Ground control research continues to be a major part of the portfolio reviewed by SMRAB. The majority of the work was carried out by Rock Mechanics Technology Ltd (RMT) but HSL and the University of Nottingham have also worked on certain aspects. Progress made during 1999 is described under various topic headings below. In some cases, the work programmes described drew on more than one project.
During 2000 RMT completed and submitted the final report on the following research project related to mine safety:
This was a collaborative Project with other European partners and the joint final report is being compiled by the Project co-ordinator (DMT, Germany). Two HSE subprojects under this project were also completed and the final reports submitted. These were:
Also final reports were issued on 2 projects completed in 1999, namely:
Towards the end of the year, in November 2000, RMT commenced work on two new ECSC part funded projects. Current ongoing ECSC ground control research projects are as follows:
RMT submitted the final report on a SIMRAC (South African Safety in Mines Research Advisory Committee) research project entitled, "Assessing and evaluating acoustic techniques to test roof conditions in coal mines" and were awarded one new SIMRAC contract:
The main technical advances during the year have been made in the following subject areas, each of which is discussed in more detail below:
In-situ Integrity Testing of Rock Reinforcement
The previous research Project on "Ultrasonics for Integrity Testing of Rockbolts" concluded that improved equipment could become available, this being a portable version of the Wavemaker equipment developed at Imperial College. This new equipment was tested on rockbolts embedded in a concrete block in the laboratory and underground at Middleton Limestone Mine. Good results were obtained in the concrete block with a 2m bolt and two partially embedded bolts, but no reflection was obtained with a heavily cranked bolt. These results were independent of the crimp angle of the bolt end. At Middleton Mine good results were obtained with 1m and 1.8m bolts, but the end of a 2.4m bolt was not detected. The main problem at this length appears to be that the bolt end reflection is masked by many higher amplitude early reflections. This phenomenon is now being investigated with a view to either modifying the amplification system to reduce the masking effect or to eliminate the early reflections in some other way.
Earlier work on the Radio Frequency Resonance system for detecting broken cable and flexible bolts in-situ achieved good results in limestone and slate but not in shaley mine roofs. A further series of tests was carried out at Meyreuil Colliery in France which has a limestone roof. This was very successful with a total of 4 broken flexible bolts being detected. This work indicated that the problem in shaley roofs is likely to be related to the frequency range of the instrument used. A new instrument has been designed and is now under construction which should overcome this problem. The target range for the new instrument is 20kHz to 20 MHz in three switched and tuneable ranges.
Ground Control Instrumentation
Remote Reading Telltale System
Underground application of the system was delayed for the first nine months of the year whilst awaiting European Intrinsic Safety Approval. During this period the user handbooks were written and further software development undertaken. Shortly after approval, a system comprising ten transponders, an underground communications unit, a surface barrier unit and a surface communications unit was installed in a working mine in Germany. Transducer installation was very successful, showing that they are sufficiently robust and easily installed. However two separate communication problems were identified. The first was associated with the Underground Communications Unit locking onto individual transducer outputs. This was easily remedied by a capacitor change in the unit, to adjust the phase locked loops. The second problem was associated with the distance over which the Underground Communications Unit had to communicate with the surface computer. This resulted in lower received signal levels than originally designed. This problem was resolved by boosting the sensitivity of the receiving circuits in the surface and underground communications units and was possible due to the low levels of noise on the cable. Both modifications were achieved without any requirement to modify the IS certification. The modified system will be fully commissioned underground during the first semester of 2001.
New Remote Reading System Transducers
Several alternative displacement transducer configurations have been designed to interface to the standard remote reading system. These are all variants upon the theme of a single wire extensometer with various gearing arrangements to allow a variety of displacement ranges and sensitivities. They are suitable for measuring such geotechnical parameters as rib displacement and roof to floor closure, both of which can be indicators of ground control hazards in different circumstances. These transducers utilise exactly the same electronics as the remote reading dual height telltale system and could therefore be included into the Intrinsic Safety Certification relatively easily.
