Status of roof bolting research at UNSW, Australia
B K Hebblewhite
School of Mining Engineering,
The University of New South Wales, Australia
ABSTRACT
The design, installation and performance of roof support systems in underground coal mines is of vital importance to both the safety and productivity of the mining operation. Inadequate or inappropriate roof support systems can lead to unsafe roof conditions threatening both personnel safety and production capacity. This paper provides a summary of a number of recent and current research activities in this field, at UNSW. These projects are being conducted on behalf of the Australian coal industry, as well as part of an in-house research program. The projects are a combination of laboratory and field investigations into four areas of bolt performance:
- behaviour of weak strata
- rock bolt anchorage systems and load transfer
- premature rock bolt failure
- rock bolt performance under shear loading conditions.
INTRODUCTION
The Australian coal industry produces over 330 Mtpa of black coal, resulting in 266 Mtpa of saleable product, of which 73% is exported. This dominance of export markets result in a priority focus on consistent, reliable supply of product, together with tight constraints on mine operating costs.
Approximately 28% (94 Mtpa) of the total national production is mined from underground mines, predominantly (81%) from retreating longwall operations. There are currently 29 producing longwall mines in Australia. Typically, each of these mines would develop in excess of 10 km of development drivage per year. This longwall focus further drives the need for efficient, safe and cost-effective rapid and stable roadway drivage. Inadequate or inappropriate roof support systems can lead to unsafe roof conditions threatening both personnel safety and production capacity.
Throughout the Australian underground industry, roof support in both primary and secondary development roadways is achieved through the use of fully-encapsulated resin-anchored bolts installed as primary support, together with varying combinations of plates, straps and steel mesh, usually installed at or near the face during a cyclic in-line cut and bolt development system. This primary support is generally made up of high tensile strength X-bar 22mm diameter bolts, and is often pre-tensioned. Primary support density can vary from as few as 2 or 3 bolts per metre of roadway, to in excess of 15 bolts per metre (including rib bolts). In addition, where required (such as in major roadways and intersections, or regions of poor ground conditions or additional mining-induced stresses (such as tailgates)), secondary support is also installed in the form of cable bolts and other longer bolts and tendons, plus standing support, where necessary.
In the light of the importance of good reliable development roadway stability, and the extent and density of roadway drivage and roof bolting used in Australian coal mines, this paper provides a summary of a number of recent and current research activities in this field, at UNSW. As a review paper, the extent of detailed results included is limited, but the paper seeks to demonstrate the range of investigations, typical results and bolt-related research capabilities available.
The research projects reviewed here are being conducted by a team of UNSW researchers, on behalf of the Australian coal industry, as well as part of an in-house research program. The projects are a combination of laboratory and field investigations into four areas of bolt performance:
- behaviour of weak roof strata
- rock bolt anchorage systems and load transfer
- premature rock bolt failure
- rock bolt performance under shear loading conditions.
The project on behaviour of weak strata was recently completed and included a major site investigation at Angus Place Colliery, west of Sydney. The behaviour of the relatively weak and laminated maingate roof strata was extensively monitored both during initial development and subsequent longwall retreat, in an effort to quantify the mechanisms of roof deformation and potential failure regimes during the various mining-induced loading cycles.
The second project on bolt anchorage performance has involved the development of a fully instrumented, servo-controlled laboratory bolt drilling, installation and pull-out facility. This enables bolt anchorage performance and anchorage load transfer characteristics to be fully evaluated. The paper describes briefly the capabilities of this facility to evaluate parameters such as bolt deform pattern effectiveness, resin annulus and other related variables.
The third project on premature bolt failure has been a collaborative research project with two other industry consultants – SCT Operations and Ground Support Services. The project has conducted an industry survey at a selected number of sites where premature bolt failure has been observed. Stress Corrosion Cracking (SCC) has been identified as the cause of such failures. The results of the survey and the subsequent metallurgical analyses are discussed, together with more recent experience.
The fourth project has only recently commenced, and involves the development of laboratory facilities to evaluate the installed performance of bolts acting in shear resistance across different types of rock mass discontinuities. The effect of different levels of pre-tensioning on bolt shear performance will be included in the testing program.
