Materiały konferencyjne SEP 1992

464 Underground Exploitation School '92 In t he p1 ane of t he seam (SH-po 1 ar i sat i on, Love wave), and the third tracę shows the particie velocity vertical to the seam (SV-po Iari sat ion, Rayle i gh wave agai n). The d i stance from t he event source to the geophone probe was about 148 m. The onset of the p-wave, refracted at the hi gh ve ł oc i ty łayers, is too smali to be recognized in the traces. The main signal begins with the refracted s-wave. The dispersion of the waves is obvious within the whole signal: at the beginning the signal contains Iow freąuencies, towards the end of the signal the freąuencies become much higher. The biggest amplitudes are due to the airy phase of the first Love wave mode. But also Rayleigh waves (P/SV polarisation) contribute to the seismograms. On the right side figurę 13 shows the freąuency contents of the signals. The Rayleigh waves (first and third tracę) have signłficant energy from 100 Hz up to 1000 Hz. Higher freąuencies are filtered out by the anti-alias-filter. The energy of the Love wave component (second tracę) is concentrated around 5S0 Hz, what is some what higher than the average airy phase freąuency of 450 to 500 Hz. More information about the dispersion of the seam waves can be obtained from dispersion analyses uslng multiple filter techniąue. In figurę 14 the contourlines very clearly show the dispersion behaviour for the first Love wave mode, which contributed mostly to the energy of the seismograms during the whole measurement. By comparing numerlcally calculated dispersion curves, based on a modela with dispersion analyses of real data, it is posslble to find a reliable velocity model for the measuring site. The informations about the seismic velocity of the different layers and about the kind of waves to be recorded are needed to localize the events. Because of the great number of events, the localization process has to be done automatically by a computer. Therefore programs had to be developed for the extraction of wave arrival times out of the digital seismograms. Afterwards classical localization rout ines could be used to determine epicentres. Figurę 15 shows a typical epicenter distribution for one day of registration. Like in former investigations with localization of only few events nearly all of the epicenters are