Improving coal seam model accuracy for longwall automation by in-seam borehole radar imaging and data integration

Longwall automation depends on the development of automatic coal seam (horizon) tracking and lateral guidance systems to maintain mining only the coal seam while steering towards the desired target. It requires spatial accuracy for local changes in a horizon's depth in the order of 10 cm, and plans positions of potentially impeding structures (faults and rolls) with a spatial resolution or wavelength of a few metres.

This project ran borehole radar (BHR) in in-seam drill holes to accurately locate the seam's roof or floor. This will fill the gap between coarse exploration data (drilling, borehole logging and seismic) and the detailed seam knowledge (ground penetrating radar - GPR, thermal and optical) at the longwall face and gate roads to secure an accurate horizon model that is useful for longwall machine guidance.

Both numerical modelling and field data (from South Africa and China) was used to demonstrate the feasibility of using in-seam BHR survey for imaging the seam floor/roof and associated structures. The results clearly show:

• Seam floor and roof can be confidently/accurately estimated by in-seam borehole radar surveying. In general, the estimated errors of seam positions reduce with the increase of BHR frequency, increase with the increased complexity of the model such as the existence of faults, and are not significantly affected by dipping or curved boreholes;

• To maximise the separation of the direct arrival and the seam reflections, and hence improve the seam floor and roof estimation accuracy, the following BHRs mapping - the central frequency of BHR > 100 MHz, and the Tx-Rx offset < 1 m - should be used for seam mapping.

• The existence of a roadway may affect the accuracy of the estimated floor and roof near the collar due to the diffractions from the roadway, but they can be improved by suppression of the diffractions through data processing. In addition, these diffractions can be used to estimate the coal seam EM wave velocity, which can be used for conversion of the roof/floor reflection times to distances for reconstruction of the coal seam floor and roof.
• For low frequency BHR data, the data processing techniques such as deconvolution can be used to increase the data frequency bandwidth, improve the data resolution and reveal possible reflections of the seam floor and roof.
• The variation of the direct arrival time is a result of EM wave velocity change and can be used to correct the estimation accuracy of the seam floor and roof along the in-seam hole.

These results demonstrate in-seam borehole radar surveying can be used for accurate mapping of coal seam roofs and floors, which can be used for both longwall automation and in-seam borehole drilling.