Figure 3.14: Garford Dynamic Cable Bolt strand and anchor configuration for samples 78-80. 91 ...... Threadbar may also be referred to as Gewi bar, a.
MINERALS AND ENERGY RESEARCH INSTITUTE OF WESTERN AUSTRALIA (MERIWA)
REPORT NO. 287
DYNAMIC TESTING OF GROUND CONTROL SYSTEMS Results of research carried out as MERIWA Project No. M349A at the WA School of Mines, Curtin University of Technology
by
E Villaescusa, A Thompson, J Player and E Morton
December 2010
Distributed by: MERIWA Mineral House 100 Plain Street EAST PERTH WA 6004 Crown copyright reserved ISBN 1 920981 48 9
To which all enquiries should be addressed
EXECUTIVE SUMMARY The original MERIWA M349 research project was initiated in response to the recognition that ground conditions were becoming increasingly more difficult as mines became deeper. One of the main technical problems is violent rock failures caused by high in situ and mining-induced stresses and the related seismicity and dynamic loadings on ground support. It is necessary to measure and understand the dynamic response characteristics of the reinforcement and support systems in order to design effective ground support schemes. Prior to this project, there was no facility in Australia capable of performing the dynamic tests required to provide the design data. In addition, the test facilities used overseas were thought not to simulate the loadings expected from violent rock failure and dynamic loadings and therefore would not provide the data required for design. A new loading concept involving momentum transfer was conceived by the principal researcher Professor Ernesto Villaescusa. A prototype was designed and constructed and tests performed to demonstrate the validity of this loading concept prior to developing the proposal to conduct this research project. The project, prior to commencement, was divided into two main phases: • Phase 1 (M349) was to design, build and commission a test facility and instrumentation and to test rock reinforcement systems comprising various elements, internal fixtures, external fixtures and face restraint. • Phase 2 (M349A) was anticipated to undertake any modifications required to the test equipment and instrumentation and to perform tests on reinforcement and surface support systems used by the Western Australian mines. MERIWA Project M349 commenced in June 2002 and was completed in 2005. The achievements included: • The WASM Dynamic Test Facility was designed and constructed. • The state-of-the art instrumentation and monitoring system was designed, purchased and commissioned. • Development of computer software for the efficient analysis of the large quantities of data collected during tests. • A preliminary database was established for reinforcement system responses to various levels of dynamic loading. The investigations undertaken in MERIWA Project M349A commenced in July 2006 and were completed in the first half of 2010. This report documents the outcomes. The main achievements have been generally: • Modifications to the test facility to make test preparation safer and more efficient. • On-going modifications to the instrumentation and enhancements to the data analysis software. With specific regard to reinforcements systems, the following achievements were made: • A procedure was developed to better simulate a borehole in rock, particularly for reinforcement systems where the interactions between the borehole wall and element are important. • Completion of over 80 further tests on reinforcement systems.
i
• The facility has been used to economically evaluate proposed new reinforcement systems; in some cases, the tests showed that the systems as proposed would not perform as anticipated. In other cases, the reinforcement system was able to be modified and its dynamic performance improved. Additionally, a program of work on static testing of surface support systems was undertaken. This work was funded internally by WASM and separate to any MERIWA and industry contributions. However, the results from these investigations are reported in the Appendix because the static testing program was designed to formulate and evaluate different surface support system specimen preparation techniques that could then be adopted for the dynamic test facility. The major project outcomes with regard to dynamic testing of support systems have been: •
Enhancement of the facility to include dynamic testing of mesh and shotcrete panels.
•
Testing of woven and welded mesh samples.
•
Development of a new ‘punch’ test configuration to evaluate adhesion of shotcrete to rock as well as the shear and bending resistance.
•
Testing of shotcrete slabs.
The perceived benefits are: •
For sponsors: o
Access to a facility by suppliers and users of ground support technology for testing and more timely development of new and improved ground support systems.
•
For the mining industry in general: o
A database of ground support system responses that can be used to assist in the design of appropriate ground support schemes.
o
Improved robustness of ground support schemes can improve safety and reduce the need for time consuming and expensive rehabilitation with adverse effects on productivity and economics of mining.
o
Exposure of undergraduate students to systematic investigations resulting in more indepth knowledge and understanding of ground support performance.
•
For the national and international research and development community: o
Demonstration of an improved testing methodology that can be adapted and implemented at other research institutions.
o
Improved understanding of the behaviour of ground support systems in response to dynamic loading and comparisons of dynamic responses with responses to static loadings.
•
For the wider community: o
Improved design and reliability of ground support schemes will result in a safer mining environment with consequent economic and social benefits.
ii
In summary, the combination of funding from MERIWA, the mining industry companies and suppliers, and the WASM Rock Mechanics Group, has resulted in the establishment of a national testing facility that is unique in concept and design. It is anticipated that the WASM Dynamic Test Facility will continue to be used into the future for the measurement and evaluation of the response of ground support systems. In particular, it has been identified that further investigations would be required to demonstrate the full capabilities of the WASM Dynamic Test Facility. The future work would involve modifications to the test equipment and instrumentation in order to perform tests on integrated ground support schemes comprising reinforcement systems and surface support systems used by the Western Australian mines. The outcome of these investigations would be to determine the relative energy absorption capacities of the individual ground support systems and to estimate for the purposes of design the fully-integrated ground support scheme capacity.
iii
iv
CONTENTS 1
BACKGROUND INFORMATION
1
1.1 PROJECT OBJECTIVES, SCOPE OF WORK AND TASKS 1.2 GROUND SUPPORT SCHEMES AND TERMINOLOGY 1.2.1 Load Transfer Concept for Ground Support Schemes 1.2.2 Reinforcement System Load Transfer 1.2.3 Support System Load Transfer 1.2.3.1 Mesh Load Transfer 1.2.3.2 Shotcrete Load Transfer 1.3 ROCK MASS LOADINGS 1.4 WASM DYNAMIC TEST FACILITY AT THE COMPLETION OF M349
1 2 3 4 7 8 11 14 17
2
21
WASM DYNAMIC TEST FACILITY – CURRENT STATUS
2.1 INFRASTRUCTURE 2.1.1 Building 2.1.2 Storage 2.1.3 Gantry Crane 2.2 HARDWARE 2.2.1 Release Mechanism 2.2.2 Drop Beam 2.2.3 Drop Frames 2.2.3.1 Mesh 2.2.3.2 Shotcrete 2.2.4 Buffers 2.2.5 Simulated Rock Loading 2.2.5.1 Reinforcement 2.2.5.2 Mesh 2.2.5.3 Shotcrete 2.3 FORCE TRANSFER AND DISPLACEMENTS IN TEST FACILITY 2.3.1 Reinforcement 2.3.2 Mesh 2.3.3 Shotcrete 2.3.4 Relative Displacement, Velocity and Acceleration 2.4 DATA ACQUISITION SYSTEM AND SENSORS 2.4.1 Data Acquisition System 2.4.1.1 Time window 2.4.1.2 DAQ connection 2.4.1.3 Schematic of final data acquisition equipment 2.4.2 Sensors 2.4.2.1 Component Accelerations 2.4.2.2 Laser Breaks 2.4.2.3 Beam/Buffer Displacements 2.4.3 Strain Gauge Based Sensing 2.4.3.1 Load cells 2.4.4 High Speed Camera 2.4.5 Physical Measurements 2.4.6 Data Recording 2.5 WASM DATA ANALYSIS SOFTWARE 2.5.1 Menu 2.5.2 Data Visualisation 2.5.3 Data Filtering 2.5.4 High Speed Video Analysis 2.5.4.1 Smoothing Video Track 2.5.5 Intermediate Data File Processing
v
21 21 21 22 25 25 26 28 28 30 31 33 33 34 35 36 36 38 38 39 40 40 41 41 41 43 43 44 45 47 47 50 52 53 53 53 54 56 58 59 62
2.5.6 Data File STORAGE 2.6 ENGINEERING ANALYSIS 2.6.1 Assumptions for the Analysis 2.6.2 Input Data Files 2.6.3 Energy-Time Chart 2.6.3.1 Buffers and Energy Dissipation 2.6.3.2 Total Error Analysis 2.7 STANDARD TEST REPORT 2.7.1 Test Summary Section 2.7.1.1 Reinforcement System Specification 2.7.1.2 Test specifications 2.7.1.3 System Performance 2.7.1.4 Overall system performance 2.7.2 Detailed Analysis and Interpretation Section 2.8 ASSESSMENT OF THE WASM TEST FACILITY 2.8.1 Features of the Facility 2.8.2 Limitations of the Facility 2.8.3 Concluding Remarks
62 63 63 64 66 66 67 69 71 71 72 73 74 74 75 75 76 77
3
79
DYNAMIC TESTING OF REINFORCEMENT SYSTEMS
3.1 SIMULATED BOREHOLES 3.1.1 Pipe Radial Stiffness 3.1.2 Standard Borehole 3.1.3 Rough Borehole 3.1.3.1 Material Selection 3.1.3.2 Creation of a Simulated Rough Borehole 3.2 ELEMENTS 3.2.1 Friction Rock Stabilisers 3.2.1.1 Split Tube 3.2.1.2 Expanded Tube 3.2.2 Threadbar 3.2.2.1 High Strength and Low Strength Threadbar 3.2.2.2 Comparison of Element Mechanical Properties 3.2.3 Cone Bolt 3.2.4 Modified Cone Bolt 3.2.5 Plain Strand 3.2.6 Garford Bolts 3.2.6.1 Dynamic Cable Bolt 3.2.6.2 Dynamic Solid Bolt 3.3 CONTINUOUSLY FRICTIONALLY COUPLED SYSTEMS 3.3.1 Split Tube Bolts 3.3.1.1 Sample Preparation 3.3.1.2 Static Test Results 3.3.2 Expanded Tube Bolts 3.3.2.1 Sample Preparation 3.3.2.2 Static Test Result 3.3.3 Summary for CFC 3.3.3.1 Dynamic Test Results for Split Tube Bolts 3.3.3.2 Dynamic Test Results for Omega Bolts 3.3.4 Comparison of CFC Systems 3.4 CONTINUOUSLY MECHANICALLY COUPLED SYSTEMS 3.4.1 Threadbar Encapsulated with Cement Grout 3.4.1.1 Sample Configuration 3.4.1.2 Cement Grout Encapsulation 3.4.1.3 Dynamic Test Results 3.4.1.3.1 HS Threadbar
vi
79 80 81 81 82 82 87 87 87 87 88 88 89 89 90 91 91 91 92 92 92 92 93 95 95 96 99 100 101 104 105 105 105 105 106 106
3.4.1.3.2 LS Threadbar 3.4.1.3.3 Influence of confinement 3.4.1.3.4 Comparison of responses 3.4.1.3.5 Influence of the number of buffers 3.4.2 Threadbar Encapsulated with Resin Grout 3.4.3 Plain Strand 3.4.3.1 Sample Preparation 3.4.3.2 Surface Hardware 3.4.3.3 Testing Results for Australian Strand 3.4.3.4 Surface hardware 3.4.4 Chilean Plain Strand 3.4.4.1 Sample Preparation 3.4.4.2 Dynamic Testing Results 3.4.4.3 Influence of number of buffers 3.4.4.4 Influence of confinement 3.4.4.5 Comparison between Plain Strand Cable Bolts 3.5 DISCRETE MECHANICALLY AND FRICTIONALLY COUPLED SYSTEMS 3.5.1 Partially Decoupled HS Threadbar 3.5.1.1 Sample Preparation 3.5.1.2 Surface Hardware 3.5.1.3 Dynamic Test Results 3.5.2 Toe Anchored HS Threadbar 3.5.2.1 Sample Preparation 3.5.2.2 Surface Hardware 3.5.2.3 Dynamic Test Results 3.5.3 Cone Bolt 3.5.3.1 Sample Preparations 3.5.3.2 Surface Hardware 3.5.3.3 Dynamic Test Results 3.5.3.4 Cone bolt performance in high strength grout 3.5.3.5 Cone bolt performance in low strength grout 3.5.3.6 Confinement and radial expansion of the pipe 3.5.3.7 Surface hardware 3.5.3.8 Summary for Cone Bolts 3.5.4 Modified Cone Bolt 3.5.5 Partially Decoupled Bulbed Strand 3.5.5.1 Bulb encapsulation and performance 3.5.6 Garford Dynamic Cable Bolt 3.5.6.1 Sample Preparation 3.5.6.2 Testing Results 3.5.6.2.1 Version One installed in cement grout 3.5.6.2.2 Version Two installed in cement grout 3.5.6.3 Summary for Dynamic Cable Bolt 3.5.7 Garford Dynamic Solid Bolt 3.5.7.1 Testing Results 3.5.7.1.1 Version One installed in cement grout 3.5.7.1.2 Version Two installed in cement grout and resin 3.5.7.2 Dissection of simulated rough boreholes 3.5.7.3 Dependence on loading velocity 3.5.7.4 Multiple loadings 3.5.8 Summary for DFMC Systems 3.6 SUMMARY OF PERFORMANCE OF REINFORCEMENT SYSTEMS
108 110 113 115 115 118 119 119 119 121 123 123 123 126 126 128 131 131 131 132 133 135 135 136 136 137 137 137 137 139 141 142 143 144 145 145 147 149 149 150 150 150 155 155 155 155 156 159 161 161 162 163
4
165
DYNAMIC TESTING OF SUPPORT SYSTEMS
4.1 MESH TESTING 4.1.1 Mesh Samples
165 165
vii
4.1.2 Mesh Dynamic Testing Results 4.1.3 Comparison of static and dynamic results 4.1.3.1 Welded wire mesh 4.1.3.2 Chain link mesh 4.1.3.3 Concluding Remarks for Mesh Testing 4.2 SHOTCRETE TESTING 4.2.1 Sample Configuration and Preparation 4.2.2 Dynamic Testing 4.3 SUMMARY FOR TESTING OF SUPPORT SYSTEMS
166 168 168 170 171 172 172 172 176
5
177
SUMMARY OF OUTCOMES AND CONCLUDING REMARKS
5.1 5.2 5.3 5.4 5.5 5.6
WASM DYNAMIC TEST FACILITY REINFORCEMENT SYSTEM DYNAMIC TESTING MESH TESTING SHOTCRETE TESTING STAFF AND STUDENTS RESEARCH SKILLS DEVELOPMENT TECHNOLOGY TRANSFER 5.6.1 Published Papers 5.6.2 Seminars/Workshops/Courses 5.6.3 Progress Reports 5.6.4 Sponsors’ Meetings 5.7 ACKNOWLEDGEMENTS 5.8 CONCLUDING REMARKS AND FUTURE DEVELOPMENTS
177 177 179 179 179 179 179 181 181 181 181 182
6
185
REFERENCES
APPENDIX Static Testing of Surface Support Systems (Separate document)
viii
LIST OF FIGURES Figure 1.1: Load transfer from surface support to surrounding reinforcement systems.
3
Figure 1.2: Load transfer between surface support and the surrounding rock surface.
3
Figure 1.3: Arch shaped surface support that does not require reinforcement.
4
Figure 1.4: Arch performance improved by road base and reinforcement.
4
Figure 1.5: Reinforcement load transfer from unstable rock to stable rock
5
Figure 1.6: The components of a reinforcement system.
5
Figure 1.7: Schematic showing the forces involved in load transfer for reinforcement systems.
6
Figure 1.8: Schematic showing the different element force distributions within each of the three classes of reinforcement systems.
7
Figure 1.9: Deformed mesh with higher forces shown in darker colours.
10
Figure 1.10: Deformed mesh with higher forces shown in darker colours.
10
Figure 1.11: Forces acting on a segment of a curved arch section.
11
Figure 1.12: Failure mechanism of shotcrete (Barrett and McCreath 1995)
12
Figure 1.13: Updated shotcrete failure mechanisms.
13
Figure 1.14: Yield line pattern for (a) centrally loaded square slab supported at the edges and (b) centrally loaded circular slab supported at the three equally spaced points.
14
Figure 1.15: Schematic representation of the behaviour of rock when subjected to remotely generated seismic loading.
15
Figure 1.16: Schematic representation of rock failure loading surface support between reinforcement
16
Figure 1.17: Schematic representation of rock failure loading both surface support and reinforcement.
16
Figure 1.18: WASM Prototype dynamic loading of ground support scheme.
17
Figure 1.19: WASM Dynamic Test Facility.
19
Figure 2.1: Storage racks for reinforcement system samples before and after testing.
22
Figure 2.2: Mezzanine with original static crane beam and unpainted guide rails.
23
ix
Figure 2.3: Installation of gantry crane and mezzanine change, October 2005.
23
Figure 2.4: Gantry crane view from the mezzanine floor
24
Figure 2.5: Painted guide I-beams and mezzanine change.
24
Figure 2.6: WASM Dynamic test facility following upgrade (February 2006).
25
Figure 2.7: Release Hook.
26
Figure 2.8:Lifting hook on top of the drop beam.
26
Figure 2.9: Luders lines on grey drop beam.
27
Figure 2.10. Stiffened beam for attachment of the frames required for mesh and shotcrete testing.
28
Figure 2.11: Dynamic mesh test configuration.
29
Figure 2.12: Increase frame connection bolt holes.
30
Figure 2.13: Dynamic fibrecrete test equipment.
31
Figure 2.14: Drilling additional buffer connection bolts.
32
Figure 2.15: Connection beam for buffers.
32
Figure 2.16: Schematic of load transfer rings and integration with the steel pipe.
33
Figure 2.17: Deformation with 990kg load pre test on the Geobrugg Tecco Mesh.
34
Figure 2.18: Curved loading mass for mesh panel testing.
35
Figure 2.19: Loading mass used for shotcrete panel testing.
36
Figure 2.20: Free-body diagram showing load transfer mechanisms within a reinforcement system.
37
Figure 2.21: Free-body diagram showing load transfer mechanisms within a mesh panel test.
38
Figure 2.22: Free-body diagram showing load transfer mechanisms within a shotcrete panel test.
39
Figure 2.23: Assumed displacement mechanism within a shotcrete panel test.
39
Figure 2.24: Schematic of instrumentation connections.
42
Figure 2.25: Laser breaks.
45
Figure 2.26: Ultrasonic displacement sensor location relative to other test components.
46
x
Figure 2.27: Potentiometer mounted on buffer to measure piston displacement.
47
Figure 2.28: Collar load cell application.
48
Figure 2.29: Anchor load cells.
49
Figure 2.30: Anchor load cell calibration.
49
Figure 2.31: Camera angle test.
51
Figure 2.32: Output from camera during test.
51
Figure 2.33: Toe displacement measurements.
52
Figure 2.34: Displacement measurements at simulated discontinuity.
52
Figure 2.35: Menu for dynamic mesh and fibrecrete test data analysis.
54
Figure 2.36: Display of raw data file (.au2) details.
55
Figure 2.37: Several sets of data displayed on the one chart.
55
Figure 2.38: Channel Description form.
57
Figure 2.39: Data storage interface form.
57
Figure 2.40: Midas track of targets.
58
Figure 2.41: Averaging displacement from a video track.
60
Figure 2.42: Velocity from average displacement tracks.
60
Figure 2.43: Acceleration from displacement tracks.
61
Figure 2.44: Displacement of buffers and video track during the test.
62
Figure 2.45: Filtered input data example showing start of video data synchronised with the channel data.
64
Figure 2.46: Filtered waveforms with synchronised start times.
65
Figure 2.47: Interface used to select channels for analysis.
66
Figure 2.48: Error analysis - stable reinforcement system test.
68
Figure 2.49: Error analysis - unstable reinforcement system test.
69
xi
Figure 2.50: Typical test report - page 1.
70
Figure 2.51: Test report sheet – reinforcement system specification.
72
Figure 2.52: Test report sheet – test specification.
72
Figure 2.53: Test report sheet - reinforcement system performance.
73
Figure 2.54: Test report sheet - overall system performance.
74
Figure 2.55: Snapped 15.2mm plain strand cables.
76
Figure 3.1: Schematic of a rough simulated sample.
83
Figure 3.2: Simulated boreholes ready for filing.
84
Figure 3.3: Poured cement mix and tested cement sample.
85
Figure 3.4: Airleg drilling of the sample to form the simulated rough borehole.
85
Figure 3.5: Custom built borehole probe.
86
Figure 3.6: Typical borehole profile results.
86
Figure 3.7: Schematic of split tube.
87
Figure 3.8: Schematic of inflatable tube.
88
Figure 3.9: Threadbar diameters.
88
Figure 3.10: Comparison of mechanical responses of Australian and Chilean threadbar.
89
Figure 3.11: 22mm diameter cone bolt.
90
Figure 3.12: Modified cone bolt (figure from Mansour Mining website).
90
Figure 3.13: Modified Cone Bolt.
91
Figure 3.14: Garford Dynamic Cable Bolt strand and anchor configuration for samples 78-80.
91
Figure 3.15: The Garford Dynamic Solid Bolt.
92
Figure 3.16: Installation of a split tube bolt into a simulated rough borehole.
93
Figure 3.17: Static collar load and displacement for 47mm diameter split tube bolts.
94
Figure 3.18: Striations along the steel due to sliding.
95
xii
Figure 3.19: Omega bolt installation.
96
Figure 3.20: Swellex sample configuration and static pull test results for one metre embedment.
97
Figure 3.21: Static pull tests on Omega bolts prior to dynamic testing.
98
Figure 3.22: Static pull test on Omega bolts after dynamic testing.
98
Figure 3.23: Dynamic force-displacement responses for split tube bolts.
100
Figure 3.24: Energy dissipated and bolt displacement from impact load.
101
Figure 3.25: Peak sliding velocity and displacement for split tube bolts.
101
Figure 3.26. Dynamic force-displacement responses for initial tests on Omega bolts.
102
Figure 3.27. Energy dissipated and bolt displacement from repeated loading of Omega bolts.
103
Figure 3.28. Peak sliding velocity and displacement for Omega Bolts.
103
Figure 3.29: Comparison of static and dynamic performance for friction stabilisers.
104
Figure 3.30: Fully encapsulated thread bar configuration.
105
Figure 3.31: Dynamic force-displacement responses for HS threadbar.
106
Figure 3.32: Fully encapsulated HS threadbar – critical and non-critical loading conditions.
108
Figure 3.33: Threadbar that has slid out of the grout at the third dynamic axial load.
108
Figure 3.34: Dynamic force displacement response for LS threadbar.
109
Figure 3.35: Thin wall pipe diameter pre-test and post-test.
110
Figure 3.36: Dissection of simulated boreholes 146 and 147.
111
Figure 3.37: Comparison of confinement on mild steel threadbar.
112
Figure 3.38: Dissection of simulated borehole 135.
113
Figure 3.39: Dynamic force displacement responses for cement encapsulated threadbars.
114
Figure 3.40: Resin reporting to collar of simulated borehole.
115
Figure 3.41: Dynamic force displacement response for resin encapsulated threadbar.
116
Figure 3.42: Bolt 148 failure surface.
117
xiii
Figure 3.43:Dynamic force-displacement responses for cement and resin encapsulated threadbar.
118
Figure 3.44: Dynamic force-displacement response for Australian plain strand.
120
Figure 3.45: Ruptured strand or sliding strand dependent on the embedment lengths.
121
Figure 3.46: Plain strand with no surface hardware prior to testing.
122
Figure 3.47: Outcome from strand test with no surface hardware.
122
Figure 3.48: force displacement response for Chilean plain strand.
123
Figure 3.49: Snapped non-centralised strand from sample 128.
125
Figure 3.50: Damaged strand from cutting the thin wall pipe compared to no damage.
125
Figure 3.51: Damage to wires from cutting the simulated discontinuity, sample 131.
126
Figure 3.52: Thin wall pipe diameter pre-test and post-test.
127
Figure 3.53: Dissection of simulated borehole 128.
128
Figure 3.54: Detail of grout damage about shear pins, sample 128.
128
Figure 3.55: Dynamic force displacement responses for Australian and Chilean plain strand.
129
Figure 3.56: Decoupled threadbar configuration.
132
Figure 3.57: Surface hardware used on decoupled threadbar.
132
Figure 3.58: Dynamic force-displacement response for decoupled threadbar.
133
Figure 3.59: Shear of threads inside short length ‘mine nut’.
134
Figure 3.60: Shearing of the thread on the bar with integrated nut and spherical based washer.
135
Figure 3.61: Jumbo installation of toe anchored resin bolt.
135
Figure 3.62: Surface hardware configuration.
136
Figure 3.63: Dynamic force-displacement responses for cone bolts.
138
Figure 3.64: Grout pulverisation from cone displacement.
140
Figure 3.65: Decoupling of the bar from the grout.
140
Figure 3.66: Surface fixture sample 60, three impacts.
141
xiv
Figure 3.67: Grout connection collar side of discontinuity.
141
Figure 3.68: Thick wall pipe deformation post cone bolt test.
142
Figure 3.69: Rubber plate plus two dome plates.
143
Figure 3.70: Two dome plates for surface hardware.
144
Figure 3.71: Single dome plate on a cone bolt.
144
Figure 3.72: Force displacement responses for bulb strand with central decoupling.
145
Figure 3.73: Rupture through the strand from the wedges.
146
Figure 3.74: Wedge location pre- and post-test sample 142.
147
Figure 3.75: Dissection of anchor encapsulation bulb strand.
148
Figure 3.76: Garford Dynamic Cable Bolts and Decoupled Bulb Strand.
150
Figure 3.77: Garford Dynamic Cable Bolt measured response, sample 78 and 80.
151
Figure 3.78: Dynamic force-displacement response for the Garford Dynamic Cable Bolt.
152
Figure 3.79: Sliding of strand through barrel and wedge, sample 144.
153
Figure 3.80: Raw stick slip response of the anchor cells, sample 143.
153
Figure 3.81: Filtered instrumentation response sample 143.
154
Figure 3.82: Enlarged replacement king wires recovered after testing.
154
Figure 3.83: Comparison of the Garford Dynamic Solid Bolt with the 22mm Cone Bolt.
156
Figure 3.84: Force-displacement responses from Garford Dynamic Solid Bolt (V2).
157
Figure 3.85: Cup and cone fracture, bar from bolt 99 after 3rd impact.
158
Figure 3.86: Resin mixing mechanism on the Garford Dynamic Solid Bolt.
159
Figure 3.87: Dissection of toe length from sample 98.
160
Figure 3.88: Short encapsulation length below anchor and displacement of the anchor.
160
Figure 4.1: Chain link and welded wire mesh, respectively.
166
xv
Figure 4.2: Typical dynamic force – displacement responses for welded wire mesh and chain link mesh.
167
Figure 4.3: Summary of force - displacement responses for welded wire mesh and chain link mesh.
167
Figure 4.4: Rupture energy results for welded wire mesh and chain link mesh.
168
Figure 4.5: Static and dynamic force – displacement response for welded wire mesh.
169
Figure 4.6: Comparison of static and dynamic force and displacement properties for welded wire mesh.
169
Figure 4.7: Static and dynamic force – displacement reaction curves for chain link mesh.
170
Figure 4.8: Comparison of static and dynamic force and displacement properties for chain link mesh.
171
Figure 4.9: Crack pattern after second test on sample 1.
173
Figure 4.10: Crack pattern after test sample 2
174
Figure 4.11: Punch through failure for sample 3.
174
Figure 4.12: Punch through failure for sample 4.
175
Figure 4.13: Cracking and punch through failure for sample 4.
175
Figure 5.1: WASM Prototype dynamic loading of ground support schemes.
183
xvi
LIST OF TABLES Table 3.1: Equivalent Rock Radial Stiffness for Steel Pipes
81
Table 3.2: Summary of static pull test results on Omega bolts.
99
Table 3.3. Fully encapsulated HS threadbar summary results.
107
Table 3.4: Summary of results for fully encapsulated LS threadbar.
109
Table 3.5: Summary of results for resin encapsulated high strength steel threadbar.
116
Table 3.6:Selected properties for Australian 7-wire steel strand.
118
Table 3.7: Selected properties for Chilean 7-wire steel strand.
119
Table 3.8: Summary of Australian Plain Strand Test Results
120
Table 3.9: Summary of Chilean plain strand test results.
124
Table 3.10: Summary of performance indices strand cable.
130
Table 3.11: Decoupled threadbar summary results.
134
Table 3.12: Summary of resin encapsulated toe anchored bolts.
137
Table 3.13: Fully encapsulated 22mm cone bolt summary results.
138
Table 3.14: Bulb strand centrally decoupled summary result
146
Table 3.15: Summary of dynamic testing on Garford Dynamic Cable Bolt responses.
152
Table 3.16: Summary of dynamic testing on Garford Dynamic Solid Bolts.
157
Table 3.17: Summary of yielding reinforcement system responses to dynamic loading.
162
Table 4.1: Shotcrete layer specifications.
172
Table 4.2: Shotcrete dynamic test specifications and summary of performance.
172
xvii
xviii
1
BACKGROUND INFORMATION
This report assumes that the reader is familiar with the objectives of the overall dynamic testing of ground support project and has read the MERIWA Report No. 249 and Addendum that were prepared at the completion of M349. This report provides: • A restatement of the primary objectives and the scope of work and tasks that evolved during Phase 2 of the project. • A background to ground support technology, schemes and their mechanisms of response to rock mass and laboratory loadings. • A review of the status of the WASM Dynamic Test Facility at the completion of Phase 1. • An account of the design modifications to the WASM Dynamic Test Facility and their implementation. • Documentation of the modifications to the instrumentation and monitoring system. • The development of software to analyse the results obtained from dynamic tests of surface support systems. • Detailed results obtained from the dynamic testing of reinforcement systems. • Comparisons and evaluations of the responses of different reinforcement systems. • Details of a program of static testing of surface support systems. • Details of dynamic testing of surface support systems. • Comparisons and evaluations of the responses of different support systems. • A summary of the current status of the WASM Dynamic Test Facility and suggestions for further developments and testing to benefit the Australian mining industry.
