Celsius. The adhesion is determined from the force required to pull a pin from the specimen divided by the area of the pin’s head. A typical loading rate is two MPa/s (Erck, 1994)

Force

Coating

Pin

Epoxy

Support

Support Substrate

Figure 2.4.12

Schematic diagram of pull-test apparatus (Erck 1994)

ASTM D4541-02 (Standard test method for pull-off strength of coatings using portable adhesion testers) specifies a method for a pull-off test. The test is based on pulling a fixture normal to the surface of the coating. An adhesive is used to attach the fixture to the 34

coating being tested. The area around the fixture is cut before load at a typical rate of one MPa per second is applied. Lewis (2001) performed adhesion tests on TSLs that were applied to granite slabs. He glued steel dollies to the liner with epoxy and then over-cored the test area. The adhesion was measured using a direct pull-off method. Tannant and Ozturk (2003) propose a direct pull-off procedure whereby a so-called elevator bolt is glued to the surface of a TSL. The coating is over-cored before the elevator bolt is attached. Load-deformation characteristics are monitored through the application of a constant displacement rate and a recording of both load and deformation. As this is a destructive test, the exposed fracture surface lends itself to visual observation so that the nature of failure can be described. The glued dolly test only yields the adhesive strength of the liner/substrate interface when de-bonding actually takes place along this interface. If failure occurs elsewhere, the test only provides a lower bound on the bond strength.

2.4.3.5 Baggage capacity test Swan & Henderson (1999) performed a TSL baggage capacity test to measure adhesive strength at different curing times. Their testing apparatus includes an open-ended steel frame, of 1,1 m x 1,1 m x 0,3 m deep (dimensions of a typical bolt spacing pattern), and loaded with actual slabs of unwashed 100 mm loose rock debris. The “loose” surface is roughly levelled and a liner is sprayed with a pre-determined six mm average thickness (see Figure 2.4.13). Since the surface is discontinuous, some penetration occurs between rock debris, making this test more conservative than one where a large intact or interlocking surface is used. After curing for the required time, the frame is inverted and placed in a loading machine. This machine applies a distributed compressive load to the “loose” rock, in this way deforming the liner, which eventually ruptures. The test measures rupture load and maximum deformation of the distending liner. Load capacity is found to increase by 115 per cent, when the curing period is increased from eight hours to 24 hours. The authors also agree that tension and adhesion strength properties contribute to the loose-rock-supporting capacity of a deformable TSL. Repeatability of this test is questionable since the distribution of rock debris varies for each test. Preparation for a test appears to be difficult and time consuming because of the size involved. With this type of test, it is also debatable to what extent the adhesive strength plays a role, as the load-bearing capacity is controlled mainly by the tensile properties of the liner. This test should therefore not be classified as an adhesion test.

Figure 2.4.13

Baggage Load Test Frame & Set-Up (Swan and Henderson, 1999)

35

2.4.4 Punch type of testing 2.4.4.1 2.4.4.1 Introduction TSLs are expected to support the weight of loose rock. TSLs’ adhesive strength and tensile capacity are necessary for bridging and carrying loads from the loose rock to the surrounding stable rock and, therefore, are inhibiting loose rock movement away from coated surfaces. Figure 2.4.14 illustrates how both adhesion and tensile stresses are involved in supporting a loose block of rock, which might otherwise be able to become detached.

Loose Rock

Tension

Adhesion TSL

Figure 2.4.14

Liner supports the load from loose rock

2.4.4.2 Plate pull test Tannant et al. (1999) designed a plate pull test to simulate, in a simple manner, the scenario shown in Figure 2.4.14. This test can be performed either in the laboratory, using slabs of concrete, or in situ, where the plate is placed on the wall of a drift (Figure 2.4.15). In both cases, the plate is coated with the TSL. The test consists of placing a solid circular plate of steel (with an attached central threaded stud) on either a concrete block or rock surface and then spraying the test material over the plate and the substrate surrounding the plate with a uniformly thick TSL. No TSL is placed between the substrate and plate, as it is not the aim of this test to measure the direct bonding strength of the TSL.

36

Figure 2.4.15

Final stages of a plate pull test showing tensile rupture in the TSL and adhesion loss with the rock (Tannant et al., 1999)

Tannant et al.’s (1999) plate pull test consists of the following procedures: ?

Select a relatively planar area on the rock or concrete surface with a diameter greater than the pull plate.

?

Place the pull plate on the concrete or rock surface or secure the plate to the rock wall.

?

Coat the pull plate and the area surrounding the plate with TSL test material (the thickness should be in the order of 5 to 8 mm) and let the material cure. The coating should extend at least 250 mm beyond the edge of the pull plate.

?

Attach a loading device, load cell, and displacement transducer to the central stud on the pull plate.

?

After the required curing time, slowly pull the plate perpendicular and away from the substrate while measuring loads and displacements. (A rate of 10 to 50 mm/minute is suggested.)

The test is completed when the load begins to drop or when the plate is pulled free of the substrate. The nature of the failure process and damage to the TSL should be documented. The mode of failure should not involve shear through the TSL. Instead, a combination of adhesion loss and tensile rupture is the expected and desired ultimate failure mode. TSL performance is sensitive to its applied thickness and continuity over the rock surface; for this reason, Tannant et al. (1999) draw attention to the importance of achieving a continuous and uniformly thick material over the sprayed surfaces during the application of TSL for ground support purposes. Archibald (2001) also assessed the material pull strength to demonstrate the combined strength properties of TSLs that can be mobilised both by liner adhesion to rock surfaces and by the tensile strength of the liner material. He describes that the bonding strength at the perimeter of the displacing block provides a liner anchoring capability, as the liner will be attached to both the displacing block and the adjacent solid rock. TSL thickness and 37

tensile strength assist in providing support resistance about the perimeter of the displacing block. Archibald’s test procedure is similar to Tannant et al.’s procedure with only minor differences. Archibald used a 30,5 cm diameter by 3,80 cm thick circular concrete paving stone as the substrate. A minimum of 20 sample tests was conducted for each TSL material. Pull plates were manufactured to be ten centimetres diameter by 3,2 mm thick, with a 19 mm diameter threaded pulling collar centrally attached (Figure 2.4.16 and Figure 2.4.17).

Figure 2.4.16

Pull strength test assembly schematic and test equipment (Archibald, 2001)

A minimum coating layer thickness of two millimetres is recommended. The lining layer is applied so that the entire surface area of the concrete substrate is covered, leaving ten centimetres of attached liner surface beyond the perimeter of the pull plate. Curing of liner material is permitted for a minimum of two days after pull plate placement. The plate is displaced for a maximum distance of 50 mm at a recommended rate of five mm/min.

Figure 2.4.17

Test assembly before and after the pull strength test (Archibald, 2001)

38

Table 2.5

Pull strength properties of TSL materials (Archibald, 2001)

Liner material

Liner thickness and range (mm)

Initial pull force and range (KN)

Initial peak pull stress and range

Average deflection @ peak pull stress

Average deflection @ release

Mineguard

2.97

3.15

3.28

2.60

37.0

(2.13-3.71)

(1.68-4.66)

(2.16-4.58)

(1.13-5.0)

(5.0-50*)

2.85

2.40

2.65

1.58

24.6

(1.84-3.98)

(1.41-4.22)

(2.16-4.52)

(0.81-3.23)

(7.0-50*)

4.85

2.38

1.56

1.47

48.9

(3.53-6.27)

(1.20-3.40)

(0.79-2.32)

(0.48-2.26)

(42.0-50*)

4.33

1.42

1.22

2.38

35.8

Rockguard Rockweb Masterseal

(1.69-8.96) (0.77-2.03) (0.44-3.74) (0.94-6.45) (9.0-50*) * - indicates plate displacement of 50 mm maximum travel without liner release from plate The pull strength results listed in Table 2.5 and shown in Figure 2.4.18 indicate that considerable variability occurs in both load-deformation response and the mechanisms of pull failure between the different TSL materials. The initial pull strength is determined by dividing the initial peak pull force by the liner material area existing about the perimeter of the pull plate at the start of the test. The liner area is calculated as the plate-perimeter length times the measured thickness of the liner at the plate perimeter.

39

Mineguard

Rockguard

RockWeb

Masterseal

Figure 2.4.18

Typical views showing pull test failure conditions at the completion of failure tests (Archibald, 2001)

The failure modes observed during Archibald’s (2001) pull test can be seen in Figure 2.4.18 and are summarised as follows: 1 Failure of adhesion bonds about the plate perimeter and/or tensile failure through the material. Then, immediate and progressive shearing follows with a gradual loss of pull resistance. 2

Tensile shearing and gradual adhesion bond release beyond the plate perimeter.

3 Gradual and progressive release of material adhesion with no tensile shearing. Total loss of pull capacity develops only after very large plate displacements and full release of the layer coating from the outer perimeter surface of the concrete paving stone. 4 Shearing failure of the pull plate through the overlying TSL coating due to strong adhesion bonding and brittle tensile deformation behaviour.

2.4.4.3 Large-scale plate pull test Espley et al. (1999) assessed the load-carrying capacity of a TSL by coating an interlocking series of 50 mm thick hexagonal concrete paving blocks. The TSL is applied 40

onto the concrete blocks from above and left to cure for about an hour to test reactive TSLs and between four and eight hours for most non-reactive TSLs. A pull-type loading is applied by a 300 mm square steel plate located in the centre and underneath the assembled paving blocks until the TSL has failed, as illustrated in Figure 2.4.19. These researchers decided that TSL is able to enhance the interaction between the loose blocks, and thus a significant portion of the supporting function arises from block-to-block interaction. The TSL also functions like a suspension bridge in order to carry load while deforming.

Square steel plate

In principle, this test has quite a lot in common with the baggage capacity test of Swan and Henderson (1999). As indicated previously, that test also involves a combination of tensile and bonding stresses and is not strictly an adhesion test.

reaction legs

Reaction legs

Figure 2.4.19

Test set-up and typical test results from a large-scale pull test on Mineguard coated concrete blocks (Espley et al., 1999)

2.4.4.4 Punch-through test Spearing et al. (2001) developed a laboratory load-displacement test method to account for the combined effects of tensile, tear, shear and bond strengths, including elongation properties of TSLs and to provide performance data for evaluating such TSLs. The socalled MDT Method (Membrane Displacement Test), provides load and displacement data on TSL performance. The main components of the test frame, as shown in Figure 2.4.20, are a concrete slab of 610x610x51 mm size supported on four concrete cylinders of 102 mm diameter by 204 mm in height, a loading plunger, and a sleeve to hold the plunger in place during frame set-up. The TSL is punched by the plunger at the end of a hole in a concrete slab as illustrated in Figure 2.4.20. This test is very similar to the plate pull test in terms of the movement of the TSL, i.e. punching or pulling effectively results in the same TSL behaviour. However, one important difference is that the liner is not able to penetrate between pull plate and substrate in the plate pull testing, while this possibility is not necessarily excluded in the punch-through test. Both short- and long-term tests can be performed with the test set-up.

41

Plunger Concrete slab TSL

Support cylinder

Figure 2.4.20

Punch testing set-up (Spearing et al., 2001)

2.4.5 Compression failure tests on coated samples

Uncoated

TSL-coated cylinders of concrete and rock were tested by various researchers (Espley et al., 1999; Archibald and DeGagne, 2000) to demonstrate a TSL’s ability to contain and reduce the damage resulting from potential pillar-bursts. Tests were done under axial-axial loading conditions and the results demonstrated significant benefits at the laboratory scale in terms of non-violent post-peak failure response, and the liner’s ability to absorb some of the stored strain energy (Figure 2.4.21).

Figure 2.4.21

Coated and uncoated cylinders are tested to failure (Espley et al., 1999; Archibald and DeGagne, 2000). 42

While the results of these tests may be more difficult to interpret in terms of TSL design, they demonstrate how limited support could have a relatively large effect in terms of stability and load-carrying capacity of the supported rock.