A series of prototype solid-state load cells suitable for interfacing to the remote reading system has been tested in the laboratory to establish the basic design. The proposed load cell is essentially a parallel plate capacitor with a deformable solid dielectric. The target parameters are a 500kN range with a diameter of 200mm. This has been achieved, after initial problems with creep and plate deformation were overcome. It is proposed to construct a small batch for underground trials. In an alternative approach, the feasibility of incorporating an inductive transducer into a standard Bourdon-type pressure gauge is being investigated.
A further application of the remote reading telltale transducer principle is currently being investigated for monitoring the height of caving from surface boreholes for sub level caving and longwall systems. This would be an alternative to the use of Time Domain Reflectometry.
Neural Networks Analysis of Rockbolt Data
HSL's contribution to ECSC projects in this field has been an investigation of the role that artificial neural networks might play in the analysis of rockbolt data. A literature survey revealed that neural networks had not been applied previously in this area. Subsequently rockbolt data were collected from a number of mines in the United Kingdom. These data were manipulated and presented to a series of neural networks. As a result, three neural networks with differing roles were developed: one was used to identify distinct clusters of data and roof behaviours associated with the relevant sensors; another was able to predict subsidence without the need for geomechanical data; the third was able to predict, from roof movement measurement, whether additional support would be required. In this last case, the accuracy of the prediction was almost 80%.
Risk assessment techniques have been investigated under previous ECSC funded projects to provide a tool for identifying falls of ground hazards. Over the last year these have been developed as a means of assisting operational personnel to select appropriate remedial support when required.
A risk assessment procedure was developed and applied to a new area of a UK coal mine. The assessment includes the potential hazard of roadway instability associated with rib failure and the potential for rib collapse as well as roof fall hazards. The assessment also provides information relevant to Regulation 10 of the Control of Ground Movement in Mines Regulations 1999, regarding the adequacy of ground control measures. The main risk factors which were identified as being associated with potential rib failure at the site were excessive roadway height and poor visual rib condition.
An underground survey was undertaken in a newly driven maingate by the mine rockbolting personnel using specially designed survey sheets. The sheets were then passed to RMT for analysis and risk classification. The results from this analysis were intended to allow the colliery to re-assess any individual areas identified as at risk and, if necessary, take remedial action in order to reduce the likelihood of ground failure. Several areas at risk for rib failure were identified and remedial action comprising spot bolting, monitoring and appropriate signage was recommended. As a result of the assessment the rib bolting pattern for future rockbolted drivages in this area of the mine was amended to account for possible increased roadway height.
Laboratory testing of support systems and reinforcement consumables
Revision of British Standard 7861:1996
The Rock Bolt Research Liaison Committee has set up a Working Party to draft a revision of British Standard 7861(1996), 'Strata Reinforcement Support System Components Used in Coal Mines'. Such a revision is necessary due to several factors, including:
The Working Group has accepted in principle that the Double Embedment Test (DET) for system performance should be replaced by the Laboratory Short Encapsulation Pull Test (LSEPT) and RMT have drafted the new Appendix describing the LSEPT test procedure and made initial recommendations on appropriate acceptance criteria. Much of RMT's work on laboratory testing of support systems during the year has been directed at achieving a repeatable and well documented procedure and undertaking a large number of tests to determine appropriate acceptance criteria.
Analysis of previous LSEPT tests on AT bolts and resin, all or which had been undertaken at 250mm embedment length, indicated that, in normal circumstances, the system can reach an applied load of 200 kN prior to yield of the bond. As 200 kN is usually adopted as the maximum test load, it was necessary to undertake a testing programme to examine the effects of reducing the embedment length in order to determine an optimum embedment length for the test. LSEPT tests have now been completed at embedment lengths of 200mm, 175mm, 160mm and 150mm. Analysis of the tests indicated that the optimum embedment length would be 160mm. This would ensure that bond yield was achieved with a standard AT system but was long enough to ensure adequate resin mixing and good hole rifling. This method has been recommended to the Working Party.
Considerable work was also undertaken to achieve a satisfactory method of ensuring repeatability of the core used in the tests. Several alternative methods were trialled. The adopted system uses a high yield threaded bar embedded in hand mixed AT resin, with a rifled hole drilled with standard consumables. It was found that this system ensures bond failure at the resin/rock interface and so achieves a test of the bonding properties of the rock. The core is acceptable if the results of a series of tests fall within a pre-determined envelope on the load /displacement curve.