1.1
PROJECT OBJECTIVES, SCOPE OF WORK AND TASKS
Three main objectives were identified in the M349 research project proposal, namely: • To establish a permanent dynamic testing facility in Kalgoorlie, WA. • To establish databases of measured dynamic responses for different types of reinforcement and support systems.
1
• To establish guidelines for expected energy absorption of various types of reinforcement and support systems. The ultimate aim was to establish criteria for selection of ground support schemes based on rock mass characteristics and expected energies associated with seismic loadings. It was identified prior to the commencement of the project that these overall objectives would need to be divided into at least two phases. The Phase 1 objectives became: • To design, build and commission the test facility and instrumentation • To test rock reinforcement systems comprising various elements, internal fixtures, external fixtures and face restraint. The Phase 2 objectives became: • To undertake any modifications required to the test equipment and instrumentation • To perform tests on surface support systems as currently used, or could potentially be used, by the Western Australian mines. Subsequent to the commencement of Phase 2, several new reinforcement systems became available commercially. These systems were tested and evaluated at the request of sponsoring companies.
1.2
GROUND SUPPORT SCHEMES AND TERMINOLOGY
In order to better understand the rationale for the development of the WASM Dynamic Test Facility and the test specimen configurations and loadings, it is important to have a limited, consistent terminology and some basic principles of the action of ground support schemes. In general, a ground support scheme consists of rock reinforcement systems and surface support systems. A rock reinforcement system consists of a single element fixed within a bore hole, drilled into the rock mass, and an exterior face plate with external fixture. A pattern of reinforcement systems is often used to support large blocks and prevent large scale ground deterioration. The type of element, and the spacing of each reinforcement system, depends on the prevailing geological conditions; namely, the geometry of the potentially unstable blocks and the loading conditions. Surface support systems are used between reinforcement systems to prevent smaller scale instability and unravelling of the rock mass. The surface support systems are restrained by the face plates used with the rock reinforcement systems to form an integrated ground support scheme. Surface support systems include simple rolls or sheets of steel wire mesh, or sprayed layers such as shotcrete and membranes that harden and apply a reactive force to the rock face.
2
1.2.1
LOAD TRANSFER CONCEPT FOR GROUND SUPPORT SCHEMES
Examples of ground support schemes comprising reinforcement systems and/or surface support systems are shown schematically in Figure 1.1 and Figure 1.2. The load transfer concept for reinforcement and support systems involves considering the response of these systems to rock movement.
Unstable Block
Mesh, Strap or Sprayed Layer or Coating
Restraint
Restraint
Figure 1.1: Load transfer from surface support to surrounding reinforcement systems.
Unstable Block
Adhesion Required
Adhesion Required
Sprayed Layer or Coating
Figure 1.2: Load transfer between surface support and the surrounding rock surface.
3
aa
In the case of support, it can often be assumed that the support is locally slab like and transfers force to points of restraint such as rock bolts or cable bolts (Figure 1.1) or zones of restraint provided by adhesion between the support system and the rock surface (Figure 1.2).
In other cases such as those shown in Figure 1.3 where the excavation profile is concave shaped, the
surface support does not necessarily have to be supplemented by reinforcement. However, in case b. the
performance of the arch is enhanced by having horizontal restraint provided by the floor of the drive or reinforcement low in the walls as shown in Figure 1.4.
a a a. Vertical Loading
b. Vertical and horizontal loading
Figure 1.3: Arch shaped surface support that does not require reinforcement.
Road base and reinforcement used to improve shotcrete arch performance
Figure 1.4: Arch performance improved by road base and reinforcement.
1.2.2
REINFORCEMENT SYSTEM LOAD TRANSFER
In the case of a reinforcement system, it is assumed that the reinforcement transfers force across a distinct interface or zone between unstable and stable rock as shown in Figure 1.5.
4
Excavation
Unstable Surface Region
Stable Interior Region
Figure 1.5: Reinforcement load transfer from unstable rock to stable rock
A reinforcement system can be considered to consist of 4 components as shown in Figure 1.6; namely: 0.
The rock.
1.
The element.
2.
The internal fixture.
3.
The external fixture.
Figure 1.6: The components of a reinforcement system. The response of the reinforcement system to rock loading involves several modes of load transfer between the various components as shown schematically in Figure 1.7.
5
Rock i Fi Internal Fixture i Ti P i+1
Pi Element i
Figure 1.7: Schematic showing the forces involved in load transfer for reinforcement systems. The modes of load transfer between the element and rock lead to a simple classification system described by Windsor and Thompson (1996). This classification system resulted in only three basic classes of reinforcement systems; namely: 1.
Continuously Mechanically Coupled (CMC) Systems.
2.
Continuously Frictionally Coupled (CFC) Systems.
3.
Discretely Mechanically or Frictionally Coupled (DMFC) Systems.
It can be easily demonstrated that all commercial reinforcement systems can be considered to fit within one of these three classes, even those systems that have yet to be developed. The load transfer and distribution of force for reinforcement systems within each of the three classes differ greatly in their responses and abilities to sustain dynamic loading; Figure 1.8 shows conceptually the expected force distributions within each class. These conceptual force distributions can be used as the basis for analysis and to identify where to instrument reinforcement systems in both the field and in the laboratory (Thompson and Windsor, 1993).
6
Stable Region
Unstable Region
Movement Vector
CMC
CFC
DMFC
Figure 1.8: Schematic showing the different element force distributions within each of the three classes of reinforcement systems. From previous experience with testing and theoretical considerations, it is known that reinforcement systems within the CMC class are ‘stiff’, those within the CFC class are low strength and may displace excessively under moderate loads and those within the DMFC class are less stiff than those in the CMC class because the element is able to extend over the decoupled element length. It is also within the DMFC class that most newer reinforcement systems have been proposed for use where high energy absorption is required. All these systems rely on displacement of the element relative to the internal fixture within the toe embedment length. Within the body of the report, the results from testing will be presented within one of the three classes. 1.2.3
SUPPORT SYSTEM LOAD TRANSFER
Support systems and their modes of action may be very different. The obvious difference is the areal coverage and accordingly support systems can be classified as: • Point Support (i.e. plates).
7
• Strip Support (i.e. flat, profiled or mesh straps). • Areal Support (i.e. mesh sheets and rolls, shotcrete and thin sprayed liners or TSLs as they have become known). There are a number of measures that may be used to assess differences between the various support systems, particularly within each of the three classes. Some measures (other than areal coverage) which are important in terms of maintaining the integrity of the rock are: • Increasing the strength of discontinuities at the boundaries of unstable rock. • Reaction in terms of both immediacy and stiffness to transverse loading. • Membrane action in tension in terms of both strength and stiffness. • Membrane action in compression in terms of both strength and stiffness. • Toughness in response to transverse loading and in plane distortion (shear). • Time dependent creep and relaxation associated with the different materials. Membrane action results from loading within the plane (extension and compression) while transverse loading causes shear and bending. Areal support systems which are restrained by reinforcement will involve both transverse loading and membrane action. The mechanisms by which mesh and shotcrete transfer load are quite different as will be shown and discussed in the following two sections. 1.2.3.1
Mesh Load Transfer
Thompson (2003) identified the following variables as being able to influence the performance of mesh when subjected to lateral loading: •
Variable wire diameters.
•
Variable wire spacings.
•
Mesh lay relative to wire loading.
•
Non-linear stress-strain properties for the wire.
•
Weld strength.
•
Variable bolt spacings.
•
Variable bolt tensions.
8
•
Slip of the mesh at the plates and bolts.
•
Variable mesh orientation relative to a restraint pattern.
•
Variable load types and areas.
•
Large mesh displacements.
These requirements were identified during systematic test programs (Thompson et al., 1999). Most of these requirements are self-explanatory; however, some are not. For example, the ‘mesh lay’ refers to the location of the cross-wires relative to the longitudinal wires. The relative location of the wires will influence whether the forces at a particular intersection in the mesh will produce tension or compression combined with shear in the weld. Variable ‘load types’ and areas’ refers to whether the mesh is displaced at a number of discrete points within the mesh to simulate a large, single block or by distributed loading of an area of the mesh to simulate a number of small blocks formed in closely-spaced, jointed rock. Lastly, large mesh displacements require that the changes in geometry must be able to be taken into account. It is impossible to conduct tests to examine all the possible variables listed above. Therefore, computer software was developed to analyse for the distribution of forces in welded mesh of any size restrained by a specified arrangement of rock bolts and subjected to an arbitrary shaped defined loading. The mesh is assumed to be comprised of cross wires and long wires connected by welds at the intersection points. The plates and rock bolts are simulated by fixed or deformable restraint at specified wire intersection points. The loading may be specified as forces and/or displacements imposed at the intersection points between wires. Two simple examples are used to demonstrate how load is transferred through the mesh to the restraint locations: •
Square pattern of restraint with central rigid loading.
•
Oblique pattern of restraint with central loading.
The deformed mesh for each case is shown in Figure 1.9 and Figure 1.10, respectively. The wire segments with high tensile forces have the darker shades of grey. These figures show clearly the path of force transmission from the simulated rock loading to the restraint. The computer simulations, whilst able to examine all the variables, need to be complemented by actual tests on mesh samples with known boundary constraints and loadings. However, the simulations do provide guidance on what to expect and this has assisted in designing the equipment and test configurations for both static and dynamic testing.
9
Figure 1.9: Deformed mesh with higher forces shown in darker colours.
Figure 1.10: Deformed mesh with higher forces shown in darker colours.
10
1.2.3.2
Shotcrete Load Transfer
In general, the reaction of a shotcrete layer in response to loading (w) is provided by a complex combination of in-layer forces (P), bending (M) and shear (V) as shown schematically in Figure 1.11. Note that P may be compressive as shown or tensile. In the case of very thin layers (i.e. TSLs), very little resistance might be expected to develop for concave surfaces without excessive displacement. It is also possible that the element of shotcrete may be flat or convex. The latter two shapes are less efficient in providing support action, particularly when the layer is brittle and may crack due to flexure. P
M
V w
V+
dV ds ds
M+
P+
dM ds ds
dP ds ds
Figure 1.11: Forces acting on a segment of a curved arch section. The basic support theory proposed by Deere et al. (1969) has gained the most recognition over the past three decades, but the complexity of the interaction between the shotcrete and the rock mass, and the difficulty in measuring this reaction, means further development of the support mechanism theory has not occurred. Studies by Holmgren (1976, 2001) and Fernandez-Delgado et al. (1976) showed that adhesion loss and flexure are the primary modes of shotcrete failure. A further review conducted by Barrett and McCreath (1995) identified that shotcrete capacity in blocky ground, under static conditions, is governed by six mechanisms: namely, adhesion loss, direct shear, flexural failure, punching shear, compressive and tensile failure (Figure 1.12). Adhesion loss occurs where the bond between the shotcrete and the rock is broken, often due to poor surface preparation prior to spraying or due to shrinkage of the shotcrete during curing. It may also be that certain rock types contain minerals that cannot sustain adhesion forces. Flexural failure is bending failure of the shotcrete and can only occur after the adhesion is broken. For flexural failure to occur, the shear strength of the material must be higher than the flexural strength. The writers disagree with the definitions provided for ‘direct shear’ and ‘punching shear’. Direct shear (or shear failure) occurs over a single planar interface, typically represented as a line. Punching shear, shown as direct shear in Figure 1.12, is direct shear that occurs over a complex three-dimensional surface. The
11
punching shear failure as shown in Figure 1.12 is a combined mechanism of flexural failure and shear failure. The updated failure mechanisms are shown in Figure 1.13. These failure mechanisms are generally not well understood and further research is required to understand the complexities of the rock / shotcrete interaction, particularly under the influence of dynamic loading.
Adhesion Loss
Flexural Failure
Punching Shear Failure
Direct Shear Failure
Compressive Failure
Tensile Failure
Figure 1.12: Failure mechanism of shotcrete (Barrett and McCreath 1995)
12
Adhesion Loss
Flexural Failure
Punching Shear Failure
Direct Shear Failure
Compressive Failure
Tensile Failure
Figure 1.13: Updated shotcrete failure mechanisms. In contemplating how to analyse the data for shotcrete panel testing in the WASM Dynamic Test Facility it became apparent that a simple model would be required to take into account the inertia of the shotcrete layer, similarly to the incorporation of the inertia of the steel pipes associated into the simulation of dynamic loading of reinforcement systems. For dynamic mesh testing, it was assumed that the mass of the mesh sample was insignificant and could be ignored. Yield line theory was chosen to describe the mechanisms of failure for flat slabs of shotcrete. Yield line theory is based on the premise that a layer that is loaded to failure will develop ‘yield lines’ in the most highly stressed areas. Yield lines are continuous ‘plastic hinges’ that are associated with the failure mechanism. It is found that, for the purposes of design, the patterns conform with a number of simple rules. These rules applied to a shotcrete layer are: •
Yield lines must end at a layer boundary.
•
Yield lines are straight.
13
•
Axes of rotation generally lie along lines of continuous support (e.g. a square slab supported on four sides as in the EFNARC (1996) test) and pass along side point support (e.g. round panel test described by ASTM (2002) and Bernard (2003)).
•
Yield lines between adjacent rigid regions must pass through the point of intersection of the axes of rotation of those regions.
Once the yield line pattern is defined, then relationships can be developed between the applied forces and the internal resisting moments. In yield line analysis, it is usual to define resisting moment (m) in units of force times distance per unit length (e.g. Nm/m). The shotcrete moment-rotation relationship depends on a number of factors such as: •
the thickness of the layer.
•
the type and distribution of any internal reinforcement.
•
the mechanical properties of the materials.
The simple rules can be used to define possible yield line patterns for complex loadings and support geometries. For example, yield line patterns for shotcrete panel testing are given in Figure 1.14. The latter case has been analysed in detail (Tran et al, 2001, 2005). Unpublished investigations at WASM have shown that theoretical relationships based on yield line theory correlate well with the results from static testing of shotcrete using different sample and loading configurations involving flexural response.
Figure 1.14: Yield line pattern for (a) centrally loaded square slab supported at the edges and (b) centrally loaded circular slab supported at the three equally spaced points.
1.3
ROCK MASS LOADINGS
In order to properly design testing facilities and associated procedures it is first necessary to understand the in situ interaction between ground support schemes and the rock mass. An important aspect of the ground
14
support scheme response is the amount of rock mass deformation and the rate at which it occurs. Two mechanisms of dynamic loading are proposed. Firstly, seismic waves may originate from an instability (an intact rock failure or slip on a structure) and travel through the rock mass. The instability (seismic source) is assumed to be located at a distance from the excavation greater than the fractured zone surrounding the excavation. The basic seismic source parameters of Seismic Moment and Radiated Energy are calculated as functions of the P (compression) and S (shear) waves and rock mass properties. The two wave components have differences in amplitude, frequency, velocity and travel direction. As the waves approach the excavation, the stable rock (the rock that will remain behind after the event), the unstable rock (the rock that will be ejected by the event), and the ground support scheme will all be ‘excited’ by the energy in the seismic wave. This process is represented schematically in Figure 1.15.
Ground Support Scheme Seismic Source
Seismic Wave Amplitude Frequency Velocity
Rock and reinforcement to be ejected Rock mass
Figure 1.15: Schematic representation of the behaviour of rock when subjected to remotely generated seismic loading. It is also possible that a volume of intact rock may fail violently, immediately adjacent to the surface support as shown schematically in Figure 1.16 and Figure 1.17. That is, failure may occur between reinforcement systems or it may be in the volume of rock surrounding a borehole in which reinforcement has been installed. In the former case, the dynamic load is transferred through the surface support into the reinforcement through the external fixture. In the latter case, the support and reinforcement respond
15
simultaneously and combine to attempt to sustain the dynamic loading. In the former case, it is possible for the surface support to fail and not transfer load to the reinforcement system. This is often observed where rock simply pushes through mesh, particularly at the locations of overlaps at edges of sheets. In the latter case, the proportion of load and energy absorbed by the reinforcement and surface support will be complex and related to the relative stiffnesses of their responses. In theory, it would be possible for the reinforcement system to fail, either internally or at the collar, and the surface support to transfer load to adjacent reinforcement systems in the ground support scheme. The WASM Dynamic Test Facility was developed with both mechanisms in mind; that is, a panel of surface support system restrained by reinforcement in the corners or a panel of surface support system restrained at the edges with a single reinforcement system in the centre of the panels as shown in Figure 1.18 for the prototype test facility.
Mesh or Shotcrete
Failed Volume of Rock
Figure 1.16: Schematic representation of rock failure loading surface support between reinforcement
Failed Volume of Rock
Mesh or Shotcrete
Figure 1.17: Schematic representation of rock failure loading both surface support and reinforcement.
16
Release point
Box
Guides and legs
Figure 1.18: WASM Prototype dynamic loading of ground support scheme.
1.4
WASM DYNAMIC TEST FACILITY AT THE COMPLETION OF M349
There are two requirements for laboratory simulations. Firstly the reinforcement system needs to be representative of the field condition. To simulate this, a reinforcement system test must simulate the response of both the collar and toe regions as implied previously in Figure 1.5 and Figure 1.6. The double embedment test developed in Australia in the early 1980s demonstrated the applicability of the configuration for static testing in the laboratory. It is also necessary that the test configuration allows for tension to be applied to the surface restraint. This generates the correct load transfer to the reinforcement system and allows for failure to occur at whichever is the weakest location: that is, at the surface hardware connection to the reinforcement element, in the reinforcement element at the simulated discontinuity or slip from the toe embedment length. Secondly, in both cases of initiated dynamic loading, a susceptible excavation may involve a detachment process in which a single block or fragments of rock may attempt to eject from the surrounding rock mass into the excavation and load the ground support. This ground support loading is unlikely to be instantaneous, but rather will take a finite time that will, for a remote event, be related to the seismic wave velocity, amplitude, the acceleration pulse of the dominant frequency and fracture velocity within the rock mass. For a local event, the loading will be a ‘pulse’ of loading that will result in a mass (M) of rock
17
moving at a particular velocity (v); that is, the failure volume will have a change in momentum that is related to a force (Ft) acting over a short time (t) as suggested by the following impulse equation:
∆Mv = ∫ Ft dt
Eq. 1.1
With these models of ground support loading, it is clear that prior to the seismic event, the rock mass and ground support are stationary. After the event, the rock mass is accelerated to a certain velocity in a short, but finite time. Hence, in a test facility, it is not appropriate to apply the load instantaneously, but rather it should be applied quickly and the rate of application should be measurable and similar to that which is considered to cause damage to underground mine ground support. It is also not appropriate to apply the load simply to the external fixture at the collar of the reinforcement system. At the WASM Dynamic Test Facility loading is provided by the relative displacement between the loading mass and drop beam. After impact, the beam is decelerated and the momentum of the loading mass causes a relative velocity to develop between the mass and the beam, causing a force to develop in the reinforcement element. Under the influence of this force, the reinforcement system will yield, slide or break in the process of dissipating the dynamic loading. It will be shown in later sections that all these mechanisms of behaviour have been observed and quantified and will depend on the design and material properties of the reinforcement system, the encapsulation material and the interaction with the borehole. The WASM Dynamic Test Facility is shown set up for a reinforcement system dynamic test in Figure 1.19. At the completion of Phase 1 (MERIWA Project M349), the status of the WASM Dynamic Test Facility could be summarised as follows: •
The infrastructure for the facility was successfully constructed.
•
The instrumentation, monitoring and data storage system were implemented.
•
Software was developed to analyse the large amounts of data.
•
The concept of momentum transfer was proven to be capable of providing sufficient energy in a single test to rupture reinforcement systems with force capacities in excess of about 300kN and energy absorption capacities in excess of approximately 40kJ.
In this final report for MERIWA Project M349A, the enhancements made to the facility will be described together with the current status of the facility with respect to dynamic testing of ground support systems, in particular the testing of mesh and shotcrete support systems.
18
Figure 1.19: WASM Dynamic Test Facility.
19
20
2
WASM DYNAMIC TEST FACILITY – CURRENT STATUS
Section 1 of the report provided an overview of the WASM Dynamic Test Facility as it was at the end of Phase 1 of the project. Section 2 now provides detailed descriptions of the significant modifications and enhancements made during Phase 2 and the current attributes and status of the test facility.
2.1
INFRASTRUCTURE
A number of changes were made to the WASM Dynamic Test Facility infrastructure to improve efficiency, general working conditions, materials handling and safety. 2.1.1
BUILDING
The drainage, exterior to the building, was improved after a heavy rainfall event caused water to flow into the pit of the test facility. The walls of the test facility were sealed to reduce the ingress of dust and wind turbines were installed to exhaust hot air and to reduce temperatures within the building. Pest exterminators were contracted to rid the shed of a spider infestation. 2.1.2
STORAGE
Storage racks were designed, constructed and covered at one of the end of the test facility. The racks were used for storage of reinforcement system specimens before and after testing as shown in Figure 2.1.
21
Figure 2.1: Storage racks for reinforcement system samples before and after testing. 2.1.3
GANTRY CRANE
The initial design used a mezzanine floor to enable the drop beam to be manoeuvred on to the guide rails. This process involved firstly lifting the drop beam up through a gap to the right hand side of the drop pit as indicated in Figure 2.2. This process proved to be inefficient. Also, the original fixed crane beam was limited in operational reach and, due to changes in the design of the test configurations, did not allow a fully assembled support system test to be lifted sufficiently high to pass over the guide rails. The fixed beam crane was removed, the mezzanine floor changed (Figure 2.3) and a moving gantry crane was installed (Figure 2.4). The guide rail I-beams were also painted to stop the surface oxidisation as part of the upgrade as shown in Figure 2.5. Figure 2.6 shows the test facility after the completion of the changes that were funded entirely by WASM.
22
Mezzanine Floor Gap
Test Pit
Figure 2.2: Mezzanine with original static crane beam and unpainted guide rails. Roof removed to extract static crane beam, internal shed modifications then lift in gantry crane support posts, runners, and then modified crane beam.
Mezzanine removed from here
Mezzanine filled here
A new mezzanine support post required at either side
Figure 2.3: Installation of gantry crane and mezzanine change, October 2005.
23
Gantry crane allows travel along the high section of the shed
Figure 2.4: Gantry crane view from the mezzanine floor
Figure 2.5: Painted guide I-beams and mezzanine change.
24
Figure 2.6: WASM Dynamic test facility following upgrade (February 2006).
2.2 2.2.1
HARDWARE RELEASE MECHANISM
A 4,536kg capacity helicopter release hook (model 10K-001 from Canam Aerospace) is used as the release mechanism to initiate a test. It is designed to undertake this task with high load capacity, accepted standard of safety and ease of working. The hook is positively locked until power is supplied to the release solenoid. The hook has been connected to an isolation circuit that requires arming and initiation of release from a location remote from the area of testing. The hook is matched with a shock absorber (SA-P-10) to reduce dynamic loading on the crane beam and building after release of the drop beam.
25
Centre of mass
Rotation point
Replacement solid connection from hook to beam.
Figure 2.7: Release Hook. 2.2.2
DROP BEAM
The drop beam is connected to the helicopter release hook through an eye frame as shown in Figure 2.10. This frame replaced the chain used initially as this created unwanted steel to steel impact that influenced the response of the sensitive accelerometers used in the monitoring of tests.
Replacement solid connection from hook to beam.
Figure 2.8:Lifting hook on top of the drop beam.
26
Dynamic Engineering, a consultancy firm specializing in dynamic load calculations for engineering structures, was commissioned to specify the original drop beam size, reinforcing flanges and webs. The design criterion was 1mm centre deflection at the maximum load expected during reinforcement testing. The design load was specified as approximately 500kN (corresponding to the nominal force capacity of two 7-wire steel strands). The original design of the facility anticipated that two separate beams dropping onto four buffers would be required for mesh and shotcrete panel testing. However, it was found to be more efficient to bolt frames (to be described in detail in the following section) to a single beam. It was noted prior to upgrade that the grey beam had Luders lines shown in Figure 2.9.through the paint on the lower flange at points corresponding with where the new frames would be attached. Luders lines are indicators of plastic deformation caused by local stresses greater than the yield stress of the steel. The steel in between each of the Luders lines remains elastic. This in itself was not cause for alarm as steel is known to work harden and still behave elastically up to a new, higher yield stress. As expected, the lines were only present on all four lower flanges with none observed on the upper flanges. However, the plastic yield did indicate that significant strengthening and stiffening would be required in order to sustain the loads from the proposed support testing frames. The beam modifications were designed in-house and were largely aimed at stiffening the beams in both bending and torsional resistance locally where the frames were to be connected to the beams as shown in Figure 2.10. These modifications have been implemented successfully with no Luders lines being observed during subsequent support testing.
Web
Additional stiffening webs
Luders lines
Flange
Reinforcing Flange
Figure 2.9: Luders lines on grey drop beam.
27
Additional web stiffening components
Figure 2.10. Stiffened beam for attachment of the frames required for mesh and shotcrete testing. 2.2.3 2.2.3.1
DROP FRAMES Mesh
For mesh panel testing, a frame to support the mesh is bolted to the drop beam. The mesh is held in place using threaded bar, shackles and eye bolts in the same configuration (Player et al ,2008) as the standard static test arrangement as described in the Appendix. The basic principles for the dynamic testing of reinforcement systems also apply to dynamic testing of support systems. The drop beam is stopped by the buffers and the momentum of the loading mass is restrained by the test element and transferred through the frame to the drop beam as shown in Figure 2.11.
28
Mass chocked in place
Loading mass Support frame boundary.
D-shackles.
Figure 2.11: Dynamic mesh test configuration. It was necessary to marginally increase the hole diameter of the connection bolt holes in each of the drop beams to accommodate the slight differences in dimensions that resulted after fabrication of the fibrecrete and mesh frames.
29
Bolt connection holes to be machined out.
Figure 2.12: Increase frame connection bolt holes. Figure 2.12 shows the process used to increase the hole diameters from 22mm to 32mm. The beam was rolled over onto its side so that a magnetic drill with rotabroach and cooling could be used to ream out the existing bolt holes. 2.2.3.2
Shotcrete
A support frame and sample configuration were developed to enable testing of panels of fibrecrete. These panels are of similar configurations to those used for static testing. Essentially the samples consist of a rock substrate with a 500mm diameter disk cut from the centre over which a shotcrete layer is sprayed. The test geometry shown in Figure 2.13 results in a form of ‘punch test’ that can cause bending and shear loads in the shotcrete layer and separation forces at the interface between the shotcrete and rock. The sandstone and fibrecrete slab is clamped in place with a lower frame to provide continuous edge support similar to that used in the static test frame. In addition, the layer can be restrained by bolts near the edges in one of three configurations; four bolts located near the corners (square pattern), four located at the centres of the edge spans (oblique pattern) and at all eight locations.
30
Drop beam
Bolts from support frame
Loading mass
500mm diameter punch
Sandstone and fibrecrete slab
Fibrecrete support frame
Fibrecrete clamp frame
Figure 2.13: Dynamic fibrecrete test equipment. 2.2.4
BUFFERS
As indicated in the previous section, the original design of the facility anticipated that two beams dropping onto four buffers would be required for mesh and shotcrete panel testing. It was found that two buffers were capable of withstanding the impact with the additional mass of the frame and reduced simulated loadings used in mesh testing. However, during the initial shotcrete test commissioning, the mass of the frame and shotcrete panel (approximately 1100kg) in addition to the drop beam and loading mess resulted in impact forces close to or in excess of the capacity of two buffers. In fact, it is strongly suspected that the buffer pistons may have reached their maximum displacement. In order to retain the concept of a single beam, it was deemed necessary to use four buffers to share the impact loads. It was decided to minimise disturbance to the original configuration of buffers by using the existing four securing bolts and adding four new bolts as shown in Figure 2.14. A short, stiff reinforced Ibeam was then placed on to the buffer pistons as shown in Figure 2.15 to distribute the impact from the beam during a test.
31
Figure 2.14: Drilling additional buffer connection bolts.
Connection beam to link buffers during impact.
New tie down bolt locations.
Box section for grout
Figure 2.15: Connection beam for buffers.