43

3. The functions of surface support 3.1 Introduction As was indicated in the introduction in chapter 1, surface support can fulfil various functions. These functions depend on loading conditions and the structure of the immediate, and potentially unstable, rock mass skin around the excavation. The stability of this rock mass skin is initially controlled by its own competence and loadbearing resistance. Sliding and opening of existing fractures, relaxation of confinement, weathering, etc. can cause the competence and load-bearing capacity of the discontinuous skin to reduce with increasing time and deformation. The application of surface support can prevent these processes from taking place, either by effectively sealing the skin or by providing a minimum amount of resistance to the surface. No large support resistance is required at this stage and the main requirement is that the surface support remains in place. The sliding and opening of fractures only occurs near the surface and is not associated with general skin failure. By preventing this type of immediate surface failure, further unravelling can be avoided. It is possible that relatively large and unstable blocks are part of the skin. Such blocks can be stabilised by the surface support through bridging and filling the discontinuities that are exposed around the blocks on the excavation surface. In this case, the requirements for the surface support can be quantified in terms of strength, stiffness, and bonding capacity. A special case is the creation of a composite beam, consisting of the liner and one or more layers of rock fragments. The presence of bedding planes promotes the formation of discontinuous beams in the hanging walls of tunnels and stopes. Surface support can greatly enhance the stability and resistance against deformation of such (discontinuous) beams. If deformations of the rock mass cannot be prevented, and fragments start losing their interlock, it is still possible for the surface support to limit the loss of integrity and in addition act as reinforcement and support. Once complete loss of interlock has taken place, the surface support acts in total isolation and its (limited) support capacity is easily determined. A conceptual model, which caters for the interaction between the various support components and the fragmented rock mass is presented here.

3.2 Maintaining initial integrity Fragmented rock can still have a relatively large resistance against deformations. Terms such as “unravelling”, “interlock”, and “integrity” are often used to describe the mechanisms associated with discontinuous rock mass behaviour. As these concepts may not be very well defined and can be interpreted in various ways, an attempt is made here to demonstrate the principles associated with rock mass integrity. The scaling of rock around an excavation is believed to be closely related to the issues of rock mass unravelling and rock mass integrity. It may be of interest, therefore, to investigate the scaling operation in more detail. If loose rock is removed from the surface of an excavation, the potential for rock falls is obviously reduced. In an interesting experiment on scaling, Kuchta et al. (2003) found that “the longer one scales, the more material can be brought down”. In their experiment, they used a combination of hand scaling with a scaling bar and water-jet scaling. While the hand-scaled material was 44

relatively large, the water-jetted material contained a significant amount of fines. By successively applying these methods, more and more material could be scaled down at each subsequent operation, even though the area appeared to be properly scaled after each scaling operation. The unravelling process may be compared to a slower, more natural and more complex scaling process, but the basic mechanics must be very similar. As a suggestion, it could be very useful to conduct a controlled scaling operation in which the scaled material is monitored in such detail that the original fragmented geometry of the excavation skin can be reconstructed in a three-dimensional model. In such a way, better insight could be obtained with respect to the mechanics of unravelling and the associated surface support requirements. Numerical models could also provide insight into the relationship between unravelling/disintegration and competency. Such models can be expected to be sensitive with respect to the size, shape and distribution of fragments. The issue of weathering is also closely related to scaling and unravelling. By effective sealing of the rock surface, weathering processes may be inhibited and initial conditions may be maintained. While this is considered an important aspect of the functioning of a TSL, it has not been further explored within this particular project, which concentrates on structural behaviour.

3.3 Single-block stability Providing stability to a single block appears to be the most straightforward application of surface support. The weight of the block or, more generally, the momentum of the block needs to be balanced by the surface support. In its simplest form, this is achieved by the total shear resistance offered by the surface support along the circumference of the block. For example, a surface coating with a thickness of 5 mm and shear strength of 2,5 MPa, would be able to offer a shear resistance of 12,5 KN/m. A rectangular block with base dimensions of a and b and a density of two-and-a-half thousand kilograms per cubic metre can be stabilised if its height is less than: (a+b)/ab. The graph in Figure 3.3.1 shows the relationship between maximum block thickness and the ratio b/a. 10

b/a=0.2

9

b/a=0.5

Height (m)

8

b/a=1.0

7 6 5 4 3 2 1 0 0

0.5

1

1.5

2

Base length a (m)

Figure 3.3.1

Maximum block height in relationship to base dimensions (dead weight only)

45

From Figure 3.3.1, it can be appreciated that relatively large blocks can theoretically be supported by a thin surface-support liner. This example illustrates the potential of TSLs. However, the assumptions used for this example may be too simplistic for many applications, and additional complexities may need to be considered. Firstly, the stress distribution around the circumference of a block is most likely not uniform, so that stress concentrations occur. Depending on the properties of the TSL, such stress concentration can induce local failure that may or may not be contained. In addition, the shear strength may be subject to creep, in which case lower values are applicable. Furthermore, the liner may be subject to de-bonding and associated bending and stretching. This could cause premature failure if the resistance against bending and/or stretching is smaller than the resistance against shear. The possible response of a liner that is subjected to punching under axi-symmetric conditions was analysed as part of this project (Kuijpers and Toper, 2003) and the results can be related to mechanisms associated with block stability. The analysis consisted of laboratory tests and numerical models in which TSL properties such as material stiffness, interface stiffness, material strength and bond strength were varied. Figure 3.3.2 depicts the geometry of the model. Post-failure behaviour was assumed to be perfectly plastic and no hardening or softening was allowed for. On the basis of these studies, it was found that the de-bonding process could be unstable if the membrane stiffness is relatively high and the de-bonding length small. If no de-bonding is allowed, the post-failure behaviour is directly associated with the post-failure response of the liner itself.

100

100

5 Controlled displacement

Figure 3.3.2

Interface Membrane

Cross section of the axi-symmetric model (half symmetry)

Some of the results are shown in Figure 3.3.3 and Figure 3.3.4, where it can be seen how bonding affects the load-deformation characteristics during punching. If no de-bonding is allowed, the punch resistance is completely controlled by the membrane strength and deformation capacity (Figure 3.3.3). However, during de-bonding, the resistance against punching can increase, because the contact area is increasing during the de-bonding process (in the axi-symmetric situation). This case is shown in Figure 3.3.4

46

18.00 16.00

Force (kN)

14.00 12.00 10.00 8.00 6.00 4.00 2.00 0.00 0.00

5.00

10.00

15.00

20.00

Displacement (mm)

Figure 3.3.3

No de-bonding (perfectly plastic material)

30.00

Force (kN)

25.00 20.00 15.00 10.00 5.00 0.00 0.00

5.00

10.00

15.00

20.00

Displacement (mm)

Figure 3.3.4

Partial de-bonding (perfectly plastic material)

47

The results are summarised in Figure 3.3.5 and Figure 3.3.6, where it can be seen how material strength, bond strength, material stiffness, and interface stiffness affect the loaddeformation characteristics. The model is still relatively simple, as all these properties are assumed to be independent of deformation and time. When these results are interpreted, it should be recognised that the stiffer materials are in general more brittle and that the softer, more ductile materials often show time-dependent behaviour. bond =10MPa; mem=10MPa

Low membrane stiffness (70MPa)

bond=10MPa; mem=5MPa

25

bond=&; mem=10MPa bond=3MPa; mem=5MPa

Force (kN)

20 15 10 5 0

0

5

10

15

20

Displacement (mm)

Figure 3.3.5

Punch resistance with varying membrane properties (soft membrane)

It was found that the highest bond strength is always associated with the highest (initial) punch resistance, whereby stiffer liner material in combination with low interface stiffness offers a larger resistance than a softer material in combination with a stiffer interface. This is associated with the distribution of bond stresses as illustrated in Figure 3.3.7, where relatively high stress concentrations are associated with a soft membrane and a stiff interface. These results suggest that the stiffer liner materials offer a better punch resistance. The modelling results confirm this finding as well. A similar conclusion was reached by Kuchta et al. (2003), who found that “poor interface strength can be overcome by added thickness of shotcrete”. Their conclusion was based on the results of numerical modelling, but no physical interpretation of the results was provided. High membrane stiffness (7GPa)

bond=10MPa; mem=10MPa bond=10MPa; mem=5MPa bond=3MPa; mem=5MPa

35

bond=&; mem=10MPa

Force (kN)

30 25 20 15 10 5 0 0

5

10

15

20

Displacement (mm )

Figure 3.3.6

Punch resistance with varying membrane properties (stiff membrane) 48

Once de-bonding has initiated, the stiffer membranes (Figure 3.3.6) show a pronounced decrease in resistance, while that is not the case with the softer membranes (Figure 3.3.5). The decreasing resistance is associated with the de-bonding process and the associated stress distributions.

Bond stress

1. Stiff membrane, soft interface rock liner

Bond stress

2. Soft membrane, stiff interface rock liner

Figure 3.3.7

Bond stress distribution affected by stiffness

The numerical modelling results were also used to establish relationships between interface stiffness, material stiffness, material strength, and bond strength. The following graphs show these relationships.

Bond strength 10MPa 60

1.00E+08 1.00E+09

50

Yield (kN)

1.00E+10 40

1.00E+11

30 20 10 0 0

1

2

3

4

Log E (MPa)

Figure 3.3.8

Effect of membrane and interface stiffness on punch resistance (yield)

49

The graphs in Figure 3.3.8 show the punch resistance at the initiation of de-bonding (yield strength) as a function of membrane stiffness for a variety of interface stiffnesses. From this graph, it can be observed that the membrane stiffness affects the yield strength only if the interface stiffness is relatively low. Figure 3.3.9 shows how the critical bond strength, i.e. the bond strength required to prevent de-bonding, depends on all the parameters that have been considered thus far, namely the interface and membrane stiffness and strength. As can be appreciated from Figure 3.3.9, the effect of membrane strength increases with increasing interface stiffness, as does the critical bond. As was indicated earlier, a relatively stiff interface generates larger maximum bond stresses and thus requires larger bond strength as well. If the membrane strength increases, the critical bond strength increases proportionally, as increased membrane strength will be associated with larger adhesive forces.

Critical bond strength

Bond strength (MPa)

30 25 20 15 10 E=7GPa 5

E=700MPa E=70MPa

0 7

7.5

8

8.5

9

9.5

10

10.5

11

Interface stiffness (log Kn )

Figure 3.3.9

Effect of membrane and interface properties on critical bond strength

If high stiffness is associated with more brittle material behaviour, it is likely that stress concentrations will trigger premature failure in such brittle materials. On the other hand, if long-term loading is of relevance, the softer materials may fail at relatively low loads because of their tendency to creep. The response of two liner materials in a laboratory punch test is shown in Figure 3.3.10. Material A is relatively soft, while material B is stiffer and more brittle. Both were subjected to punching similar to that simulated with the numerical models. As can be observed, the stiff material, B, can resist a punch load of 15 to 20 kg, after a deformation of around two-tenths of a millimetre. No de-bonding was observed and any additional deformations must be associated with post-failure behaviour of the material itself. The softer and more ductile material, A, was observed to de-bond, which enabled it to develop gradually more resistance until final material failure. During the test of material A, the punch was unloaded after approximately ten millimetres of deformation and subsequently reloaded. The stiffer and stronger response may be associated with the faster rate of loading during reloading. The bond strength of material A was found to range between 1,8 MPa and 2,0 MPa and that of material B between 2,4 MPa and 3,0 MPa. The tensile strength of material B is around 1,5 MPa, while the strength of material A depends on the loading rate and can be expected to range between 2,0 MPa and 3,0 MPa in this particular test. 50

Punch material B

100 90 80 70 60 50 40 30 20 10 0

25 20

Load (kg)

Load (kg)

Punch material A

15 10 5 0

0

5

10

15

0

20

0.5

Displacement (mm)

Figure 3.3.10

1

1.5

2

2.5

Displacement (mm)

Results of laboratory punch tests: A) soft, B) stiff

3.4 Composite beam The composite beam is used to illustrate a simple reinforcing mechanism of a surface support. When the excavation is curved and convex, beam action is irrelevant as the rock mass skin has the shape of a shell in which compressive stresses are generated. However, in the case of straight excavation walls, any external load will lead to bending and the induction of both compressive and tensile stresses. The surface layer of a fragmented excavation wall can be considered as an assembly of blocks, which cannot transmit any tensile stresses. This layer is particularly unstable without any additional support or reinforcement, as it is not able to resist bending moments. If the surface layer is coated with a TSL, the situation can be substantially improved, because the TSL is able to absorb the tensile forces, as is shown in Figure 3.4.1.

F M Slab

T

w

t Yielding membrane Figure 3.4.1

F

Stress distribution through a section of a coated layer

The maximum bending moment depends on two parameters: namely the maximum tensile force that can be supported by the membrane and the thickness of the layer (slab):

w? ? M max ? Fy ?T ? ? , 3? ? whereby:

3.4.1)

Fy represents the maximum tensile force in the membrane, T is the thickness of the surface layer, and

w equals the height of the compressive zone in the surface layer

51

The height of the compressive zone depends on the relative stiffness of membrane and rock. With a relatively large rock stiffness, this height can conservatively be estimated to be half that of the thickness of the surface layer. Therefore:

M max ?