FIELD STUDIES – WEAK ROOF STRATA
Project Background
Angus Place Colliery is located in the southern section of the Western Coalfield of New South Wales, approximately 110km due west of Sydney, Australia. At this colliery, longwall mining has been the primary source of production since 1979. Mining conditions have progressively deteriorated as more severe geological/geotechnical conditions have been encountered. It is claimed by some (Doyle [1]) that Angus Place experiences arguably the most difficult coal mining strata conditions in Australia. The mine roof changes from a combined Lithgow - Lidsdale Seam section (convergence zone) in the north and to the east, through a seam split zone, where the seams diverge. Combinations of three joint sets, namely NW-SE, NE-SW and N-S, can affect the coal roof. Coupled with the structural discontinuities caused by jointing in the roof, the laminated stratigraphic nature of the strata – composed of varying layers of coal, mudstone, sandstone and claystone – further contributes to a roof that exhibits both weak and relatively soft geomechanical characteristics.
A research investigation by UNSW was carried out for a number of mining companies, with government financial support. Results of the Angus Place work were recently reported by Hebblewhite & Lu [2]. The major objective of this investigation was to develop an understanding of the detailed roof deformational behaviour and the performance of the installed rock bolting reinforcement system used in a longwall gateroad at Angus Place Colliery.
Monitoring System
The geotechnical monitoring instrumentation consisted of multiple wire and sonic extensometers, plus strain gauge instrumented bolts. Figure 1 illustrates the monitoring configuration in heading 2, maingate 22, Angus Place Colliery. (The chain pillar is located on the left of this diagram (looking inbye), while the future longwall block is on the right).
This instrumentation pattern was all installed on development, within 2m of the roadway face, in order to obtain the maximum component of the development-related deformation, as soon as the face advanced away from the site. A key feature of the instrumentation layout – apart from the considerable mine logistical organisation and assistance provided to enable installation of such a large array of instruments at the face – was the large number of monitoring points across the full roadway width (or roof span). Unlike routine geotechnical monitoring conducted at most mines, where instrumentation is usually restricted to the roadway centreline, it was considered important in this instance to gain a complete profile of the roof deformation across the full span of the roof, to assist in defining and interpretting the roof behavioural mechanisms. In this sense, the comprehensive nature of the dataset obtained from this investigation was quite unusual, if not unique, for the Australian coal industry.
In order to gain an absolute, stable reference point in the roof, above the height of roof dilation and sag, a wire extensometer was installed in the centre of the heading roof. Previous experience at Angus Place had indicated that the 7.5m upper limit of the standard sonic extensometers was at times only marginal for acting as a non-displaced reference point. The reference anchor for the wire extensometer was located 15m above the roof horizon with three other measurement anchors located at 12m, 9.5m and 7.5m above the roof respectively. (The weak and deformable nature of the floor strata, combined with mine traffic obstruction and potential damage, prevented the use of any floor monitoring stations).
Based on the monitoring arrangement adopted, the relative deformation within the 7.5m to 15m region of roof could be determined. There were also five sonic extensometers installed, each with 20 anchors at up to 500mm spacing over a 7.5m length. These were positioned in a single row across the roadway width, with a centrally located station immediately adjacent to the wire extensometer to enable the two datasets to be combined to create an effective 23 anchor, 15m long extensometer.
Apart from extensometers, six strain gauge instrumented bolts, 2.1m in length, (9 gauges per side at 200mm spacing), were also installed across the roadway in the same plane as the extensometer instrumentation. This replicated the normal roof bolt support pattern in the mine, forming a single row in the normal support system.
Figure 1: Detailed instrument layout in heading 2, maingate 22
(vertical section view, all dimensions in metres)
(WE-wire extensometer, SE-sonic extensometer, SG-strain gauged bolt)
Summary of Results
Figures 2 to 6 present a range of the results obtained from this comprehensive site investigation. They not only provide data to illustrate the nature and geomechanical behaviour of the roof strata, but also illustrate some useful alternative means of result presentation, to assist in understanding the ground behaviour.
Figure 2 presents the raw extensometer deformation data plotted against time, for the centre of the roadway. It illustrates the response of the roof strata to several different loading regimes, generated by the initial development face moving away from the line of instrumentation (days 15 – 60), followed by the subsequent approach of the retreating longwall (days 250 – 280). The results also illustrate a component of ongoing, time-dependent deformation, under quasi-static loading conditions, between these major two loading regimes.