32
2.2.5
SIMULATED ROCK LOADING
Several different configurations of loading mass have been developed to simulate rock loading for reinforcement systems and panels of mesh and shotcrete. These are detailed in the following three sections. 2.2.5.1
Reinforcement
It is important that the interaction between the rock mass and a borehole is correctly simulated. In particular, it is very important that the loading mass and the simulated borehole collar pipe are integrated into a single unit as are the toe end pipe and the beam. Figure 2.16 shows schematically the arrangement used to couple the loading mass to the collar pipe and to couple the anchor pipe to the beam. Load Cells between Pipe Flange and Beam
Toe embedment / Stable zone
Drop Beam
Buffers
The simulated borehole discontinuity
Collar embedment / Unstable zone
Circular steel plates – loading mass Split Plate to couple the mass of steel rings to the load transfer ring welded onto the collar pipe Load transfer ring welded onto collar pipe. Solid base plate for surface hardware and collar load cell
Connection bolts through steel mass
Not to scale diagram of the drop beam, reinforcing element, and the rock mass
Figure 2.16: Schematic of load transfer rings and integration with the steel pipe. A thick wall steel pipe is used to simulate the rock surrounding a borehole. Two types of boreholes are simulated; these are designated standard and rough. Standard boreholes are primarily used for reinforcement elements that are encapsulated with resin or cement grout. In this case, the elements are grouted directly into thick wall steel pipes. On the other hand, rough boreholes are used mainly for reinforcement systems that rely on friction for load transfer directly between the element and the rock (i.e. friction rock stabilisers). A rough borehole is a new concept that was developed in the M349A
33
investigations and will be described in more detail in Section 3. Rough boreholes are also used for resin encapsulated bolts installed by a jumbo. The test reinforcement system is coupled to the steel pipe; either mechanically through cement or resin grout in the annulus between the reinforcement element and pipe or by friction between the element and a simulated rough borehole. In both instances, shear pins are installed through the wall of the pipe to stop any potential for the grout to slide relative to the steel pipes. The loading mass is bolted to a steel ring welded onto the collar pipe. A base plate butts to the edge of the collar pipe. The reinforcing element protrudes through the base plate, and is tensioned with the appropriate surface hardware to complete the reinforcement system. Note that tensioning of the reinforcement causes a compressive force to develop at the interface between the collar and anchor pipes. This means that the in situ condition where a discontinuity would be compressed by reinforcement tensioning is simulated. Suitable toe and collar embedment lengths are selected to simulate the likely combination of lengths that might result in a mining environment. 2.2.5.2
Mesh
Two types of loading mass may be placed into the centre of the restrained mesh. The first mass, shown in Figure 2.17, consists of a pyramid shaped bag filled with a known mass of steel balls (0.5 or 1 tonne). The loading area of the bag is 650mm x 650mm. A wooden prop is placed between the loading mass and the drop beam to prevent the loading mass ‘floating’ during the initial free fall period. The drop beam and attached mesh frame assembly are dropped from a specific height to generate dynamic loading on the mesh sample.
Figure 2.17: Deformation with 990kg load pre test on the Geobrugg Tecco Mesh.
34
The second loading mass, shown in Figure 2.18, may be used to test the mesh directly or integrated with additional mass to provide a higher dynamic loading demand for a simulated ground control scheme of mesh plus bolts.
Figure 2.18: Curved loading mass for mesh panel testing. 2.2.5.3
Shotcrete
Dynamic testing may be performed on a continuous panel or as a punch test with a similar sample configuration to that developed for static testing. In either case, the loading mass consists of a steel cylinder as shown in Figure 2.19. The steel loading mass has been manufactured with a centrally located hole through which a reinforcement system may pass in anticipation of ‘ground support scheme’ testing in which both surface support and reinforcement are loaded simultaneously.
35
Bolts from support frame Loading mass
500mm diameter punch
Sandstone and fibrecrete slab
Fibrecrete clamp frame
Figure 2.19: Loading mass used for shotcrete panel testing.
2.3
FORCE TRANSFER AND DISPLACEMENTS IN TEST FACILITY
The selection of instruments and their locations are based directly on the force transfer and displacements in the various test configurations as presented and discussed in the following sections. In all cases, ‘freebody’ diagrams were used. The equilibrium of each component in the test configuration forms the basis for the ‘engineering analysis’ of a test and results in the reinforcement or surface support system performance being expressed in terms of a force-displacement response and energy absorption. 2.3.1
REINFORCEMENT
The locations and symbols of force transfer in a reinforcement system test are shown in Figure 2.20. The equilibrium analysis was presented both by Thompson et al. (2004) and in the final report for M349.
36
mA ½PA
½TEA
½PA
½TEA
mP
FRA ½mAg
½TA
PS
PS
m Bg
m Bg
½PJ
PB
½PJ
loading mass (including collar pipe).
mB -
mass of buffer piston.
FRA -
element force at a discrete internal fixture.
FRJ -
element force at the interface between anchor and collar zones.
FRC -
element force at the collar fixture.
TEA -
load transfer between element and wall of pipe (or borehole) at a discrete anchor.
TA
-
load transfer between element and wall of pipe (or borehole) in anchor zone.
TC
-
load transfer between element and wall of pipe (or borehole) in collar zone.
TEC -
load transfer between element and fixture.
PJ
force transfer at the interface between the anchor and collar zones.
PB
FRJ ½PJ
½PJ
½mCg
COLLAR ZONE
½mCg ½TC
½TC
FRC
½PCP
½PCP
Collar plate ½PEP
-
½PEP
PCP -
force transfer between the collar zone and plate.
PEP -
force transfer between the fixture and the plate at the collar.
Surface Fixture ½TEC
½TEC
mass of anchor pipe.
mC ½mAg
½TA
ANCHOR ZONE
-
mass of beam.
Figure 2.20: Free-body diagram showing load transfer mechanisms within a reinforcement system.
37
2.3.2
MESH
A free body diagram similar to that shown in Figure 2.20 for a reinforcement system was developed for the mesh test and is shown in Figure 2.21. The force-displacement response for the mesh is derived from the deceleration of the loading mass and its displacement relative to the frame.
½MABg
½MABg
½PAB
½PFB
½PFB
½PAB
½MBg
½MBg
PB
PB ½MFg
½MFg
MCg
½PFS
PCS
½PFS
Figure 2.21: Free-body diagram showing load transfer mechanisms within a mesh panel test. 2.3.3
SHOTCRETE
The force and displacements for a shotcrete test are similar to those for a mesh test with the major difference being the self-weight of the shotcrete panel needing to be included in the analysis of loading and energy calculations. In this case, the free body diagram is as shown in Figure 2.22 and the assumed displacement mechanism is as shown in Figure 2.23. The force-displacement response for the mesh is derived from the deceleration of the loading mass and its displacement relative to the frame.
38
½MABg
½MABg
½PAB
½PFB
½PFB
½PAB
½MBg
½MBg
PB
PB ½MFg
½MFg
½MCg
½MCg
½PFS
½PCS
¼MSg
½PFS
½PCS
½MCSg
½MCSg
¼MSg
Figure 2.22: Free-body diagram showing load transfer mechanisms within a shotcrete panel test.
Loading Mass
Figure 2.23: Assumed displacement mechanism within a shotcrete panel test. 2.3.4
RELATIVE DISPLACEMENT, VELOCITY AND ACCELERATION
In all three types of tests, the impact velocity is not the velocity that the simulated loading mass will impose on the reinforcing or support system. During a test, relative displacements, velocities and accelerations develop when the drop beam is slowed by the buffers while the loading mass is slowed by the reinforcement or support system. The relative velocity between the loading mass and drop beam prior to impact is zero, and will be zero again once the buffers have reached maximum compression for a particular test and the ground support system has been able to sustain the dynamic load and dissipate the associated energy. If the ground support system fails, then the mass will continue moving with some velocity and acceleration due to gravity.
39
2.4
DATA ACQUISITION SYSTEM AND SENSORS
The objective of the instrumentation is to determine the parameters defining the motion at any time of all components involved in a test. That is, the objective is to be able to measure directly, or derive, the acceleration, velocity and displacement of all components. Knowing the mass of all components, it is then a relatively simple process to calculate the forces, momentum and energy of all the components at any time in a test. The objective is achieved by ensuring that all components are monitored by at least one instrument during a test and that separate sources of data are synchronised in time. Instrumentation was designed or selected to record force, displacement, acceleration and strain in small time increments. The design and selection of the instrumentation and data acquisition system involved a number of key requirements: •
a large number of channels.
•
high speed acquisition of data per sensor channel.
•
relatively high speed digital video capture.
•
integration of digital video with sensor data.
•
support strain gauge, integrated circuit protocol (ICP), and direct voltage output instrumentation.
•
software control of sensor data acquisition and digital video capture.
To understand the required locations for the instrumentation and to design appropriate instrumentation it was firstly necessary to define the component interactions, load transfer mechanisms and forces for reinforcement systems (shown previously in Figure 2.20) and mesh and shotcrete panels (Figure 2.21 and Figure 2.22, respectively). 2.4.1
DATA ACQUISITION SYSTEM
The Data Acquisition System (DAS) and sensor accessories selection process was controlled by the need to be able to purchase an ‘off the shelf’ hardware and associated software system to perform the high speed data acquisition and digital video capture and storage. This market is relatively small and therefore took precedence over the selection of particular sensors, knowing that the DAS could be adapted to measure outputs from all types of sensors. A National Instruments PCI6071E data acquisition (DAQ) board controls the acquisition of data from all sensors. The DAQ card is configured for 32 differential input channels using 12 bit sampling. The facility utilises all of the available channels to support the upgraded sensors for testing of support systems and to allow a standardised configuration of instrumentation with minimal rewiring of strain gauge signal conditioning boards.
40
All channels are sampled simultaneously at a rate of 25,000 samples per second. The sample rate determines the smallest time interval recorded (40 microseconds), but the highest recordable frequency for cyclic analogue signals is half the sample rate according to the Nyquist Theorem. The underlying data acquisition software is from National Instruments, but the video control software is Midas 2.0 from Xcitex. Midas software provides a front-end operating system for recording the data. All data can be output to an Excel spreadsheet for displaying the sensor data along with point analysis of object locations via auto-tracking of the video record. A video file is also generated. Sensor data and video information are both time-coded and interlinked to allow for combined analysis. 2.4.1.1
Time window
The DAQ system has a maximum recording window of two seconds. The sensors and video are continuously monitored but a trigger is required to initiate data recording. In the WASM Dynamic Test Facility, the trigger is a voltage pulse that occurs when a laser beam is broken by the drop beam. Section 2.4.2.2 provides the details. The actual duration of data recording depends on the type of test performed (reinforcement or support system) and whether the test results in arrest of the loading mass or failure of the system being tested. 2.4.1.2
DAQ connection
The wiring for the DAQ instrumentation was customised by Xcitex to WASM requirements for the analysis of displacement from the high speed video. In the test configuration the software depends on the trigger voltage being detected on a Xcitex customised National Instruments BNC2110 block. When WASM decided to not use the customised BNC2110 but rather use a NI-SC2043SG (to allow the use of strain gauge based instrumentation) the wiring protocols became very important. It was during this process that it was found that the manuals for the SC2043SG board had multiple errors causing confusion over wiring protocols. The addition of the second SC2043SG board increased the complexity of the connections again. 2.4.1.3
Schematic of final data acquisition equipment
Dividing the strain gauges into two banks of eight channels provided for 16 strain gauge channels (signal conditioning by National Instruments SC2043SG board). One bank was connected directly into the DAQ board while the other bank was connected via the BNC2115 connector block. In addition there were 12 channels configured for accelerometers connected via a PCB Piezotronics 483A signal conditioner unit with BNC connections to the BNC2115 connector block. The remaining 12 channels were DC voltage channels with BNC connections onto the BNC2115 connector block and onto the DAQ board. Figure 2.24 shows a schematic of the DAQ boards, computer and instrumentation.
41
6 channels of load cells monitoring load and frame interaction for mesh testing.
10V power supply for strain gauges and load cells
SC2043SG (1)
Trigger, 5V and ground
Laser break trigger
Video Screen for data visualization
Second laser for velocity calculation SC2043SG (2)
5 strain gauges on the drop beam and 2 strain gauge based load cell for reinforcement systems
NI-DAQ card (PCI6071E– 32 channels) BNC2115 connection board Computer
Two potentiometers, one connected to each buffer side, to confirm impact time and displacements
PCI card – Motionscope 1000 5 Accelerometers (500g triaxial and 5000g shock uniaxial), on the support frame and/or drop beam and mass.
12channel PCB-483A signal conditioning for accelerometers
Redlake digital video – high speed camera for displacement tracking.
Figure 2.24: Schematic of instrumentation connections.
42
2.4.2
SENSORS
A combination of primary and secondary sensors for each measurement was used to acquire or calculate the data. Duplicate monitoring of critical components provided by the secondary sensors allowed an improved success rate in obtaining the necessary measurements. All sensors were selected to be able to provide an output voltage that would be detected by the data acquisition card. Commercial equipment was supplied with normal production calibration factors. Purpose built equipment was calibrated as part of the project. The quality of the measured data is a function of the combined accuracy and precision of both the sensor, the allocated voltage range for the channel, inbuilt channel or sensor filtering and the DAQ board. 2.4.2.1
Component Accelerations
All accelerometers were purchased from PCB Piezotronics. Three model 356A02 triaxial 500g accelerometers were selected to use on the drop beam and loading mass with the potential for determination of any out of balance component. The units have a frequency range of 1Hz to 5kHz. The addition of a mechanical filter (PCB Model 080M150) protects the accelerometers from high frequencies and sensor saturation due to metal to metal contact. The mechanical filter reduces the upper range to approximately 2kHz. Two model 350A14 uniaxial shock accelerometers with a 5,000g range and 0.4Hz to 7.5kHz frequency response were selected for mounting on top of the drop beam and loading mass. These units contain inbuilt electronic filtering and generate cleaner raw responses. Accelerometers are mounted in strategic locations on the test components; for example: •
5000g accelerometer on top of the drop beam above the buffer,
•
500g accelerometer on top of the beam at approximately 1/3 the distance from the end to the middle to stop saturation of the sensor,
•
500g and 5000g accelerometers on top of the loading mass. These allow for direct comparison of signals and assist in understanding any differences in sensor responses.
The use of multiple accelerometers on single components allows the filtered response signal to be averaged both with the same and different sensor type. It also allows for failure of one sensor without the loss of data required to complete the analysis of a test and having to resort to more complex and less reliable differentiation or integration of other data.
43
2.4.2.2
Laser Breaks
Laser breaks are used for two purposes; namely, triggering of data logging and impact velocity measurement. Triggering A laser beam is aligned a short distance above the unloaded position of the buffer piston. As mentioned previously, the drop beam moves through the laser beam (emitter to a receiver) to trigger the recording process. Breaking the laser beam changes the input voltage to the SC2043SG board from 0.2volts to approximately 5volts, for as long as the laser beam remains broken as shown in Figure 2.25. The rising trigger signal stops the over-writing of the two-second continuous memory buffer. The software is configured for a 30% pre-trigger to ensure sufficient data are obtained prior to and following the completion of a test. Velocity Measurement The release hook direction is orthogonal to the length of the drop beam. If the crane trolley is not directly above the centre of mass, it is expected to introduce a slight lateral displacement into the dropped test beam, reinforcement system and loading mass. The lateral displacement is corrected by the sliders mounted on the drop beam that fit onto the guide rails. A small rotation of the beam occurs and the friction between the rails and sliders result is a slightly lower velocity at impact than expected from a calculation based on the drop height. Hence, a second laser was located above the first in the fall path of the drop beam. Both laser outputs were routed through the BNC2115 to the DAQ board to provide a travel time for the beam over a known distance (+/-1mm) to calculate the velocity of the beam just before impact. The rise time for the voltage response from the laser break is faster than the 40 microsecond sample rate on the board.
44
drop beam, falling / sliding Laser emitter 2
Laser receiver 2
Laser emitter 1
Laser receiver 1 buffer
Power supply
Two laser beam breaks recorded. The time difference allows velocity calculation.
Figure 2.25: Laser breaks. 2.4.2.3
Beam/Buffer Displacements
Two types of instruments have been used for measurement of the displacements of the beam and the buffer piston; namely, non-contacting ultrasonic sensors and linear potentiometers with mechanical connection. Ultrasonic Measurement An ultrasonic motion sensor was first selected to measure compression of the buffers during a test. The selection of the ultrasonic device was based on the following criteria: •
no direct contact to beam, hence no damage to sensor.
•
ability to measure the maximum buffer compression of 114mm.
•
a DC volt output compatible with the DAQ board range.
45
A HydePark SM606A02 ultrasonic motion sensor was selected for this purpose. However, the selected unit had a comparatively slow sample rate of 1.5milliseconds (ms). The ultrasonic unit also had a limited resolution of 0.69mm. The digital sampling technique used by the DAQ card results in a step function record of the buffer compression that required filtering to result in a smoothed response prior to further analysis. The sensor was mounted in a bracket on the side of the buffer as shown in Figure 2.26. The accelerometer mounted on top of the beam above the buffer was used for determining impact with the buffer due to the potential error of 1.0mm to 1.5mm in beam displacement from using the ultrasonic device, as the beam
Guide rails
initially is moving into the far limit of its detection range. Accelerometer mounted to beam, records impact Drop beam contacts buffer, and comes into ultrasonic range
Piston moves into buffer, reaction force upwards
Ultrasonic probe records beam movement as the buffer is compressed and recovers, direct voltage output to DAQ.
Figure 2.26: Ultrasonic displacement sensor location relative to other test components. Potentiometers An alternative to the ultrasonic sensor was identified as a linear potentiometer (SLS130) from Penny and Giles. This model of potentiometer is specifically designed for measuring the displacement of racing car shock absorbers and was attached to mounting plates connected to the piston and base of the buffer, as shown in Figure 2.27. The compression of the buffers and displacement of the beam is measured by this potentiometer. Initially just one buffer was measured and compared with the ultrasonic result.
46
Potentiometer mounted on the side of buffer
Figure 2.27: Potentiometer mounted on buffer to measure piston displacement. The first potentiometer voltage output had a single pole resistor-capacitor (RC) filter installed late March 2006. The single pole resistor-capacitor filter was set to function as a low pass filter on the voltage to the DAQ thereby removing high frequency ripples and spikes caused by the rapid acceleration of the potentiometer. The ultrasonic device was removed and a second linear potentiometer was added to the other buffer at the beginning of the March 2006. This unit has not had the single pole RC filter applied to its voltage output because the response was cleaner than the first potentiometer. The use of two potentiometers allows the start times for the buffer compression to be rapidly and accurately detected. The differences in the response of the buffers to the load are determined. The analysis software allows the outputs of the buffer compression to be averaged. 2.4.3 2.4.3.1
STRAIN GAUGE BASED SENSING Load cells
A single load cell (Transducer Techniques LWO-80 load cell, Figure 2.28) is used to measure the collar force; this load cell has 356kN (80,000lbs) capacity and a manufacturer’s estimated 5% accuracy during dynamic loading. This was the preferred load cell due to its low profile (allowing greater application for bolts with short exposed collar lengths) and relatively low price when purchased direct from the supplier. The collar load cell is primarily used to measure performance of the surface hardware and determine the load transfer in the reinforcement system between the simulated discontinuity and the collar surface fixture. It is worth noting that the force-displacement responses are derived for the force in the element and the
47
displacement across the simulated discontinuity; also, the force in the element at the discontinuity is usually greater than that measured at the collar. Simulated discontinuity lower pipe length with bolt
Surface hardware and washer
Paint marker targets for video to track
LWO-80Load cell Nut
Figure 2.28: Collar load cell application. The anchor force is recorded by four purpose built 300kN capacity, hollow core load cells with spherical seats supplied by Fenixx Pty Ltd. The four load cells shown in Figure 2.29 are wired as a single full bridge. They are located between the top of the drop beam and the flange plate welded to the anchor pipe and held in position by bolts that are tensioned to result in a small initial compression force in the load cells. The anchor cell results are not directly used for the assessment of the energy dissipation by the reinforcement system, but are very useful for defining load duration, and understanding the force and momentum transfer from the loading mass to the drop beam via the reinforcement system. Note that the anchor load cells measure a force greater than the element force at the discontinuity due to the inertia of the anchor pipe.
48
Anchoring plate for thick wall pipe
Load cells (4 of)
Load cell pushed down into the beam, buffers push up against drop beam.
Anchor length of thick wall pipe
Drop beam
Junction box to wire load cells in full bridge
Simulated discontinuity
Figure 2.29: Anchor load cells.
Add equation / result and what I used in the lab
Figure 2.30: Anchor load cell calibration.
49
2.4.4
HIGH SPEED CAMERA
The displacement of the loading mass and the surface hardware is calculated from a digital video camera record taken during a test. The Red Lake (PCI-1000S 1108-0008) camera is fitted with a Pentax 1:1.8 and 4.8mm lens, and has a pixel resolution of approximately 3.3mm at the viewing distance. The camera was usually mounted at 15° above the horizontal (measured and recorded each set-up), which allowed viewing of the plate and surface hardware during each test. A geometric correction factor (i.e. parallax correction) was developed to account for the camera mounting angle to the horizontal, and was presented in the final report for Phase 1. The physical requirements for the high speed photographs are strong lighting from underneath the mass. This is achieved by using two by 2500Watt lights located in recessed foot-wells at the base of the pit and a 500Watt fill light from above to balance the exposure. Balanced exposure assists in the automated tracking by the Midas software. The capture rate is 250 frames per second. Higher sample rates of 500 and 1000 frames per second are possible; however, there is a significant loss of frame area and no testing has been undertaken at the higher rates. Time coding of the video allows synchronisation with the sensor data. The auto-tracking software for the camera requires targets marked on the loading mass and surface hardware. The software can resolve displacements to half of one pixel or about 1.7mm. The analysis primarily uses the points on the centre line of the drop. Pearce (2007) found that there was barrel distortion towards the edge of the lens image and an error in the optical axis of 4° up from the horizontal (see Figure 2.31 and Figure 2.32). The error in the optical axis was accounted for by including it in the customised analysis spreadsheet for displacement. The use of the correction reduced the error between the video track and physical post test displacement measurements. When velocities exceed the equivalent of ½ pixel / frame rate, comparatively large and unrealistic steps can occur in the predicted displacements of the loading mass. Therefore, in order to obtain a regular displacement versus time response, the data require smoothing across multiple targets from which velocity and acceleration may be derived.
50
The camera is setup level.
General view of the setup for the test. The optical axis of the camera is 3.6cm above the steel cupboard.
Figure 2.31: Camera angle test.
The optical axis is high by 13cm; this equates to 4° over the length of the test.
Figure 2.32: Output from camera during test.
51
2.4.5
PHYSICAL MEASUREMENTS
Physical measurements are taken before and after each test in order to confirm sensor measurements and provide an understanding of the deformation of the reinforcement system. Measurements are routinely made for: •
toe of bolt displacement, particularly for yielding bolts (Figure 2.33), or if multiple drops are going to occur without post sample dissection.
•
separation displacement at the discontinuity (Figure 2.34).
•
torque applied to the nut or estimated tension induced during barrel and wedge anchor installation.
•
post test surface plate deformation. Distance to bolt measured down the hole drilled in the grout.
Anchoring plate
bolt
Load cells
grout
Upper pipe length to the anchoring plate
Figure 2.33: Toe displacement measurements.
Upper separation for additional test, either yield mechanism or bolt stretch. 2
Centre separation from first test, represents the edge of the upper and lower pipe lengths on the bolt, measured with each test Lower pipe length including integrated mass and surface hardware
1
Bottom separation, from additional test, generally bolt stretch Two dynamic loads on mild steel threadbar
Figure 2.34: Displacement measurements at simulated discontinuity.
52
2.4.6
DATA RECORDING
The data that the instruments need to measure and the Data Acquisition System needs to monitor and retain must enable the following information to be derived: •
Velocity of the beam prior to impact (laser breaks).
•
Displacement of the head of the reinforcement and the mass before and after impact (high speed digital video).
•
Velocity of the loading mass before impact (high speed digital video).
•
Deceleration of the drop beam following impact (accelerometers on the beam).
•
Deceleration of the loading mass following impact (accelerometers on the mass).
•
Closure of the buffers following impact (linear potentiometers).
With this information, it is possible using the WASM developed software to quantify the performance of the reinforcement and support systems in terms of their force-displacement responses and energy absorption capabilities.
2.5
WASM DATA ANALYSIS SOFTWARE
A prime requirement for the WASM Dynamic Test facility was the development of associated filtering and processing of the data from the acceleration and displacement sensors to calculate the dynamic forcedisplacement responses for reinforcement and support systems. The WASM Data Analysis Software was developed in-house and reported in the final report for M349. A User Manual for the software has been written and is updated as new options and enhancements are added. In particular, the software has been expanded to enable the analysis of the responses of mesh and shotcrete panels. 2.5.1
MENU
The menu for all operations in the analysis of the dynamic test data is shown in Figure 2.35. The basic steps in the analysis are as given in the menu items. That is: 1.
Open Data File.
2.
Examine Raw Data (both text and charts).
3.
Filter Data (using the most appropriate filter) and save to processed data file.
4.
Set up processed data for analysis.
5.
Perform engineering calculations.
6.
Display the results as charts.
53
Figure 2.35: Menu for dynamic mesh and fibrecrete test data analysis. There are many additional steps within each of these broad steps and these are detailed within the following sections. This following sections will present and discuss how the sensor data and test details are acquired and analysed for incorporation into a Standard Test Report. 2.5.2
DATA VISUALISATION
The raw data for all sensors is read from a file with a standard header section and a variable number of sensors and a variable total number of data records. The software is able to automatically read this data file without user intervention regarding the number of channels of data and the duration of the test. An example of raw data is shown in Figure 2.36. The list of data channels is automatically created and the user may visualise single or multiple line charts as shown in Figure 2.37. Since there are extraneous data from before the start of the test and following the end of the test, the user may specify either by a mouse controlled window or text input the required range of time for data filtering.
54
Figure 2.36: Display of raw data file (.au2) details.
Figure 2.37: Several sets of data displayed on the one chart.
55
2.5.3
DATA FILTERING
Filtering of accelerometer and other data is necessary because the raw data can not be directly analysed. A number of filters have been implemented and are available for the analysis of any data channel. In general, the way the filters work is to first perform a frequency analysis from which it can be decided at which frequency to truncate the data. Then an inverse transform is performed to create the data that will be used in the engineering analysis of the test to produce the charts on which performance of the various ground support system are assessed. A number of options are available for creating the final inverse transform. These options include a phase shift, signal offset and signal sense change. The phase shift is required in order to be able to synchronise the starting times for all signals that will be used in subsequent analyses. The signal offset is required to account for the recorded data not being ‘zeroed’ prior to the test. The sense change is required to account for signals being positive instead of negative and vice versa. The Fast Fourier Transform is the most commonly used filter. Considerable experience has been gained in analysing signals from various types of instruments and interpreting the results with regard to causes and identifying methods to prevent these in future tests. For example, unwanted high signal frequencies associated with metal to metal noise were identified early in the testing programs. This noise was significantly reduced by changing the crane lifting mechanism and placing rubber mats on top of the buffer pistons. After filtering, the Channel Description form shown in Figure 2.38 is used: • To show the label that will be used to identify the type of signal; that is, either the actual channel number (e.g. Ch06) or a derived value such Dis, Vel or Acc. • To select a Test Component (i.e. beam, simulated rock mass or ground support system component). Each channel is associated with a component in the test to assist with selection later of channels to be included in the engineering analysis • To select the Data Type (i.e. displacement, velocity, acceleration or force). • To show the Filter Type and its Settings. The parameter-time record is transferred to the ‘Data Storage’ interface shown in Figure 2.39. After all required data have been transferred, a new data file is created with the same structure as the original file (.au2) but with a different file extension (.flt).
56
Figure 2.38: Channel Description form.
Figure 2.39: Data storage interface form.
57
2.5.4
HIGH SPEED VIDEO ANALYSIS
The high speed video analysis is performed in customised Excel spread sheets. The purpose of the analysis is to calculate the location of the loading mass and head of the reinforcement for the duration of a test by tracking a line of targets on the centre-line of the drop for the ground support element. The camera frame rate is 250 frames per second; that is every 4milliseconds. At this sample rate the loading mass travelling at 6m/s will move 24mm between frames. Various targets are used for the video camera to record. For example, for reinforcement tests there are painted crosses on the two nuts that hold the loading mass together. These are in the same vertical plane as the web of the drop beam. Also, one or two painted markers on the head of the reinforcement system may be used. Other target arrangements are used for mesh and shotcrete panel testing. A Midas software track of targets is shown in Figure 2.40. On completion of the tracking, the configuration and target tracking information are exported to a customised Excel spread sheet. The tracking information includes the raw x-y screen coordinates determined from a grid system with an origin in the lower left hand corner of the camera frame. The co-ordinates for the first frame of the video track are entered by the user.
Figure 2.40: Midas track of targets. As indicated earlier, a calibration procedure was developed to improve the accuracy of the displacement tracks. This was necessary to reduce the error associated with using the software based calibration. The
58
corrections required to adjust apparent measurements in the camera picture plane to the actual displacements in the plane of the targets were described in detail in the M349 final report. 2.5.4.1
Smoothing Video Track
For the testing geometry used it was found that when the displacement of the targets exceeds the velocity of 0.43m/s (equivalent to ½ pixel divided by camera frame rate) comparatively large and unrealistic steps could occur in the data. Therefore, the raw captured video data requires analysis to obtain a realistic, smooth displacement-time curve. Smoothing is achieved for each target by using a rolling three point average of the prior, current and next value for the current time. The result of the three point averaging is shown in Figure 2.41. It was found that five point averaging of the displacement data removed the maximum velocity which occurs at impact. The individual, averaged displacement tracks can be used to derive the velocity (by differentiation of displacement) with the results shown in Figure 2.42. In this case, target number two stopped tracking the frame before impact. The acceleration of the tracks is derived by differentiation of the velocity with the results shown in Figure 2.43. Small errors in the indicated rate of change of displacement lead to large variations in the predicted accelerations. Multiple targets can be averaged as appropriate and compared with other independent data sources. Figure 2.43 show the accelerations derived from three point averages from individual tracks compared with the average of all tracks and the Fast Fourier Transform of the response from the 500g accelerometer on the loading mass. The acceleration estimated by double differentiation of the displacement to within 20% of the accelerometer output is deemed an acceptable result.
59
EffectAffect of three of three point point averaging averaging on displacement-time on displacementdata fromderived auto track fromsoftware auto track
0.03 Target 1 3 point on target 1
Change between frames (m)
0.025 0.02 0.015 0.01 0.005 0 -0.005 -0.01 -0.05
0
0.05
0.1
0.15
0.2
Test time (seconds)
Figure 2.41: Averaging displacement from a video track.