5 Fy T 6

(3.4.2)

In the case of a simply supported slab or beam, gravitational loading will involve the following relationship:

l 2 ? 2.67 Fy ,

3.4.3)

whereby l is the span of the slab or beam and a rock mass density of 2 500 kg/m 3 was assumed. Using a yield force of nine Newton per millimetre (yield strength three MPa and coating thickness three millimetres), the maximum stable span is 4,9 m. Under dynamic conditions, the slab will be subjected to accelerations and, depending on the stiffness of the coating, the maximum stable span will decrease. With a relatively stiff coating, higher forces will be induced during dynamic loading, as the momentum of the slab needs to be absorbed more abruptly. Softer and tougher coatings will therefore be more effective under such conditions as smaller forces are required for energy absorption. The reinforced slab or beam can further be treated as a single block in a similar way to that discussed in Section 3.3, and the shear resistance of the TSL can thus be investigated.

3.5 Highly fragmented dilating rock The effect of a limited amount of confinement on rock that is dilating, in response to excess loading, was observed on quartzite discs in laboratory experiments (Wojno et al., 1998). From these observations it was concluded “that, in principle, it is possible to stabilise the post failure response of brittle pillars by the use of a structural membrane”. While it was also emphasised that scale effects may reduce the benefits for larger, in situ applications, this reinforcing mechanism may be of practical interest. Theoretical limit equilibrium analyses (Barron, 1982) demonstrate how the residual strength of broken (fragmented) rock affects the vertical stress distribution in the failed zone of a pillar abutment. In fact, in the absence of a residual stress, no vertical stress can be supported by the broken rock in the failed zone. However, even a limited residual strength leads to an exponential build-up of resistance with depth into the failed zone of the pillar abutment: Vertical stress= UCSB *W ,

(3.5.1)

where UCSB is the residual uniaxial compressive strength of the broken, fractured rock,

W ?

?1 ? sin ? B ?e ?

? 1 ? sin ? B , ? B is the friction angle of the broken rock, 2 sin ? B 52

? 1 ? sin ? B ? 1 ? sin ? B

and ? ? tan ? B ??

?? x ? ??? ? , where 2 h is the height of the excavation and ?? h ?

x is the distance into the abutment. This equation is graphically displayed in Figure 3.5.1 for an excavation height of one metre and for internal friction angles of 30 and 45 degrees.

Ratio of vertical stress to residual strength

Limit equilibrium 160 140 120 100 80 60 40

30 degrees

20

45 degrees

0 0

0.5

1

1.5

2

2.5

3

Depth into abutment (m)

Figure 3.5.1

Development of vertical stress into a pillar abutment (after Barron, 1982)

The importance of the residual strength ( UCSB ) can be appreciated from the above. Even a minimal strength at the surface of the excavation would rapidly lead to the build-up of substantial resistance further into the rock mass. Using the example in Figure 3.5.1, a residual strength of one MPa would result in a vertical resistance (strength) of 37 MPa at a depth of two metres, if the internal friction angle of the broken rock is equal to 30 degrees. With an internal friction angle of 45 degrees, a strength in excess of 250 MPa would already be reached at a depth of one metre. However, the residual strength of broken rock is often not a reliable parameter, because of weathering and scaling. This means that, in practice, residual strength cannot be used as a design parameter. By installing support and/or reinforcement at the surface of a fragmented excavation wall, it is possible to generate a more permanent residual strength. This is achieved by inducing confining stresses that in turn provide additional material strength according to:

UCSC ? ?

c

1 ? sin ? B 1 ? sin ? B

(3.5.2)

where UCSC is the additional (residual) material strength and ? induced by the support.

c

is the confining stress

The effect of tendons for instance can clearly be evaluated in this way. Although the exact reinforcing mechanism of a TSL may be more complicated and debatable, the effect of providing some initial residual strength at the surface of an excavation can be substantial and therefore of great practical relevance. In this respect, it is important to recall that a more permanent residual strength is required. This implies that the support should 53

function under all circumstances, even when large deformations need to be accommodated. This is where the more flexible liners can contribute, as relatively stiff support systems would fail under similar conditions. Dilating pillars would cause stretching of the surface support and it is essential that the support remains operative. Therefore, the exact nature of the stretching needs to be known; i.e. if it occurs locally, across individual cracks, the amount of crack opening is important for determining induced strains. Numerical models (Wojno et al., 1998) have confirmed the analytical model and this model can be used to investigate the need for and potential of any surface reinforcement and/or support.

3.6 Conceptual model 3.6.1 Introduction In order to provide stability around an underground excavation, two principal support mechanisms are available. The first is based on retaining and containing the unstable rock mass skin, while the second is based on the creation of a stable shell. Retaining is achieved by ensuring anchorage into stable rock, beyond the limit of unstable rock. Effectively, the unstable, fractured and fragmented rock mass is pinned to a substrate of solid rock. This is the most common support philosophy in a hard rock mining environment and the underlying assumption is that, beyond a relatively limited skin of unstable rock, stable elastic conditions prevail. The creation of a stable shell can be achieved in various ways. Such a shell can be an artificial structure, which might be required when the rock mass itself is so incompetent that is cannot be effectively reinforced. Such conditions are mainly encountered in weak rock and soils. Tendons, for instance, can reinforce rock that is more competent and a stable shell may be generated within the skin of the excavation. These shells are capable of providing an effective support pressure and are, in principle, suitable for controlling unstable closures in an otherwise unstable excavation. In most support systems, some combination of these two principal support mechanisms may be encountered (Haile, 1999). The concept of the creation of a structurally competent reinforced rock mass arch is the basis of most of the empirical rockbolt design methods (Stillborg, 1986) and thus generally applies in South African gold mines as well. The retaining mechanism allows for a straightforward and relatively simple support design methodology. In its simplest form, it is assumed that the tributary area theory can directly be applied to the tendons, which are supposedly anchored in stable rock beyond the limits of the fractured and fragmented skin (Wojno and Jager, 1987; Stilborg, 1986). Rockbursts being the most severe loading condition, the tendons are designed to absorb the energy associated with the arresting of a certain volume of unstable rock that has been accelerated to a certain velocity. The same principle is applied in stope support design (Wagner, 1983; Wojno, Jager and Roberts, 1987). While this method allows the selection of appropriate tendons with the required characteristics, case studies (Haile, 1999; Güler et al., 2001) suggest that the typical support failure is not associated with failure of tendons, but rather with the unravelling and ejection of rock in between such tendons. The tendons are not subjected to the assumed tributary area loading and hence even nonyielding tendons can survive rockbursts. Although the presence of mesh and lacing provides a local support function in the areas between tendons, it is obvious that this support function is relatively limited when compared to that of the tendons themselves. The support components within a typical tunnel support system are not compatible, which 54

often results in premature failure of the mesh, being the weakest link in the system (Haile, 1999; Güler et al., 2001). Previous studies on the performance of mesh and lacing (Güler et al., 2001; Kaiser et al., 1996; Ortlepp, 1983; Stacey and Ortlepp, 1997) have provided relevant data. It remains important, however, to consider the boundary and loading conditions that were applied in these studies. Mesh and lacing are relatively simple components whose performance is controlled entirely by their tensile strength and in-plane stiffness. Shear forces are not transmitted directly, as bending moments cannot be accommodated. The bending stiffness and bending resistance of these support components is negligible in comparison. In the Canadian Rockburst Support Handbook (Kaiser et al., 1996), recommendations are made with respect to the requirements for tunnel support under rockburst conditions. The function of the holding elements (i.e. the tendons) includes the absorption of energy under severe conditions, according to this handbook. It is also mentioned that individual holding or retaining elements are ineffective in preventing rock falls between directly supported areas. The requirements for the containing elements are not quantified, but the ability to absorb kinetic energy and withstand impact loads is mentioned. Pull tests were conducted to measure the load-deformation characteristics of mesh and shotcrete. Mesh-reinforced shotcrete was found to be more efficient at absorbing energy than mesh alone and, due to its initial high stiffness, has a better load-bearing capacity compared to regular mesh at small deformations (Kaiser et al., 1996). As high stiffness and energy-absorption capacity are mutually exclusive, it is unclear which property, under which conditions, is thought to be desirable. From interpretation it appears that an initial stiffness and strength is useful to cater for less severe conditions, while large energy-absorption capacity, associated with large bulking of the rock mass, is recommended for more severe cases. The tributary area theory can be applied to the design of tunnel and stope support if the rock mass between the individual support units is sufficiently strong and competent. This is generally the case if the skin of potentially unstable rock can be viewed as a solid continuous shell or plate. However, if this skin of potentially unstable rock is fragmented, it is necessary to provide additional support in order to stabilise the rock mass in between the primary retaining support units. Within the conceptual model, the primary, or retaining support units are “conventionally” designed, based on the tributary area. However, in addition to the design of the primary support, the surface, or containing support, is also designed. This design is based on the supporting capacity of the fragmented excavation wall, as well as on the design of the primary support units. In an extreme scenario, the skin of potentially unstable rock is so highly fragmented that is does not offer any resistance against shear and tensile deformations. An extremely highly fragmented and unravelled rock mass may closely resemble such a scenario, and the case is therefore not completely academic. Under these conditions, the secondary surface support has to be designed to absorb and transmit all the unbalancing forces, which are generated within the volume of unstable skin between the primary retaining units. Under such conditions, a TSL will offer limited benefit, as its stiffness and strength are relatively small. A relatively thick layer of shotcrete can, in principle, act as a shell or plate, and thus transmit shear forces by itself. Under less extreme conditions, the TSL can be envisaged to interact with the fragmented rock mass and in such a way create an effective shell or plate consisting of a layer of TSL bonded to the fragments that are exposed at the excavation surface. In addition, the TSL may penetrate open cracks and in such a way create bonding of fragments underneath the surface as well.

55

3.6.2 Description and examples Figure 3.6.1 is similar to Figure 1.2.1 and shows the distribution of (inertia) forces through a system comprising primary holding or retaining elements, surface or containing support and the potentially unstable rock mass itself. Loading is assumed to be in an orthogonal direction and is associated with inertia due to dynamic excitation or even plain and simple gravity (hangingwall). Within the conceptual model, the internal distribution of forces is quantified so that the loading of the surface support can be expressed as a percentage of the total inertia forces. The tributary area principle can be used to design the primary retaining elements, and the selection of the secondary surface support can be directly based on that design. Typically, the potentially unstable volume of rock is determined from an empirically established instability depth in combination with the actual spacing between retaining elements.

contributory area

contributory area

tendon spacing

fracture zone depth fabric forces tendon forces Figure 3.6.1

Distribution of inertia forces in the unstable rock mass skin

Within the conceptual model, both the load distribution and the energy absorption can be taken into account. The crucial component with the conceptual model is the contribution of the (fragmented) rock mass. In order to simplify matters, this contribution can be neglected and the integrity of the rock mass is thus assumed to be completely lost. This is clearly a “worst-case” scenario for which surface support requirements will reach an upper limit. Referring to Figure 3.6.1, the disintegrated rock is assumed to be located within the “fracture zone” (the potentially unstable volume of rock) and is supported via the fabric, or surface support, by the tendon forces. In order for the surface support not to fail prematurely, its strength has to match the tendon strength. According to the tributary area concept, each tendon supports a mass of:

h *l 2 * ? ,

(3.6.1)

where h is the fracture zone height, is the density of the rock mass.

l is the spacing between the tendons, and ?

During dynamic excitations, this mass may be subjected to accelerations that cause the tendons to yield. The required yield forces are calculated on the basis of an energy balance, whereby it is assumed that the mass of unstable rock has been accelerated to a certain maximum velocity and needs to be arrested within a certain limited distance: 56

mv 2 mv 2 ? F y? , or Fy ? , 2 2?

(3.6.2)

where ? is the maximum deformation distance, or yield length. In this example, it is assumed that gravitational loading does not take place (side wall) and that the yielding force remains constant during the deformation process. An empirical value of v = 3 m/s is often used as a maximum velocity, while ? ? 0,5m is seen as a practical value for maximum allowable deformation. With a support spacing of one metre, a density of 2 500 kg/m 3 and a fracture zone height of one metre, the minimum required yield force would be 22,5 KN. In practice, larger yield forces are easily obtained and it is important to realise that this directly affects the design of the surface support. Assuming that the surface support and the fragmented rock mass are subjected to uniform deformations, the maximum shear traction and maximum bending moments within the surface support can be determined: Maximum shear stress:

Fy 4Dt

,

(3.6.3)

where Fy is the tendon yield force, D is the dimension of the bearing plate and t is the thickness of the liner. Maximum bending moment:

Bending stress: ? ?