A second series of LSEPT tests was also undertaken on Galvanised AT bolts/AT resin at 250mm embedment. This confirmed that the load transfer characteristics were comparable or better than the standard AT system which was tested at the same time as a control. Further tests are to be undertaken at 160mm embedment. Double embedment tests were also undertaken and these confirmed that the galvanised system satisfied the current BS7861 performance criteria for the axial double embedment test.
Enhanced Standard for Mining Rockbolts
HSL is also contributing to the work for the proposed modifications to the British Standard for mining rockbolts. This involves research on the derivation of fracture toughness from Charpy impact test data and on establishing the suitability of galvanized rockbolts for use in corrosive applications.
The work has concentrated on the two common types of rockbolt used in the UK for roof support, known as "Polish" and "Co-steel or AT (type 2)". However, a limited amount of work has been carried out on continuously threaded German (SAT KT) and induction hardened Tempcore rockbolt materials. A substantial amount of Charpy impact and fracture toughness data has been generated for all materials and, when combined with data from a previous project, it has been possible to derive and validate a correlation between the two parameters (Charpy energy and fracture toughness).
Based on the fracture toughness values obtained during this work and the appropriate stress intensity factor solution, values of fracture stress were calculated for various critical defect depths. Assuming rockbolts are subjected to stresses approaching the yield strength in service, critical defect sizes were calculated for each level of toughness. It was then possible, using the Charpy/fracture toughness correlation, to relate the average or lower bound critical defect size to the Charpy impact energy. In this case a material having a toughness in the region of 27J was found to exhibit a critical defect size of ~3mm. This value has been recommended for inclusion in the revised standard.
The use of galvanized bolts in corrosive environments was recommended in previous work. The current project has confirmed that the pickling and galvanizing process itself does not affect the fracture toughness adversely and hence the use of galvanized bolts in corrosive environments is to be included in the revised British Standard.
Alternative Reinforcement Consumables
A fully threaded reinforcement bar (SAT KT) is currently being used at a number of UK mines for rib reinforcement and trials are planned by the manufacturer and operator to assess the bar's suitability for roof reinforcement. The dimensions of the bar are very similar to AT bar allowing it to be installed in the same diameter holes. Three LSEPT tests were performed by RMT on the system together with an AT system control. The results indicated that the performance compares favourably with the AT system, both in terms of stiffness and bond strength.
Double embedment tests were also performed on the SAT KT system plus an AT system control test. These results satisfy the requirements of BS7861 for double embedment tests. Also they indicate a much higher system stiffness (greater than 3 times) for the new bar than for the AT control. This indicates that large discrepancies between systems shown by double embedment tests do not necessarily reflect differences in load transfer actually achieved in rock
A study was undertaken of cable bolting systems currently being applied in South African coal mines where approximately 25,000 are installed each year. The study found that most systems use plain single strand (7 wire) which may have one or two 'bulbs' per cable bolt length. The bulbs are located at the top end of the cable to ensure proper mixing of the resin anchor. The cables are pre-tensioned and may be pre- or post- grouted relative to this operation. Laboratory testing of one of these systems was undertaken to assess whether they might be suitable for European conditions.
It was concluded that the system was a poor hybrid of the birdcaged and flexible bolt systems and did not benefit from the advantages of either. This system would not be appropriate for European high stress situations.
Standing Support Systems
Further laboratory compression tests were undertaken on Link-N-Lock cribs. These were of standard dimensions and constructed from Sycamore. This follows suites of tests in late 1999 on a number of alternative configurations and materials (English Ash, English Oak and English Beech). The cribs performed in a similar manner and reached an average maximum load of 207 Tonnes. The mean load at 50mm displacement was 133 Tonnes. This is a similar performance to previously tested standard Link-N-Locks constructed from English Ash.
Rib Control Systems
Work has recently begun on investigating the applicability of sprayed polymer systems for coal rib control. The first phase of the study is confined to the desktop and will examine the published properties of the materials to determine if they are likely to be suitable and, if so, in what type of configuration. If the conclusion is positive then the next step would be laboratory testing, followed by underground trials if operators believed that a system could be cost effective.