Velocity calculation 3pt averaging for the single tracks 7.00 6.00 3p y1 vel
Velocity (m/s)
5.00
3p y2 vel 3p y3 vel
4.00 3.00 2.00 1.00 0.00 -1.00 -2.00 -0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
Test time (seconds)
Figure 2.42: Velocity from average displacement tracks.
60
0.35
0.4
Accelerations from target tracks in comparison to accelerometer response 60 40
Acceleration (m/s2)
20 0 -20 -40 -60
3p y1 acc
-80
3p y2 acc 3p y3 acc Average all tracks
-100
flt file on accel on mass -120 -0.050
0.000
0.050
0.100
0.150
0.200
0.250
Test time (seconds)
Figure 2.43: Acceleration from displacement tracks. The average values from the tracks are linked to a single spreadsheet in Windows Excel. The data from the spreadsheet can be exported as a ‘.csv’ (comma separated variable) file to the WASM custom software for further cumulative smoothing of the displacement data. The WASM software uses the Savitzky-Golay algorithm (Savitzky and Golay, 1968). The algorithm performs a least squares fit of a small set of consecutive data points to a polynomial and takes the calculated central point of the fitted polynomial curve as the new smoothed data point. The smoothed displacement is imported to a separate page of the same Excel spreadsheet. The difference between the smoothed displacement track of the loading mass and the buffer closure (the average of two linear potentiometers smoothed using a Butterworth filter) provides the elastic and plastic deformation of the reinforcement system shown in Figure 2.44. The final displacement for a stable test can be compared with the measured yield of the reinforcement system at the simulated discontinuity and deformation of the surface hardware. In this example, the physically measured displacement post test at the simulated discontinuity was 146mm. This is in close agreement with the value that can be derived from Figure 2.44.
61
Displacement (meters)
average track of all tags
buffer closure
all tags minus buffer closure
0.28 0.26 0.24 0.22 0.2 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 0
0.1
0.2
0.3
0.4
0.5
0.6
Test time (seconds) Figure 2.44: Displacement of buffers and video track during the test. 2.5.5
INTERMEDIATE DATA FILE PROCESSING
After the raw accelerometer signals were filtered a common time base needs to be established for the test. The process of applying the FFT adds a phase shift to the processed signal, causing all waveforms to have different apparent impact times. This could be corrected individually by applying the time phase shift as the waves are created, or by post processing in an Excel spreadsheet by shifting the rows within the columns of sensor data to provide the common start time given by the impact time onto the buffers determined by the video record. The data in the columns for the waveforms prior to impact are changed to be equal to zero, and all rows are made the same length as a 2n window. The file is reformatted within Microsoft Word to enable the custom software to read the format. The time difference is maintained between the impacts on the linear potentiometers. 2.5.6
DATA FILE STORAGE
The files that are generated from a dynamic test and the data analysis are: •
sensor raw data
•
high speed video
•
configuration file
•
processed video tracks (displacement, velocity and acceleration)
•
filtered instrumentation data
•
analysis output derived from the WASM software
62
•
the analysis spreadsheet
All these files are stored in a sub-directory created for each test. All files have a common root name with different file extensions.
2.6
ENGINEERING ANALYSIS
The engineering analysis uses as input the file of filtered data created by the WASM software and the file saved from Excel for the smoothed video displacement track of the loading mass and head of the reinforcement system. 2.6.1
ASSUMPTIONS FOR THE ANALYSIS
The assumptions required to analyse the filtered data are: •
The drop beam and support testing frames behave stiffly. There are strain gauges on the beam and there are elastic deformations during a test but the deformations of these can be considered to be small compared with the reinforcement and support displacements.
•
The beam displacement, velocity and deceleration after impact are considered to be equal to the corresponding parameters for the buffers. This means that parameters may be measured for either the beam or the buffer piston.
•
Vertical movements only are considered in the analysis so only the vertical components from the accelerometers are used and the rotational and horizontal components of acceleration (and displacements) are ignored.
•
Filtering of the data does not influence the overall outcomes from the results or the relative performance of the classes of reinforcement system classifications but it will affect details of the charts.
•
The shift of the filtered data to a common start time from the digital video does not significantly influence the overall result; however, it does change the early shape of the response charts.
•
The number of buffers in a test does not significantly affect the results. This was validated by testing samples of the same reinforcement system impacting onto either two or four buffers and comparing the resultant dynamic force-displacement responses.
Other than already noted, the other assumptions have also been validated as part of the testing program of reinforcement systems and can be expected to also apply to the testing programs of mesh and shotcrete panels.
63
2.6.2
INPUT DATA FILES
The filtered data (.flt) file is imported first and all the data are added to a text list box and into an array in computer memory. The video (.vid) file is then imported. Because the frame rate for the video file differs by a factor of 100, during the import process additional data for the missing time stamps is interpolated assuming a linear variation. Also, the video file often has a later start time than the filtered data. This is automatically taken into account and the data then added to the text list box and the array in computer memory. Figure 2.45 shows an example of the synchronised combination of channel and video data.
Figure 2.45: Filtered input data example showing start of video data synchronised with the channel data.
The FFT waveforms have a common start time applied by zeroing the data before the impact to remove artefacts that result from the application of the FFT to the complete waveform. Waveforms with the synchronised start times are shown in Figure 2.46. The WASM software uses the processed waveforms with the common start time. The channels for analysis are selected using the interface form shown in Figure 2.47. It is possible to ‘drag and drop’ more than one channel into a text box. The software automatically determines the number of combined channels and then averages the data for that parameter.
64
The analysis of data is performed over a specified time period. The displacement, velocity, acceleration momentum, energy and force variations with time for all components can be reviewed. The forcedisplacement responses for the ground support component and buffers are also calculated. These data are saved automatically as a csv (comma separated variable) file that can be imported into Excel for further analysis or to enable charts to be created for comparison of several different systems.
Video displacement
Potentiometer displacement
Figure 2.46: Filtered waveforms with synchronised start times.
65
Figure 2.47: Interface used to select channels for analysis. 2.6.3
ENERGY-TIME CHART
The Energy-Time chart is one of the key measures used to characterise and reconcile the dynamic ground support system tests. The following sections discuss this chart and how it is used. 2.6.3.1
Buffers and Energy Dissipation
Recognizing all dynamic test facilities have energy losses, it is not adequate to report the maximum kinetic energy at impact of a free moving body onto a stationary body as implied by Eq. 2.1. 2 KEabsorbed = 1 2 mload vrelative
where
Eq. 2.1
mload = the mass of loading (i.e. steel rings and collar pipe) vrelative = the relative velocity is the difference in the velocity between the loading mass and the beam.
An energy balance at any time during the test is required to understand the way energy is dissipated by the test specimen and facility. This expressed by Eq. 2.2 as:
E T = E R + E B + KE M
Eq. 2.2
66
where ET is the input energy comprising the kinetic energy of all components immediately prior to impact plus the loss of potential energy associated with downward displacements after impact. ER is the energy absorbed by the reinforcement or surface support. EB is the energy absorbed by the buffers. KEM is the kinetic energy of the loading mass. The force-displacement responses provide the best measures of the energy absorbed by the buffers and ground support system. 2.6.3.2
Total Error Analysis
A comparison of input energy with dissipated energy is used to assess the errors that might arise during data processing. Errors are associated with the measurement of the masses, accelerations and displacements, calculation or measurement of the impact velocity, measurements of unrecorded elastic and plastic deformations and filtering of the instrumentation signals. The methodology and equations for these calculations were detailed by Thompson et al. (2004). Figure 2.48 shows the energy summary for a test. The step changes in the energy of the mass are due to the position of the loading mass only being calculated by the changes from the video track which occur every 4ms. Figure 2.48 illustrates the close agreement between the measured input energy and the sum of the energy dissipated by the buffers and the reinforcement system. In this case the input energy (ET) was 50.5kJ and the sum of the dissipated energy in the reinforcement (ER)) and the buffers (EB)) was also 50.5kJ. In general, there will be some errors. For a test in which the ground support system does not fail, this is expressed as the ratio of the difference between input and absorbed energies to input energy. That is:
E T − (E R + E B ) ET
Eq. 2.3
67
EI=50.5kJ
100%
ER=15.5kJ EB=35.0kJ
Figure 2.48: Error analysis - stable reinforcement system test. For tests that result in failure of the ground support system, the loading mass will have kinetic energy and momentum following the time of failure. In this case the error is given by:
E T − KE M − (E R + E B ) ET
Eq. 2.4
The energy values shown in Figure 2.49 are for failure of a reinforcement system at 46.0ms after impact. The input energy should be equivalent to the energy dissipated into the buffers and reinforcement system plus the kinetic energy of the rock mass. However, there is an error of about 8%.
68
EI=52.7kJ 108% KEM=10.3kJ
ER=21.7kJ EB=24.9kJ
Figure 2.49: Error analysis - unstable reinforcement system test. Tests in which failures occur are generally more difficult to reconcile due to errors associated with the shorter analysis period (stable tests have longer durations and an averaging effect to reduce error), and inaccurate representation by the Fast Fourier Transform (FFT) of the deceleration of the mass and acceleration of the buffers at the beginning and end of the test. The error in the energy balance has been typically less than 20% for each test in which the element did not fracture and is included in the Standard Test Report.
2.7
STANDARD TEST REPORT
The results from all tests are given in the form of a Standard Test Report, an example of which is shown in Figure 2.50. The report has two major sections: •
Test Summary that is used for general distribution and discussion of the test.
•
Detailed Analysis and Interpretation Section that contains details on how the analysis was performed, errors encountered, and the shapes of the filtered and processed waveforms.
The following sections show examples of sections of the Standard Test Report for reinforcement systems. However, the same format is also used for presenting the results from dynamic testing of mesh and shotcrete panels.
69
Figure 2.50: Typical test report - page 1.
70
The Test Summary includes: •
energy dissipation by all system components.
•
load-displacement curve at the simulated discontinuity.
•
acceleration-time curve of the loading mass.
•
the velocity-time curve of the loading mass relative to the stiff beam.
A number of indices result to characterise the performance of reinforcement systems. These indices allow for comparison of different reinforcement systems. In particular it has been found that it is insufficient to only report dissipated energy. Other parameters such as total displacement, peak velocity and acceleration are required to assess performance. The testing also shows the significant differences in performance by undertaking multiple loadings on a single reinforcement system to the failure point compared with a single loading that fails the reinforcement system. Most importantly the energy dissipated from multiple small loads cannot be summed to conclude that the total will be the ultimate capacity for a single loading event. This is due to progress damage that changes the condition of the load transfer mechanism and the system configuration. Therefore, only the results from the first loading are considered in characterising the response of the reinforcement system. 2.7.1
TEST SUMMARY SECTION
The main parts of the Summary Section of the test report are the reinforcement system specification and configuration, the test specification, the performance spreadsheet graphs (dynamic force-displacement responses, time based results, summary and peak values), overall system performance (physical measurements and response to the load), photographs and analysis of the failure or yielding mechanism. 2.7.1.1
Reinforcement System Specification
An example of the information recorded for a reinforcement system specification tested is shown in Figure 2.51. The reinforcement system specification documents: •
The test operator, data and the type of reinforcement system under test.
•
Classification of the system – that is CFC, CMC or DMFC.
•
The encapsulated or coupled distance for the collar and anchor lengths. The total length of the reinforcement system. The presence and length of any decoupling component.
•
Information about the steel grade.
•
A description of the simulated borehole.
•
A description of the surface hardware.
•
The applied collar tension or torque applied to a nut (from which tension can be estimated using software developed for the purpose).
71
* WASM Static Testing
Figure 2.51: Test report sheet – reinforcement system specification. 2.7.1.2
Test specifications
An example of the information recorded about the test specifications is shown in Figure 2.52.
Figure 2.52: Test report sheet – test specification. The test specifications document: •
The conditions of the borehole that may be expected to influence the test results - the internal diameter, variation in diameter (only documented for the rough simulated boreholes) and equivalent radial stiffness.
•
The masses of components that contribute to the energy calculations - loading mass and the borehole collar mass (reinforcement system plus the simulated borehole).
•
The length of the collar pipe to provide the simulated discontinuity.
•
The impact velocity of the beam onto the buffers, assessed by comparison of the high speed video to the laser break trigger calculations of the impact velocity.
•
Nominal input energy that is equal to the kinetic energy of the combined loading mass plus collar pipe mass at impact.
•
The presence of a load transfer ring defines whether the loading mass is integrated with the outside of the borehole.
72
2.7.1.3
System Performance
Four charts shown in Figure 2.53 are used to characterise and summarise the system performance.
52.7kJ
94% 21.6kJ 27.8kJ
Figure 2.53: Test report sheet - reinforcement system performance. The Energy Summary shows the total input energy and the energy dissipated by the loading mass plus collar pipe mass, buffers, beam and reinforcement system. The energy summary is also used to determine the error between the input energy and the energy dissipated by the buffers and reinforcement system. The peak energy dissipated by the reinforcement system is reported in the summary table as ERS. The dynamic force-displacement response is calculated at the simulated discontinuity; as indicated previously, it can be quite different from the anchor and collar load cell results due to load transfer within the reinforcement system. The calculated peak force (Fpk) and displacement (Smax) from this curve are reported in the summary table. The loading mass acceleration-time chart displays the average of the filtered responses of the loading mass as the reinforcement system acts to bring the mass to rest from the impact velocity. The peak deceleration (GMpk) is reported in the summary table.
73
The reinforcement element relative velocity-time chart displays the velocity difference between the loading mass and the drop beam. The peak relative velocity is reported as (Vpk) in the summary table. 2.7.1.4
Overall system performance
The physical response of the reinforcement system and influence of the potential energy change of the loading mass are reported in Figure 2.54 for the overall system performance. This is followed by a description of the behaviour mechanism. The report may include a series of photos of the reinforcement system, simulated discontinuity, yielded length, surface hardware performance, a dissection of the simulated borehole examining the encapsulation medium, the simulated borehole measurements pre- and post-test to determine if there was plastic deformation, and, when appropriate, comparison with the static force-displacement responses.
Figure 2.54: Test report sheet - overall system performance. 2.7.2
DETAILED ANALYSIS AND INTERPRETATION SECTION
The Detailed Analysis and Interpretation Section of the test report provides a mechanism to ensure that necessary or unusual information that could explain the performance of the reinforcement system tested is gathered. The data and information that are recorded in this section include: •
The performance of the instrumentation channels (sensor saturation, sensor or cable damage).
•
The video time of impact on the buffers (this becomes the zero time for the test for which the filtered instrumentation channels are set).
•
A summary of the buffer performance – type and number, non-uniformity of impact (time and compression), maximum deceleration and time after impact calculated form the analysis software.
•
Peak acceleration components in x, y, z-axis for the triaxial accelerometers on the beam and mass, although just reporting peaks doesn't give enough detail of out of balance forces because the duration is not specified.
74
•
Examination of the filtered and non-filtered waveforms.
•
A comparison of the calculated impact velocity between the video track and the laser beam triggers.
•
A graphical comparison of the loading mass deceleration from the high speed video track compared to the filtered accelerometer results.
•
A comparison of the displacement of the loading mass relative to the beam at the end of a test with the physical displacement measurement of the simulated discontinuity post test.
Some examples of the information recorded in the Detailed Analysis and Interpretation Section are extracted when presenting and discussing results from various tests on reinforcement and support systems.
2.8
ASSESSMENT OF THE WASM TEST FACILITY
2.8.1
FEATURES OF THE FACILITY
The features of the WASM Test Facility are its ability to: •
Test full scale systems (both reinforcement and support).
•
Integrate a simulated rock mass with the reinforcement system to be tested.
•
Simulate the effects of different borehole wall conditions and equivalent rock mass radial confinement.
•
Input energies sufficient to fail high strength reinforcement and support systems with a single impact (e.g. Figure 2.55).
•
Replicate dynamic loading caused by a large block of rock with high momentum.
•
Provide data from extensive instrumentation within the facility and systems tested.
•
Use data analysis methodology and software to understand the critical loading conditions.
•
Use extensive analysis techniques to develop dynamic force-displacement response curves, velocity, deceleration and energy time graphs for the ground support systems that have been tested.
These features have been used to establish a database of mechanical properties for both reinforcement and support systems. This database is documented as an Addendum to this final report. The first version (Ver 0) of the Addendum was created after the completion of M349. The Addendum is currently being finalised and will be made available to project sponsors following its completion.
75
Figure 2.55: Snapped 15.2mm plain strand cables. 2.8.2
LIMITATIONS OF THE FACILITY
Some perceived limitations in the facility relate to the time and cost considerations for tests in the facility: •
The facility is no doubt the most expensive mining dynamic test facility ever to be constructed.
•
It is likely to have the highest unit test cost.
•
It probably has the longest test set-up time.
However, these limitations need to be considered relative to the costs associated with ‘getting it wrong’. The development and marketing costs associated with new reinforcement systems are very high. In looking back over many years, reinforcement systems have come and gone for a variety of reasons, but one reason has been the failure to be able to address unforeseen technical problems as they arose. As a consequence, potentially large and lucrative markets within Australia disappeared along with the ground support system. In the current project, the WASM Dynamic Test Facility has been used to evaluate several systems. The tests have been able to identify and resolve technical issues. In one case, the results suggested that the reinforcement system was unreliable and consequently further development and promotion to the Australian mining industry was discontinued. In another case, a system was modified and resulted in improved performance. It is suggested that the costs of testing are relatively small compared with the costs associated with rectifying technical issues after many reinforcement systems have been installed and potentially have to be replaced. All prior mining dynamic test facilities as discussed previously have had serious deficiencies. The writers would argue that much of the costs associated with establishing the WASM Dynamic Test Facility resulted from the uniqueness of the facility and attempting to better simulate loading conditions representative of those caused by violent rock failure. It is believed that should any subsequent facility need to be constructed, it could draw on much of the work completed by WASM in being able to correctly load the
76
systems being tested and being able to provide calculations of the energy absorbed by the systems that have been tested and understanding the parameters that contribute to their performance. In common with most other testing facilities, WASM static and dynamic tests are carried out under axial load conditions that are not necessarily representative of the loading that can occur underground. In particular, rock movements may cause shear loading of a reinforcement system. It would be difficult, though not impossible, to design a test in which shear loading of reinforcement systems could be simulated. At this time, there are no plans to attempt this work. 2.8.3
CONCLUDING REMARKS
The WASM Dynamic Test Facility has been independently assessed in two ways. Firstly, Professor Ted Brown, an internationally recognised authority in rock mechanics stated in his 2004 keynote presentation at the international ground support conference held in Perth that: ”The most advanced dynamic testing system known to the author is that developed recently at the Western Australian School of Mines (WASM), Kalgoorlie”. Secondly, the facility and research team received a Special Commendation for Research and Development at the 2005 Western Australian Engineering Excellence Awards. It is worth noting these acclamations were made prior to the improvements and enhancements made during MERIWA Project M349A.
77
78
3
DYNAMIC TESTING OF REINFORCEMENT SYSTEMS
This section describes the preparation and dynamic testing of samples of various reinforcement systems. The reinforcement systems were classified into the three classes defined previously; namely: •
Continuously Frictionally Coupled
•
Continuously Mechanically Coupled
•
Discrete Mechanically or Frictionally Coupled
The reinforcement systems tested within each class were as follows: Continuously Friction Coupled •
Split Tube Bolts
•
Expanded Tube Bolts
Continuously Mechanically Coupled •
Encapsulated Threadbar
•
Encapsulated Strand
Discrete Mechanically or Frictionally Coupled
3.1
•
Partially Decoupled Threadbar
•
Partially Decoupled Bulbed Strand
•
Cone Bolt
•
Modified Cone Bolt
•
Garford Dynamic Cable Bolt
•
Garford Dynamic Solid Bolt
SIMULATED BOREHOLES
An important component of an in situ reinforcement system is the rock surrounding the borehole. In particular, the interface between the internal fixture and the borehole wall influences the rate of load transfer. In laboratory tests with static loads it has been found that load transfer for reinforcement systems encapsulated with cement grout is controlled mainly by the interface between the grout and the element. For reinforcement systems such as those within the Continuous Frictionally Coupled category, the load transfer is directly between the element and the borehole wall. These systems have not been tested in the laboratory to the same extent as encapsulated systems. Therefore, it was necessary to develop a method to better simulate the performance of friction rock stabilisers than previously in any static or dynamic facilities world-wide.
79
In the WASM Dynamic Test Facility, there are two basic types of simulated borehole: •
A standard borehole where the main function is to provide confinement to a cement encapsulated reinforcement element.
•
A rough borehole where the main function is to simulate the profile of a borehole drilled into rock.
The following sections detail issues related to the simulated boreholes and particularly, the creation of a rough borehole into which reinforcement elements are driven or rotated in the field by a drilling jumbo. 3.1.1
PIPE RADIAL STIFFNESS
It is well known that the performance of reinforcement systems is influenced by the confining effects provided by the rock surrounding a borehole. The radial stiffness provided by steel pipes can be varied by changing the wall thickness to simulate different rock confining conditions. Hyett et al., (1992) used thick wall pipe theory to derive Eq. 3.1 for the ratio of the internal radial outwards deformation (ui) due to the applied internal pressure (pi),
2 ( 1 + νP ) u i ri (1 − νP ) r i = + pi EP 2 r o 2 1 − r i 1 − r i r r o o where
EP
=
Young’s modulus for the pipe material
νP
=
Poisson’s ratio for the pipe material
ri
=
internal radius of the pipe
ro
=
external diameter of the pipe
Eq 3.1
The ratio of displacement to pressure is a measure of compliance (the radial stiffness is the inverse of compliance and will be expressed in units of MPa/mm). If the external diameter is assumed to be very large (i.e. ro = ∞) to simulate rock surrounding a borehole, Eq 3.1 can be re-written as Eq 3.2:
u i ri = (1 + νr ) pi E r
Eq 3.2
where Er and νr are, respectively, the Young’s Modulus and Poisson’s Ratio for rock. Eq. 3.1 and Eq. 3.2 may then be used to estimate the radial stiffnesses for different dimensions of the steel pipes used in laboratory tests in terms of equivalent rock stiffness Er.
80
3.1.2
STANDARD BOREHOLE
Table 3.1 shows the selected pipe dimensions together with the equivalent rock stiffness. The pipe internal diameters are consistent with the borehole diameters used in mining practice.
20mm HS Threadbar (Australian)
49.5
60.3
5.4
767
50
22mm LS Threadbar (Chilean)
45.4
50.5
2.5
457
25
22mm Cone Bolt – High Strength Cement Grout
45
63.6
9.3
1389
80
22mm Cone Bolt – Rock Yield Cement Grout
45
63.6
9.3
1389
80
15.2mm Plain Strand (Australian)
62.1
73.2
5.5
516
40
15.2mm Plain Strand (Chilean)
56.7
62.2
2.8
326
25
Typical Hard Rock
45
Infinite
1174
65
(GPa)
Rock Stiffness
Equivalent
(MPa/mm)
Stiffness
(mm)
Wall Thickness
Diameter (mm)
External
Diameter (mm)
Reinforcement Element
Internal
Table 3.1: Equivalent Rock Radial Stiffness for Steel Pipes
The very high equivalent rock stiffness used for the 22mm Cone Bolts in high strength grout was to minimise the possibility of the expansion of the pipe due to the high lateral forces generated from the action of the cone ploughing through the grout. This was not always successful and expansion of the thickest wall pipe occurred in the region of cone displacement. The standard simulated boreholes also included shear pins through the steel pipe into the grout to prevent, or at least minimise, the grout annulus from sliding inside the steel pipe. Two shear pins 180º apart were located nominally 300mm from the simulated discontinuity. The pins, typically 15mm long by 8mm diameter bolts, were passed through unthreaded holes drilled through the pipe wall and then held in place after the poured cement grout had cured. In other cases, holes were drilled after the grout had cured and the pins were put in place prior to the test. 3.1.3
ROUGH BOREHOLE
Extensive development was required to create a consistent, rough borehole located centrally within a pipe. Rough boreholes were required for rock bolts where performance was known to be sensitive to the wall conditions of the borehole and the equipment used for installation. It was not practical to simulate the borehole deviation that occurs during the drilling of boreholes. The following sections describe the materials trialled for the construction of the simulated rough boreholes and the construction process.
81
3.1.3.1
Material Selection
The development of a suitable material and methodology to create a simulated borehole with roughness and strength similar to a borehole in rock was an evolutionary process and resulted from three decisions: 1.
The simulated hole should be constructed as a single piece and not two half shells clamped together.
2.
The hole was to be drilled by an airleg drill rather than poured about a polystyrene mould of the required dimensions that was later removed.
3.
The selection of a low-shrinkage, high strength construction grout with addition of basalt aggregate. This resulted in a concrete core with higher strength and stiffness than neat cement grout.
As a result of the second and third decisions, the initial 11 split tube bolt tests that used low heat cement grout cast about a polystyrene mould needed to be ignored. In these tests, the measured forces in responses to dynamic loading were very poor, despite the static test pull resistance being comparable with underground pull test results. 3.1.3.2
Creation of a Simulated Rough Borehole
A simulated rough borehole consists of a hollow steel bar (80mm internal and 100mm external diameter) that contains a concrete annulus and a centrally located hole. A polystyrene tube was used to create a pilot hole for the airleg drill to follow as shown schematically in Figure 3.1. The polystyrene tube was passed over an 8mm diameter, threaded steel bar. A 4mm thick square end plate was welded to the toe end of the pipe to stop the fluid material flowing out. At the collar end, an 8mm thick bar was clamped in place by tensioning the threaded steel bar to hold the polystyrene centrally in the steel pipe as shown in Figure 3.2.
82
Dimension of rough borehole once drilled out by airleg
80*100 hollow bar
Poured concrete annulus
8mm steel guide rod to hold polystyrene in place
20mm diameter polystyrene guide.
Figure 3.1: Schematic of a rough simulated sample. The materials were mixed and poured into the steel pipes as shown in Figure 3.3. At the same time, test cylinders of the concrete were poured for uniaxial compressive strength testing of the products. After allowing the material to cure for at least 28 days, the threaded steel bar was removed and the hole then reamed out to the required size by an airleg drill used by an experienced driller as shown in Figure 3.4. Significant skill on the part of the operator was required to successfully ream out the polystyrene from a diameter of 20mm to the required nominal 45mm diameter. Nominal 32mm diameter boreholes, associated with creating resin encapsulated reinforcement samples for testing, could be drilled carefully without the use of the polystyrene guide. It is worth noting that the range of simulated borehole diameters are limited to available knock on drill bits for an airleg drill steel.
83
Collar
80*100Hollow bar
Toe
Top bar
Figure 3.2: Simulated boreholes ready for filing. A number of materials to represent the rock surrounding a borehole were trialled prior to achieving a satisfactory outcome. Initially, the cement based construction grouts used were Masterflow 880 and 870 manufactured by Degussa Construction Chemicals. These grouts had 80-90MPa compressive strength and were non-shrink. The 880 grout is supplied with iron filings. This was initially thought to be necessary to reduce the potential for tensile fractures during the drilling process. However drilling into a standard 880 mix was still difficult. Trials were then undertaken with basalt aggregate chips added to Masterflow 880 grout. The basalt aggregate chips (sized 8-10mm), cement and water were mixed at the recommended water/cement ratio for the recommended time in a conventional mixer. The aggregate significantly improved the airleg operator’s ability to drill the borehole. The next material trialled used a combination of basalt aggregate chips with a 1:2 ratio of 8-10mm and 1012mm particle sizes. The aggregate comprised 20% by weight of the total mix. It was noted subsequently that the iron filings in the Masterflow 880 promoted corrosion of split tube bolts when they were driven into the simulated boreholes by a jumbo using hyper-saline flushing water. Consequently, it was decided to use the Masterflow 870 grout. Also, it was decided to only use the larger size 10-12mm basalt aggregate chips. The results from core compression testing determined the Young's Modulus was 25GPa for the Masterflow products without aggregate and 31.4GPa with aggregate. A typical equivalent rock mass modulus of 36GPa was calculated for the nominal 45mm diameter simulated rough boreholes.
84
Figure 3.3: Poured cement mix and tested cement sample.
Figure 3.4: Airleg drilling of the sample to form the simulated rough borehole. The simulated rough boreholes were thoroughly washed to remove any drill chips in order to maximize the grip available between the installed bolt and the side of the simulated borehole and to prevent dust or grit acting as a layer of weakness. This was also necessary as drilling the simulated borehole horizontally does not result in the same amount of flushing that would occur normally from drilling inclined holes in the underground mining environment. The rough borehole internal diameters were measured with a custom built borehole probe shown in Figure 3.5 in order to assist with interpretation of variations in the test results. The device has a hole diameter measurement range of 40mm to 47mm. The information obtained was used to produce the borehole diameter profiles shown in Figure 3.6 which can be compared with results from boreholes drilled at underground mining operations. The important aspect of Figure 3.6 is the similarity in variation of the diameter for the simulated and actual boreholes rather than the mean diameter being assumed to be the same as the bit size.
85
Figure 3.5: Custom built borehole probe.
46
45.5
Diameter (mm)
45 Sample 104 Sample 108
44.5
Sample 111 Sample 106 UG1
44
UG2
43.5
Samples for Omega bolts, underground holes for split tube tube bolts.
43
42.5 42 0
0.5
1
1.5
2
2.5
Distance from Toe (m)
Figure 3.6: Typical borehole profile results. The final stage was to weld on an appropriate sized steel end plate at the collar end of the pipe. This end plate enabled the simulated borehole assembly to be taken into the field and placed inside a large diameter borehole prior to driving in split tube friction rock stabilisers or spinning in bars to encapsulate them with resin grout. The plate at the end of the pipe corresponds with the collar of the borehole and is used to simulate the rock surface to allow installation of the reinforcement system surface hardware. This feature of the simulated rough borehole will be described in more detail in later sections.