Fy l

(3.6.4)

12

Fy

(3.6.5)

2t 2

Of interest is the fact that these stresses are independent of the span between the tendons. This can be explained from the fact that the maximum tendon forces are independent of the span, and the (uniformly) distributed load thus reduces proportionally with increasing span. These are, however, theoretical calculations, based on elastic stress distributions. Results that are more realistic may be obtained from laboratory tests on panels, similar to the EFNARC test (1996) Example 1 Assuming the maximum tendon force is ten ton (100 KN), the tendon span is one metre and the dimension of the bearing plate is 50 mm, then Table 3.1 is applicable:

57

Table 3.1

Induced stresses associated with “beam” behaviour

Liner thickness (mm)

Maximum shear stress (MPa)

Maximum bending stress (MPa)

5

100

2000

10

50

500

25

20

80

50

10

20

100

5

5

From this (simplified) example, in which the fragmented rock mass offers no resistance, it follows that at least a 100 mm-thick liner is required to match the yield strength of a typical tendon under uniform deformation conditions. If the rock mass is allowed to deform nonuniformly, the liner acts as a bag, similar to mesh and lacing. Under such conditions, and again assuming that the fragmented rock mass offers no resistance against deformation, the load-bearing capacity of the liner is determined by its ability to accommodate deformations and its tensile strength. Equation 3.6.3 is still applicable to the maximum shear stress, while the following equation roughly reflects the tensile stresses

Fy l Maximum tensile stress: 12? Dt ,

(3.6.6)

where d is the centre deflection of the panel. Example 2 Assuming maximum tendon force 100 KN, span one metre, and a bearing plate of 50 mm, then Table 3.2 applies.

Table 3.2

Induced stresses associated with “bag” behaviour

Liner thickness (mm)

Centre deflection (mm)

Maximum shear stress (MPa)

Maximum tensile stress (MPa)

10

100

50

167

25

100

25

67

50

100

10

33

100

100

5

17

10

300

50

56

25

300

25

22

50

300

10

11

100

300

5

5

58

From these results it is clear that at least a 100 mm-thick liner would be required to match the strength of a typical tendon under the assumptions of a non-resisting fragmented rock mass and non-uniform, bag-like rock mass deformations. The bearing plate dimensions play an important role here and an increase in bearing plate dimensions would reduce liner requirements. This example is based on simplifying assumptions and appropriate testing would be required to obtain results that are more realistic. Nevertheless, these examples do give a reasonable indication of the liner potential under extreme conditions. If the contribution of the fragmented rock mass could be quantified, the requirements for the surface support can be further reduced. In what follows, an attempt is made to quantify the resistance of the fragmented rock mass, as well as the interaction between the fragmented rock mass and the liner. By expressing the contribution of the fragmented rock mass as a percentage of the total inertia forces, the loading that needs to be transmitted via the liner can simply be reduced accordingly and proportionally. The interaction between the fragmented rock mass and the liner is simply viewed as the formation of a composite plate that consists of one layer of rock fragments unto which the liner is coated. The bending resistance of such a composite plate is strongly affected by the fragment size, as this controls the effective plate thickness. The tensile stress in the liner can be expressed by:

(100 ? ? ) Fy 1000ht

,

(3.6.7)

where h is the fragment size and ? is the fragmented rock mass contribution. It is hereby assumed that final failure occurs when the liner breaks at mid-span after having cracked initially around the tendon. The maximum shear stresses occur around the bearing plate and conservatively it is assumed that the fragmented rock mass does not assist the liner in resisting shear stresses. Therefore, the maximum shear stress is expressed as:

(100 ? ? ) Fy

(3.6.8)

400 Dt

Using the same parameters as in the previous examples, namely a span of one metre between the tendons, a bearing plate dimension of 50 mm x 50 mm and a maximum tendon force of 100 KN, the results listed in Table 3.3 are obtained.

Table 3.3

Induced stresses in composite beam and contributing rock mass

Liner

Fragment

Fragmented R.M.

Maximum shear

Maximum

thicknes

size

contribution (%)

stress (MPa)

tensile stress

s (mm)

(mm)

5

100

0

100

20

10

100

0

50

10

25

100

0

20

4

(MPa)

59

Liner

Fragment

Fragmented R.M.

Maximum shear

Maximum

thicknes

size

contribution (%)

stress (MPa)

tensile stress

s (mm)

(mm)

5

300

0

100

6,7

10

300

0

50

3,3

25

300

0

20

1,3

5

100

50

50

10

10

100

50

25

5

25

100

50

10

2

5

300

50

50

3,3

10

300

50

25

1,7

25

300

50

10

0,7

(MPa)

The results from Table 3.3 suggest that the maximum shear stress would be the critical factor. However, these results may be far too conservative in that the potential of the fragmented rock mass may be underestimated. In order for failure to occur under shear, the tendons have to effectively “punch” into the fragmented rock mass. The actual resistance against such punching can be expected to increase with increasing fragment size. It is important that such parameters are properly quantified by means of laboratory testing and numerical modelling. In addition, the overall contribution of the fragmented rock mass needs to be quantified in a more realistic way. Some form of rock mass classification, in which the fragmentation is related to the load-bearing capacity, seems to be the most appropriate. These issues are, however, beyond the scope of the current project.

60

4. Proposed testing procedures 4.1 Introduction The proposed testing procedures are based mainly on the literature review presented in chapter two. Although many testing procedures were considered and analysed, it was felt that only two of those meet the requirements for standard testing. Many of the other testing procedures involve multiple mechanisms of failure and/or are too complex. In addition, most of these procedures are not well known and recognised. The proposed testing procedures are based on ASTM standards and are thus well established. They offer a practical means to quantify relevant material properties and, as such, allow a relative comparison between various products. A large amount of additional work has been directed towards investigating alternative testing methods. Members of the project team have published most of this work and the relevant papers are included in the appendices. While these investigations may have demonstrated the merits of alternative tests, it has been decided to restrict the proposed testing procedures to two simple and basic tests, namely a direct tensile test and an adhesion test that is designed to obtain a direct bond strength under both laboratory and underground conditions. These two tests were not arbitrarily selected by the project team, but are also strongly supported by the majority of experts who expressed their views during the three-year series of international seminars and workshops on surface-support liners. In fact, these two tests are the only ones that complied with the objectives as set out in Section 2.3 In the following sections, the test procedures are specified in the format of the ISRM guidelines for standard test procedures. These two sections form the major deliverable of the SIMRAC project titled, “required technical specifications and standard testing methodology”

4.2 Tensile strength test 4.2.1 Introduction The proposed tensile test is based on ASTM D 638, which covers the determination of the tensile properties of Thin Sprayed Linings (TSLs) by the use of standard dumbbell-shaped specimens. Tannant et al. (1999) recommended the use of a specimen with Type I dimensions (ASTM D 638) based on experience with property testing of TSL. However, it is advised in this same standard that “The Type IV specimen should be used when direct comparisons are required between materials in different rigidity cases”. The Type IV specimen dimensions as shown in Figure 4.2.1 therefore seem to be applicable to TSLs. This test method can be used for testing TSL products of any thickness between two and four millimetres. This testing method was selected and used by most of the researchers (Tannant et al., 1999; Archibald, 2001; Spearing and Gelson, 2002) as a primary linercharacterisation test. Other material properties that can be assessed using this test are the in-line stiffness (modulus) and elongation capacity or degree of tensile extension at failure.

61

D=65 mm C

D

B

E

18.7 mm

18.7 mm

A

R 1=14 mm

R1=14 mm

F

R 2=25 mm

WN= 6 mm

WO=20 mm

WO=20 mm

LG=25 mm R2=25 mm

LN=33 mm L

R 1=14 mm

18.7 mm

R1=14 mm

G 18.7 mm

K

H J

I LO=115 mm

Coordinates

A

B

C

D

E

F

G

H

I

J

K

L

X (mm)

0.0

18.7

41.0

74.0

96.3

115.0

115.0

96.3

74.0

41.0

18.7

0.0

Y (mm)

27.0

32.0

34.0

34.0

32.0

27.0

7.0

2.0

0.0

0.0

2.0

7.0

Where: W O: Overall Width, LO: Overall Length, D: Distance between grips, LN : Length of Narrow section, LG: Gage Length, WN : Width of Narrow section, R 1: Inner Radius, R 2: Outer Radius

Figure 4.2.1

Proposed tensile strength test specimen dimensions for testing of TSL material

However, the proposed dimensions as shown in Figure 4.2.1 were adjusted in order to account for the relatively large grain size in certain TSL products. A specimen width of 12 mm instead of six millimetre with a corresponding larger grip is therefore suggested. Owing to the high degree of sensitivity exhibited by many TSL products to loading rate, data obtained from a test with a specific loading rate cannot be considered representative for applications involving load-time scales that vary widely from that loading rate. This sensitivity to loading rate necessitates testing over a broad load-time scale (including very fast (dynamic) and very slow (creep) loading). In determining realistic loading rates, the deformation capacity of the material should be considered. Brittle cementitious materials typically have limited deformation capacity and show limited time-dependent behaviour. These materials could therefore be tested at a nominal, practical loading rate. With a maximum deformation of around one per cent and an associated displacement of not more than half a millimetre, a loading rate of half a millimetre per minute or less seems to be reasonable. Products that do demonstrate time-dependent behaviour are generally more deformable and could enable maximum deformations in excess of 100 per cent. With a loading rate of half a millimetre per minute, a typical tension test could take more than one hour. However, it is considered important to include such low loading rates together with faster loading rates, as this would be indicative of the loading rate sensitivity of a particular product. As a guideline, it is recommended to use three different loading rates for products that show time-dependent behaviour (besides curing time). A fast rate would be in the order of ten millimetre per minute, a slow rate - as mentioned before - around half a millimetre per minute, and an additional loading rate of one millimetre per minute in order to qualify the nature of the loading-rate sensitivity. Tensile properties of TSL products may vary, depending on the method used for sample preparation. When comparative tests are desired, the greatest care must be exercised to ensure that samples from different products are prepared in a very similar way (if not the identical way). Similarly, for referee purposes or comparisons within any given series of specimens (i.e. the same batch from a particular product), care must be taken to secure the maximum degree of uniformity in details of sample preparation. Specimens prepared 62

by moulding may have different tensile properties from those prepared by machining or die cutting. This effect may be more pronounced in specimens with narrow sections. Pouring the mix into the moulds may lead to the trapping of undesired air bubbles in the specimens. On the other hand, some TSL products are not suitable for machining or die cutting, as they are prone to be damaged during the process. In order to standardise the specimen-preparation process, it is recommended that the TSL product be poured into a standard mould. If this is not practical for any particular type of TSL product (e.g. a fastsetting or reactive product) then machining or die cutting of the sprayed coating may be the only solution. Whatever method is used, it should be clearly indicated when the test results are presented. Tensile properties of many TSL products may also vary significantly depending on the characteristics of the environment (i.e. temperature and humidity). In addition, the curing time may also be affected by these environmental parameters. Every TSL product should be tested at different curing times in a predefined environment (with constant temperature and relative humidity). Time-dependent behaviour implies creep and thus a decreasing load-carrying capacity with increasing time. In order to determine the long-term load-carrying capacity of a particular material, creep tests are recommended. As a suggestion, materials should be tested at a fixed load that is equal to 50 per cent of the average strength that was obtained from specimens subjected to the standard loading / deformation rate. If no creep is observed, the material may be considered insensitive to creep. If time-dependent deformations are observed, the test should be continued for a period of at least 24 hours and also be conducted on specimens that are subjected to loads of 25 per cent and ten per cent of the average strength. TSL products that are targeted to provide support in seismically active areas may need to be subjected to additional tests. It is important that these products are tested under conditions that are representative of seismically induced rock failure. For this purpose, larger deformation rates may be required and are recommended. The practical aspects of this recommendation have not yet been addressed, but it is envisaged that the load could be applied by a weight that is dropped from a certain height. The arresting distance would indicate the resisting force, the material stiffness, and the duration of the impact.

4.2.2 Scope This method of testing is intended to provide a consistent measure of relevant parameters that are associated with the deformation in tension of any Thin Sprayed Liner. These parameters include tensile strength, tensile stiffness, and deformation capacity.

4.2.3 Apparatus A testing machine of the constant-rate-of-crosshead-movement type is required. The drive mechanism imparts a uniform and controlled velocity to a movable member with respect to a stationary member. Grips, holding the test specimen between the fixed member and the moveable member, must be self-aligning in such a manner that they will move freely into alignment as soon as any load is applied. Both the load-indicating mechanism and the displacement-indicating mechanism must be free of inertia lag at the specified rate of testing. The load indicator must indicate the load with an accuracy of ±one per cent of the indicated value, or better. A suitable displacement-measuring instrument (extensometer) must be used for determining the distance between two designated points within the gage length of the specimen as it is stretched. For referee purposes, the extensometer must be set at the full gauge length of the specimen. It is desirable to record this distance (or any 63

change in it) as a function of the load on the test specimen and the elapsed time from the start of the test. The modulus of a material is determined from the slope of the linear portion of the stress-strain curve. For some TSL materials, this linear portion may be very small and may occur very rapidly. In order to determine the modulus of a TSL material, the extensometer system must monitor the displacement with a maximum error of 0,001 strain or ±one per cent of the indicated strain, whichever is greater. However, for making measurements at elongations greater than 20 per cent, measuring techniques with an error of less than ±ten per cent of the measured value are acceptable. 1.

