Underground » Strata Control and Windblasts
This document reports on the work undertaken during 1996 as the second part of a project aimed at developing new technology to aid the reliable and objective evaluation of roof conditions in underground coal mines. Of particular concern are mines with conglomerate roofs, as found in the Great Northern Seam in the Newcastle area of New South Wales where a number of serious accidents have occurred in recent times. Stage 1 of this project was reported in September 1993 (NERDDC Project C1620). That work aimed at demonstrating the effectiveness of an acoustic resonance technique supported with Digital Signal Processing (DSP) techniques that could later be implemented with the latest DSP technology into a real-time analysis tool to reinforce the miner's own subjective evaluation of the local roof condition. Although some degree of discrimination was achieved during Stage 1, the report (Jacka et al. 1993) recommended that further work was required in order to obtain results that could indicate whether the acoustic resonance technique of roof evaluation was worth further investigation.
CSIRO funded some further investigations in 1994, including modifying the instrument built during Stage 1 to allow separating the impactor from the accelerometer, as suggested in the Stage 1 report. The results obtained provided enough confidence for a further $50,000 to be sought from ACARP to extend the data analysis already done.
This funding was made available, leading to the work reported in this Stage 2 document. The anticipated outcome of Stage 2 was to give some degree of certainty whether a truly portable instrument using DSP technology could be built to adequately provide the required information on roof conditions in conglomerate roof as found in the Great Northern coal seam.
The initial part of Stage 2 concentrated on a detailed analysis of the data acquisition instrument.
New electronics were designed and built to enable greater dynamic range and sensitivity in our recording. Also, a simpler scheme was devised to replay and edit the data collected from the field trials.
Further investigations led to our replacing the original accelerometer which was built at the CSIRO, with a commercial unit, which is readily available, suffers less from internal resonances and which can be more easily characterised. All of these modifications to the basic data acquisition equipment enabled us to better analyse and characterise the system.
After much effort, we developed a sufficiently simple and reliable method of acoustic coupling between the roof and transducer. This included two sets of field trials at Wallarah Colliery.
The final field trial at Wallarah provided us with reliable data that have been used to further develop signal processing techniques. These now clearly indicate correlations between our results and those obtained by subjective roof evaluation by the local mine geologist.
Three parameters: morphology, decay time, and resonant frequency, are shown to relate to 'assessed safety'. The bulk of this report gives details of these findings, and recommendations are made for further research.
Previous work in this area is largely summarised in the 1985 report by Hanson (1985). In this report, Hanson postulated that decoupled slabs should have resonant modes and that when the slab is struck each mode should result in a 'harmonic disturbance decaying at an exponential rate'. He concluded that 'detached blocks in a mine environment
should have a preponderance of low-frequency resonances as compared with solid rocks'.
His experimental results showed that 'energy content in the 200 - to 1000-Hz frequency range was greater in drummy rock than in solid ones'. This work resulted in the design of a rock stability tester that relied on comparing the energy in two frequency bands.
Some improvements on this design have been proposed (Zamel & Kent 1990) but no commercially viable rock stability tester has resulted from this work.
Results
Techniques have been developed to measure accurately the response of a mine roof structure to an impact - normally a hammer blow. The measurement method promises to be easy to do in a mining environment.
Numerous digital signal processing techniques have been applied to the measured data with 'assessed' safety. The techniques include Fourier transforms applied to the full data set and to the tail of the impulse response, energy-time plots and the windowed Laplace transform - a technique developed explicitly for analysing the data in this report. The results of these analyses are summarised by three parameters:
- morphology - whether a resonance is detected and, if so, the number and type of resonances;
- decay time - which measures the rate at which energy in the resonance decays;
- frequency of the fundamental resonance - where the fundamental resonance is defined to be that with the lowest frequency.
Each of these parameters shows a correlation with 'assessed' safety. Of the three, the frequency of the fundamental resonance promises to be the most reliable indicator of roof stability. This conclusion is reinforced by the fact the at this result agrees with the theory presented in this report.
Hypotheses
- Non-detection of resonances implies safe roof
- If a fundamental resonance is detected then the competence of the roof increases as the frequency of the fundamental frequency increases
- The fundamental resonance of roof structures is proportional to its thickness and inversely proportional to its length squared
Conclusions
- For useable results the hit must be to, and the accelerometer attachment point must be on, the same rock structure.
- Good coupling between the rock and accelerometer is needed; poor coupling leads to the introduction of extra resonances which corrupt the measurement.
- Hanson's method, as originally proposed, can lead to incorrect conclusions.