Numerical modelling for mine design
RMT's research in this area has involved development of numerical models to examine the effects of multiple seam interaction, fracturing of strata above longwall panels leading to water inflows and interaction due to underworking. Each of these mine design problems requires a different modelling approach.
The method of analysis adopted for multiple seam interaction involves the use of a large scale three dimensional boundary element model to examine the stress distributions below worked areas. These stress distributions have then be used as boundary conditions for a detailed two dimensional FLAC analysis of roadway behaviour. This approach has now been applied to several different sites with success.
When examining the extent of fracturing above longwall faces, the strength and failure properties of the rock must be taken into account and so the large scale three dimensional boundary element model is not suitable. For this reason the computer code selected was FLAC. Geometrically, the modelling was restricted to two dimensional cross-sections, mainly transverse sections across one or more longwall panels. Using this method it was feasible to generate and run models simulating the extraction of several panels in sequence. Site specific strata sequences, including cross dips and unconformities could be incorporated. The results showed failure of strata extending large distances into the roof strata above the panels. The extent of strata failure and fracturing generated by the model was sufficient to explain the ingress of groundwater encountered at the site considered. For constant strata properties, the extent of fracturing was strongly related to panel width. Reducing the panel width provides the primary way to reduce the extent of fracturing and the risk of groundwater. This is consistent with existing guidelines.
Examination of the effects of underworking requires a combination of the 2 methods described above. It is necessary to apply a 2 stage modelling technique, moving from the large scale to the smaller scale to determine more precise roadway behaviour. However the large scale behaviour must take account of rock fracturing and so the use of FLAC is necessary. For the smaller scale model FLAC can also be used but alternative codes such as UDEC which deal with the behaviour of discrete blocks may be appropriate. This methodology has been applied to a current UK mining case and feedback experience of actual conditions whilst driving above the old goaf is now awaited.
In-Situ Stress Characterisation through the Kaiser Effect
The Kaiser Effect is an Acoustic Emission (AE) phenomenon briefly defined as the absence of detectable AE until the previously applied stress level is exceeded. By measuring the increases in Acoustic Emission activity during sample loading it is possible to determine the previously applied maximum in-situ stress in the direction of loading. The testing of core obtained from a vertical borehole would therefore give the maximum previously applied vertical stress. For the horizontal stress, rock samples could either be obtained from horizontal holes or samples could be sub-cored in the desired horizontal orientation to obtain information of the stress magnitude in that direction. With this, the opportunity for the rapid and economic determination of the in-situ stress in rock could become possible.
Samples from two collieries were initially tested by RMT. Both sets of samples came from vertical roof cores and were in close vicinity to a previous stress measurement site. Three samples from Longannet Colliery were tested, all from the Calmy Limestone found in the near roof strata above the Upper Hirst Seam. The measured resolved vertical stress was 11.6MPa. Two of the three AE samples gave results within 1MPa of the measured value, one of these two samples generated a clear 'good' Kaiser Effect, the other was less clear and considered 'fair'. Three samples of sandy siltstone were tested from Riccall Colliery above the Stanley Main seam. The measured resolved vertical stress was 26.7 MPa. A value of 27MPa was indicated in one AE test which was considered to be 'fair'.
Two stress measurements were undertaken at the 83 Level of a South African gold mine indicating a resolved vertical stress of 78.7-83.3MPa, maximum horizontal stress of 49.5-48.9MPa and minimum horizontal stress of 23.4-24.7MPa. Core was obtained from the stress measurement borehole at the horizon of the measurements. A total of 11 samples were obtained from the vertical stress direction. Four of these showed clear 'good' Kaiser Effects indicating a vertical stress of 81-88MPa. Two samples were obtained from the maximum horizontal stress direction, one indicating a 'fair' Kaiser Effect at 49MPa. Samples sub-cored along the direction of the minimum horizontal stress component were heavily disked and therefore unsuitable for testing.
The initial results from the coal measures and gold mine strata indicate that the Kaiser Effect can indicate the stress to which the samples have been subjected, however it was not observed in all samples and further research is necessary.