A suitable machine should be used to apply a tensile load to the specimen. The machine should have the capacity to apply a tensile force of one kilo Newton at a constant displacement rate of up to ten millimetres per minute. The displacement rate should not vary by more than five per cent during the testing procedure. The equipment should also be able to apply loading rates as small as half a millimetre per minute. A typical test configuration is shown in Figure 4.2.1

2.

Load applied must be measured using an in-line load cell with a resolution of 0,05 kg or better. The test machine shall be calibrated against a known and accepted standard.

3.

The deformation of the sample must be measured with an accuracy of at least 0,05 mm.

4.

The machine should have two sample holders comprising a lower plate and an upper plate that are held together by bolts. One of the plates must be recessed in order to accommodate the sample.

5.

The system must have facilities for electronic recording of the load and deformation, using suitable data-acquisition devices.

4.2.4 Procedure 1.

The sample is to be cast in a “dog-bone” shape. The preferred sample dimensions are given in Figure 4.2.1

2.

The samples are to be poured into a suitable mould and removed once sufficient setting time has elapsed. The samples are then left to cure under constant temperature and humidity conditions.

3.

Sample dimensions (thickness and width) have to be recorded to the nearest 0,1 mm by averaging three readings on the narrow section of the specimen before loading. Curing time and temperature must be recorded.

4.

The sample shall be loaded with a displacement rate of between 0,5 mm and 10,0 mm per minute until failure or attainment of the machine stroke.

5.

The load and deformation must be recorded at least once per second.

6.

Failure occurring other than within the central thin section of the dog-bone specimen must be reported, but results should be discarded.

7.

Specimens should be tested after one, three, seven and fourteen days of curing at a constant temperature and humidity. The recommended temperature is 25 degrees Celsius and the recommended humidity is 50 per cent. A minimum of five samples is to be tested for each curing time.

64

8.

It is recommended that ductile materials, such as polyurethanes and acrylics, are subjected to various deformation rates, in particular the minimum and maximum rate. In addition, it is recommended that such materials are subjected to creep testing.

4.2.5 Calculations 1.

The tensile strength of a tested product must be calculated by dividing the maximum load to which the specimen was subjected by the original cross-sectional area of the specimen. This cross-sectional area is determined from the sample dimensions as obtained under Section 4.2.6.4 - 3.

2.

The maximum stiffness can be determined by detecting the steepest part of the loaddeformation curve and by quantifying the slope at that location in terms of stress (MPa) per displacement (mm).

3.

The maximum deflection is determined at the point where the specimen has lost most of its load-bearing capacity. It is suggested that 25 per cent of the peak strength may be considered an acceptable cut-off level.

4.2.6 Reporting of results The following must be reported: 1.

Type of Thin Sprayed Liner

2.

Curing time, temperature, humidity and specimen dimensions

3.

Ultimate tensile strength in MPa

4.

Maximum stiffness in MPa/mm

5.

Maximum deformation capacity in mm

6.

Deformation rate in mm/minute

7.

Failure mode

8.

Graph of tensile stress versus nominal displacement in MPa versus mm.

9.

In case of any non-compliance with these guidelines such non-compliance should be reported

65

Figure 4.2.2

Tensile loading apparatus and recording equipment 18.7 33.0 35

40

12.0

R50.0 R28.0 44.6

Figure 4.2.3

Recommended dimensions for the tensile testing specimen

4.3 Proposed adhesion/bond strength test 4.3.1 Introduction On the basis of the tests that were described in the literature survey, a direct pull test with a glued dolly appears to be the most practical and acceptable method for assessing adhesion. This method complies with existing standards (ASTM D4541) and has the added advantage that it can also be applied in situ. The main disadvantage of this testing procedure is the fact that the pull-off strength may be limited by (premature) failure other than the adhesive failure between the substrate and the liner. The interface between epoxy 66

and liner or between epoxy and dolly could be relatively weak. The epoxy itself could fail in tension or shear, or the lining material could fail in tension. However, if an adequate epoxy is selected, the only remaining possibility is tensile failure of the liner. This is a potential problem with the more brittle materials and the monitored pull-off strength will be an underestimation of the adhesive strength in that case. As has been indicated by Tannant and Ozturk (2003), the test will then only provide a lower bound for the adhesion.

4.3.2 Scope This method of testing is intended to provide a consistent measure of the bonding strength between any TSL and the substrate onto which it has been applied. The test can be conducted in a laboratory as well as underground, on an existing application. The primary use of the results from the laboratory tests is an objective comparison between various products, while underground results can be used mainly for quality control on the actual application.

4.3.3 Apparatus 1. A steel disc, with a flat contact surface and a suitable shape (Figure 4.3.1) for connection to the pulling equipment is glued to the surface of the liner. The recommended glue is an epoxy under the name of Araldite 4076, which is locally available 2. The pulling equipment should be able to apply a tensile stress of at least 5 MPa to the steel disc and consists of a hydraulic cylinder that acts against a reaction frame in such a way that only tensile forces are transmitted. The testing frame should be positioned in such a way as not to interfere with the de-bonding process. A typical pulling machine is shown in Figure 4.3.2. 3. Loading is to be applied at a constant pressure rate. The maximum hydraulic pressure is to be recorded and the machine is to be calibrated so that the monitored pressure can be related to the applied tensile load.

4.3.4 Procedure 1

With the laboratory test, the TSL is brushed to a standard surface. The recommended surface is the unpolished side of a tile of Norite (trade name “Rustenberg Black Granite”) of dimensions 300 by 300 by 10 mm. A minimum thickness of 20 mm is recommended. The surface should be cleaned with water and left to dry.

2

With the laboratory test, the TSL is let to cure under constant conditions for four days and the curing temperature and humidity are to be recorded. It is recommended to conduct this test at a temperature of 25 degrees Celsius and a humidity of 50 per cent if possible.

3

The disc of known diameter needs to be cleaned with rough sand paper.

4

Underground, the substrate needs to be cleaned of dust and grease with a wire brush and water. Chemical cleaners are not recommended.

5

Once the surface is dry again, the epoxy is generously applied to the disc. It is recommended to mix the epoxy on a 2:1 volume ratio between resin and hardener. The epoxy should be left to cure for a short while, until it is sufficiently easy to handle. 67

6 Place the disc with the epoxy gently onto the substrate until the steel starts touching the liner. Do not push further. Underground, it is recommended to secure the disc and the epoxy with a quick setting putty to prevent sliding of the disc and leaking of the epoxy. 7

In the laboratory, the epoxy is allowed to cure for three days, while underground curing should be at least 24 hours.

8

Underground, the position of discs should be adequately recorded relative to each other and relative to a calibrated position such as a survey peg.

9

The test frame is placed in position and connected to the dolly. This connection is to be hand tightened. Underground, a second person is required to hold the frame in position.

10

The sample is to be gradually loaded by increasing the pressure constantly at a rate of less than one MPa per five seconds until a maximum pressure has been reached

11

This maximum pressure is to be recorded and the dolly is to be removed and stored.

12

Underground, a photo is to be taken of the exposed area underneath the dolly.

13

The failure modes and associated failure areas must be recorded and estimated. Recognised failure modes are internal failure of the substrate, de-bonding between substrate and liner, internal failure of the liner, de-bonding between epoxy and liner, internal failure of the epoxy and de-bonding between epoxy and steel. The associated failure area is to be expressed as a percentage of the total failure surface. It is sufficient to estimate this percentage to the nearest five per cent.

4.3.5 Calculations The failure stress is calculated by dividing the maximum applied load by the area of the steel disc. The adhesive strength only equals the failure stress if the failure mode is 100 per cent de-bonding between substrate and liner. The adhesive strength will exceed the recorded failure stress in all other cases.

4.3.6 Reporting of results The following must be reported: 1.

Type of TSL (underground, only if known)

2.

Curing time, temperature and humidity (underground, only if known)

3

Ultimate failure stress in MPa

4.

Failure modes and relative size of associated failure area

5.

In case of any non-compliance with these guidelines such non-compliance should be reported

68

18.7 0.70 R9.0 2.0

9.0 R5.0

8.0 4.0 5.0

13.36 39.0

Figure 4.3.1

Recommended dimensions for (axi-symmetric) steel dolly

Figure 4.3.2

Example of pull testing equipment connected to glued dolly

4.4 Conclusions Two testing procedures were proposed and described. Both procedures are aimed at quantifying basic and relevant TSL properties in an effective and efficient way. The tensile test allows assessment of material strength and stiffness as well as postfailure behaviour. The influence of environmental factors such as temperature and

69

humidity can be taken into account. In addition, the effect of time, both during curing as well as during loading, can be quantified. The adhesion test allows an assessment of the bonding capacity between a TSL and the substrate to which it is attached. This test requires minimum disturbance of the liner and is therefore suitable for in situ testing as well. Under laboratory conditions, the influence of environmental factors and curing time can be evaluated. These tests are relatively simple to conduct and allow quantification of essential TSL parameters. The tests are therefore very suitable for comparative purposes. In addition, the properties that were analysed with these two tests are sufficient to predict the performance of a particular TSL in actual and more complex applications. This could be achieved in practice with appropriate numerical models, for instance. It is strongly recommended that both the tension as well as the adhesion test be conducted under conditions that are equivalent to in situ conditions, in order that realistic values can be obtained. High temperatures and humidity are typically associated with a deep level mining environment and are very likely to affect liner performance.

70

5. Results from laboratory testing 5.1 Introduction The laboratory programme consists of the tensile testing of dog-bone-shaped specimens and the adhesion testing of the liner material. The dog-bone-shaped specimens were tested for various curing times and at various loading rates in a specially developed testing rig (Figure 5.1.1). In addition, the specimens were subjected to constant deformation and constant loads, so that the associated time-dependent responses could be monitored. The bonding strength of various materials was evaluated by coating the liner on a standard substrate from which it was pulled off by means of metal disc. These discs were glued to the liner with a specially selected epoxy. This epoxy not only has a high strength, but it is especially able to act along an irregular interface where a relative high thickness may be required. The selected epoxy was Araldite 4071 from Ciba-Geigy. The bond strength of this epoxy was measured to be in excess of three MPa for typical interfaces between steel disc and liner. This value is deemed more than the expected bond strengths between liner and substrate so that the contact between disc and liner would remain intact during an adhesion tests.

Figure 5.1.1

Test rig for dog-bone specimens

5.2 Tensile tests 5.2.1 Introduction Figure 5.1.1 shows the testing rig that was designed and built as part of this project for the purpose of testing dog-bone-shaped specimens of typical TSL materials. While the recommended shape of the specimens was described in Section 4.2, the tests were 71

conducted on large specimens with a width of 20 mm, as well as on small specimens with a width of 12 mm. Calculated tensile stresses are based on the measured crosssectional area of these specimens. A representative variety of TSLs including materials with a cement base, an acrylic base, and a polyurethane base, were tested. The effects of temperature and humidity were not explicitly quantified as all tests were conducted at room temperature and humidity. Appendix A1 contains the detailed results of all tensile tests that were analysed. Only a summary of these results and a brief description of the various products are included in the section that follows.

5.2.2 Product A Product A is cement based and required careful preparation because of its brittleness. Some specimens were lost through premature failure associated with sample handling. Figure 5.2.1 shows the monitored strength of all the tests that were conducted.

Maximum stress (MPa)

Product A 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

1 day 3 days 7 days 14 days

1

2

3

4

5

Test number

Figure 5.2.1

Variation in tensile strength for different curing times (product A)

Product A is relatively stiff and brittle, with a maximum strain of around three per cent and a maximum stiffness of up to one GPa. Owing to this high stiffness, only slow loading rates were applied. It is believed that this product does not demonstrate time-dependent behaviour, and the effect of loading rate was therefore not investigated explicitly.

5.2.3 Product B Product B is a cement-based coating and proved extremely difficult to handle with respect to specimen preparation because of its brittle nature. Some shrinkage occurred during curing and this caused many of the specimens to fracture prematurely inside the mould. Because of these problems, only a limited number of specimens could be tested. Figure 5.2.2 shows the variation in strength that was monitored.

