A Vibratory Seismic System For High-Resolution Applications

A vibratory seismic system has been developed by the authors, utilizing an engineering seismograph a&<br>a low-cost vibratory source. The project was a collaboration between the Geophysics Group at the Waterloo Centre<br>for Groundwater Research (WCGR) and the Audio Research Group in the Department of Physics at the University<br>of Waterloo. The system was developed in the winter of 1995 and preliminary field work was completed in the<br>spring of 1995. The source consists of an electrodynamic vibrator driven by a portable car audio amplifier. Source<br>excitation signals are generated with a sound card for a personal computer (PC). Software was developed to<br>generate excitations and perform crosscorrelation on the data. The source was well instrumented during preliminary<br>field tests, with accelerometers on both the reaction mass and moving coil, and a strain gauge between the coupler<br>and the moving coil to monitor the force acting on the coupler. Three different coupling devices were tested: a<br>short auger, a long auger and a baseplate. The preliminary testing was done at the Bauer Warehouse area on the<br>University of Waterloo’s North Campus, where both surface and downhole data were acquired. Continuous core<br>from a drill hole at the site shows that the stratigraphy consists of a series of glacial tills and alluvium underlain<br>by limestone of Silurian age below 68 metres. The water table was about 1 metre below ground surface at the time<br>of the survey.<br>For the surface tests, 24-channel common shot records were recorded. Twenty-four 100 Hz geophones<br>were spaced 2 metres apart with a source offset of 2 metres from the end of the array. For the downhole tests, an<br>8-channel hydrophone array was used to acquire a vertical seismic profile (VSP) from ground surface to 70 metres<br>depth. Hydrophones were deployed every metre with a source offset of 20 metres from the borehole. A number<br>of acquisition parameters were used and compared: sweep range, sweep duration, sampling rate and stack count.<br>Cosine-tapered linear sweeps as wide band as 50-1000 Hz were used. As well, pseudo-random maximal length<br>sequences (MLS) were tested. The 14-bit sequence used was “clocked out” at rate of 5 kHz, then bandpass filtered<br>from 100-1000 Hz prior to being used as an excitation. The results have been very encouraging; the vibrator<br>compared favourably to both a 12-gauge shotgun and a hammer and plate as a source. The baseplate was found<br>to be the preferred coupler, although the long auger generated a correlation wavelet with a higher centre-frequency<br>than the baseplate on VSP recordings. The short auger produced relatively poor records. Analysis of the strain<br>gauge response suggests the short auger applied the least force to the earth. Very good correllograms were obtained<br>using: 8 second sweeps from 50-1000 Hz, stacked 16 times, and recorded at a sampling rate of 2 kHz. The filtered<br>MLS excitation was found to be just as effective as sweeps of similar duration, stack count and frequency content.