86
3.2
ELEMENTS
The following sections describe the elements that have been used in preparing reinforcement systems that have been tested in the WASM Dynamic Test Facility. 3.2.1
FRICTION ROCK STABILISERS
The testing of friction rock stabilisers has presented challenges to understanding the performance characteristics of this category of reinforcement system bolts. Two types of bolts have been tested; split tube and expanded tube. 3.2.1.1
Split Tube
A split tube bolt is shown schematically in Figure 3.7. The original split tube bolt was the proprietary product Split Set. In recent years, many similar products have been made available to the mining industry. The products may look similar but have been manufactured with different dimensions and from different grades of steel with consequent variable performance. For example, a lower yield strength material results in a higher contact area but lower radial forces compared with a higher strength steel as shown in Figure 3.7a and b, respectively.
b. Split-tube steel elastic, reduced contact area, increased contact force
a. Yielded split-tube, high contact area reduced contact force
Figure 3.7: Schematic of split tube. 3.2.1.2
Expanded Tube
There are currently two expanded tube bolts available in the Australian mining industry. They are both manufactured in the same factory and marketed under the names Swellex and Omega Bolt. Figure 3.8. shows schematically the cross-section of these bolts after inflation. The radial contact forces are related to the borehole size, rock stiffness and strength, and the pressure used for expansion.
87
Inflated bolt, consistent higher contact force
Figure 3.8: Schematic of inflatable tube. 3.2.2
THREADBAR
Bar with a continuous thread has a long history of use in civil engineering projects, particularly for long ground anchors where the ability to join short straight lengths is achieved using threaded couplers. This has become known as ‘threadbar’ in mining applications. Threadbar may also be referred to as Gewi bar, a proprietary product with a left-hand thread developed by Dywidag Systems International (DSI) many years ago. Two types of threaded bar have been tested; one designated High Strength (HS) as supplied in Australia and the other designated Low Strength (LS) as used in Chile. 3.2.2.1
High Strength and Low Strength Threadbar
The Australian HS threadbar is nominally 20mm in diameter with a coarse 10mm pitch thread and geometry shown in Figure 3.9. The Chilean LS threadbar was designated with a diameter of 22mm by the manufacturer. However, it has the same bar area as the Australian 20mm HS threadbar.
On the outside of the threads
Across the flat
Figure 3.9: Threadbar diameters.
88
3.2.2.2
Comparison of Element Mechanical Properties
The mechanical properties of the Australian HS threadbar are specified by Australian Standard 4671:2001, Steel Reinforcing Materials. The threadbar used in the fully encapsulated tests were hot dip galvanized. Static tests undertaken on samples of both types of threadbar are shown in Figure 3.10. The LS threadbar has a significantly lower yield stress of 280MPa, compared with the Australian HS threadbar at 500MPa, but a more prominent profile. An important aspect of the lower yield stress steel is that it is expected to have a higher strength increase when subjected to dynamic strain than higher yield stress steels (Malvar and Crawford, 1998). Galvanised high strength steel threadbar from Australia
Threadbar Pull Tests
250 Black high strength steel threadbar from Australia
200
ET-1B ET-1A ET-2A ET-2B DSI-1A DSI-1B DSI-2A Data-1B
Load (kN) 150 Black mild steel threadbar from El Teniente
100
50
0 0
50
100
150
200
Displacement (mm)
Figure 3.10: Comparison of mechanical responses of Australian and Chilean threadbar. 3.2.3
CONE BOLT
The original cone bolt was designed in South Africa to be installed in a cementitious grout. The bolt was available in two bar diameters (16mm and 22mm) with a forged cone at the end of the bar. The threads at the collar of the bolt are M18 or M24 respectively. The bar and the cone had a saponofied wax coating to assist with the decoupling of the steel from the cement grout encapsulation. The cone bolt steel conforms with the properties defined by SABS 1408 – 1987.
89
Figure 3.11: 22mm diameter cone bolt. 3.2.4
MODIFIED CONE BOLT
The Modified Cone Bolt (MCB) shown in Figure 3.12 was developed in Canada and is designed to be encapsulated with resin grout that is mixed by spinning the bar through a two-component cartridge. The concept of using resin for encapsulation medium rather then cementitious grout was discussed originally by Jager (1992). The idea was further developed with the modified cone bolt (MCB) as described by Gaudreau et al. (2004) and St-Pierre (2007). The MCB has been designed to use a blade/paddle (Figure 3.13) on the end of the cone to act as the mixing device for the resin. The use of resin provides for immediate reinforcement of the rock. The MCB has been specified to be used only with a particular encapsulation medium (i.e. Dupont Coneloc resin). The MCB has been further modified with the inclusion of a plastic sheath to ensure more effective decoupling between the resin and the shaft of the bolt. The plastic sheath was not used in the WASM tests.
Figure 3.12: Modified cone bolt (figure from Mansour Mining website).
90
Figure 3.13: Modified Cone Bolt. 3.2.5
PLAIN STRAND
Two types of plain strand have been used in the dynamic tests; Australian and Chilean nominal 15.2mm diameter, 7 wire steel strand. 3.2.6
GARFORD BOLTS
Garford supplies two reinforcement systems designed specifically to have high energy absorbing capacities; these are designated as Garford Dynamic Cable Bolt and Garford Dynamic Solid Bolt, with the former product being developed prior to the latter. 3.2.6.1
Dynamic Cable Bolt
Three sets of Dynamic Cable Bolts have been tested. The samples were configured as shown in Figure 3.14 with a 700mm sliding capacity. The third set of yielding cables (samples 143-145) had a shorter total length at 3.4m in length with 400mm of sliding capacity. 3.6m long 0.6m long
0.76m long
Sheathed / decoupled section of compact strand Anchor that remains fixed in the grout, but allows the strand to slide
Figure 3.14: Garford Dynamic Cable Bolt strand and anchor configuration for samples 78-80.
91
3.2.6.2
Dynamic Solid Bolt
Figure 3.15 shows the Garford Dynamic Solid Bolt. The bolt consists of a 21.7mm diameter 5152 AVH grade steel solid bar with a rolled M24 thread. The patented sliding mechanism involves a ferrule being swaged on to the bolt. A resin mixing device consisting of a 350 mm long, 43 mm steel tube with a coarse thread that is crimped onto the toe end of the bolt. A polyethylene sleeve is passed over the bar between the sliding mechanism and the thread at the collar end; this sleeve is used to decouple the bar from the encapsulation material which may be either cement or resin grout. The ferrule is held in the borehole by the encapsulation material and sliding of the bar relative to the ferrule is activated when loading exceeds the resistance to sliding.
Mixing Device
Debonding sleeve
Yielding Device
High Capacity Nut with breakout insert Plate
M24 Thread
Figure 3.15: The Garford Dynamic Solid Bolt.
3.3
CONTINUOUSLY FRICTIONALLY COUPLED SYSTEMS
Continuously Frictionally Coupled (CFC) systems tested are a type of split tube bolt and a type of expanded tube bolt. As indicated previously, there is now more than one supplier of each generic element type. 3.3.1
SPLIT TUBE BOLTS
3.3.1.1
Sample Preparation
The simulated rough boreholes were taken to mine sites to have the friction rock stabilisers installed using WA mining practices. It is known that these bolts are sensitive to the equipment and operator skills and this approach was preferred to installation under ideal laboratory conditions.
92
A large hole was drilled into the tunnel wall to enable placement of the simulated borehole assembly. The jumbo was then used to drive the bolt into the simulated rough borehole as shown in Figure 3.16. The simulated borehole containing the installed bolt was recovered and transported back to the laboratory.
Figure 3.16: Installation of a split tube bolt into a simulated rough borehole. 3.3.1.2
Static Test Results
Some of the split tube bolts had pull rings installed to allow static pull testing. Static pull tests were undertaken prior to dynamic testing to provide a comparison with loads measured from pull tests for bolts installed in hard rock. The friction rock stabilisers have specific test equipment for undertaking static pull tests. In both cases the pressure is read off a gauge inline with a hydraulic cylinder and converted to an applied load at the collar of the bolt. With each pressure reading the extension of the ram on the jack is measured to obtain the displacement at the collar. Although there is the potential for the short displacement that occurs in static tests to influence the dynamic test result, it is important to validate the installed quality of the reinforcement system before proceeding to dynamic testing. This may allow a poor dynamic test to be explained if it is known that the particular sample also had a low resistance to static loading. The testing program involved split tube bolts from two manufacturers. Both galvanised and ungalvanised split tubes were tested. The split tubes were 47mm diameter and driven into a nominal 45mm diameter hole. The testing program involved: •
4 tests on galvanised split tubes from Manufacturer A.
93
•
3 tests on galvanised split tubes from Manufacturer B.
•
4 tests on non-galvanised split tubes from Manufacturer A. None of these bolts were tested statically.
In practice, when 2.4m long split tube bolts are subjected to quality assurance testing, loading is terminated between 80 and 120kN. Static pull test results for split tube bolts in simulated boreholes are shown in Figure 3.17. All tests were stopped once a collar load of 120kN was reached or when sliding had started to occur. The values are consistent with in situ pull tests. In order to compare the results from different test configurations, resistance to sliding is typically reported as force per metre of embedment.
140
Applied Collar Load (kN)
120 Bolt 96, Type A Galv, ave 44.5mm 100 Bolt 95, Type A galv, ave 44.3mm 80
Bolt 94, Type A Galv, ave 44.3mm Bolt 116, Type B Galv, ave 44.3mm
60
Bolt 118, Type B Galv, ave 44.6mm 40 Bolt 115, Type B Galv, ave 44.8mm 20
0 0
2
4
6
8
10
12
14
16
Ram Extension (zeroed at first reading) (mm)
Figure 3.17: Static collar load and displacement for 47mm diameter split tube bolts. The variation in contact area between the split tube and side of different diameter simulated rough boreholes can be seen clearly in Figure 3.18. In a small diameter borehole, the steel yields and has a greater contact area than for a larger diameter borehole.
94
yielded split tube, wide contact area reduced contact force
split tube steel elastic, reduced contact points, increased contact force
Figure 3.18: Striations along the steel due to sliding. 3.3.2
EXPANDED TUBE BOLTS
As mentioned previously, two types of expanded tube bolts were involved in the testing program. Swellex Mn24 bolts and Omega Bolts. Omega bolts are non-galvanised. However, the second set had a corrosion protection coating for transportation to Australia. The Swellex Bolts were only involved in a static testing program while the Omega bolts were subjected to both static and dynamic testing. 3.3.2.1
Sample Preparation
The bolts were installed in nominally 45mm diameter holes representative of those used in practice. The bolts were inflated by a jumbo as shown in Figure 3.19. The bolts were installed by simply pushing them into a simulated rough borehole and attaching the inflation wand; the pump on the jumbo was used to inflate the bolts to a set pressure of 300bar which was held for times ranging from one second to 15 seconds.
95
wand
Strain gauge
wand
monitoring
Figure 3.19: Omega bolt installation. Static testing of the expanded tube bolts was performed in two stages; an initial program of four tests was followed by a second program of three tests. The first program had strain gauges installed onto the outside of the steel pipes to measure the strain change during inflation. These changes were also monitored during dynamic tests. 3.3.2.2
Static Test Result
The first stage of testing was used as an evaluation of the simulated rough boreholes. An Atlas Copco Swellex Mn24 bolt was installed at 280bar and held for 4 minutes. After inflation the simulated rough borehole was cut with a 200mm free length and a 1.0m embedded length as shown in Figure 3.20 to allow examination of the toe of the bolt as sliding occurred. The bolt was pulled with readings taken for collar load, ram extension and toe displacement. The 200mm of free length allowed elastic stretch, followed by sliding of the bolt at a consistent force of 160kN. The purpose of this trial was to evaluate embedment capacity and whether consistent load transfer was achieved once sliding started to occur. Frequent ‘cracks and pops’ were heard during the ram extension for this test; these were considered to indicate shearing of asperities on the surface of the simulated rough borehole or ‘flexing’ of the inflated bolt. The results from the first stage were considered sufficient to show that the static capacity was comparable to what would be expected from tests in rock. Accordingly, samples were then prepared for dynamic testing.
96
In-line gauge
Hydraulic Jack
Ram and cylinder
200,mm free length
1000mm embedment
180 160
Collar Load (kN)
140 120 100 80 60 40 20 0 0
10
20
30
40
50
60
70
Displacement (mm)
Figure 3.20: Swellex sample configuration and static pull test results for one metre embedment. It was decided, similarly to the split tube bolts, that static testing would also be performed prior to and after dynamic testing. Static testing was possible for the first four Omega bolts. However, only two static tests were performed prior to the dynamic tests. The static tests performed after the dynamic tests required spacer bars to be welded between the separated toe and collar embedment lengths. Load was then applied to the collar of the Omega bolt. The static tests conducted prior to dynamic loading were limited to a maximum force of 160kN to minimise disturbance to the toe embedment length. The results are shown in Figure 3.21. The slope of the curve for two bolts is the important aspect and the two bolts can be considered to have the same performance. The application of the first 30kN takes up slack in the pull test equipment and aligns the bolt with the loading direction. The measure displacement comprises elastic extension and straightening and some relative movement between the bolt and the simulated borehole near the collar rather than sliding of the entire bolt. Note that the embedment lengths in Figure 3.21 are the total lengths in contact with the simulated borehole material.
97
Bolt 111 Before Drop (2.25m)
Bolt 104 Before Drop (2.25m)
180 160
Force (kN)
140 120 100 80 60 40 20 0 0
10
20 30 Displacement (mm)
40
50
60
70
Figure 3.21: Static pull tests on Omega bolts prior to dynamic testing. The results for the static tests performed after dynamic testing are shown in Figure 3.22; the indicated embedment lengths are the toe lengths. Table 3.2 summarises the static results. In this table, the measured total loads have been converted to load transfer per metre of embedment length to facilitate comparisons of performance. A key feature from the static tests performed after the dynamic tests is that bolts that had displaced farther during the dynamic test exhibited significantly lower resistance. This is considered to be due to either shearing of asperities or deformation of the tube away from the side wall of the simulated borehole. Bolt 111 After drop (1.06m)
Bolt 104 After drop (0.95m)
Bolt 106 After Drop (1.5m)
180 160
Force (kN)
140 120 100 80 60 40 20 0 0
10
20
30
40
50
60
Displacement (mm) Figure 3.22: Static pull test on Omega bolts after dynamic testing.
98
70
Table 3.2: Summary of static pull test results on Omega bolts. Initial Performance
Bolt
Embedment (m)
Collar Force (kN)
Post Dynamic Load Performance
Capacity (kN/m)
Displacement
Total
Collar
Sliding
in the dynamic
Embedment
Force
Capacity
test (m)
(m)
(kN)
(kN/m)
104
2.25
>160
>70
0.82
1.43
50 - 80
34 – 56
106
2.25
-
-
0.26
1.99
120
60
111
2.25
>160
>70
0.71
1.54
50 - 80
32 – 52
3.3.3
SUMMARY FOR CFC
Tests were undertaken on 19 friction rock stabiliser samples in the simulated rough boreholes with a total of 36 loadings. Observations show increased variability from the primary loadings to the second or third loadings. This could be caused by a change in the condition of the bolt, changes in the roughness of the borehole, or the fact that these types of friction stabiliser do not have a consistent response to dynamic load. This final concept is derived from the scatter of the primary load test results. Therefore only the primary loading results are considered for further analysis and discussion. The friction rock stabiliser program testing results have been summarised in three charts. The same vertical axis scale was used in the graphs to enable comparison between the split tube bolts and the Omega bolts. A circled data point indicates that the loading mass and hence the reinforcement system failed and the loading mass impacted the bottom of the test pit.
1. The dynamic force-displacement responses (Figure 3.23). In these charts, the force has been divided by effective embedment length. The effective length takes into account the reduction due to displacement. The displacement axis is arbitrarily truncated at 500mm. The cyclic nature of the response is due to the FFT filter used to process the loading mass accelerometer data.
2. Energy dissipated by the reinforcement system versus the pipe separation (Figure 3.24). The dynamic performance of reinforcement systems must consider displacement. A high energy dissipation could be associated with a large displacement which may be in excess of the deformation capacity of the support system or the limit of deformation required for excavation serviceability.
3. A summary graph of the relative velocity versus displacement of the loading mass relative to the drop beam (Figure 3.25).
99
Dynamic Test Results for Split Tube Bolts
3.3.3.1
Figure 3.23 shows the dynamic force-displacement responses for the split tube bolts. The results show a wide range of variability in the response of the reinforcement system; there is also a trend for a higher resistive response at lower input energies and a lower resistive response at higher input energy. However,
Force per meter of embedment (dynamic) (kN/m)
the most important fact is that the dynamic frictional resistance is lower than the static resistance.
140
Bolt 93-1 94-1 95-1
120 100
96-1 115-1 116-1 118-1
80 60
119-1 120-1 121-1 122-1 123-1
40 20
KEimpact 15.5 15.6 Type A 12.6 28.7 14.9 11.9 5.6 12.1 11.7 11.8 11.8 23.6
Galv
Type B Galv
Type A Non-galv
0 0
100
200
300
400
500
Displacement (mm)
Figure 3.23: Dynamic force-displacement responses for split tube bolts. Figure 3.24 shows the energy dissipated by the frictional resistance to sliding. Manufacturer A galvanised bolts (short dotted line) appear to have a different relationship to the others that were tested. The kinetic energy of the loading mass is also listed. It is possible for dissipated energy to be greater than the input kinetic energy because of the additional potential energy from displacement of the loading mass following impact. Hence a bolt that allows a large amount of displacement also has to dissipate more energy. Figure 3.25 shows the peak relative velocity for the loading mass. The velocity at impact ranged from 4.0m/s to 6.0m/s depending on the test. The main result is that a bolt that allows a high sliding velocity is likely to also allow a high displacement, as it does not provide sufficient resistance to slow the loading mass.
100
Energy Dissipated (kJ)
50 45
Bolt
40
119-1 120-1 121-1 122-1 123-1 93-1 94-1 95-1 96-1 115-1 116-1 118-1
35 30 25 20 15 10
KEimpact 15.5 15.6 Type A 12.6 Non-galv 28.7 14.9 11.9 Type A 5.6 Galv 12.1 11.7 11.8 11.8 23.6
Type B Galv
5 0 0
200
400
600
800
1000
1200
Displacement (mm)
Figure 3.24: Energy dissipated and bolt displacement from impact load. 6 Bolt 119-1 120-1 121-1 122-1 123-1 93-1 94-1 95-1 96-1 115-1 116-1 118-1
Peak Relative Velocity (m/s)
5
4
3
2
1
Impact Vel 4.84 4.86 Type A 5.89 Non-galv 5.97 5.9 5.95 Type A 4.08 Galv 6.01 5.9 5.94 5.93 5.95
Type B Galv
0 0
200
400
600
800
1000
1200
Displacement (mm)
Figure 3.25: Peak sliding velocity and displacement for split tube bolts. 3.3.3.2
Dynamic Test Results for Omega Bolts
Figure 3.26 shows the dynamic force-displacement responses for Omega bolts are more consistent than for the split tube bolt results. Bolt 110 stands out as significantly different. However, analysis indicated minor shear loading due to the centre of gravity of the loading mass not being in line with the centre line of the bolt 110. Figure 3.26 also shows that resistive force decreased as the displacement increased. Repeat load results on the Omega bolt are included in Figure 3.27. The comparison of initial and secondary tests in Figure 3.27 shows that a higher displacement occurred on a repeat test. They also show a second
101
loading on a bolt that has already displaced 200mm allows significantly more displacement for equivalent
Force per meter of embedment (dynamic) (kN/m)
energy input. This is because sliding occurs at a lower resistive load on the secondary loading.
140 120 100 80 60 40 20 0 0
50
100
150
200
250
300
350
400
450
Displacement (mm) 104-1 106-1 108-1 111-1
23.9 13.3 33.1 23.8
105-1 24.0 110-1 30.0 113-1 32.0
Program One
Program Two
Figure 3.26. Dynamic force-displacement responses for initial tests on Omega bolts.
102
500
50
35 30 25 20 15 10 5 0 0
200
400 600 800 Displacement (mm)
1000
104-1 104-2 106-1 106-2 106-3 108-1 111-1 111-2
23.9 23.9 13.3 13.2 13.3 33.1 23.8 23.9
105-1 105-2 105-3 110-1 110-2 113-1
24.0 22.9 9.9 30.0 25.6 32.0
Program Two
Dissipated Energy (kJ)
40
Program One
Bolt # KEinp
45
1200
Figure 3.27. Energy dissipated and bolt displacement from repeated loading of Omega bolts. Figure 3.28 shows the relationship between peak relative velocity and displacement for all tests on the Omega bolts. This chart supports the concept that axial sliding velocities up to 3.5m/s were stabilised by the bolt, particularly if it was an initial loading. However, sliding velocities of 5m/s or more will result in
Bolt # Velimp 104-1 5.94 104-2 5.94 106-1 5.93 106-2 5.90 106-3 5.93 108-1 5.98 111-1 5.93 111-2 5.94 105-1 5.93 105-2 5.89 105-3 3.88 110-1 6.75 110-2 6.23 113-1 6.96
5 4 3 2 1 0 0
200
400
600
800
1000
1200
Displacement (mm)
Figure 3.28. Peak sliding velocity and displacement for Omega Bolts.
103
Program Two
Peak Relative Velocity (m/s)
6
Program One
failure of the system.
3.3.4
COMPARISON OF CFC SYSTEMS
A comprehensive static and dynamic testing program has been undertaken on friction rock stabilisers. The investigations included the development of high strength rough simulated bore-holes that would provide conditions similar to those in rock. The dynamic loading methodology, instrumentation and analysis techniques promote a comprehensive understanding of acceleration, displacement, force and time for components of the test facility and sample. This allows the calculation of energy dissipation, forcedisplacement curves, acceleration and velocity time curves. Figure 3.29 enables comparisons of the axial static versus dynamic performance for the split tube bolts and the Omega bolts. The Omega bolt even when sliding has an equivalent or greater capacity than the static capacity for the split tube bolt. The Omega bolt and split tube bolt have a reduced resistance to sliding on repetitive loadings. The split tube bolt shows a greater variation in performance than the Omega bolt.
Emedment Capacity (kN/m)
180 160 140 Averaged Dynamic Splitset Performance
120
Averaged Dynamic Omega Performance
100 Static Swellex MN24 1m embedment test
80
Normalised Static Splitset Capacity
60 40 20 0 0
100
200
300
400
500
Displacement (mm) Figure 3.29: Comparison of static and dynamic performance for friction stabilisers.
The split tube bolt should be expected to have an axial static capacity of 50kN/m of embedment, with a reduction to 30-25kN/m as the bolt slides under dynamic load. This can be calculated to be an average energy dissipation of 2.7kJ (sd+/-1.0) per 100mm of slip normalised to per metre of anchor embedment. The Omega bolt (equivalent to Swellex Mn24) can be expected to have an axial static capacity of 160kN/m of embedment with a reduction to 70kN/m of embedment that will further reduce to 50kN/m as it slides under dynamic load. This can be calculated to an average energy dissipation of 6.2kJ (sd+/-1.3) per 100mm of slip normalised to per metre of anchor embedment.
104
3.4
CONTINUOUSLY MECHANICALLY COUPLED SYSTEMS
A number of systems were tested in which the elements were coupled over their entire lengths using cement or resin grout. The elements were threaded steel bar (threadbar) and 7-wire, steel strand; two types of each element were used. 3.4.1
THREADBAR ENCAPSULATED WITH CEMENT GROUT
Two, nominally 20mm diameter threadbars were tested; one that is supplied and used in Australia and the other is from Chile. As indicated previously, these are designated as high strength (HS) and low strength (LS), respectively. The 20mm diameter threadbars were fully encapsulated with cement grout. The testing also involved various collar plate, washer and nut combinations. The effects of changes in embedment lengths and surface fixtures on the performance of the threadbars in response to dynamic loading were investigated. 3.4.1.1
Sample Configuration
The threadbars were 2.4m long, and all samples were configured to have a 1.0m collar section as shown in Figure 3.30. Due to variations on installation, the exposed collar length from the pipe varied between 0.1m and 0.2m, to give toe embedment lengths of 1.2m to 1.3m.
Exposed thread +/0.15m 1.0m collar
1.3m toe length
Simulated discontinuity 49.5ID, 60OD pipe
Collar load transfer ring
Toe end loading flange
Figure 3.30: Fully encapsulated thread bar configuration. To develop an understanding of the possible influence on the threadbar results of the lower confinement offered by the thin wall steel pipe, two additional samples were prepared in very heavy steel wall pipes with equivalent rock stiffness of 85GPa. These samples had a collar length of 990mm and a toe length of 1195mm. 3.4.1.2
Cement Grout Encapsulation
The threadbar was encapsulated with a 0.45 water/cement ratio grout with 0.2% Methocel to control segregation. Low heat cement was used. Testing was carried out a minimum 28 days after grouting; one sample was tested five years after grouting. Wood (1992) showed for air-cured concrete samples that a consistent strength was obtained after 28 days to greater than 5 years.
105
Dynamic Test Results
3.4.1.3 3.4.1.3.1
HS Threadbar
Typical force-displacement responses for fully encapsulated HS threadbar are shown in Figure 3.31. The responses resulted in both steel yield and fracture of the bar at the simulated discontinuity. Table 3.3 summarises the results. The encapsulated threadbar required plastic deformation of the steel bar at the simulated discontinuity to dissipate the input energy from the dynamic load. The dynamic axial loading and partial threads of the bar allow the grout to interlock and the bar to fail under some loading conditions. The loading conditions causing rupture are related to the rate at which the energy is consumed in plastic deformation of the steel bar compared with the fracture growth between the steel bar and grout interface. At a lower loading velocity, the plastic deformation along the shaft of the bolt and the fracturing of the grout allow a free length to develop away from the simulated discontinuity towards the toe and collar ends of the element. The fracture process causes the grout to be pulverised at the outsides of the raised thread. This effectively increases the element length over which elongation can occur.
300
Dynamic Force (kN)
250 200
Threadbar 5 Threadbar 10 Threadbar 6 Threadbar 9 Threadbar 11
150 100 50 0 0
20
40
60
80
100
120
Displacement (mm)
Figure 3.31: Dynamic force-displacement responses for HS threadbar.
106
Table 3.3. Fully encapsulated HS threadbar summary results.
Bolt Number
Load Time (ms)
Displacement (mm)
Peak Deceleration (g)
Peak Force (kN)
Peak Ejection Velocity (m/s)
Energy Absorbed (kJ)
11
56
92
-12
248
2.2
14.8
Bar stretched
9
26
62
-12
256
3.0
10.9
Bar fractured at simulated discontinuity
6
28
69
-12
260
3.1
13.9
Bar fractured at simulated discontinuity
10
56
100
-13
270
3.2
20.8
Bar stretched, no surface hardware
17.5
Bar stretched and pulled in grout, 5 year old grout
5
100
91
-13
235
2.4
Results
The differences in response to loading of the fully encapsulated threadbar are clearly shown in Figure 3.32. In the first instance, rupture of the bar occurred. In the other case, plastic deformation of the bar is evident from the cracking of the zinc coating. Figure 3.33 shows the condition of the cement grout retained within the threads of the bar that has effectively become decoupled from the surrounding grout annulus. The bolt can then withstand loads that would have caused it to fail when it was fully coupled.
107
Figure 3.32: Fully encapsulated HS threadbar – critical and non-critical loading conditions.
Figure 3.33: Threadbar that has slid out of the grout at the third dynamic axial load. 3.4.1.3.2
LS Threadbar
The two stage program consisted of six fully encapsulated ungalvanised LS threadbar installed in thin wall steel pipes. The results are shown in Figure 3.34. The LS threadbar reinforcement systems have a very consistent response. The results are summarised in Table 3.4. The force-displacement responses vary less than those for the Australian threadbar due to the consistent plastic deformation of the mild steel threadbar. It was found that tests in which bolts exhibit a plastic behaviour are easier to analyse than tests in which the HS threadbar fails.
108
250
Dynamic Force (kN)
200
Threadbar 137 (mild steel)
150
Threadbar 138 (mild steel)
100
Threadbar 136 (mild stee) FOUR Buffer Threadbar135 (mild steel) FOUR buffer
50
0 0
50
100
150
200
Displacement (mm)
Figure 3.34: Dynamic force displacement response for LS threadbar.
Table 3.4: Summary of results for fully encapsulated LS threadbar.
Bolt Number
Load Time (ms)
Displacement (mm)
Peak Deceleration (g)
Peak Force (kN)
Peak Ejection Velocity (m/s)
Energy Dissipated (kJ)
135
77
151
- 9.7
207
3.5
27
Bar stretched
136
80
145
-10.1
215
3.4
30
Bar stretched
137
82.5
158
- 9.7
206
3.5
28.2
Bar stretched
138
77.5
163
-10
212
3.3
29.3
Bar stretched
146
46.5
120
- 9.7
208
3.0
21.8
Bar fractured at simulated discontinuity
147
89.6
224
- 8.7
189
3.5
38.4
Bar stretched
109
Results
3.4.1.3.3
Influence of confinement
Plastic deformations of the thin wall steel pipes used for the LS threadbar samples were small but measurable with an average increase in diameter of 0.2mm across all samples as indicated in Figure 3.35. No deformation of the steel pipe was recorded for testing on galvanised HS threadbar. Threadbar 138 Pipe Diameters Nominal 27.4GPa rockmass confinement 51.00
Average diameter pre 50.57mm post 50.73mm
Pipe Diameter (mm)
50.90 50.80 50.70 50.60 50.50 50.40
Pre Post
50.30 0
500
1000
1500
2000
2500
3000
Distance from the collar (mm)
Figure 3.35: Thin wall pipe diameter pre-test and post-test. To assist in understanding the possible influence of confinement from the thin wall steel pipe on the threadbar results, two additional samples were prepared in thick wall pipes. The thick wall samples had a collar length of 990mm and a toe length of 1195mm. One bolt snapped as expected but the other bolt stretched and survived as indicated in Figure 3.36. The stable response appears to have been caused by a lower strength grout that allowed for a significantly longer decoupled length to be formed.
