Pseudo-random sweeps for built-up area seismic surveys

Abstract

Optimized pseudo-random sequences proved effective, both theoretically and empirically, in increasing data quality while reducing possible damage to buildings by vibratory sources during seismic reflection data acquisition. The linear sweeps commonly used for vibratory measurements can cause resonance in nearby infrastructure; hence, there is a potential for property damage. A pseudo-random sweep signal can be designed to decrease resonance effects and therefore reduce damage thresholds. The sweep sequences produced by simple random number generators, like the ones given by the vibrator manufacturers, have marked disadvantages, such as high correlation noise, large fluctuations in spectral amplitudes, and reduced seismic energy. To overcome some of these problems, an optimization process can be developed to produce pseudo-random sweeps for site-specific conditions. Two strategies are considered for preprocessing: crosscorrelation and deterministic deconvolution. Analysis of examples for an optimum pseudo-random sweep and simple pseudo-random sweep demonstrates that the total seismic energy is increased, while side-lobe energy is decreased and spectral fluctuations are reduced when the pseudo-random sweep is optimized. A limited-scope field test reveals that the peak particle velocity values are lowered substantially, while correlated and deconvolved records generated by the optimized pseudo-random sweep are of similar quality to a linear-sweep record. Application of optimized pseudo-random sweeps has the potential to increase productivity, while maintaining data quality and reducing resonance in surveys of built-up areas.