Figure 2: Cumulative displacement with time recorded by extensometer SE3 (centre)
Figure 3 presents the central roadway extensometry data (incorporating the 15m wire extensometer, combined with the adjacent sonic extewnsometer), relative to roof depth, for a selection of time intervals. It is clear from this result that the majority of roof deformation is occurring within 5m of the roof horizon, but there is a low level of movement extewnding beyond the 7m sonic extensometer height, up to at least 12m into the roof. A further observation is that differential movement (hence tensile strain) is occurring within the first 2.1m bolted horizon, with a significant step in deformation close to the top of the bolts.
Figure 3: Roof deformation at centre of heading.
Figure 4 provides a simplified graphic interpretation of the extent of the roof deformation at different horizons into the roof, utilising data from the full set of roof extensometers. This illustrates the shape of the overall roof relaxation zone, often referred to as a zone of softening. It also defines the bending profile of different strata units, behaving as stacked beams within the roof, subjected to an imposed dead weight vertical load, and also significant end load due to a high in situ horizontal stress field. The height of this zone of softening is considerable in this weak laminated roof material, and illustrates the need for secondary support, using cable bolts or equivalent, if further control of this roof material is required.
Figure 5 is another alternative data presentation plot. In this case the data from an individual extensometer is presented as a calculated bay strain, or average strain within each bay between adjacent anchors. A strain bar is shown for each of a nominated series of time slices through the data. Using these strain histograms for each roof horizon, it is possible to clearly see which horizons are active over time, as well as observing any progression of increasing strain upwards into the roof. Similar results can be presented for strain rate.
Figure 4: Roof deformation at day 277 (after mining extraction)
Figure 6 is another alternative presentation of the combined roof extensometer data, based on an assumption that all deformations measured between each successive anchor horizon are concentrated in the one location, resulting in a display of potential bed separation zones, with magnitudes noted. This presentation provides an indication of any differential movements on particular horizons – across the heading width, as well as further highlighting the extent of deformations occurring at significant heights into the roof, in spite of a relatively high density of immediate roof reinforcement. This once again suggests that there may be some benefit gained by the use of some longer tendon secondary reinforcement.
Figure 5: Strain between adjacent anchors along extensometer (SE1)
Figure 6: Bed separation/horizontal fracturing within roof strata (day 277)
ROCK BOLT ANCHORAGE PERFORMANCE
The laboratory bolt anchorage test facility at UNSW has been developed to provide a research and product evaluation test facilty, under controlled laboratory conditions. Further details of the facilty and some recent test results have been provided by Hagan [3]. In particular, Hagan describes the use of the facility to quantify the magnitude and the distribution of load transfer from the rock, through resin to the rock (and vice-versa). Effective load transfer is a key to achieving optimum performance from fully encapsulated bolts. Figues 7 and 8 illustrate the test rig in the laboratory, and the bolt installation facility within a biaxial cell, where sample constraint, as well as bolt loading can be fully servo-controlled.
Figure 7: Laboratory bolt anchorage test facility.
Figure 8: Schematic of biaxial confinement cell (after Hagan [3]).
It is widely understood that the profile of load transfer, away from the point of load application, is at a relatively constant rate, eg 30 tonnes per metre of bolt length. However, other views are that the rate of load transfer (or decay in bolt load along the bolt length, away from a fixed applied axial load (or displaced discontinuity)), will always distribute exponentially through the bolt to the extremity of the bolt – regardless of bolt length.
Preliminary results indicate that this latter type of load transfer behaviour can be observed, but further testing is required to validate this. Test results also demonstrate very different behaviour when a displacement driven bolt loading is achieved, similar to the in situ bed separation behaviour, as opposed to a conventional pull test situation, often used for anchorage assessment.
