In this paper, we consider the problem of estimating the apparent slowness vector p of a plane P wave caused by a regional seismic event and recorded by a small-aperture seismic array. The case is considered when strong non-stationary and non-Gaussian random interferences act on the array sensors. In this case, the well-known estimate of wideband frequency-wave-number analysis (WFK) becomes ineffective due to large estimation errors. We have proposed three new algorithms for estimating the vector p that are robust i.e. resistant to changes in the statistical properties of the random interferences. They mainly use information about the slowness vector contained in the phases of the spectra of seismograms recorded by the array sensors. An intensive Monte Carlo simulation was carried out to compare the accuracy of the proposed phase-based estimates with the accuracy of WFK-estimate in the case when non-stationary and non-Gaussian anthropogenic interferences act on the array sensors. In the simulation we used synthetic mixtures of signals caused by a seismic event and anthropogenic seismic interferences. These signals and interferences were recorded by real small-aperture seismic arrays. It was shown that the proposed phase-based estimates of the vector p provide significantly better accuracy than traditional WFK-estimate in the case of strong anthropogenic interferences, and have approximately the same accuracy in the case of white Gaussian noise.
相似文献