110
Bolt 146 = 260mm to collar and 470mm to toe = 530mm of decoupling 260mm Collar
simulated discontinuity
Toe
470mm Bolt 147 = 280mm to collar and 710mm to toe = 990mm of decoupling
Toe
710mm
280mm Collar
simulated discontinuity
Figure 3.36: Dissection of simulated boreholes 146 and 147. Figure 3.37 shows the thin wall pipe to the left with the non-centralised bolt samples that did not snap and the thick wall pipe to the right with centralised threadbar that did break. Failure of the LS threadbar appears similar to the fully encapsulated galvanised HS threadbar failure.
111
3mm wall, 27GPa Eq
10mm wall, 84.4GPa Eq
Figure 3.37: Comparison of confinement on mild steel threadbar. Dissection of four samples showed the principal location for grout fracturing occurred between the shear pin location and the simulated discontinuity as shown in Figure 3.38. This has been interpreted as resulting from a reduction in confinement (elastic and plastic deformation of the steel pipe) that allowed additional stretch of the threadbar and minor sliding of the grout inside the pipe between the simulated discontinuity and the shear pin location. This softened the response of the reinforcement to dynamic load. Hence the midpoint of the results between sample 146 and the average of samples 135-138 may be a more appropriate or conservative indicator of the mild steel threadbar reinforcement system performance.
112
simulated discontinuity
A Shear pin location
Collar
B
A
Toe
simulated discontinuity
B Shear pin location
Figure 3.38: Dissection of simulated borehole 135. 3.4.1.3.4
Comparison of responses
The dynamic force-displacement responses for the LS threadbar in Figure 3.34 can be compared with the galvanised HS threadbar results shown in Figure 3.31 (with a higher equivalent radial stiffness of 49GPa versus 27GPa). The LS threadbar has a higher displacement capability at a lower resistive load. The LS threadbar when heavily confined (84.4GPa) maintained the same dynamic resistance of approximately 200kN, but the magnitude of displacement was variable. This is interpreted to be caused by the grout strength with failure in higher strength grout at 120mm of displacement and in lower strength grout failure at a larger displacement of 190mm.
113
The dynamic force-displacement responses for fully encapsulated threadbars are compared in Figure 3.39. The difference in performance is clearly evident and can be related directly to the static mechanical properties of the element given in Figure 3.10.
Threadbar 137 (mild steel LC)
Threadbar 138 (mild steel LC)
Threadbar 136 (ms LC) FOUR Buffer Threadbar 146 (ms HC)
Threadbar135 (ms LC) FOUR buffer Threadbar 147 (ms HC)
ThreadBar 5-galv Threadbar 6-galv
Threadbar 10-galv Threadbar 9-galv
Threadbar 11-galv
300
Dynamic Force (kN)
250
200
150
100
50
0 0 Rupture
50
100
150
200
250
Displacement (mm)
Figure 3.39: Dynamic force displacement responses for cement encapsulated threadbars. * LC – low confinement, HC – high confinement, ms – mild steel, galv – galvanised, FOUR buffer – impact on four buffers rather then the standard of two buffers.
114
3.4.1.3.5
Influence of the number of buffers
Sample numbers 137 and 138 were tested with impact of the drop beam onto the standard configuration of one buffer on either side of the pit. Samples 135 and 136 had impact onto two buffers on each side of the pit. This gave a stiffer response as there was a quicker deceleration of the drop beam. It can be seen in the tests that increasing the number of buffers (Figure 3.39) from two to four and increasing the rate of deceleration of the drop beam (peak deceleration increased from 31.7g to 50.1g) and reducing buffer compression (87.8mm to 78.3mm) has not influenced the resultant force-displacement responses for the mild steel threadbar with low to moderate confinement. The total energy dissipated by the buffers increases slightly from 24.1kJ to 25.6kJ with the change from two to four buffers. 3.4.2
THREADBAR ENCAPSULATED WITH RESIN GROUT
A program was completed for testing of three galvanised HS threadbar encapsulated with resin grout in simulated rough boreholes. The simulated rough borehole assemblies were taken to a mine site and the bolts were installed using a drilling jumbo. Minova medium set resin was used. There was no leakage from the toe of the pipe assembly and resin was observed at the collar of all three holes as shown in Figure 3.40.
#150
#149
#148
Figure 3.40: Resin reporting to collar of simulated borehole. The test program used progressively reducing impact energies of 35.3kJ, 24.5kJ, and 15.7kJ for samples 148, 149 and 150, respectively. This was achieved by reducing the impact velocity. The third impact energy was found not to be sufficient to break the reinforcement system on the first loading. The results are shown in Figure 3.41 and summarised in Table 3.5. Both loadings on sample 150 are reported. The short loading time to failure or completion of the load transfer does make analysis difficult, as there are few frames from the digital video camera for the analysis of displacement. Reinforcement system performance was characterised by higher peak deceleration of the mass and the estimated plastic deformation length at the simulated discontinuity from connecting the failed halves together averaged 27mm.
115
300
Dynamic Force (kN)
250 200
Threadbar 148-1 Threadbar 149-1 Threadbar 150-1 Threadbar 150-2
150 100 50 0 0
20
40
Rupture
60
80
100
120
Displacement (mm)
Figure 3.41: Dynamic force displacement response for resin encapsulated threadbar.
Dissipated (kJ)
Velocity (m/s)
Peak Ejection
149
24.5
28
41
-13.2
261
1.7
7.3
Bar snapped
150-1
15.7
33
24
-12.0
240
1.7
3.8
Bar stretched
150-2
23.9
23
30
-14.1
299
1.2
6.5
Bar snapped
Results Bar snapped
Energy 5.7
(kN)
2.5
Peak Force 298
(g)
-15.2
Deceleration
29
Peak
12
(mm)
35.3
(ms)
Input Energy
148
(kJ) Load Time
Bolt Number
Displacement
Table 3.5: Summary of results for resin encapsulated high strength steel threadbar.
The interaction between the resin and threadbar provided very effective load transfer and resulted in small displacements at the simulated discontinuity prior to failure as shown in Figure 3.42. This is interpreted to be yield of the steel bar in preference to the development of a short decoupled length.
116
Anchor side of the simulated discontinuity.
Collar side of the simulated discontinuity
Stretch of the steel bar up to failure – 4mm wide cut in pipe prior to test, therefore maximum 27mm of plastic deformation.
Figure 3.42: Bolt 148 failure surface. The results for tests on threadbar with resin and cement encapsulation are shown in Figure 3.43. The responses of the fully encapsulated resin threadbar are more consistent when compared with those for the HS threadbar in cement grout. The higher strength resin prevents development of a decoupled length as observed for cementitious grout encapsulation. This means that it is important to specify the encapsulation material for reinforcement systems with threadbar elements.
117
300
Threadbar resin 148-1
Dynamic Force (kN)
250
Threadbar resin 149-1
200
Threadbar resin 150-2
150
HS Galv Threadbar Grout 5
100
HS Galv Threadbar Grout 10
50
HS Galv Threadbar Grout 6 HS Galv Threadbar Grout 9
0 0
20
40
Rupture
60
80
100
120
Displacement (mm)
Figure 3.43:Dynamic force-displacement responses for cement and resin encapsulated threadbar. 3.4.3
PLAIN STRAND
Two types of strand were used in the test program: Australian and Chilean strands. The properties are slightly different as shown in Table 3.6 and Table 3.7, respectively. This is consistent with various reinforcement products for which codes specify minimum mechanical properties rather than physical properties or grades of steel. Table 3.6:Selected properties for Australian 7-wire steel strand. Parameter
Minimum
Typical
N/A
15.2
Cross Sectional Area (mm )
N/A
143
Yield Force 0.2% (kN)
212
235
Tensile Force (kN)
250
265
Elongation on 600mm length (%)
3.5
6.5
Core Diameter (mm) 2
118
Table 3.7: Selected properties for Chilean 7-wire steel strand. Parameter
Typical
Yield Force (tonnes)
23.9
Tensile Force (tonnes)
26.5
Elongation (%)
6.5 2
3.4.3.1
Cross Sectional Area (mm )
140
Linear Weight (kg/m)
1.13
Steel Grade
A416 - 270
Sample Preparation
Twelve samples were prepared with plain strand encapsulated in cement grout within steel pipes. Each simulated bore hole was 2.58m in length with a discontinuity located either 0.62m or 1.0m from the collar. 3.4.3.2
Surface Hardware
Surface fixtures examined were 200mm square, 8mm thick plate with barrel and two- or three-part wedge anchors. In some cases, a rubber pad was placed underneath the steel plate or no surface hardware was used. 3.4.3.3
Testing Results for Australian Strand
The dynamic force-displacement responses for five tests are shown in Figure 3.44. The responses included: •
ability to survive the loading.
•
fracture of the strand at the simulated discontinuity.
•
yield and slip of the strand from within the toe embedment length.
•
sliding of the strand (in the absence of surface hardware) from within the collar pipe.
Table 3.8 summarises the results in the order of testing. These results were the first indication that reporting on the performance of the reinforcement system only in terms of energy dissipated does not provide enough detail and that other indices are required. Another sample similar to bolt number 23 was tested with an impact velocity of 8m/s. No details were obtained due to an instrumentation malfunction. However, the strand slid completely out from the toe embedment. This was a result of doubling the input energy and, possibly, grout embrittlement through aging. The additional input energy meant that a significant amount of momentum had not been dissipated by the time the decoupling front had reached the toe end of the pipe.
119
350 300
Plain Strand 19 Plain Strand 22 Plain Strand 23 Plain Strand 26 Plain Strand 27
1.96m toe
200
1.5m toe 150
1.96m toe
Strand Rupture
100 50
Unplated cables high deformation to >650mm 0 0
50
100
150
200
250
300
Displacement (mm)
Figure 3.44: Dynamic force-displacement response for Australian plain strand.
Results
(kJ)
Energy Dissipated
Velocity (m/s)
Peak Ejection
Peak Force (kN)
Deceleration (g)
Peak
(mm)
Displacement
Load Time (ms)
Table 3.8: Summary of Australian Plain Strand Test Results
Bolt Number
Force (kN)
250
23
73
103
-11
228
2.2
18
Stable, sliding as less then 2.0m of toe embedment
27
48
85
-14
302
2.8
18
Rupture of cable at simulated discontinuity.
22
38
91
-12
254
3.8
18
Rupture of cable at simulated discontinuity.
19
130
644
-3
78
5.5
34
Unstable, mass sliding off the strand as no surface hardware to restrain displacement.
26
288
650
-6
126
4.2
40
Unstable, mass sliding along the strand as no surface hardware
120
Typical photos from the first tests are shown in Figure 3.45. In the left hand photo, rupture occurred at the simulated discontinuity with 1.5m of toe embedment and surface hardware to control the displacement of the loading mass; the right hand photo shows some sliding of the strand from toe end grout annulus.
Rupture
Sliding
Figure 3.45: Ruptured strand or sliding strand dependent on the embedment lengths. 3.4.3.4
Surface hardware
The role of the surface hardware (plate and barrel and wedge anchor) is to restrain the loading mass. In all tests, it is possible for decoupling to progress from the simulated discontinuity towards the collar. Without the surface hardware, the decoupling occurs for the entire embedment length and may result in large displacements prior to the loading mass being arrested or complete pull out from the grout annulus. It is worth noting that the additional displacement results in more energy being required to be absorbed. Figure 3.46 shows sample 19 set up for testing without surface hardware; it is clearly stable statically but, as shown in Figure 3.47, becomes unstable when subjected to dynamic loading. The large displacement makes analysis and interpretation difficult for although a large amount of energy is dissipated, the system still fails due to the lack of surface hardware. The high dissipated energy value is due to the significant additional potential energy change from the larger displacement of the loading mass.
121
Sample 19
Sample 19
Figure 3.46: Plain strand with no surface hardware prior to testing.
Sample 19 Sample 19
Sample 26, 650mm of sliding
Figure 3.47: Outcome from strand test with no surface hardware.
122
3.4.4 3.4.4.1
CHILEAN PLAIN STRAND Sample Preparation
Six plain strands were encapsulated by cement grout within thin wall steel pipes and transported from Chile for testing. The water/cement ratio was not specified. The embedment lengths (2.0m toe and 1.0m collar) and surface hardware were the same for all specimens. 3.4.4.2
Dynamic Testing Results
The dynamic force-displacement responses for the Chilean plain strand samples are shown in Figure 3.48.
Dynamic Force ( kN)
400 350
Chilean Plain Strand Two Buffer #128
300
Chilean Plain Strand Two Buffer #129
250
Chilean Plain Strand Four buffer #130
200
Chilean Plain Strand Four buffer #133
150
Damaged Plain Strand two minor nicks # 132
100
Damaged Plain Strand sliding # 131
50
Ruptured Strand 0 0
20
40
60
80
100
120
Deformation (mm)
Figure 3.48: force displacement response for Chilean plain strand. The summary of the results from the dynamic testing program on the Chilean plain strand cables is given in Table 3.9. The results are listed in the order of testing and may be compared with results for Australian plain strand given in Table 3.8.
123
Velocity (m/s) Energy Dissipated
Peak Force (kN)
15.9
327
2.5
12.6
Snapped cable at discontinuity – two buffer
129
33
92
15.7
326
3.7
17.8
Snapped cable at discontinuity – two buffer
130
28
84
17.6
363
3.4
18.4
Snapped cable at discontinuity – four buffer
133
17.7
73
15.5
323
3.1
12.7
Snapped cable at discontinuity – four buffer
132
16
37
14.7
303
2.5
6.1
Minor damaged strand, snapped at discontinuity
131
17.8
48
6.3
142
3.8
4.1
Moderate damaged strand, internal sliding of the strand wires.
(kJ)
Result
67
Peak Ejection
29.9
(mm) Peak
Load Time (ms)
128
Displacement
Bolt Number
Deceleration (g)
Table 3.9: Summary of Chilean plain strand test results.
All strands snapped at the simulated discontinuity similarly to those shown in Figure 3.49. However, it should be noted that two of the strands (sample 131 and 132) were damaged during cutting of the thin wall pipe to create the simulated discontinuity. This occurred because the strand was located adjacent to the wall of the pipe. Sample 132 had a slight notch on two of the seven wires in the strand. This resulted in premature and non-uniform rupture in comparison with undamaged strand (e.g. sample 133 shown in Figure 3.50). Sample 132 had the shallower cuts of the two damaged samples. Its force-displacement response was consistent with the undamaged plain strand cable but all wires ruptured at a smaller displacement and with less energy dissipation.
124
Figure 3.49: Snapped non-centralised strand from sample 128.
Sample 132
Sample 133
Figure 3.50: Damaged strand from cutting the thin wall pipe compared to no damage. The three wires on sample 131 were cut deeper than sample 132 and only these wires ruptured. The other four strands slipped through the grout and barrel and wedge anchor; consequently the force-displacement response for this particular test is difficult to interpret.
125
Figure 3.51: Damage to wires from cutting the simulated discontinuity, sample 131. 3.4.4.3
Influence of number of buffers
The undamaged strand samples had two tests conducted at the standard nominal 36kJ input energy with the drop beam being decelerated by two buffers or four buffers to assess any variation in results caused by the rate of deceleration of the drop beam. There is a greater variation in the results when comparing the results of these tests with the results from the threadbar tests. However, the variance is considered to be a feature of the mechanism of behaviour rather than an influence of the buffers because the variance is less than that documented for the testing of Australian plain strand cable which all impacted onto two buffers. The force-displacement responses show that increasing the number of buffers (Figure 3.48) from two to four increases the rate of deceleration of the drop beam (peak deceleration from 38.9g to 48.5g) and reduces buffer compression (78.9mm to 50.4mm). The total energy dissipated by the buffers reduced from 24.9kJ to 19.7kJ. 3.4.4.4
Influence of confinement
Figure 3.52 shows that plastic deformation of the thin wall steel pipe (equivalent rock stiffness of 23.9GPa) was small and perhaps at the limit of accuracy of measurement with an average increase of 0.14mm across all samples. The elliptical shape of the pipe is shown by the 0.5mm variance between the minimum and maximum measured diameters. No deformation was discernable for the tests on Australia strand in the thicker wall pipe with an equivalent rock stiffness of 43GPa.
126
Bolt 130 Pipe Diameters Nominal 24.8GPa rockmass confinement Average diameter pre 62.23mm post 62.33mm
63.50
Pipe Diameter (mm)
63.00
62.50
62.00
61.50 Pre Post
61.00 0
500
1000
1500
2000
2500
3000
Distance from the collar (mm)
Figure 3.52: Thin wall pipe diameter pre-test and post-test. Samples 128-131 were dissected to enable examination of the grout. Figure 3.53 shows the grout was fracturing between the shear pins and the simulated discontinuity. The shear pins have a significant role in influencing how the samples behave by stopping the grout from slipping inside the steel pipe as shown in Figure 3.54. This may be interpreted as a reduction in radial confinement that has allowed additional stretch of the strand and sliding of the grout inside the pipe between the discontinuity and shear pin location with a resultant softening in the response of the reinforcement system. In future, it will be appropriate to use additional shear pins located closer to the simulated discontinuity to minimise the effects of the low radial confinement provided by thin wall pipe.
127
Collar length - damage section is where the shear pin was located. Simulated discontinuity
collar
Simulated discontinuity Photos of toe embedment length no damage observed from 1500 to 2510mm tensile cracks at the very toe possibly due to 'rebound' effect.
Figure 3.53: Dissection of simulated borehole 128.
Figure 3.54: Detail of grout damage about shear pins, sample 128. 3.4.4.5
Comparison between Plain Strand Cable Bolts
A comprehensive double embedment test program has been taken on 15.2mm plain strand cable from both Australia and Chile. The cables were all encapsulated in cementitious grout. The test work identified that plain strand cable bolt responses to dynamic loading conditions were highly dependent on the collar and toe embedment lengths and the use of surface hardware.
128
Figure 3.55 enables a direct comparison between Australian plain strand and the Chilean plain strand with equivalent rock stiffness confinement of 43.0GPa and 23.9GPa, respectively. The results indicate a similar early response but the Chilean strand has a slightly higher capacity; this could be attributed to the slightly larger diameter and longer strand embedment length. It is therefore concluded that the difference between the Australian and Chilean strands is minimal, especially when compared with all the other systems.
Dynamic Force (kN)
400 350
Chilean Plain Strand Two Buffer#128
300
Chilean Plain Strand Two Buffer #129
250
Chilean Plain Strand Four buffer #130
200
Chilean Plain Strand Four buffer #133
150
Australian 15.2mm strand 2.0m toe #22 Australian 15.2mm strand 2.0m toe #19 Australian 15.2mm strand 1.5m toe #23 stable Ruptured Strand
100 50 0 0
20
40
60
80
100
120
Deformation (mm) Figure 3.55: Dynamic force displacement responses for Australian and Chilean plain strand. The dynamic axial loading conditions of the facility are favourable to sliding of the strand through the grout. It would be difficult to design a short embedment to act as a sliding mechanism to strong ground motion due to the shear component that would also exist from both the loading wave and the displacement of the rock mass into the excavation. The dynamic work confirmed static test work by Hutchins et al. (1990) that showed a plain strand cable does not generate sufficient load to break the cable with less than 2.0m of embedment. Testing also showed that surface hardware is required to maximise or maintain the capacity of the reinforcement system. The bulb strand centrally decoupled in the tested configuration was stable to the input loads. However, in the case of using a decoupled strand to the collar of the hole there is the potential for 41.9mm barrels to expand under the load allowing the wedges to bite deeper into the strand and rupture the strand. A summary table for fully encapsulated strand with critical embedment lengths are detailed in Table 3.10.
129
The values in Table 3.10 can be compared with the previously published test results on the dynamic performance of comparable plain strand cable.
Australian plain strand cable fully encapsulated in cementitious grout
88 +/-3
-13 +/-1
Chilean plain strand cable fully encapsulated in cementitious grout
79 +/-12
-16 +/-1
Australian bulb strand, central decoupling
106 +/-6
-14 +/-1
(kJ) Results
Energy Dissipated
Velocity (m/s)
Peak Ejection
Peak Force (kN)
(g)
Peak Deceleration
(mm)
Displacement
Type
Table 3.10: Summary of performance indices strand cable.
258
3.4 +/-0.4
18.0
1.96m of toe embedment, 0.62m of collar embedment and surface hardware rupture at simulated discontinuity.
334
3.2 +/-0.5
15.5 +/-2
2.0m of toe embedment, 1.0m of collar embedment and surface hardware, rupture at simulated discontinuity.
25.1
1.0m toe embedment of 3 bulbs, 1.7m decoupled, 0.8m collar embedment 2 bulbs (stable)
305
130
2.8
3.5
DISCRETE
MECHANICALLY
AND
FRICTIONALLY
COUPLED
SYSTEMS The more a reinforcement system moves during a dynamic loading event, the greater the energy a reinforcement system must be capable of absorbing due to the additional potential energy input from the ‘rock’ moving into the drive. Yielding systems that allow large displacements due to their softness may also result in excessive fracturing / bulking of the rock mass and have adverse loading of the support system. A soft response from a reinforcing system will also allow higher ejection velocities of the rock into the drive. 3.5.1
PARTIALLY DECOUPLED HS THREADBAR
The elongation capacity of the overall reinforcement system may be increased by creating a free length between the toe and collar regions of the bar; either by having a short encapsulation length at the toe end or by decoupling the bar from the encapsulation medium. The latter method is preferred as it provides better resistance to shear movement across the axis of the borehole. Two embedment configurations for the 20mm diameter threadbar were examined; decoupled in cement grout and jumbo installed with a resin toe anchor. The testing also involved various collar plate, washer and nut combinations. 3.5.1.1
Sample Preparation
The decoupling provided a significant free length that was available to stretch and dissipate energy under dynamic load when compared with the fully encapsulated threadbar. The partially decoupled threadbars were non-galvanised, nominally 3.0m in length with the PVC tube clamped onto the central 1.6m of the threadbar as shown Figure 3.56. Cement Grout Encapsulation The grout was prepared with 0.40 - 0.45 water / cement ratio and tested between 22 and 56 days. The unconfined compressive strength determined from tests on cylinders was >40MPa in all cases. The simulated discontinuity was located at a standard 1.0m from the collar. Due to variations in installation the configurations tested are shown in Figure 3.56 for sample numbers 83 to 86.
131
1.0m toe anchor
1.6m central decoupled 3.0m
Figure 3.56: Decoupled threadbar configuration. 3.5.1.2
Surface Hardware
Two surface hardware configurations shown in Figure 3.57 were used in the test program. The first comprised a separate 32mm long ‘mine nut’ (T20 x 10.0 pitch LH thread) and spherical based washer. The second was an integrated nut and spherical based washer with a continuous thread 45mm in length. Both types of test used a 150mm square, 8mm thick dome plate,. The system was tensioned using a torque wrench to rotate the nut.
Mine nut
Dome washer
Dome plate (150*150*8)
Integrated Nut with washer.
Figure 3.57: Surface hardware used on decoupled threadbar.
132
3.5.1.3
Dynamic Test Results
The decoupled threadbar required plastic deformation of the steel in the decoupled length to absorb the input energy. To achieve this, the collar mass needed to transfer the load through the surface hardware and the side of the simulated bore hole onto the short length of encapsulated threadbar in the collar section. The dynamic force-displacement responses at the simulated discontinuity are shown in Figure 3.58. Figure 3.31 and Figure 3.58 both show forces greater than the expected average static yield force of 165kN. Malvar and Crawford (1998) have shown that, for strain rates of approximately one strain per second, there is a dynamic increase factor of approximately 1.3 in yield and ultimate strength capacities for reinforcing bar of nominal 550MPa yield stress. For 20mm threadbar this increases the average yield load from 165kN to 213kN; a value that is in agreement with the dynamic yield load assessed in the facility.
300
Dynamic Force (kN)
250 200
Debonded # 83 Debonded # 84 Debonded # 85 Debonded # 86
150 100
Elastic recovery of the bolt – due to the debonded length
50
Nut Strips
0 0
20
40
60
80
100
120
Displacement (mm) Figure 3.58: Dynamic force-displacement response for decoupled threadbar. Table 3.11 summarises the test data for the decoupled threadbar. The critical functionality for a decoupled threadbar was the correct selection of the surface fixture. Testing with a short nut resulted in the internal threads being partially sheared (as shown in Figure 3.59) with the nut riding over the bar threads once 180kN was reached (as measured by a load cell at the collar). However, when the longer integrated nut and washer was used, this force increased to 200kN. The failure mechanism changed to either survival of the surface hardware or partial shearing of the threads along the shaft of the bolt as shown in Figure 3.60. The second most important criterion appeared to be the collar length encapsulated by cement grout. Short lengths allow faster transfer of the load to the nut promoting failure before sufficient energy had been dissipated by yielding of the central decoupled length.
133
Table 3.11: Decoupled threadbar summary results.
Bolt Number
Load Time (ms)
Displacement (mm)
Peak Deceleration (g)
Peak Force (kN)
Peak Ejection Velocity (m/s)
Energy Absorbed (kJ)
83
55
93
-11
240
2.5
18
Nut stripped
84
35
82
-10
217
2.6
13.6
Nut stripped
85
58
101
-12
244
2.6
21.8
Reinforcement system survived two drops
21.6
Reinforcement system partial thread shearing on second drop.
86
62
106
-11
226
2.6
Figure 3.59: Shear of threads inside short length ‘mine nut’.
134
Results
Figure 3.60: Shearing of the thread on the bar with integrated nut and spherical based washer. 3.5.2 3.5.2.1
TOE ANCHORED HS THREADBAR Sample Preparation
Resin encapsulated galvanised threadbars of 2.4m length were anchored at the toe into rough simulated boreholes. The simulated boreholes were taken underground and a drilling jumbo was used for installation at a mine site as shown in Figure 3.61. All used a simulated discontinuity at 1.0m from the collar. Effective resin mix was achieved from the mechanism attached to the threadbar. The simulated borehole suffered from resin bleed through the end of the sample or from the bolt having been pushed too far into the hole and breaking out the end plate on the simulated borehole. This reduced the length of anchor encapsulation and increased the length of decoupled bar at the collar of the hole as detailed in the sample description. It allowed the assessment of variable toe anchor lengths to determine if there was a minimum encapsulated length that would change how the bolt performed under dynamic load.
Figure 3.61: Jumbo installation of toe anchored resin bolt.
135
3.5.2.2
Surface Hardware
Surface hardware for the resin encapsulated bolts consisted of an 8mm thick domed plate, spherical based steel washer, flat low friction washer and nut as shown in Figure 3.62. The nut was tensioned with a torque wrench when the sample was set up in the test facility. A load cell was used between the nut and washer when space allowed.
dome
mine nut
ball washer
Friction reducing washer
Figure 3.62: Surface hardware configuration. 3.5.2.3
Dynamic Test Results
The toe anchored bolts had no encapsulation medium within the collar section of the reinforcement system, i.e. collar side of the simulated discontinuity. This meant that in a dynamic test the loading mass acted directly onto the domed plate and then through the nut on to the thread and then the shaft of the bolt. All bolts failed by early stripping of the nut and thread off the bolt. Table 3.12 summarises the results of the test program. Bolt 124 (not shown) also had the same result of nut and thread stripping. No difference in performance was noted for either the jumbo tensioned or hand tensioned bolts. The short load duration leading up to failure of the bolt made analysis quite difficult. The capacity of the nut in response to dynamic loading is a major deficiency of the system. A reinforcement system that has no load distribution capability from the rock to the reinforcing element between the end of the encapsulation medium and the surface hardware must have a fixture at the collar that can transfer load for a considerable period. The ‘mine nut’ test does not achieve this requirement. The results show that very low energy absorption will result when a full encapsulation is not achieved for this reinforcement system.
136
Results
Absorbed (kJ)
Energy
Velocity (m/s)
Peak Ejection
(kN)
(g) Peak Force
Deceleration
Peak
(mm)
Displacement
(ms)
Load Time
(kJ)
Input Energy
Bolt #
Table 3.12: Summary of resin encapsulated toe anchored bolts.
125
12.6
7
15
-26
196
2.2
0.9
Nut and thread stripped
126
36
6
12
-8
159
1.7
1.1
Nut and thread stripped
127
21.7
6
8
-8
155
1.5
0.8
Nut and thread stripped
3.5.3 3.5.3.1
CONE BOLT Sample Preparations
The cone bolts were encapsulated in two different cement grouts; the first used a 0.40 water-cement ratio with strengths greater then 40MPa (from grout cylinder testing) and aged for two months to five years (sample numbers 31 – 42); the second was a cement and limestone dust mix that achieved 25MPa at 28days (sample numbers 60 and 61). 3.5.3.2
Surface Hardware
The reinforcement systems were trialled with the simulated discontinuity at 1.0m from the collar. and a variety of surface hardware that included 150mm square, 6mm thick dome plates used alone or as a pair and in some cases with a rubber backing plate. 3.5.3.3
Dynamic Test Results
The program consisted of 14 fully encapsulated 22mm cone bolts installed in simulated standard boreholes. Details of the bolt configurations and testing results were detailed in the M349 final report; Figure 3.63 and Table 3.13 summarise the results.