72

Product B


2.5 2 1.5

2 days 8 days

1 0.5 0 1

2

3

4

Test number

Figure 5.2.2

Variation in tensile strength for different curing times (product B)

Product B is brittle and stiff, with maximum deflections of around one per cent strain and a maximum stiffness that varied from 0,5 GPa to 1,0 GPa. Owing to its stiffness, only slow loading rates were applied. It is, however, expected that loading rate is not a relevant parameter, as time-dependent behaviour is not expected in the cement-based products. No creep tests were conducted. Although the tensile strength did not appear to increase much with curing time beyond two days, the maximum stiffness did increase noticeably. The observed maximum strength may, however, be strongly related to local defects that create stress concentrations, and a larger number of tests would be required to obtain statistically relevant results.

5.2.4 Product C Product C has a polyurethane basis and the dog-bone specimens were prepared by the manufacturer. Figure 5.2.3 shows the variation in maximum strength that was observed, while Figure 5.2.4 shows how the loading rate affects this maximum strength.

73

Product C


6 5 3 days

4

10 days

3

15 days

2

17 days

1 0 1

2

3

4

5

6

Test number

Figure 5.2.3

Variation in tensile strength for different curing times

Product C 6

Load (kg)

5 3 days

4

10 days

3

15 days

2

17 days

1 0 0

2

4

6

8

Loading rate (mm/min)

Figure 5.2.4

Variation in tensile strength with respect to loading rate

The sensitivity with respect to loading rate can be attributed to the time-dependent behaviour of the polyurethane. In these particular tests, the loading rate was limited to around 0,60 mm/min. As maximum deflections of around 70 mm, which is more than 200 per cent, were achieved, tests could take as long as two hours. As can be observed from the graph in Figure 5.2.4, the maximum strength decreases with decreasing loading rates and lower values can be expected at even lower loading rates. In order to quantify an absolute minimum strength (the extrapolation to intersect with the vertical axis in Figure 5.2.4), it is recommended that creep tests of longer duration be conducted. Some creep test results are shown in Appendix A2. The maximum stiffness of this product was sensitive with respect to both the loading rate and the curing time and ranged from approximately 10 MPa to 30 MPa. 74

5.2.5 Product D Product D has an acrylic base. Only large specimens, with a width of 20 mm, were tested. Figure 5.2.5 shows the variation in strength in relation to the curing time, while Figure 5.2.6 shows the effect of loading rate on strength. Note that the one-day-old specimens were only tested at one loading rate.

Product D (large)


3 2.5 2

1 day

1.5

4 days

1

7 days

0.5 0 1

2

3

Test number

Figure 5.2.5


Product D

Maximum load (kg)

3 2.5 2 1.5 1

1 day 4 days

0.5

7 days

0 0

2

4

6

8


Figure 5.2.6


Maximum deformation was around 100 per cent and maximum stiffness varied with loading rate as well as with curing time from around five MPa to 70 MPa. This product also shows time-dependent behaviour and creep tests are recommended for determining minimum strength levels. The current equipment has a minimum loading rate of around 75

0,3 mm/min and this is considered too high for determining long-term time-dependent behaviour.

5.2.6 Product E Product E is another acrylic-based liner and Figure 5.2.7 shows the variation of maximum strength for different curing times. In Figure 5.2.8, the relationship between the loading rate and the maximum strength is displayed.

Product E Maximum stress (MPa)

4 3.5 3 2.5

1 day

2

3 days

1.5

8 days 16 days

1 0.5 0 1

2

3

4

Test number

Figure 5.2.7


Product E


4 3.5 3

1 day

2.5

3 days

2 1.5

8 days 16 days

1 0.5 0 0

2

4

6

8


Figure 5.2.8


Product E shows maximum deformations that range from more than 100 per cent for the one-day-old specimens to around 30 per cent for the 16-day-old specimens. The stiffness 76

ranges from around ten MPa for the one-day-old specimens to around 75 MPa for specimens that have cured for 16 days. The loading-rate dependency is well established for the three- and 16-day-old specimens, but less so for the others. It is most likely that a larger database is required to obtain relationships that are more reliable and less subject to individual specimen variations. Nevertheless, these results clearly indicate a trend that was also observed in products C and D.

5.3 Adhesion tests 5.3.1 Introduction The pulling equipment that is used for the adhesion tests is shown in Figure 5.3.1. This equipment was used for both the laboratory tests and the underground tests. For the laboratory tests commercially available granite tiles with the dimensions 300 mm x 300 mm x 10 mm were used as standard substrate. These tiles were cleaned with water, after which they were coated with different lining materials. After a curing time of seven days, a number of metal discs were glued to the liner. These discs were first cleaned and roughened with sandpaper and the Araldite epoxy was allowed to cure for at least three days. Using the pulley, each disc was pulled away from the tile and the maximum pressure and the failure surface were recorded. This data gives an indication of the bonding potential of typical liners and is not only useful for comparing various products, but can also be used as a benchmark for underground applications. Appendix A2 contains the detailed results of all adhesion tests that were analysed. A summary of these results is described in the following sections.

Figure 5.3.1

Pulling equipment used for adhesion tests in the laboratory

77

5.3.2 Product A Adhesion product A liner de-bond Pull strength (MPa)

7

failure before de-bond

6 5 4 3 2 1 0 1

2

3

4

5

6

7

8

Test number

Figure 5.3.2 Table 5.1

Adhesion results for product A

Adhesion results for product A

Test number

Age of liner

Age of epoxy

Maximum pressure (MPa)

Remarks

A1 A2 A3

15 days 15 days 15 days


4,0 6,2 4,4

A4

15 days

3 days

3,0

A5 A6 A7 A8 A9 A10

15 days 11 days 11 days 11 days 11 days 11 days

3 days 3 days 3 days 3 days 3 days 3 days

4,6 4,6 4,4 4,8 5,0 5,6

50% de-bond; 50% liner 90% liner; 10% epoxy/liner 60% liner; 20% epoxy/liner; 20% epoxy/steel 10% liner; 40% epoxy/liner; 50% epoxy/steel 90% liner; 10% epoxy/liner 70% de-bond; 30% liner 60% de-bond; 40% liner 10% de-bond; 90% liner 70% liner; 30% epoxy/liner 90% liner; 10% epoxy/liner

Product A is cement based and shows a superb resistance against adhesion, as well as large internal membrane strength. The bond strength appears to range from about 4,5 MPa to values in excess of six MPa. The low result of test A4 is most likely associated with inadequate preparation of the steel disc. As the average resistance against pulling is 4,65 MPa, the actual bond strength of the product can be expected to exceed that value.

5.3.3 Product B

78

Table 5.2

Adhesion results for product B

Test number

Age of liner

Age of epoxy


Remarks

B1 B2 B3 B4 B5 B6 B7

10 days 10 days 10 days 10 days 10 days 10 days 10 days 10 days 10 days 10 days

3 days 3 days 3 days 3 days 3 days 3 days 3 days 3 days 3 days 3 days

1,6 1,6 1,4 1,6 1,2 1,6 1,4 1,2 1,4 1,4

100% liner failure 100% liner failure 100% liner failure 100% liner failure 100% liner failure 100% liner failure 100% liner failure 100% liner failure 100% liner failure 100% liner failure

B8

B9 B10

Pull stress (MPa)

Adhesion product B 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

failure before de-bond

1

2

3

4

5

6

7

8

9

10

Test number

Figure 5.3.3

Adhesive test results for product B

Product B is a cementitious product and showed consistent tensile failure of the liner itself. The steel dolly became detached by pulling a very thin layer from the coating. The thickness of this layer was probably of the order of the grain size of the material. The actual bond strength between the liner and the substrate was therefore not reached and no de-bonding was achieved. De-bonding strength is therefore in excess of the maximum pressures that were recorded. The apparent limited material strength may be associated with a relatively low tensile strength as measured in the dog-bone specimens (Section 5.2.3). In addition, the brittleness of the material may also affect these adhesion tests, as stress concentrations due to defect and irregular surfaces cannot be avoided. The average pull resistance for product B is 1,45 MPa and the actual bond strength will be larger than this value.

79

5.3.4 Product C The adhesion tests for product C, which is polyurethane based, were not without problems. Tile number one was not adequately cleaned, and the remaining solvent affected the bond strength negatively. The second tile was well prepared, but fractured during the pull tests. This may have affected the bond strength, as fractures intersected the area underneath the dollies. While the results from the first tile can be disregarded, the results from the second tile should be regarded as a lower bound for the adhesive strength.

Table 5.3

Adhesion results for product C

Test number

Age of liner

Age of epoxy

Thickness (mm)


Remarks

C1

14 days

4 days

2.5

1.0

95% de-bond; 5% liner

C2

14 days

4 days

2.5

1.0

50% de-bond; 50% epoxy/steel

C3

14 days

4 days

2.5

1.0


C4

14 days

4 days

2.5

1.4


C5

14 days

4 days

2.0

4.0

De-bond and peel and tile fracturing

C6

14 days

4 days

2.0

4.4

100% epoxy/steel

C7

14 days

4 days

2.0

3.5


C8

14 days

4 days

2.0

3.2


C9

14 days

4 days

2.0

3.2


C10

14 days

4 days

2.0

2.6


80

Pull stress (MPa)

Adhesion product C 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

tile 1 failure before de-bonding

1

2

3

4

5

6

Test number

Figure 5.3.4

Adhesive test results for product C

The thickness of the application was relatively small (about two millimetres), and this may have promoted the fracturing of the tile. It is therefore recommended that tiles that are thicker than the current ones of ten millimetres be used. Using the results of the second tile only, the average pull resistance is 3,5 MPa. The bond strength of product C most likely exceeds that value.

5.3.5 Product D Product D is an acrylic-based product and hardly de-bonded, as internal material failure in tension, as well as in shear, occurred prematurely. It can therefore safely be assumed that the actual bond strength is in excess of the maximum values that were recorded. Besides internal liner failure and de-bonding, failure of the epoxy-liner interface and the epoxy-steel interface were occasionally observed. The average pull resistance is 1,75 MPa and the actual bond strength of product D must be higher than that.

Table 5.4

Adhesion results for product D

Test number

Age of liner

Age of epoxy


Thickness (mm)

D1 D2

13 days 13 days

3 days 3 days

1.5 2.0

4.0

D3 D4 D5



1.6 2.2 1.9

5.0 81

Remarks

Tensile failure of liner Substrate/liner 15%; liner 50%; epoxy/liner 10%; epoxy/steel 25% Substrate/liner 15%; liner 85% Tensile failure of liner Liner 75%; epoxy/steel 25%

Test number

Age of liner

Age of epoxy


Thickness (mm)

D6 D7 D8 D9 D10

13 days 13 days 13 days 13 days 13 days

3 days 3 days 3 days 3 days 3 days

1.6 1.8 1.8 1.6 1.5

-

Remarks

Liner 75%; epoxy/steel 25% Tensile failure of liner Liner 95%; epoxy/steel 5% Tensile failure of liner Liner 90%; epoxy steel 10%

Adhesion product D

Pull stress (MPa)

2.5

failure before de-bonding

2 1.5 1 0.5 0 1

2

3

4

5

6

7

8

9

10

Test number

Figure 5.3.5

Adhesive test results for product D

5.3.6 Product E Product E is an acrylic-based product and generally de-bonded in the adhesion tests. When the applied thickness was in excess of four mm, the de-bonded area was substantially smaller than the actual disc size because the liner sheared off along an inclined failure surface. In the case of the four millimetre thick liner, this shear failure surface was approximately perpendicular to the plane of the liner, which resulted in a debonded area that was equal to the size of the steel disc. However, the shearing of the liner appears to be a post-de-bonding effect and the actual de-bonding values do not seem to be affected by the post-failure behaviour. The recorded values can therefore be considered as representative of the adhesive strength of product E. As the average adhesive strength of all these tests is exactly 2,5 MPa, this value represents the bond strength of product E.