PREMATURE ROCK BOLT FAILURE
A recently completed research project conducted by UNSW, SCT Operations and Ground Support Services (Crosky et al. [4]) has identifed the existence of a number of premature bolt failures across the industry, many of which have been attributed to a metallurgical phenomenon known as stress corrosion cracking (SCC). The project was commenced in 1999 in response to ‘piecemeal’ evidence of rock bolt failures in a number of mines, many in relatively benign loading environments. This led to the recognition of the problem of “stress corrosion cracking” as a factor in many of these instances. The problem appeared to be linked to characteristics of either the bolt metallurgy or the bolt manufacturing process (or both), coupled with a source of corrosion. The problem had been recognised by various parties in Australia, including the three organisations involved in this research project. Similar experiences had also been reported from the UK coal industry, and subsequently also the US coal industry.
Stress corrosion cracking is defined, in layman’s terms, as
“Stress corrosion cracking is slow, progressive crack growth under the application of a sustained load (either residual or applied) in a mildly corrosive environment, with failure occurring below the ultimate tensile strength of the material”.
Figure 9 shows a typical bolt failure surface, with the characteristic SCC thumbnail initiation fracture evident. Such initial fracture surfaces can be caused by the original bolt manufacturing process, or stress concentrations induced at or near the bolt surface. Figure 10 shows a microscopic image of crack propagation spreading through a bolt from such a source location.
 |
 |
Figure 9: Characteristic SCC failure surface
Figure 10: Magnified image of crack branching
The results of a survey of a limited number of mines (12) reporting problems of this nature found that:
· The failures at all of the mines were limited to specific areas and time over a 6 year period (1994 through to 2000). With the exception of 1 mine, no broken bolts installed prior to 1994 have been reported.
· The only failures identified are those that fall out of the roof or that are significantly displaced out of the roof. There is a strong possibility that:
· Bolts have failed within the encapsulated section of the bolt and remain in place. (Confirmed by at least two incidents of broken bolts exposed in falls).
· Additional bolts prematurely broken within their free length but in roofs exhibiting shear, do not significantly displace as they are held in place by mechanical interlock.
· The mines showing higher frequency of bolt failure do not achieve full encapsulation and typically are noted in environments that would not have sufficient roof shear to prevent the bolts falling from the roof.
· Although detailed records are not routinely kept, there is evidence that premature failure may occur between weeks and years of installation. The oldest recorded bolts were placed in 1986 with failure probably occurring in the later 1990’s. There are other reports of bolts prematurely failing within weeks of installation.
· The level of visible corrosion of the roof bolts can be negligible at the time of failure of the bolt. This would be consistent with some failures occurring on defects that are either pre-existing in the bolt or else develop rapidly at a localised position on the bolt.
· With the exception of 2 mines the number of broken bolts identified is limited. The two mines where more frequent occurrences were noted both have thick coal roof strata and clay bands are present.
· 4 other mines with similar geology (thick coal or carbonaceous mudstone roof with clay bands) also reported some broken bolts.
· The issue of clay bands and thick coal may be a potential indicator of elevated probability of premature failure occurring however:
· There are mines in close proximity to those where significant failure has occurred, mining in very similar geology and stress conditions that have reported no failures.
· The clayband/coal lithology is noted for difficulty in achieving full encapsulation due to loss of resin into the strata. Typically bolts in this environment are not encapsulated for up to 500mm into the roof.
· Groundwater and chemistry issues have been identified as factors influencing the level of corrosion and stress corrosion. This includes alkaline and H2S environments as well as those that are acidic. Groundwater samples have been collected by the mine sites, but no systematic survey has been carried out.
· The geochemical environment related to clay bands is also conducive to various bacteria that can cause various forms of corrosion.
· Within the limitations of the survey already stated, it is evident that the bolt loading in areas of broken bolts include the possibility of shear with associated bolt bending. This is not to infer that the prematurely broken bolts uniquely result from shear deformation or bending. A percentage of collected bolts do not show any permanent strain that would indicate high bending loads. (Strains that exceed the plastic limit of the steel).
· The presence of clay bands, noted in several failure sites, are commonly associated with shear deformation of the strata. The shear deformation resulting in bending of the bolt that maybe within the elastic range of the steel. Bending of the bolts would result in higher tensional strains on one side of the bolt balanced by lower strains or even compression on the opposite side. Even where the axial load in the bolt is low, insufficient to deform a bearing plate, it is possible to have axial strains close to or even exceeding the plastic limit of the steel resulting from shear deformation.