137
Dynamic Force (kN)
250 Cone bolt 60 (LS Grout)
200
Cone Bolt 61 (LS Grout) 150 Cone Bolt 32 (HS grout) 100
Cone Bolt 41(HS grout)
50
Cone Bolt 35 (Old HS Grout) (High Impact velocity)
0 0
100
200
300
400
Deformation (mm)
Figure 3.63: Dynamic force-displacement responses for cone bolts.
-7.5
167
3.5
35.0
100%
32
>40MPa grout
73
120
-9.5
209
3.1
19.0
80%
41
>40MPa grout
69
120
-10
217
2.9
20.5
30%
102
312
-10.5
213
5.0
55.6
98%
>40MPa grout, 35
8m/s impact velocity, 5yr old grout
138
in separation
273
cone movement
120
Proportion of
25MPa grout
Absorbed (kJ)
61
Energy
100%
Velocity (m/s)
32.8
Peak Relative
3.8
(kN) 157
Peak Force
-7
(g)
288
Deceleration
137
(mm) Peak
25MPa grout
Displacement
Load Time (ms)
60
Bolt Number
Specification
Table 3.13: Fully encapsulated 22mm cone bolt summary results.
3.5.3.4
Cone bolt performance in high strength grout
The mechanism of energy dissipation by the cone bolt in high strength cement grout (>40MPa) was a combination of cone movement and plastic deformation of the shaft of the bolt. The plastic deformation of the steel shaft of the bolt occurs at an elevated yield stress; this is consistent with the work reported by Malvar and Crawford (1998). The relative split between cone movement and plastic deformation of the shaft appears to have at least a secondary dependence on the collar tension applied to the bolt. For low or no collar tensions, particularly on repeat loading where decoupling of the shaft from the grout resulted in the loading mass not being in intimate contact with the surface hardware prior to impact, there was consistently minimal to no cone movement and primarily plastic deformation of the shaft of the bolt. The low displacement of the cone was accompanied by heavy deformation of the surface hardware. The cone bolt in the high strength grout with the largest displacement relative to the deformation of the shaft was the five year old sample 35 with the 8m/s impact velocity. An explanation for this mechanism of cone movement or shaft deformation has not been fully investigated. An example of the pulverisation of the grout about the cone from radial dissection of sample 32 is shown in Figure 3.64. The decoupling of the bolt shaft from the grout through plastic deformation is shown in Figure 3.65.
139
Pulverised grout
Slice where the cone has pulled through grout End of the cone
Section that cone bas pulled through
Figure 3.64: Grout pulverisation from cone displacement.
Figure 3.65: Decoupling of the bar from the grout.
140
3.5.3.5
Cone bolt performance in low strength grout
The mechanism of energy dissipation in the low strength (25MPa) cement and limestone dust grout was from the displacement of the cone pulverising the grout; there was no plastic deformation of the shaft of the bolt to dissipate the input energy. This occurred at a lower resistive force than that of the cone bolt in high strength grout as indicated in Figure 3.63. It also permitted repeat impacts to have similar loading conditions as the first impact on to the surface restraint as shown in Figure 3.66. This was because no decoupling developed between the shaft and the grout in the collar pipe as shown in Figure 3.67.
Bolt 60 test 1
Bolt 60 test 3
Bolt 60 test 2
Figure 3.66: Surface fixture sample 60, three impacts.
Bolt 60 – collar side second impact
Bolt 60 – collar side first impact
Figure 3.67: Grout connection collar side of discontinuity.
141
Three impacts were performed on each of samples 60 and 61; the results indicated a slightly lower resistive force for each test; suggesting there was a frictional component of resistance on the shaft in addition to the cone resistance. The test program showed that the ability of the cone to penetrate through the 25MPa grout was independent of the collar tension. Confinement and radial expansion of the pipe
3.5.3.6
The tapered faces of the cone generate significant radial forces on the wall of the pipe as the cone moves through the grout. The forces are greater in the stronger grout because there is more resistance to fracturing. The high forces are characterised by deformation of the 9.7mm thick steel wall pipe as shown in Figure 3.68. The average pipe diameters were measured at intervals along the simulated borehole after the tests had been performed. Figure 3.68 raises interesting questions regarding the expected field performance of the cone bolt. For example; would the radial forces be sufficient to fracture or deform the rock mass about the cone? Or, would a greater displacement of the cone result because a rock mass is easier to deform than the thick wall steel pipe?
Post Test Average Pipe Diameter (mm)
65.2 65.0 64.8 64.6 Bolt 61 (LS)
64.4
Bolt 60 (LS)
64.2
Bolt 42 (HS)
64.0
Bolt 38 (HS) Bolt 35 (HS)
63.8
Bolt 41 (HS)
63.6 63.4 63.2 0.0
0.3
0.6
0.9
1.2
1.5
1.8
2.1
2.4
Distance from Collar (m) Figure 3.68: Thick wall pipe deformation post cone bolt test.
142
3.5.3.7
Surface hardware
The most important factor influencing the performance of the surface hardware was whether the cone could move through the grout or not on the first impact. The pre- and post-test photos of the surface hardware are given in Figure 3.69 to Figure 3.71. For subsequent impacts on a sample, the loading mass could initially lag behind the surface hardware and then cause a heavy shock load onto the surface hardware that would not represent reality. The effect of the rubber plate was to soften the surface fixture and dampen the direct deceleration to which the nut was exposed; it also appears to have increased survivability of the complete reinforcement system, as shown in Figure 3.69.
Sample 32 (pre) 125kN collar tension.
Sample 32 (post) slight doming of outer plate and compression of rubber plate causing it to dome, 84mm of cone movement and 18mm stretch.
Figure 3.69: Rubber plate plus two dome plates. The effect of the two dome plates shown in Figure 3.70 was to stiffen the response of the surface hardware and to improve the survivability of the system. The effect of a single dome plate was to allow heavy deformation of the plate shown in Figure 3.71 when the cone only moved 30% of the total deformation from the first impact,. The plate deformation increased on successive drops as the cone did not continue to move through the grout. Therefore, strong surface hardware is appropriate when stiff and strong cement grout are used.
143
Sample 35 (before) 60kN of collar tension
Sample 35 (post) 64kJ impact and 304mm of cone movement and 4mm of stretch.
Figure 3.70: Two dome plates for surface hardware.
Sample 41 (pre) 120kN of collar tension
Sample 41 (1st drop) 33mm of cone movement 65mm of bar stretch
Sample 41 (2nd drop) 4mm of cone movement 80mm of bar stretch, independent impact on plate
Figure 3.71: Single dome plate on a cone bolt. 3.5.3.8
Summary for Cone Bolts
Cone bolt performance was primarily affected by grout strength with a significant difference in performance between the 25MPa and >40MPa grouts as shown clearly by the force-displacement responses in Figure 3.63. This means that cone bolts require the selection of a suitable grout to be effective. This is contrary to prior work by Gillerstedt (1999) that states "the cone bolt (22mm) is not very sensitive to grout strength variation between 28MPa to 47MPa" for axial quasi-static testing of the cone bolt. Jager (1992) also states that the influence of grout strength between 20MPa and 40MPa was minimal, but a reduction in grout strength to 10MPa resulted in a 50% decrease in sliding resistance.
144
3.5.4
MODIFIED CONE BOLT
A program of eight samples was established by a third party that wanted independent testing of the MCB and assumed responsibility for the preparation of the simulated boreholes and installation of the MCBs within the simulated boreholes. The MCB has been extensively tested by CANMET and its performance has been reported by Gaudreau et al. (2004) and St-Pierre (2007). St-Pierre (2007) reports "…a qualitative analysis revealed the two energy absorption mechanisms of the bolt; the steel elongation and bolt sliding in the resin. Both mechanisms are present during every test, but their proportions vary significantly." Testing at WASM resulted in similar findings with the variation from incomplete mixing of the resin where the MCB slid easily out of the resin to complete mixing of the resin with small movement of the cone when subjected to load. The detailed results are not presented herein. 3.5.5
PARTIALLY DECOUPLED BULBED STRAND
Three bulb strands that were centrally decoupled from cement grout were tested. The samples are shown together with the Garford Dynamic Cable Bolts in the following section. The dynamic force-displacement responses are shown in Figure 3.72 and summarised in Table 3.14. All cables yielded within the 1.69m decoupled length. 350
Displacement (mm)
300 250 Centrally decoupled cable #141 200 Centrally decoupled cable #140 150
Centrally decoupled cable #142 with strand rupture
100
Rupture
50 0 0
20
40
60
80
100
120
Dynamic Force (kN)
Figure 3.72: Force displacement responses for bulb strand with central decoupling.
145
Table 3.14: Bulb strand centrally decoupled summary result
Bolt Number
Load Time (ms)
Displacement (mm)
Peak Deceleration (g)
Peak Force (kN)
Peak Ejection Velocity (m/s)
Energy Dissipated (kJ)
140
60
112
-13.7
285
2.9
25.1
Bar stretched
141
60
100
-15.7
325
2.8
25.2
Bar stretched
Results
One sample 142 had incorrect welding of the load transfer ring to the collar pipe. During the test the weld failed with detachment of the load transfer ring. This resulted in complete load transfer to the surface hardware followed by rupture of the strand initiated by the teeth of the wedges being pushed into the strand as shown in Figure 3.73.
Failure of the weld on the load transfer ring
Bite marks from the wedges into the cable strands clearly visible. The king wire remains undamaged. Radial fractures through the 3 piece wedge set show the fractures coincide with the contact point of the strand with the wedge.
Figure 3.73: Rupture through the strand from the wedges.
146
Measurement of the barrel diameter after the test showed an increase from a uniform 41.95mm diameter to a taper ranging from 42.70mm diameter where the barrel wall was thinner to 42.16mm where it was thicker. It is postulated that the increase in diameter occurred due to the wedge being drawn deeper into the barrel by high measured force of 300kN. The displacement of the wedge is shown in Figure 3.74.
Pre-test tensioned surface fixture.
Post test wedges drawn into the barrel.
Fracturing of the leaves.
Figure 3.74: Wedge location pre- and post-test sample 142. However, this would be very unlikely to happen in practice because load transfer associated with the strand would reduce the force at the barrel and wedge anchor. However, it would represent the case where the collar length was decoupled from the grout. 3.5.5.1
Bulb encapsulation and performance
Figure 3.75 shows the simulated borehole after it was dissected from the anchor flange to the toe. This was undertaken to assess the grout condition and load transfer by the bulbs. The collar sections were not dissected. The leading bulb (bulb number three) was located at the anchor restraining flange and it was not practical to dissect the pipe at this location to assess encapsulation. In the encapsulated toe above the flange, there was no damage to the grout (apart from air voids introduced during grouting). As the grout was chipped away to reveal bulb number two, it became clear that it was fully encapsulated within the grout and the bulb did not appear to have taken any of the dynamic load from the test. This implies the first bulb of the anchor has taken the entire load.
147
2600
Toe @ 3400
2600
Toe @ 3400
Figure 3.75: Dissection of anchor encapsulation bulb strand.
148
3.5.6
GARFORD DYNAMIC CABLE BOLT
Testing was undertaken on two versions of the strand. Only one test was performed for the first version as it was replaced before the testing program commenced. Six samples of the later version of the strand were tested. The dynamic cable bolt relies on a stick slip mechanism as the oversized centre king-wire and compact strand are drawn through the anchor. The strand is decoupled from the grout between the anchor and the surface fixture. The surface fixture has a significant influence on the performance of the reinforcement system. 3.5.6.1
Sample Preparation
The simulated discontinuity was located at 1.0m from the collar for all samples. All Garford Dynamic Cable Bolts had duct tape wrapped around the end of the decoupling tube to reduce the potential for grout ingress. All samples were encapsulated with a 0.4-0.45w/c ratio grout. The first set (samples 43-54) used the first version of the anchor installed in 2.58m long simulated boreholes. The second set (samples 78-80) were installed in 3.0m long simulated boreholes and used the second version of the anchor. The samples were configured as shown in Figure 3.14 with a 700mm sliding length of strand. The third set (samples 143-145) had a shorter 3.4m total length with 400mm of slide capacity; these strands were installed in simulated boreholes 3.4m in length. Testing with the longer simulated boreholes were made possible by adjusting the release hook connection to the beam to allow 900mm of pipe length to be located above the beam. The finished specimens are shown in Figure 3.64. This figure also shows the decoupled bulb strand and specimens described in the previous section.
149
0.47m 2.53m 0.97m
Decoupled bulb strand
Dynamic Cable Bolt Strand cable
1.69m
Duct tape
0.4m
1.12m
Collar
3.4m
Anchor flange 0.9m Toe
Figure 3.76: Garford Dynamic Cable Bolts and Decoupled Bulb Strand. 3.5.6.2 3.5.6.2.1
Testing Results Version One installed in cement grout
One out of 12 samples was tested. The strand pulled completely through the anchor. Further analysis was not undertaken as this version of the bolt had been superseded. 3.5.6.2.2
Version Two installed in cement grout
Three samples (78-80) were prepared initially and tested. However, instrumentation problems (slower DAQ rate of 250 samples per second) reduced the quality of the results for samples 78 and 80 (Figure 3.77) and no instrumentation data was obtained for 79. Due to the poor quality of these results, three additional samples (143-145) were prepared and tested. All samples were subjected to a 36kJ input load. Cyclic force-displacement responses as shown in Figure 3.77 resulted with recorded separations at the simulated discontinuity of 455mm, 627mm, and
150
280mm, respectively. The anchor cell data shows a stick slip process. Due to the lower sample rate, the results from samples 78-80 will not be presented in further detail.
Ripples in buffer closure
Cyclic stick / slip yielding mechanism
Figure 3.77: Garford Dynamic Cable Bolt measured response, sample 78 and 80.
The tests on samples 143-145 with half the decoupled length available to yield resulted in the irregular dynamic force-displacement responses shown in Figure 3.78 and summarised in Table 3.15. Sample 144 used a barrel and two-part wedge on the compact strand; there was partial sliding of this anchor along the strand before it locked in place and sliding at the anchor was initiated. The change in exposed strand below the barrel and wedge fixture and its rotation on sample 144 are shown in Figure 3.79.
151
250
Displacement (mm)
200
150
Garford Yielding Bolt #145 Garford Yielding Bolt #144 Garford Yielding Bolt #143
100
Strand pulls out
50
0 0
100
200
300
400
500
600
Dynamic Force (kN)
Figure 3.78: Dynamic force-displacement response for the Garford Dynamic Cable Bolt.
Peak Force (kN)
44
Strand pulls through yielding device
144
36.6
120
534
-9.4
202
5.5
60
Barrel and wedge slip and hold, then strand pulls through.
145
35.7
104
412
-9.8
212
5.0
46
Strand pulls through yielding device
152
Results
4.8
Dissipated (kJ)
196
Energy
-9.1
Velocity (m/s)
353
Peak Ejection
Displacement
100
Peak
Load Time (ms)
34.4
(mm)
Input Energy
143
(kJ)
Bolt #
Deceleration (g)
Table 3.15: Summary of dynamic testing on Garford Dynamic Cable Bolt responses.
Barrel and wedge target
End of cable marker End of cable marker
Figure 3.79: Sliding of strand through barrel and wedge, sample 144.
The stick slip mechanism of the compact strand through the anchor is evident in the raw results shown in Figure 3.80 and Figure 3.81; for example, there is a force range of 140kN between the peaks and troughs for sample 143. Note the response in the latter figure was inverted during the filtering process.
Bolt pulls out of yielding device
Figure 3.80: Raw stick slip response of the anchor cells, sample 143.
153
Butterworth filtered and inverted anchor cell response
Ac02 Ac06
AC01 Ac03 Ac07
Figure 3.81: Filtered instrumentation response sample 143. The video clearly shows rotation of the barrel and wedge anchor. The rotation occurs as the strand is pulled through the anchor and is associated with the spiral nature of the outer wires of the strand. The video shows that the anchor rotation occurs at the same time as the drop in anchor and strand load. The centre pin that is used to replace the king wire is deformed as the compact strand is drawn through the anchor as shown in Figure 3.82.
Tapered pin from sample#145
Pins from sample#143
Figure 3.82: Enlarged replacement king wires recovered after testing.
154
3.5.6.3
Summary for Dynamic Cable Bolt
The stick slip mechanism of the Garford Dynamic Cable Bolt results in a cyclic response between reasonably well defined upper and lower force limits. The decoupled strand is impacted at the collar by the loading mass via the barrel and wedge anchor. The strand stretches and slips through the anchor with rotation and a reduction of load transfer. The most significant issue with the mechanism of the strand sliding through the anchor is that load transfer is not consistent. Hence, there is potential for significant variability in bolt performance when subjected to the same dynamic load. 3.5.7
GARFORD DYNAMIC SOLID BOLT
Two versions of the Garford Yielding Solid bolt were tested. The test program for the bolt was established as:
•
Dynamic testing of the first and second versions with the bolt encapsulated in cement, installed in simulated standard borehole of equivalent stiffness 82.8 and 84.4GPa to represent ideal conditions; and
•
Development of a simulated rough borehole, equivalent stiffness of 35GPa, (Section 3.1.3) to allow installation by a Jumbo in resin. The simulated rough borehole with installed bolt would then be tested, to assess bolt performance. The effectiveness of resin encapsulation and possible damage to the bolt during installation could be assessed though dissection of the simulated rough boreholes.
3.5.7.1 3.5.7.1.1
Testing Results Version One installed in cement grout
The first version of the bolt, without an end stop mechanism, was installed in cementitious grout. Two bolts, labelled sample numbers 74 and 75, were tested. The dynamic force-displacement responses of the first version of the Garford Dynamic Solid Yielding Bolt are compared with the performance of 22mm Cone Bolts in Figure 3.83 and summarised in Table 3.16.
155
220 41
200 180
32
Force (kN)
160
74
140
75
120
60
100
61
80 60 40 20 0 0
40
80
120
160
200
240
280
320
360
Displacement (mm) Cone bolt 32, 40MPa Grout
Cone bolt 60, 20MPa Grout
Garford Solid yielding bolt 74
Cone bolt 41, 40MPa Grout
Cone bolt 61, 20MPa grout
Garford Solid yielding bolt 75
Figure 3.83: Comparison of the Garford Dynamic Solid Bolt with the 22mm Cone Bolt. 3.5.7.1.2
Version Two installed in cement grout and resin
The first version of the bolt performed consistently. However, it was considered that a second version of the bolt should be produced with a higher sliding resistance to the dynamic load and an end stop mechanism. The end stop mechanism maximises the energy dissipation capability of the reinforcement system by requiring the bar to be ruptured at the completion of sliding. Version two of the bolt was manufactured and installed in high strength construction grout (sample 89 and 90) to allow early testing. Once an improved resistance to sliding was confirmed, a jumbo was used to install the bolts with resin encapsulation in simulated rough boreholes. The dynamic force displacement responses from the first impacts are shown in Figure 3.84 and Table 3.16 provides a summary of responses for the first impact on all Garford Dynamic Solid Bolts.
156
350 Bolt 89 Bolt 98
300
Bolt 97 – resin is in the yielding mechanism, higher forces recorded
250 Load (kN)
Bolt 97 Bolt 100
Bolt 90 Bolt 99
200
Bolt 89 and 90, installed in 80MPa grout (indicates, idealised performance)
150 100
Bolt 100 – 8m/s impact, reduces dynamic frictional resistance
Bolt 98 & 99, installed by jumbo in resin
50 0 0
50
100
Until the end stop mechanism is reached and plastic deformation of the bar occurs.
150
200
250
300
350
400
450
Displacement (mm) Figure 3.84: Force-displacement responses from Garford Dynamic Solid Bolt (V2).
Peak Force (kN)
Bolt in grout, idealised performance (V1)
75
117
276
-7.3
166
3.8
38.8
Bolt in grout, idealised performance (V1)
89
90
196
-9.1
198
3.2
30.9
Bolt in grout, idealised performance (V2)
90
94
183
-8.4
183
3.0
25.9
Bolt in grout, idealised performance (V2)
97
64
114
-12.0
243
2.3
22.1
High peak due to resin in the yielding mechanism
98
87
171
-9.9
205
2.8
23.5
Bolt installed in resin by jumbo good performance
99
90
179
-9.5
197
3.0
26.1
Bolt installed in resin by jumbo good performance
100
147
405
-11
240
5.5
53.1
High impact test – end stop mechanism reached and shaft of bolt yields
157
Results
34.6
Dissipated (kJ)
3.7
Energy
178
Velocity (m/s)
-7.9
Peak Ejection
268
Peak
118
(mm)
Load Time (ms)
74
Displacement
Bolt Number
Deceleration (g)
Table 3.16: Summary of dynamic testing on Garford Dynamic Solid Bolts.
500
The tests on the second version of the bolt identified a number of key elements:
•
The second version of the bolt when installed in grout had less displacement and higher resistive force to loading when compared with the first version of the bolt. The reduction in displacement also means less energy is required to be dissipated through change in potential energy of the loading mass following impact.
•
The resin encapsulated bolts behaved marginally more stiffly when compared with cementitious encapsulated bolt and with smaller displacements (Figure 3.84).
•
It was possible for the bolt to be damaged during installation; this allowed resin to leak into the anchor mechanism and increased the resistance to sliding and decreased the displacement (e.g. bolt 97, particularly on the first loading).
•
The end stop mechanism developed full capacity of the steel bar followed by ‘cup and cone’ fracture of the steel bar as shown in Figure 3.85.
Figure 3.85: Cup and cone fracture, bar from bolt 99 after 3rd impact.
158
3.5.7.2
Dissection of simulated rough boreholes
The simulated rough borehole with resin encapsulated bolts were dissected after dynamic loading. This examination showed:
•
The mixing device was very effective with good quality resin as shown Figure 3.86. The best installation performance was achieved by rotating the bolt as it was slowly pushed through the entire length of the resin.
•
Over-drilling the holes by 100mm to 150mm allowed the resin bag to move to the end of the simulated borehole hole and not wrap around the mixing device.
•
Small air voids were formed around the anchor. These allowed for a small displacement (1030mm) of the anchor until dense resin was reached as shown in Figure 3.87 and Figure 3.88.
•
A short resin length of 260mm below the anchor on the bolt was to cause rupture of the solid bar once the end stop mechanism was reached (Figure 3.88). The short encapsulation length in sample 99 resulted from significant resin bleed at the toe of the simulated rough borehole during bolt installation. The length and consistency of the resin between the anchor and the collar is key to effective Garford Dynamic Solid Bolt performance.
Minor bagging
Mixing mechanism
Figure 3.86: Resin mixing mechanism on the Garford Dynamic Solid Bolt.
159
End stop mechanism – a weld line along the side of the bolt
Powdering of the resin, and probable movement of the yield mechanism
Stain due to resin contact with steel pipe. This occurred due to drill bit contacting the pipe wall during drilling.
Bubble in the resin
Figure 3.87: Dissection of toe length from sample 98.
260mm
450mm
Figure 3.88: Short encapsulation length below anchor and displacement of the anchor.
160
3.5.7.3
Dependence on loading velocity
The dynamic load on sample 100 (8m/s at impact) shows the resistive force is dependent on the dynamic frictional resistance of the steel bar pulling through the anchor. This occurred even though it is assumed that the majority of energy is dissipated due to plastic deformation of the bar caused by displacement through the anchor. A reduction in the friction from a static to dynamic state has been reported by numerous authors across a variety of fields; these include Forrester (1946) in steel, Spurr et al. (1957) for bitumen, and Toro et al. (2004) for quartz in earthquake faults. The explanation for the process varies depending on the properties of the materials involved. The most important aspect is that the Garford bolt exhibits similar dynamic friction dependence. 3.5.7.4
Multiple loadings
The test on sample 100 also showed that cumulative addition of energy dissipation from several drops leading up to breakage is not going to be the same as the energy required to break the reinforcement system from a single impact. Sample 100 dissipated 53kJ with the end-stop mechanism taking significant load while bolts 89 and 90 apparently absorb the same amount of energy on the first two drops without reaching the end stop mechanism. The suggested total energy absorption capability from summing smaller impacts of 65-70kJ from bolts 89, 90 and 99 is shown to be a 20-30% overestimate of the bolts’ ultimate single loading capacity of 53kJ measured in the test on bolt 100. This also implies that to be able to determine the true energy absorption of a reinforcement system or support element, the test equipment must be able to fail the reinforcement system on the first impact. The basic mechanism of the Garford Dynamic Solid Bolt is deformation of a round bar sliding through a smaller aperture from the application of force at the collar fixture. The mechanical deformation of the bar and friction dissipates energy. The manufacturing process involves small tolerances for the dimensions of the components and therefore consistent and repeatable results should be obtained. Dynamic testing of the second version of the Garford Dynamic Solid Bolt indicated both a higher sliding force and an effective end stop mechanism. The higher load transfer resulted in smaller total displacements from each dynamic load cycle than the first version. The end stop mechanism resulted in rupture of the steel bar after the sliding capacity was reached. The reinforcement system can also be effectively installed in either cementitious grout or resin; however, there was an unexpected stiffening of the response when installed in resin, possibly a result of the installation practice. The summation to capacity from multiple loadings on a single sample showed an over-estimation of 2030% of the energy dissipation compared to a larger impact that has sufficient energy to fail the reinforcement system on the first impact. The Garford Dynamic Solid Bolt frictional component of the bar through the yielding mechanism is significant, as shown by the drop in resistive force with the higher sliding velocity that resulted from the 64kJ input load.
161
3.5.8
SUMMARY FOR DFMC SYSTEMS
A comprehensive double embedment length dynamic test program has been undertaken on four reinforcement system designed specifically for the application of controlling the ground deformation from seismic events; three of these are fully reported. Variations to the design of the reinforcement system, grout strength and surface hardware were investigated as part of the program. Although each system tested belonged to the DMFC class, the effectiveness of the decoupling was dependent on the design of the system and encapsulation medium. The effectiveness of the reinforcement system in total was dependent on the design of the energy dissipation mechanism and how reliable the mechanism can be made. The results from the WASM dynamic test program on the reinforcement systems subject to the standard input energy of 36kJ are summarised in Table 3.17.
Results
(kJ)
Energy Dissipated
Velocity (m/s)
Peak Ejection
Peak Force (kN)
(g)
Peak Deceleration
(mm)
Displacement
Type
Table 3.17: Summary of yielding reinforcement system responses to dynamic loading.
22mm cone bolt in >40MPa cement grout
-
-9.8
213
3.0
19.7
Stable – both cone sliding and stretch of the bar
22mm cone bolt in 25MPa cement grout
280
-7.3
162
3.6
33.9
Stable - sliding of the cone through the grout, no bar stretch
Garford Dynamic Solid Bolt (V2) in cement grout
193
-8.1
181
3.0
26.8
Stable – sliding of the bar through the anchor
Garford Dynamic Solid Bolt (V2) in resin
168
-9.2
190
22.5
Stable – sliding of the bar through the anchor
Garford Dynamic Cable Bolt (V2)
382
-9.4
204
162
4.9
45
Unstable – strand pulls through the anchor
3.6
SUMMARY OF PERFORMANCE OF REINFORCEMENT SYSTEMS
The dynamic testing of reinforcement systems has resulted in the following general observations: •
Friction rock stabilisers can sustain dynamic loading but do so with excessive displacement or complete detachment from the toe anchor length.
•
The performance of threadbar depends on both the properties of the steel and the encapsulation material. Low strength, high elongation steel bars can absorb more energy than high strength steel bars. Threadbar in cement grout can decouple more readily and absorb more energy than the same threadbar encapsulated with resin grout.
•
The energy absorbing capacity of strand based cable bolts can be increased by decoupling the strand from the cement grout encapsulation. However, it is critically important that the strand is securely restrained at the toe end and at the collar. Effective toe capacity is best achieved by using bulbed strand while at the collar a plate and properly installed barrel and wedge anchor are required.
•
Reinforcement systems specifically designed with a discrete anchor within the toe embedment length have the highest energy absorbing capacities with the least displacement. These systems typically also have an adequate static force capacity and stiffness.
163
164
4
DYNAMIC TESTING OF SUPPORT SYSTEMS
This section describes the configurations and results for support systems that were subjected to dynamic testing. Firstly, a program of work on static testing of surface support systems was undertaken. This work was funded internally by WASM and separate to any MERIWA and industry contributions. However, the results from these investigations are reported in the Appendix because the static testing program was designed to formulate and evaluate different surface support system specimen preparation techniques that could then be adopted for the dynamic test facility. The static testing program involved: • Design, manufacture, construction and commissioning of a static test frame. • Design and commissioning of instrumentation and an associated data acquisition and analysis system. • Comprehensive literature review of surface support system testing. • Program of testing on woven and welded mesh specimens. • Development of a test method involving rock slabs and sprayed layers (e.g. shotcrete, fibrecrete, mesh reinforced shotcrete and a thin sprayed liner). • Program of sprayed layer testing. Two types of mesh were involved in a comprehensive testing program. A smaller program of dynamic tests was performed on shotcrete panels.
4.1
MESH TESTING
Static and dynamic testing was completed to assess the properties of welded wire and chain link meshes. 4.1.1
MESH SAMPLES
Testing has been undertaken on two different mesh types (Figure 4.1). The standard welded wire mesh used in Western Australian mines comprises 5.6mm diameter galvanised wires welded to form a 100mm square grid pattern. The second type of mesh is 4mm diameter high strength steel wire chain link mesh provided by Geobrugg. Wires are shaped in a zigzag pattern and are then woven together to form a diamond grid pattern. The wires are joined at the ends using a specially designed process. A 1.3m x 1.3m sample size was used in both the static and dynamic testing facilities. The welded wire mesh samples were cut from larger sheets provided by various sponsors. The cross wires were marked so the sample was placed within the mesh frame using a known configuration. All samples were positioned with the cross wires oriented across the frame and upwards, in contact with the loading plate.