82

Table 5.5

Adhesion results for product E

Test number

Age of liner

Age of epoxy


Thicknes s (mm)

E1 E2 E3 E4 E5 E6 E7 E8

10 days 10 days 10 days 10 days 10 days 10 days 10 days 10 days

3 days 3 days 3 days 3 days 3 days 3 days 3 days 3 days

2.6 2.4 2.4 2.4 3.2 2.4 2.2 2.6

4.0 4.0 4.0 4.0 4.0 5.0 5.0 7.0

E9 E10

10 days 10 days

3 days 3 days

2.4 2.4

6.0 8.0

Remarks

Fully de-bonded Fully de-bonded Fully de-bonded Fully de-bonded Fully de-bonded Fully de-bonded Fully de-bonded 80% de-bonded 20% epoxy/liner failure Fully de-bonded Fully de-bonded

Adhesion product E

Pull stress (MPa)

3.5

liner de-bond

3 2.5 2 1.5 1 0.5 0 1

2

3

4

5

6

7

8

9

Test number

Figure 5.3.6

Adhesion test results for product E

83

10

6. Underground Monitoring 6.1 Introduction Underground monitoring was conducted at five different sites. The purpose of the monitoring programme was to investigate existing applications of TSLs. Adhesion tests were conducted at five sites at gold mines located in the West Rand and Far West Rand region. The main aim of these adhesion tests was to quantify the effectiveness and quality of the underground application of TSLs. The results show the variation in bond strength at each site and between the different sites. By comparing these in situ results to values derived from laboratory tests, as described in Section 5.3, an indication of the relative effectiveness can be obtained. During the conducting of the underground adhesion testing, the testing procedure was optimised. This led to the formulation of a guideline for underground adhesion testing. The underground monitoring programme also included the establishing of the crude rock mass ratings at sections of unlined excavations in close proximity to the test sites. On the basis of visual inspections and interviews with mine personnel, the effectiveness of the liner(s) was judged as well. Working plans and / or sketches of the test sites illustrate local conditions. To foster transfer of knowledge, discussions were held with mine rock-engineering personnel for the purpose of outlining the preliminary results of the adhesion tests

6.1.1 Disk installation procedure The adhesion test consists of gluing a metal dolly onto the liner, using a strong adhesive, allowing the adhesive sufficient curing time, and using a calibrated pull tester to remove the dolly and the liner from the substrate. In order for the epoxy to cure sufficiently, a minimum period of 12 hours is recommended. The process of installing and monitoring in practical terms, therefore, requires at least two underground visits. The maximum pullout pressure is recorded. The procedures adopted for the tests conducted were subdivided into a dolly (disc) installation procedure and a pull test procedure. The adhesive that is recommended for these tests is an epoxy called Araldite HV 4076. This particular epoxy was selected after a comparative study of various products on the market. The adhesive is prepared by mixing a hardener with the epoxy, using the ratios stipulated in Table 6.1. The manufacturer’s specifications for mixing were altered by trial and error to suit the ambient conditions at the test sites.

Table 6.1

Mix ratios for the adhesive

Mix ratio

By weight

By volume

Epoxy

100

100

(two part)

Hardener

44

50

(one part)

The mixture was allowed sufficient time (15 minutes) to form a paste so that the weight of the disk did not induce sliding at the interface of the disk face and the epoxy. The semihardened paste had a pot life of approximately one hour.

84

The faces of the disks were cleaned with the use of rough sandpaper to remove any grit and oxidised residue from the face before the epoxy was applied. Preparation of the substrate required the following: ?

Wire brushing of the lined rock was done to remove grease, dust, and blast coatings. The use of chemical cleaners was avoided.

?

A decoupling square slot was either chiselled or cut into the liner at the position where the disk was to be installed to create an unvarying pullout surface. The square slot had dimensions of approximately 50 mm by 50 mm (Figure 6.1.1).

?

The depth of the slot was governed by the thickness of the liner; i.e. the slotting was stopped once the rock face was intersected (Figure 6.1.1).

Figure 6.1.1

Square 50 mm by 50 mm decoupling slot

Figure 6.1.2

Removed slotted disk along pre-cut lines

85

?

The application of the disks onto the prepared substrate entailed:

?

Using the face of the disk to scoop paste evenly across the disk

?

Sliding the disk into its position onto the substrate in between the slotted area

?

Using nails below the disks to prevent sliding (better to use a quick-setting putty!)

?

Marking the disks once they are affixed

?

Measuring the location (using surveyed peg positions) and recording the disk number relative to the location

?

Leaving the affixed disk to adhere to the substrate for a period exceeding 24 hours (Figure 6.1.3). Pull tests conducted before a 24-hour epoxy-curing time failed because the disk parted at the interface of the metal and the epoxy.

Figure 6.1.3

Disc pasted onto liner and left for a 24-hour curing period

The marking of disc positions was done using the surveyed peg positions as a reference beacon. Normally, the position of the first disk was placed 0,5 m above the grade line of the tunnel or on the centre line of the gully (Figure 6.1.4 and Figure 6.1.5). The second disk was placed 0.5 m vertically above the position of the first disk. When no liner was found at the position of the second disk, the new position was located by swivelling the disk at a radius of 0.5 m from the first disk with a fixed angle of 45º to the vertical.

86

Swivel position

Fixed disks

0.5m

0.5m Grade line

Figure 6.1.4

Figure 6.1.5

Disk positioning along the sidewall of a tunnel

Marking of positions for disks above the grade line

The same procedure was adopted for the opposite tunnel sidewall. Disk positions were offset and marked relative to the position of the first installed disk. The disks were placed on a grid to avoid a bias in selection of locations. Disks were not located at the ‘best’ and ‘most suitable’ testing positions but followed the predetermined testing layout described in the preceding paragraph. Safety precautions in line with mine-specific codes of practice for making the site safe, preparing the site, and handling of materials were enforced at all times.

6.1.2 Pull test procedure The pull test procedure required the physical work of two people. One person held the frame of the puller (Figure 6.1.6) while the other attached the grip to the disk, which needed to be hand-tightened. The pump was pressurised and the pressure at which the 87

disc detached from the rock surface (with or without the liner) recorded. The pulled-out surface was photographed and the thickness of the pulled-out liner was measured with a calliper vernier (Figure 6.1.7). If not already done, the disks were labelled or relabelled, stored in protective bags, and transported to surface. The life of the stored disks was found to not exceed three months. Beyond three months the surfaces of exposed metal disk faces were contaminated through oxidation.

Figure 6.1.6

Frame of pull tester being held in position by mining technician

Figure 6.1.7

Calliper vernier used to measure slot size and thickness of liner

88

6.1.3 Quantity of testing The number of adhesion tests conducted depended on: ?

The length of available TSL sections within the designated host mines

?

The accessibility of the sites

?

The practical constraints of conducting the tests at the sites

?

The provision of mine assistance and

?

The distance for transportation of testing equipment to the sites

The number of tests conducted per site is shown in Table 6.2

Table 6.2

Quantity of adhesion tests conducted per site

Site

Placer Dome Western Area JV Savuka (1st site) Savuka (2nd site) Kloof No.8# Driefontein 1E Total

Total per site

Total successful tests (liner and disk alone or rock/liner were pulled out)

Total unsuccessful tests (Failed on epoxy)

12

9

3

20 17 20 19 88

18 17 16 19 79

2 0 4 0 9

6.2 Description of the test sites The five test sites are located in the West and Far West Rand gold fields. Some simple geotechnical parameters of the five test sites are listed in Table 6.3. The sites were chosen on the grounds of their extensive application of TSLs during the past five years and the variation in the type and sites of application.

Table 6.3 Mine site

Brief outlines of the geotechnical characteristics of the five mines hosting the test sites Operating

Depth

depth range for stoping (m)

Dominant

Estimated

below

natural

fall out

surface

failure

thickness

(m)

planes

(m)

Savuka pipe raise

2 200 – 3 200

3 300

Savuka

2 200 – 3 200

2 010

Placer Dome

1 800 – 2 650

2 600

Reefs mined

VCR, Carbon Leader

Pilloid structures,

0,9

Green bar 0,9

VCR, MB, MI, 89

Tuffaceous

1,1

Western

EC

lavas, jointing

areas JV East Driefontein

2 000 – 3 300

2 700

VCR, Carbon Leader

1# Kloof 8#

2 000 – 2 600

1 400

Tuffaceous lavas, Green

1,2

bar

VCR, Kloof

Tuffaceous

Reef

lavas

1,4

6.3 Savuka pipe raise 6.3.1 Site location The site is located in a pipe raise that was excavated at an exit-depth elevation of 3 300 m below surface in good quality quartzite. The pipe raise was established to form a material interconnection between 104 Level and 106 Level. The Carbon Leader Reef (CLR) was extensively stoped at approximately 70 m above the pipe raise position, resulting in the raise being effectively stress relieved. Two strike-stabilising pillars border the pipe raise at dip distances exceeding 100 m from the position of the raise. The raise has typical dimensions of three metres in height and two metres in width and extends from collar to toe for approximately 100 m. The regional layout depicting the pipe raise relative to existing on- and off-reef service infrastructure is shown in Figure 6.3.1.

90

N AREA PLAN – SAVUKA

INCLINE POSITION

200m

Figure 6.3.1

200m

Regional layout of the Savuka pipe raise site showing the line of stoped areas and position of strike-stabilising pillars

The local profile of the pipe raise with positions of test sites is shown in Figure 6.3.2. SAVUKA GOLD MINE 104-106 PIPE RAISE WEST

Top of incline

f

2 0 19

96m 18 1 7 16 15 14 13 12

Start of incline

11 10 9

8 7 6 4 5 3

2

1

f Start of incline

Figure 6.3.2

Local layout of the pipe raise test site indicating test site positions

91

6.3.2 History of TSL at the test site The time frame, purpose and possible reasons for applying TSL at this site are summarised in Table 6.4.

Table 6.4 Site

History of TSL application Date of first application

104-106 pipe raise

2002

Currency of application

Numbe r of tests

Reasons for application

20

Replaced the application of shotcrete because of comparatively more favourable logistics

Completed

6.3.3 Installed support The primary and secondary support installed at the test site, in addition to the TSL, is summarised in Table 6.5.

Table 6.5

Support systems in place at the site apart from TSL

Site

Primary support

Secondary support

104-106 pipe raise

2.4m long Shepherds Crooks at a 1.5 m square pattern

Sporadic application of nonreinforced 50-mm-thick shotcrete

6.3.4 Major natural and induced discontinuities A summary of the natural and induced discontinuities observed, measured and recorded at the test sites is set out in Table 6.6. A detailed geotechnical description of the test sites is contained in Section 6.8.

Table 6.6

Dominant natural and induced discontinuities

Site

Natural major geological features

Induced discontinuities

104-106 pipe raise

No major faults or dykes are located within 70 m of the test site

Stress fractures are estimated at 10 per metre, which is a relatively low frequency for this depth, i.e. raise developed in stress-

A minor fault is located at the second

92

Site



cubby from the entrance to the site

relieved ground

Bedding planes spaced 0.5 m apart and two orthogonal joint sets spaced between 1m and 0.25 m apart form the dominant failure surfaces

6.4 Savuka strike gully 6.4.1 Site location The site is a sprayed hangingwall section of a strike gully located on the Ventersdorp Contact Reef (VCR) reef horizon. The 73-31 W1 gully is the bottom-most gully in a minilongwall and abuts a partially established 30-m-wide strike-stabilising pillar (Figure 6.4.1). The site is located at 2 010 m below surface. The current mining spans for the site are approximately 250 m on strike and 400 m on dip. The gully hangingwall is developed marginally beyond the VCR in competent high-strength lava, with an approximated strength exceeding 250 MPa. The testing was conducted on the sprayed hangingwall section of the W1 gully, the positions of which are indicated in Figure 6.4.2.

TEST SITE

Figure 6.4.1

Regional layout of the gully test site at Savuka gold mine

93

SAVUKA TSL SITE LAYOUT – 73-31 W1

16 14 12 10

8

4

6

N

2

W1 Gully V 6763 18

17 15 13 11 9

7

5

3 1

15m

Figure 6.4.2

Local setting of the Savuka gully site with positions of tests indicated

6.4.2 TSL history The time frame, purpose and possible reasons for applying TSL at this site are summarised in Table 6.7

Table 6.7 Site

73-31 W1 gully

History of TSL application Date of first application 2004

Currency of application Current

Numbe r of tests 17

94


Owing to the gully abutting the stabilising pillar, ground conditions are unfavourable because of related lowangled stress fractures and frequent episodes of strong ground motion. The TSL is applied to augment the current support systems in place and is designed to keep small key blocks in place


Table 6.8

Support systems in place at the site apart from TSL

Site

Primary support

Secondary support

73-31 W1 gully

1.5-m-long split sets on a 1-m square spacing

0.75 m x 1.5 m timber packs on shoulders, OSRO straps suspended from split sets


Table 6.9


Site



73-31 W1 gully

No major faults, and dykes are located within 30 m of the gully

Persistent low-angled fractures that are associated with the strikestabilising pillar with fracture frequencies of 30 per metre. Fracture angles range between 500 and 700.

