Fundamental investigations of new source signals for seismic-while-drilling exploration ahead of the face in conventional and mechanised tunnelling

Tunnel Seismic While Drilling (TSWD) is a well established and recognised method for geophysical exploration in mechanised tunnelling. The special feature of this method is to use acoustic emissions from a tunnel boring machine (TBM) as a seismic source. In order to further develop this methodology in mechanised tunnelling and also transfer it to conventional (NATM) tunnelling, a large-scale laboratory test was carried out on a disc cutter test bench and various machines used in conventional tunnelling were tested at the Zentrum am Berg (ZaB) for their suitability as seismic sources. The tested devices included a mini excavator with an attached hydraulic hammer, a diesel-powered single-armed small drilling jumbo (Tamrock) and an electrically powered two-armed drilling carriage (Epiroc). A suitable pilot sensor (±500 g accelerometer) position to pick up the source signal was determined based on an analysis of hammer stroke signals recorded at various test positions on the excavator and the Epiroc drilling jumbo. The hammer of the respective devices provides a suitable option for a pilot sensor placement since the waveform recorded there is characterised by a short duration, high amplitude at its onset and reproducibility in terms of waveform, spectrum and polarisation. A direct comparison of different geophone types and couplings to record the waves transmitted in the rock near the source shows that the best signal quality was obtained with borehole geophones. They are characterised by high signal amplitudes and a clearly defined spectrum without any surface wave influence. The weakest results were obtained with single geophones in shotcrete. The coupling of surface geophones in plain rock and shotcrete was evaluated by examining their amplitude decrease over increasing distance. This shows that most of the transducers considered (in wooden cubes and single geophones) experience a smaller amplitude decrease than expected by spherical divergence in absence of surface waves. However, the amplitude strength decreases by a factor of 10 in shotcrete. A pilot signal analysis of a percussive source (excavator with hydraulic hammer) in contrast to rotary percussive sources (drilling jumbos) reveals clear differences. Although individual impacts are present in all signals, the level of background noise is higher for rotary percussive sources and shows parallels to a TBM pilot signal. Both minimum delay transformation and conventional cross-correlation can be used for the correlation between pilot and receiver traces, with either method delivering satisfactory results. Prior integration of the pilot sensor trace or a derivation of geophone traces is not necessary. Bandpass filters and predictive or spike deconvolution demonstrated improvements in the seismic trace quality. The frequency bandwidth transmitted through the rock mass differs for the three sources tested and ranges from 50 to 500 Hz for the excavator with a hydraulic hammer, from 100 to 400 Hz for the Tamrock drilling vehicle, and from 200 to 500 Hz for the Epiroc drilling jumbo. Thus, the tested sources offer sufficient signal bandwidth and the necessary higher frequency components for seismic exploration purposes. An overall comparison of the resulting seismic traces, taking into account the tested spatial range, showed that the excavator and the Epiroc drilling machine can be classified as suitable source signal generators, while the Tamrock device delivered less promising results. With regard to the emitted waveforms, both compressional and shear waves were detected during usage of the excavator, exhibiting clear parallels in their spatial orientation to the characteristics of a single force source. Due to a limited angular range of observation (sideways of the source), only shear waves were detected for the two drilling machines. In order to investigate the suitabi