165
The chain link samples were specifically manufactured by Geobrugg to the size required for testing purposes. The chain link samples were oriented with the direction of the stiff wires stretching across the frame.
Figure 4.1: Chain link and welded wire mesh, respectively. The results from the static test programs are provided in the Appendix. 4.1.2
MESH DYNAMIC TESTING RESULTS
Typical dynamic force - displacement reaction curves for welded wire mesh and chain link mesh are shown in Figure 4.2. A summary of all the dynamic force - displacement results is shown in Figure 4.3. The energy absorption is a function of force and displacement. Displacement is influenced, sometimes significantly, by the number of failures within the sample. For this reason, analysis of the mesh types has been undertaken at rupture. Rupture is defined as the point at which a part of the system breaks. Rupture may or may not correspond to the peak force achieved during the test. The variability of the sample response after rupture occurred means detailed analyses cannot be made post peak. The average dynamic rupture displacement for welded wire mesh is 203mm whereas the average dynamic rupture displacement of chain link mesh is 306mm. The average dynamic rupture force for welded wire mesh is 55kN. The average static rupture force for chain link mesh is 185kN. The average dynamic rupture energy for welded wire mesh is 2kJ. The average dynamic rupture energy for chain link mesh is 9kJ. The energy and displacement results are shown in Figure 4.4. As with the static testing, differences were observed in the dynamic failure mechanisms of welded wire mesh and chain link mesh. The failure mechanisms for welded wire mesh under dynamic loading conditions were the same as in the static tests; namely tensile wire failure, weld failure and failure of the wire through the heat affected zone (HAZ). Due to the rapid nature of the testing, the location of rupture was difficult to determine in all tests. In general, rupture occurs at the centre of a boundary of the sample and progresses towards a corner.
166
300 Chain Link Mesh 18kJ No Rupture Weld Mesh 2.7kJ 2 Ruptures 250
Force (kN)
200
150
100
50
0 0
100 200 300 Total Central Displacement (mm)
Figure 4.2: Typical dynamic force – displacement responses for welded wire mesh and chain link mesh. 300 Weld Mesh Dynamic Chain Link Dynamic
Rupture Force (kN)
250
200
150
100
50
0 0
100 200 300 Total Displacement at Rupture (mm)
400
Chain link squares shaded grey indicate no rupture
Figure 4.3: Summary of force - displacement responses for welded wire mesh and chain link mesh.
167
The dynamic failure mechanism for chain link mesh was different from those observed during the static tests. Generally the mesh failed at the ‘link’ between two wires. The location of the failure varied, with some tests failing close to the boundary of the sheet and some samples failing on the edge of the loading mass.
14 Weld Mesh Dynamic Chain Link Mesh Dynamic
12
Rupture Energy (kJ)
10 8 6 4 2 0 0
100
200
300
400
Total Displacement at Rupture (mm) Chain link squares shaded grey indicate no rupture
Figure 4.4: Rupture energy results for welded wire mesh and chain link mesh. 4.1.3
4.1.3.1
COMPARISON OF STATIC AND DYNAMIC RESULTS
Welded wire mesh
A comparison of typical force-displacement responses for welded wire mesh under static and dynamic conditions is shown in Figure 4.5. There is a good correlation between the static and dynamic performances of welded wire mesh. Less than 10% variation occurred in the average rupture displacement characteristics between the two facilities. A summary of all the static and dynamic results for welded wire mesh is shown in Figure 4.6. The results indicate that slightly higher forces are achieved under dynamic loading. This may be a result of the larger loading area being used for dynamic testing which causes more wires to be directly loaded.
168
70 Dynamic force - displacement result Static force - dispalcement reuslt
60
Force (kN)
50
40
30
20
10
0 0
50
100 150 Displacement (mm)
200
250
Figure 4.5: Static and dynamic force – displacement response for welded wire mesh. 70 Weld Mesh Dynamic Weld Mesh Static
Rupture Force (kN)
60
50
40
30
20
10
0 0
50 100 150 200 Total Displacement at Rupture (mm)
250
Figure 4.6: Comparison of static and dynamic force and displacement properties for welded wire mesh.
169
4.1.3.2
Chain link mesh
Figure 4.7 shows the typical static and dynamic force–displacement responses of chain link mesh. The chain link mesh results did not have as good a correlation between the static and dynamic performance as welded wire mesh. The results indicate a difference in the stiffness performance of the mesh between the two facilities. This effect may be a result of the differing loading areas between the two facilities. Figure 4.8 presents a summary of the force displacement results from the two facilities. The results show that the static test facility has more consistent force–displacement results but underestimates the maximum potential force. This is likely to be due to the influence of the failure mechanism under static test conditions. Further testing is required to improve the static test loading configuration.
300 Static force - displacement result Dynamic force - displacement result
250
Force (kN)
200
150
100
50
0 0
100 200 Displacement (mm)
300
Figure 4.7: Static and dynamic force – displacement reaction curves for chain link mesh.
170
300 Chain Link Mesh Dynamic Chain Link Static
Rupture Force (kN)
250
200
150
100
50
0 0
100
200
300
400
Chain link Total squares shaded grey at indicate no rupture Displacement Rupture (mm)
Total Displacement at Rupture (mm)
Chain link squares shaded grey indicate no rupture
Figure 4.8: Comparison of static and dynamic force and displacement properties for chain link mesh. 4.1.3.3
Concluding Remarks for Mesh Testing
Static and dynamic testing using the WASM test method has provided good comparisons for various mesh types. A comparison of static and dynamic test results has shown a good correlation between the two facilities. High tensile 4mm diameter chain link mesh has much higher force, displacement and energy capabilities than standard welded wire mesh used currently in Western Australian mines.
171
4.2
SHOTCRETE TESTING
The dynamic testing program involved 4 tests on the same concrete mix design with 2 different types of plastic fibres. 4.2.1
SAMPLE CONFIGURATION AND PREPARATION
The samples were prepared in the same manner as described in the Appendix for the samples tested statically. In summary, the samples consisted of a shotcrete layer sprayed over a sandstone slab with a centrally located, 500mm diameter disk cut from it. The sample and shotcrete layer specifications are summarised in Table 4.1 Table 4.1: Shotcrete layer specifications.
Sample
Slab Size (W x L) m
Thickness (mm)
Fibre Type
Restraint
1
1.4 x 1.6
90
Shogun
3 sides
2
1.4 x 1.6
102
Shogun
3 sides
3
1.4 x 1.5
77
Reoco Hookshot
4 sides
4
1.4 x 1.5
79
Reoco Hookshot
4 sides
4.2.2
DYNAMIC TESTING
The test specifications and summary of performance are given in Table 4.2. Sample 1 was subjected to two separate loadings. Table 4.2: Shotcrete dynamic test specifications and summary of performance.
* #
Test
Age (days)
Impact Velocity (m/s)
Energy Input (kJ)*
Peak Force (kN)
Energy Absorbed (kJ)
Failure Mechanism
1A
25
3.3
2.4
17.5
0.01
Bending/cracking
1B
25
4.45
4.4
93
1.8
Bending/cracking
2
32
5.69
7.2
96
3.5
Bending/cracking
3
11
3.3
2.4
TBA#
TBA
Punch through
4
29
3.3
2.4
TBA
TBA
Punch through
Loading mass of 446kg used in all tests To Be Analysed
172
Samples 1 and 2, although severely cracked and deformed, did not fail completely. Photos showing some typical crack patterns are given in Figure 4.9 and Figure 4.10. These crack patterns have some similarities to those expected from yield line theory for centrally loaded slabs that are supported on 3 edges. On the other hand, testing on samples 3 and 4 resulted in complete penetration of the load through the shotcrete layer as shown in Figure 4.11and Figure 4.12, respectively. Note that for sample 3, incomplete circular failure at the boundary of the loading resulted in adhesion causing some failure of the sandstone layer. For sample 4, the shear failure through the shotcrete layer occurred after some cracking as shown in Figure 4.13. More complete details for these tests are provided in the separate Addendum to this report.
Figure 4.9: Crack pattern after second test on sample 1.
173
Figure 4.10: Crack pattern after test sample 2
Figure 4.11: Punch through failure for sample 3.
174
Figure 4.12: Punch through failure for sample 4.
Figure 4.13: Cracking and punch through failure for sample 4.
175
4.3
SUMMARY FOR TESTING OF SUPPORT SYSTEMS
The behaviour of mesh may be characterised by an initial, low stiffness response with large displacements followed by a steadily increasing development of force until rupture of one or more wires or welds occurs. On the other hand, shotcrete is characterised by an initial stiff response up to the initiation of cracking followed by a steadily decreasing force when the load is applied by controlled displacement in the static test facility or potentially complete punch through in the dynamic test facility where the loading has momentum after the initiation of failure. It is worth noting that the observed punch through mode of failure is made possible by the sample configuration including the sandstone layer to which the shotcrete layer can adhere. This failure mode is unlikely in ‘standard’ tests on shotcrete slabs. The results suggest that the common practice of installing mesh over shotcrete means that the shotcrete will probably always fail due to rock movement and the mesh will be acting to control the movement of the failed shotcrete layer, but only potentially after large displacements have occurred. A potentially larger surface restraint force could be developed by having the mesh embedded within the shotcrete layer similarly to reinforced concrete. In this way, the force after cracking would be larger than is possible with fibres and less displacement would occur. Another benefit of embedded mesh would be improved resistance to the observed punch through failure mode. The caveat for embedding mesh with shotcrete is the requirement to use larger apertures than the nominal 100mm spacing between wires used widely in mines. Future testing with embedded mesh is required to demonstrate the benefits and address the concern that shotcrete will separate at the plane of the mesh layer and create a hazard similar to rock falls.
176
5
SUMMARY OF OUTCOMES AND CONCLUDING REMARKS
This section summarises some of the achievements made during the dynamic testing of ground support systems in MERIWA Project M349A. The achievements and related activities are: •
WASM Dynamic Test Facility
•
Reinforcement Dynamic Testing
•
Mesh Static and Dynamic Testing
•
Shotcrete Static and Dynamic Testing
•
Staff and Student Research Skills Development
•
Technology Transfer
Summaries and comments are provided in the following sections.
5.1
WASM DYNAMIC TEST FACILITY
The WASM Dynamic Test Facility was upgraded to improve safety and efficiency and to enable the testing of mesh and shotcrete panels. In addition, the associated WASM developed software was improved through the implementation of additional filtering capabilities and was enhanced to enable the processing and analysis of data obtained in mesh and shotcrete panel dynamic testing. It should be noted that the project indirectly involved a complementary program of investigations involving the establishment of a facility for static testing of mesh and shotcrete panels. This program of investigations was entirely funded by WASM from sources external to the MERIWA Project M349A. The work in this complementary program facilitated the development of the equipment and the sample configuration and preparation required for dynamic testing of mesh and shotcrete panels. The WASM Dynamic Test Facility remains at the forefront world-wide in dynamic testing and, particularly, in the area of data collection, processing and analysis. The facility has been demonstrated to have sufficient capacity to fail reinforcement systems and mesh and shotcrete panels within a single impact loading. Other facilities often require more than one impact to cause failure and it has been shown that the summation of energies from repeated impacts is not valid.
5.2
REINFORCEMENT SYSTEM DYNAMIC TESTING
The WASM Dynamic Test Facility essentially ranks the performance criteria (E. Fs, v(t) a(t)) of the reinforcement systems tested to survival or failure. This allows a decision to be made for mining applications based on limitations of allowable deformation to the development headings, plastic deformation of elements of the ground support systems, and the ability to install particular reinforcement and support systems to meet the operational requirements of the mine. Operators must bear in mind that all tests on reinforcement systems
177
assume an axial response only and consider the extent to which shear may influence reinforcement systems in the configurations that result underground and the loadings from the rock mass. The results from testing of reinforcement systems were classified into the three classes defined previously, namely: •
Continuously Frictionally Coupled
•
Continuously Mechanically Coupled
•
Discrete Mechanically or Frictionally Coupled
The reinforcement systems tested within each class were as follows: Continuously Friction Coupled •
Split Tube Bolts
•
Expanded Tube Bolts
Continuously Mechanically Coupled •
Encapsulated Threadbar
•
Encapsulated Strand
Discrete Mechanically or Frictionally Coupled •
Partially Decoupled Threadbar
•
Partially Decoupled Bulbed Strand
•
Cone Bolt
•
Modified Cone Bolt
•
Garford Dynamic Cable Bolt
•
Garford Dynamic Solid Bolt
The reason for presenting the results within these classes based on load transfer is that reinforcement systems are usually selected on the basis of their relative force capacity. With reinforcement systems likely to be subjected to dynamic loadings, additional considerations need to be made with respect to energy absorption and total displacement. It is generally accepted that reinforcement systems in the CFC category have lesser load capacity and therefore would not be compared with systems in the other two categories. Similarly, it is generally accepted that many CMC reinforcement systems are too stiff and have lower energy absorption capacities than the DMFC reinforcement systems. The results for testing of each system have been summarised in this report. The detailed results in the form of a Standard Test Report for each test are contained within a separate Addendum to this report.
178
5.3
MESH TESTING
Two types of mesh were tested during this project, namely welded wire mesh and woven, high strength steel mesh. The results for testing of each system have been summarised in this report. The detailed results are contained within the separate Addendum to this report.
5.4
SHOTCRETE TESTING
The results for testing of each shotcrete slab have been summarised in this report. The detailed results in the form of a Standard Test Report for each test are contained within the separate Addendum to this report.
5.5
STAFF AND STUDENTS RESEARCH SKILLS DEVELOPMENT
The following WASM staff and post-graduate students worked on various aspects of the M349A investigations: •
Ernesto Villaescusa, Alan Thompson, John Player, Ellen Morton, Moises Cordova, Brett Scott and Lance Fraser.
In addition, Andy Roth and Roland Butcher from Geobrugg assisted with setting up and testing of mesh statically and dynamically. Undergraduate students received research training through involvement in components of M349A investigations and prepared theses as part of the requirements for the final year of their mining engineering degrees. The students and their thesis title are in chronological order: •
Frederic Wallefeld 2005. Dynamic Testing Of Rock Reinforcement Systems Using The WASM Momentum Transfer Method.
•
Alex Cull 2006. Dynamic Testing of Rock Reinforcement Systems in the WASM Test Facility.
•
Martin Filar 2007. Dynamic Testing of Splitsets in the WASM Dynamic Testing Facility.
•
Dave Pearce 2007. Dynamic Testing of Rock Support Systems in the WASM Dynamic Test Facility.
•
Kyle de Souza 2008. Dynamic Testing of Inflatable Rockbolts in the WASM Dynamic Test Facility.
5.6
TECHNOLOGY TRANSFER
5.6.1
PUBLISHED PAPERS
The following papers were published on the outcomes from M349 as a means of technology transfer. •
Player, J.R., A.G. Thompson and E.Villaescusa 2004. Dynamic testing of rock reinforcement using the momentum transfer concept. Ground Support in Mining and Underground Construction, 327-340, Balkema:Leiden.
179
•
Thompson, A.G., J.R. Player and E. Villaescusa 2004. Simulation and analysis of dynamically loaded reinforcement systems. Ground Support in Mining and Underground Construction, 341358, Balkema:Leiden.
•
Villaescusa, E., A.G. Thompson and J.R. Player 2005. Dynamic testing of rock reinforcement systems. 2005 Australian Mining Technology Conference, AusIMM:Melbourne, 79-95.
The following papers were published during M349A as a means of technology transfer. •
Morton, E.C., A.G. Thompson and E. Villaescusa 2008. Static testing of shotcrete and membranes for mining applications. 6th International Symposium on Ground Support in Mining and Civil Construction, Cape Town, South Africa, 195-212.
•
Morton E.C. A.G. Thompson and E. Villaescusa 2009. The performance of mesh, shotcrete and membranes for surface ground support. RockEng09, Rock Engineering in Difficult Conditions (eds. M Diederichs and G Grasselli), Paper 4022, CIM:Montreal, 12p.
•
Morton, E C , A G Thompson, E Villaescusa and D Howard 2008 Static Testing of Shotcrete, 13 Australian Tunnelling Conference, IEAust:Melbourne, 83-87.
•
E.C. Morton, A.G. Thompson, E. Villaescusa, & A. Roth, 2007. Testing and analysis of steel wire mesh for mining applications of rock surface support. 11th ISRM Congress on Rock Mechanics V2, 1061-1064, Lisbon, July, Taylor and Francis:London.
•
Morton E.C., E. Villaescusa and A.G. Thompson 2009. Determination of energy absorption capabilities of large scale shotcrete panels. Shotcrete for Underground Support XI (Ed. F Amberg), Davos, Switzerland, Engineering Conferences International:New York, 20p.
•
Player, J.R., E.C. Morton, A.G. Thompson and E. Villaescusa 2008. Static and dynamic testing of steel wire mesh for mining applications of rock surface support. 6th International Symposium on Ground Support in Mining and Civil Construction, Cape Town, South Africa, 693-706.
•
Player, J.R., A.G. Thompson and E. Villaescusa 2008 Improvements to reinforcement systems through
dynamic
testing.
10th
Underground
Operators
Conference,
Launceston,
AusIMM:Melbourne, 79-88.
•
Player, J.R., E. Villaescusa and A.G. Thompson 2008. Dynamic testing of reinforcement systems. 6th International Symposium on Ground Support in Mining and Civil Construction, Cape Town, South Africa, 597-622.
•
Player, J.R., E. Villaescusa and A.G. Thompson 2009. Dynamic Testing Of Friction Stabilisers. RockEng09, Rock Engineering in Difficult Conditions (eds. M Diederichs and G Grasselli), Paper 4027, CIM:Montreal, 12p.
180
•
Player, J.R., E. Villaescusa and A.G. Thompson 2009. Dynamic testing of threadbar used for rock reinforcement. RockEng09, Rock Engineering in Difficult Conditions (eds. M Diederichs and G Grasselli), Paper 4030, CIM:Montreal, 12p.
•
Thompson, A.G. 2006. Energy absorption of shotcrete and reinforcement schemes. GoldenRocks, Proc 50th US Symposium on Rock Mechanics, Paper 942, 12p.
•
Varden, R., R. Lachenicht, J.R. Player, A.G. Thompson and E.Villaescusa 2008. Development and implementation of the Garford Dynamic Bolt at the Kanowna Belle mine. 10th Underground Operators Conference, Launceston, AusIMM:Melbourne, 95-104.
•
Villaescusa, E., A.G. Thompson and J.R. Player 2005. Dynamic testing of rock reinforcement systems. 2005 Australian Mining Technology Conference, AusIMM:Melbourne, 79-95.
5.6.2
SEMINARS/WORKSHOPS/COURSES
Two short courses including aspects of the dynamic testing investigations were presented: •
A Decade of Research in Ground Support at WASM - A 3 day update course. WA School of Mines, Kalgoorlie April 2007.
•
A Decade of Research in Ground Support at WASM - A 2 day short course in conjunction with the AusIMM 10th Underground Operators’ Conference, Launceston, March 2008.
5.6.3
PROGRESS REPORTS
Eighteen progress reports were prepared at three monthly intervals from July 2005 to December 2009 5.6.4
SPONSORS’ MEETINGS
A planning meeting was held in June 2005. A review meeting was held in April 2010.
5.7
ACKNOWLEDGEMENTS
The writers wish to express their deepest appreciation for the financial and moral support of MERIWA, the members of the mining advisory committee and the following organisations and their representatives
Atlas Copco
Dywidag Systems International
El Teniente
Geobrugg
MERIWA
Newmont
Placer Dome Asia Pacific
181
Strata Control Systems
WASM
5.8
CONCLUDING REMARKS AND FUTURE DEVELOPMENTS
Most of the developments of new reinforcement systems have or are being made in the DMFC category. During the close to 9 years since the MERIWA related investigations began, the reinforcement systems considered to be applicable to ground support in seismically active mines have grown from a single system (i.e. the South African developed Cone Bolt) to include at least eight. These are in approximate chronological order of development: •
Cone Bolt from South Africa
•
Garford Dynamic Cable Bolt from Australia
•
Durabar from South Africa
•
Garford Dynamic Solid Bolt
•
Durastrand from South Africa
•
Modified Cone Bolt (MCB) from Canada
•
Atlas Copco Roofex from USA/Sweden
•
D-Bolt from Finland
It is vitally important that the WASM Dynamic Test Facility continue to provide independent testing and evaluation to avoid the mistakes of history when bolts were released to the mining industry without proper due diligence with regard to their performance and resulted in failures due to a number of factors that had not hitherto been considered. Already, the WASM Dynamic Test Facility has been used to assess one reinforcement system as being unreliable and has been able to identify improvements to at least two other reinforcement systems and the manufacturer has been able to implement the required changes prior to them being made available on a commercial basis. In mining, the restraint provided by rock bolts and cable bolts is not rigid. In fact, in some circumstances, special reinforcement systems designed to absorb energy by allowing for large displacement to occur without exceeding the force capacity are used in combination with shotcrete. These ground support schemes are used in areas prone to dynamic loading from the rock after it fails. The interactions between the rock, shotcrete and reinforcement are poorly understood. In particular, it is not currently known how the relative strengths and stiffness of the ground support systems (i.e. shotcrete and reinforcement) affect their ability to absorb energy and survive dynamic events. The objectives for any future work would be to determine the response and energy absorption of fullyintegrated ground support schemes and to determine the distribution of forces and energy absorption within
182
the component systems. The upper, right hand photo in Figure 5.1 shows the results from a typical prototype test of mesh supported by four corner bolts. The lower photo in Figure 5.1 is prior to testing and shows an alternative scheme with one centre bolt as envisaged for the WASM Dynamic Test Facility. The scope of work and associated specific tasks have not yet been well defined. But it is anticipated that a research proposal involving WASM staff and a PhD student will be prepared soon and thereafter circulated to interested parties with a view to obtaining funding.
Release point
Box
Guides and legs
Figure 5.1: WASM Prototype dynamic loading of ground support schemes.
183
184
6
REFERENCES
Armelin, H.S. and N. Banthia 1997. Predicting the flexural post-cracking performance of fibre reinforced concrete from the pullout of single fibers. ACI Materials Journal, 94(1),18-31. ASTM 2002. Standard Test Method for Flexural Toughness of Fiber-Reinforced Concrete (Using Centrally Load Round Panel., West Conshohocken:ASTM, 6p. Barrett, S and McCreath, D 1995. Shotcrete support design in blocky ground: Towards a deterministic approach. Tunnelling and Underground Space Technology, Vol.10, No.1, pp. 79-89. Great Britain Elsevier Science. Bernard, S. 2003. Release of new ASTM Round Panel Test, Shotcrete, Spring, 20-23, Farmington Hills:American Shotcrete Association. Deere, U, Peck, R, Monsees, J and Schmidt, B 1969. Design of tunnel liners and support systems. Final report for office of high speed ground transportation. Washington US Department of Transportation. EFNARC 1996. European Specification for Sprayed Concrete, 30p. Fernandez-Delgado, G, Mahar, J and Parker, H 1976. Structural behaviour of thin shotcrete liners obtained from large scale tests. Shotcrete for ground support. Proceedings of the engineering foundation conference. Oct 4-8. pp. 399-442. ACI. Forrester, P. G. (1946) Kinetic Friction in or Near the Boundary Region. II. The Influence of Sliding Velocity and Other Variables on Kinetic Friction in or Near the Boundary Region. Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, 187, 439-463. Holmgren, J 1976. Thin shotcrete layers subject to punch loads. Shotcrete for ground support. Proceedings of the engineering foundation conference. Oct. pp. 443-459. ACI. Holmgren, J 2001. Shotcrete linings in hard rock. Underground mining methods. Engineering fundamentals and international case studies. Hustrulid, W and Bullock, R (ed). pp. 569-577. Colorado Society for Mining, Metallurgy and Exploration. Hyett, A, Bawdem, W, and Reichardt, R, 1992. The Effect of Rock mass Confinement on the Bond Strength of Fully Grouted Cable Bolts in Int. J. Rock Mech. Min. Sc.& Geomech. Abstr. Vol 29, pp 503-524. Li, T., Brown, E.T. Coxon, J. and Singh, U. 2004. Dynamic capable ground support development and application, Ground Support in Mining and Underground Construction – Villaescusa & Potvin (eds.). Malvar, L.J. and Crawford, JE 1998, 'Dynamic increase factors for steel reinforcing bars', in 28th Department of Defence Explosive Safety Board Seminar, Orlando, Florida, August 1998.
185
Ortlepp, W.D. 1983. Considerations in the design of support for deep hard rock tunnels. 5th international congress on rock mechanics. Vol.2, Balkema Rotterdam pp. 179 - 187. Pakalnis, V and D. Ames 1983. Load tests on mine screening. Underground support systems. Canadian Institute of Mining Metallurgy and Petroleum, Special Volume 35. Udd, J (ed) pp 79 – 83 Pearce, D. 2007. Dynamic Testing of Rock Support Systems in the WASM Dynamic Test Facility. WASM Undergraduate Final Year Thesis. Player JR, 2007. Dynamic testing of Jumbo installed Garford Yielding Solid Bolt into simulated boreholes for Barrick – Kanowna Belle, Research Report, by John Player, July 2007. Player, J.R., A.G. Thompson and E.Villaescusa 2004. Dynamic testing of rock reinforcement using the momentum transfer concept. Ground Support in Mining and Underground Construction, 327-340, Balkema:Leiden. Player, J R, Villaescusa, E, and Thompson, A G. 2005 An Examination of Dynamic Test Facilities. Advanced Geomechanics in Mines, 2005 (ed: Potvin) Section 9, Australian Centre for Geomechanics: Perth. Player, J R, Villaescusa, E, and Thompson, A G, 2008. Improvements to reinforcement systems through dynamic testing in 10th Underground Operators Conference, AusIMM:Melbourne Rabinovich, F.N. 1995. Concretes with Dispersed Reinforcement, Rotterdam:Balkema, 214p. Robins, P.J., S.A. Austin and P.A. Jones 1996. Flexural strength modeling of fibre reinforced sprayed concrete. In Sprayed Concrete Technology, ed. S.A. Austin, London: Spon:, 107-114. Roth, A, Windsor C.R., Coxon J. and de Vries R. 2004. Performance assessment of high tensile wire mesh ground support under seismic conditions. Ground support in mining and underground construction. Villaescusa and Potvin (ed). pp. 589 - 594 Soroushian, P. and C.D. Lee 1990. Distribution and orientation of fibers in steel fiber reinforced shotcrete. ACI Materials Journal, 87(5), 433-439. Spurr, R T, and Newcomb, T P, 1957 The Variation of Friction with Velocity. Proceeding of the Physics Society. Section B, 70, 198-200. Tannant. D. 2001. Load capacity and stiffness of welded wire, chain-link and expanded metal mesh. Section 3. International seminar on mine surface support liners: Membrane, shotcrete and mesh. August, Tannant, D, Kaiser, P.K, and Maloney S.1997. Load - displacement properties of welded - wire, chain - link and expanded metal mesh. International symposium on rock support - Applied solutions for underground structures. Lillehammer Norway. Broch, E Myrvang, A Stjern, G (ed). June 22 - 25, pp. 651 - 659.
186
Thompson, AG 2001. Rock support action of quantified by testing and analysis. Section 1. International seminar on mine surface support liners: Membranes, shotcrete and mesh, ACG: Perth. Thompson, A G, and Player, J R 2005. Implementation of dynamically load resistant rock bolts at Kanowna belle Gold Mine, WA School of Mines, Unpublished report. Thompson, A.G., J.R. Player and E. Villaescusa 2004. Simulation and analysis of dynamically loaded reinforcement systems. Ground Support in Mining and Underground Construction, 341-358, Balkema:Leiden. Thompson, AG, Windsor, CR and Cadby, GW 1999. Performance assessment of mesh for ground control applications. Rock support and reinforcement practice in mining. Villaescusa, Windsor, Thompson (ed). 119-130. Toro, G D, Goldsby, D L, and Tullis, T E 2004. Friction falls towards zero in quartz rock as slip velocity approaches seismic rates. Nature, 427, 436-439 Tran, V.N.G., A.J. Beasley and E.S. Bernard 2001 Application of yield line theory to round determinate panels. In Shotcrete: Engineering Developments, ed. E.S. Bernard, 245-254, Lisse:Swets and Zeitlinger. Tran, V.N.G., E.S. Bernard and A.J. Beasley 2005. Constitutive modeling of fiber reinforced shotcrete panels. ASCE J. Engineering. Mechanics., May, 512-521. Van Sint Jan, M and Cavieres P. 2004. Large scale static laboratory test of different support systems. Ground support in mining and underground construction. Eds Villaescusa and Potvin , 571-577, Balkema:Leiden. Villaescusa, E 1999. Laboratory testing of weld mesh for rock support. Rock support and reinforcement practice in mining. Villaescusa, E, Windsor, C and Thompson, A (ed). Balkema Rotterdam,155 - 159. Villaescusa E, Player JR, Morton EC and Thompson AG 2008. Ground Support at the WA School of Mines. MassMin 2008. June 9-11, Luleå, Sweden.
187
![FREE [DOWNLOAD] FEEDBACK CONTROL OF DYNAMIC SYSTEMS ...](https://m.moam.info/img/260x300/free-download-feedback-control-of-dynamic-systems-_64789b7a097c474e708cff99.jpg)