The 44 broken bolts that were examined metallurgically, were of eight different types and included at least six different chemical compositions, indicating the widespread nature of the failures. However, most of the failed bolts showed a set of similar features, these being:
· an absence of significant necking in the vicinity of the fracture
· discernible bending in the vicinity of the fracture
· a fracture surface perpendicular to the axis of the bolts
· a discoloured region at the fracture origin
Apart from the bending in the vicinity of the fracture, most of the bolts showed minimal plasticity prior to fracture, this being evident from the minimal reduction in area at the fracture and the lack of tensile elongation. The absence of plasticity is to be expected when failure occurs from stress corrosion cracking.
For stress corrosion cracking to occur, all of the following factors must exist simultaneously:
· a susceptible material
· a corrosive environment
· an applied and/or residual stress
Removal of any one of these factors will eliminate the problem. There are thus a number of different ways in which the problem may be overcome, or at least alleviated, some of which are:
· reducing the strength of the bolts
· increasing the toughness of the bolts
· reducing the pre-stress on the bolts
· galvanising the bolts
· cathodically protecting the bolts
· use of a corrosion inhibitor.
Further studies since the completion of the Stage 1 project have involved the use of a German (DMT) developed ultrasonic bolt integrity test device. This allows non-destructive testing of bolts in situ. This is particularly useful to detect bolts that may have failed within the encapsulation horizon, and therefore not fallen from the roof, as with the initial samples recovered.
Trials of this device at a number of collieries have in fact revealed a significant number of broken bolts still in place as a component of a roof suppport regime. Whilst clearly still providing some degree of roof reinforcement, such bolts are certainly not providing the length of encapsulation intended.
It must be clearly stated that premature failure due to SCC is not necessarily a widespread problem. However it has been found to exist at over 15 collieries now, in most of the major coalfields and coal seams in Australia. Only relatively small numbers of failed bolts have been detected to date, but then only a very small section of the industry’s installed bolt capacity has been surveyed. The concern is particularly with the extent of failures that remain in the roof due either to encapsulation or borehole shear deformation constraint.
BOLT PERFORMANCE UNDER SHEAR LOADING
This project aims to investigate the role of reinforcement tendons (bolts or cables) in resisting shear across discontinuity planes. Such planes exist in most rock masses, either as joint sets at a range of different angles to the bolt direction, or as bedding planes, normally close to perpendicular to the bolt orientation. The latter situation is the most common form of rock reinforcement in coal measure strata. In such a situation, bolts play a major role in laminated roof strata to achieve beam building through the axial stiffness of the bolt. Directly linked to that role, the bolts also provide shear resistance to relative shear deformation between the layers of laminated strata, thus restricting bending of the individual or composite beam rock mass. The effect of bolt pre-tensioning is also believed to contribute to clamping the discontinuity surfaces together, thereby further generating frictional resistance to shear deformation and reduction of resultant beam bending.
This project is currently at the stage of construction of a laboratory test rig to enable two blocks of rock material to be loaded in test frame, with a bolt then installed across the discontinuity between the blocks, with varying angles of plane intersection, and with the ability to apply pre-tension to the bolt. The two blocks will then be able to be loaded in direct shear, resisted by the installed bolt system. Further details on the role of bolts in shear resistance, and of this project, have been provided by Hartman & Hebblewhite [5].
REFERENCES
1. Doyle P: Angus Place Colliery Case Study - Longwall production versus roadway support. (unpublished), 1998.
2. Hebblewhite B & Lu T: Geomechanical behaviour of laminated, weak coal mine roof strata and the implications for a ground reinforcement strategy. Int.J Rock Mech. & Min. Sc., 41 (2004), pp147-157.
3. Hagan P: Observations on the differences in load transfer of a fully encapsulated rockbolt. 1st Aust. Ground Control in Mining Conf., Sydney, Nov. 2003, pp 161-166.
4. Crosky A, Fabjanczyk M Premature rock bolt failure. ACARP Project C8008 Final Report.
Gray P, Hebblewhite B UNSW Mining Research Centre Report 3/02, April 2002,
& Smith B ISBN 0 7334 1948 8, 135p.
5. Hartman W & Understanding the performance of rock reinforcement elements under
Hebblewhite B: shear loading through laboratory testing: a 30-year history. 1st Aust. Ground Control in Mining Conf., Sydney, Nov. 2003, pp 151-160.