A dominant joint set with a north south strike is traversed by the gully Lava flow structures, though none were observed, are a distinct possibility for occurrence in the hangingwall

6.5 Placer Dome Western Areas Joint Venture (PDWAJV) 6.5.1 Site location This test site is located in the mechanised trackless, large span drift within the Massive Elsburg stratigraphic package. The TSL was applied at this site approximately two years ago and is currently being monitored by the mine’s rock engineering personnel for performance and durability. The 87-2W TSL test site is situated at the intersection of six wide excavations (Figure 6.5.1). These excavations lead on to an in-stope trackless tipping point. The 2W trackless mining section is mining the Elsburg massive reefs at a depth of 2 600 m. The mining method used is drift and fill. Drifts are mined on apparent dip with excavation dimensions 95

of 5,5 m wide x 5.5 m high. Owing to large spans, (up to 8,0 m) the potential fallout block size is large. Appropriate support measures were implemented to secure the hangingwall and sidewalls of drifts. The TSL membrane was applied with the aim of consolidating smaller key blocks that help maintain overall stability.

f

2B11MR TIP

9

8

7

2B11SR

1 6 5 4 3

10

11 12

SOUTH DEEP GOLD MINE 87-2W TRACKLESS

Evermine Sidewall Evermine Hangingwall

Figure 6.5.1

2

f

Local site layout for the PDWAJV test programme

The test site was stress relieved by the extensive understoping of the EC reef horizon located some 20 m below the massives.

6.5.2 TSL history The time frame, purpose and possible reasons for applying TSL at the site are summarised in Table 6.10.

Table 6.10 Site

87-2W

History of TSL application Date of first application 2002


Number of tests

Completed but being replaced with shotcrete

12

96


Testing of suitability of TSL to replace conventional shotcrete


Table 6.11

Support systems in place apart from the TSL

Site

Primary support

Secondary support

87-2W

2.4-m-long rebars on a 2-m square pattern

4 m long rope anchors integrated with the primary support on a 2 m square pattern Sporadic application of reinforced 50-mm-thick shotcrete to replace damaged TSL


Table 6.12 Site 87-2W

Dominant natural and induced discontinuities Natural major geological features


The tip area intersects a north west – south east fault system, which displays distinct slickensided surfaces and mylonitic infilling

Low-frequency stress fractures are located at the edge of pillars that define the intersection points of drifts

Sub-parallel sympathetic joints to the fault system are located in the area

Fracture frequencies range between five and ten per metre.

The most dominant failure planes are the bedding, which display well bedded, abrupt partings at a spacing of 0,5 m

6.6 Kloof 8 Shaft 6.6.1 Site location The test site is located within a haulage and a crosscut, whereby the haulage is developed at approximately 60 m below the plane of the reef and the crosscut is developed to intersect the reef. The haulage is positioned at varying degrees within the stoping abutment of the VCR horizon.

97

The crosscut was developed in stress-relieved ground. The haulage was recently rehabilitated as a result of high abutment stresses on reducing tramming clearances, and the blocky, friable nature of the hangingwall as a result of the extensive failed rock envelope around the haulage, created by the high abutment stresses. The excavation has typical dimensions of 3,7 m high by 3 m wide. Figure 6.6.1 is an illustration of the regional setting of the haulage and crosscut relative to the mined-out VCR reef horizon. The footwall excavations are located at an approximate grade elevation of 1 400 m below surface.

Figure 6.6.1

Regional setting of the 14L Footwall Drive East at Kloof 8#

6.6.2 TSL history The time frame, purpose and possible reasons for applying TSLs at the site are summarised in Table 6.13. The footwall drive was developed during the 1960’s, 1970s and 1980s and was overstoped on the VCR horizon. The cutting of the top VCR abutment affected both the direction and intensity of the stress field, which led to typical Gothic-arch fallout in the hangingwall of the haulage. Rehabilitation of the tunnel commenced during 2002 as part of the access programme for VCR remnant pillar mining. Tunnelguard TSL was chosen to consolidate the damaged rock mass prior to the installation of wire mesh and lacing. As the tunnel is now sited under a static abutment, little or no further deformation is expected.

98

Table 6.13 Site

History of TSL application Date of first application

14L-FW Drive East

2002


Number of tests

Completed but integrated with wire mesh and lacing

20


Consolidation of loose already damaged wall rock in the haulage. Accepted support method


Table 6.14


Site

Primary support

Secondary support

14L-FW Drive East

2,4 m long rebars on a 2 m square pattern

Wire mesh and lacing integrated with the rebar support on a 2 m diamond pattern


Table 6.15


Site


14L-FW Drive East

A 20 m wide dyke is located within 50 m of the Low-frequency stress fractures are crosscut (inactive with respect to seismicity) located in the crosscut and sections of the haulage that are stress relieved No other major geological features traverse the site The excavations traverse bedding and a conjugate joint set. Bedding is spaced at 0,5 m. The dominant joint set is spaced one metre apart and the second generation set five metre apart

99


Fracture frequencies in stressed sections of the haulage range between 30 and 50 per metre

6.7 Driefontein 1 East 6.7.1 Site location The Driefontein site comprises an access tunnel being developed in the footwall shale horizon. The access tunnel is sited at a depth of 2 700 m below surface and approximately 70 m in the footwall of the Carbon Leader Reef. VCR was mined out on a higher elevation but the influence of this mining can be neglected. The regional setting of the test site and the location of test disks within the site are shown in Figure 6.7.1 and Figure 6.7.2, respectively. No stoping has occurred in the vicinity of the section of footwall haulage where the tests were performed.

Figure 6.7.1

Test site at the 36-22 FW Drive East at Driefontein No.1 East

100

N

DRIEFONTEIN No 1# East

D 1 = Calibration site

36 - 22 Footwall Drive east Disk D2 - D12 25m

Sprayed December 2003 Sprayed January 2004 Disk D13 - D20 15m

40m

Depth = Liner product Contracror Pump Type

Figure 6.7.2

Schematic illustrating the test positions at the 36-22 FW Drive East at Driefontein No.1 East

6.7.2 TSL history The time frame and possible reasons for applying TSLs at this site are summarised in Table 6.16 History of TSL application at the Driefontein site. The deep footwall drive is being developed on strike in the Jeppestown Shale horizon. The overall rock mass strength of this shale horizon is inherently lower than other Witwatersrand rock strata, so more intense support measures may be required. A TSL was chosen as a replacement for sprayed shotcrete, due to the lower bulk volumes required. A further advantage was that the TSL could be applied closer to the face than shotcrete. Current support-to-face distance is approximately ten metres.

Table 6.16 Site

36-22 FW Drive East

History of TSL application at the Driefontein site Date of first application 2003

Currency of application Ongoing, providing aerial coverage between the primary tendon support

101

Number of tests


19

Accepted aerial coverage support system


Table 6.17


Site

Primary support

Secondary support

36-22 FW Drive East

2,3 m long smooth bars on a Wire mesh and lacing integrated with the 2 m square pattern smooth bar and 4 m long rope anchor support on a 2 m diamond pattern Wire mesh and lacing follows the application of TSL


Table 6.18


Site



36-22 FW Drive East

The end of the footwall drive is being developed through a 20 m wide dyke

Fracture frequencies range between 30 and 50 per metre.

Bedding is closely spaced at approximately 0,3 m A joint set striking north west – south east with a 1,5 m spacing is traversed by the drive

6.8 Geotechnical characteristics of the test sites Scan-line geotechnical mapping was done at or close to the test sites to obtain geotechnical data that would allow for an estimation of the rock mass rating. This geotechnical data set could also by useful for input into any numerical models that may be generated to simulate the geometrical setting and the potential TSL response to block deformations. The data set that was collated is shown in Appendix A1 Using the Tunnel Assistant programme, which automatically calculates the rock mass rating and Q-value for excavations, given the appropriate input variables, an estimate of site rock mass conditions was established. These ratings used to describe the site conditions are provided in Table 6.20.

102

The purpose of accounting for geotechnical conditions is to establish a quantitative, systematic method for the design of support systems. The guidance offered by rock mass classification systems can then be superimposed on observations made underground to formulate meaningful support strategies. In order to rate and/or rank the conditions of the exposed wall rock in the excavations, an estimate must be made of the: ?

Rock strength

?

The rock quality or quality of core recovery

?

The orientation and condition of joints

?

The presence and persistence of water in the rock mass

The ratio of the strength of the rock mass to the applied stress fields The Q-system developed by Barton et al. (1974) was selected to assess the excavations, as it can provide quantitative data for the design of tunnel support systems. The basis of the system is to describe the rock mass in terms of three quotients: block size, inter-block shear strength and active stress. Although joint orientation is not included, it is considered relevant, because the description of the joints applies to those joints that are most unfavourably orientated. The overall rating (Q) is the product of the three quotients used to describe the rock mass. Where RQD is the rock quality designation, Jn the joint set number, Jr the joint roughness number, Ja the joint alteration number, Jw the joint water reduction number, and SRF the stress reduction factor. Sections of the excavations defined earlier were rated using the Q system. The descriptors assigned to the Q-values are contained in Table 6.19.

Table 6.19

Quality classification of the rock mass based on q-values Tunnel quality index < 0.01 0.01 – 0.1 0.1 – 1 1–4 4 – 10 10 – 40 40 – 100 100 – 400 > 400

Table 6.20 Host mine

Placer Dome Western areas Joint Venture Savuka

Classification Exceptionally poor Extremely poor Very poor Poor Fair Good Very good Extremely good Exceptionally good

Listing of rock mass rating values for the test sites Site name

Laubsche r RMR value

Description Qvalue

Description Average RQD

2B11SR Sth sidewall

76

Good

10,2

Good

62,5

Pipe Raise

74

Good

8,89

Fair

75

103

Host mine

Site name

Kloof 8#

14L FW Drive East 73-31 W1 gully 36-22 FW Drive East

Savuka Driefontein 1E

Laubsche r RMR value

Description Qvalue

Description Average RQD

61

Good

6,67

Fair

35

61

Good

6,67

Fair

85

45

Fair

4,4

Poor to fair

75,3

6.9 Test results 6.9.1 Monitored data The following tables show the results that were obtained from the underground adhesion tests.

Table 6.21 Disc no SVA01 SVA02 SVA03 SVA04 SVA05 SVA06 SVA07 SVA08 SVA09 SVA10 SVA11 SVA12 SVA13 SVA14 SVA15 SVA16 SVA17 SVA18 SVA19 SVA20

Table 6.22 Disc No SVB01

Adhesion test results for Savuka pipe raise Curing time Maximum (epoxy) pressure 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24

0,20 1,20 1,20 0,20 0,60 0,40 0,20 0,20 0,20 0,20 0,20 0,40 0,50 0,60 0,70 0,20 0,60 0,20

Thicknes s (mm) 2,0 3,0 4,0 2,0 4,0 2,0 4,0 2,0 2,0 2,0 3,0 2,0 2,0 4,0 3,0 4,0 3,0 2,0 3,0

Remarks Pulled by hand

Abandoned Pulled by hand

Pulled by hand Pulled by hand Pulled by hand Pulled by hand Pulled by hand Pulled by hand

Pulled by hand Pulled by hand

Results for Savuka gully Curing Time Maximum (Epoxy) Pressure 72 0,8

Thicknes s (Mm) 2,5 104

Remarks

Disc No SVB02 SVB03 SVB04 SVB05 SVB06 SVB07 SVB08 SVB09 SVB10 SVB11 SVB12 SVB13 SVB14 SVB15 SVB16 SVB17

Table 6.23 Disc No

SD 01 SD 02 SD 03 SD 04 SD 05 SD 06 SD 07 SD 08 SD 09 SD 10 SD 11 SD 12

Table 6.24 Disc no D1 D2 D3 D4 D5 D6 D7 D8 D9

Curing Time Maximum (Epoxy) Pressure 72 1,2 72 1,4 72 1,2 72 1,6 72 1,6 72 0,2 72 1,8 72 2,0 72 2,6 72 2,4 72 0,2 72 0,6 72 1,6 72 2,0 72 1,8 72 0,2

Thicknes s (Mm) 2,5 2,5 4,0 2,0 2,0 3,0 4,0 2,0 2,0 2,5 2,0 2,5 1,0 2,0 1,5 1,0

Remarks

Results for Placer Dome Westonaria Joint Venture 2B11SR South sidewall Curing Time Maximum (Epoxy) Pressure 24 24 24 24 24 24 24 24 24 24 24 24

Thicknes s (Mm)

1,00 1,80 1,20 1,80 1,60 1,90 2,10 1,10 0,80 0,00 0,00 0,00

3,0 1,5 3,0 1,5 2,5 3,0 1,5 2,0 3,0

Remarks

Oxidation on pyrite, rock failure

Failed on epoxy Failed on epoxy Failed on epoxy

Results for East Driefontein 1E 36-22FW drive east Curing time Maximum (epoxy) pressure (MPa) 24 24 24 24 24 24 24 24 24

Thickness (mm)

5,0