In this thesis, I focused on the identification of secondary waves at the DOBS seismic station, where improper installation of the station shaft leads to resonant distortions in the seismic signal. Accurate determination of the onset of secondary waves is crucial for identifying earthquake characteristics such as hypocenter location and magnitude.
To process the data and reduce the effects of resonance, I applied digital filtering. First, I analyzed several recordings of local earthquakes from the DOBS station and compared them with data from other, more reliable stations. Based on the findings, I designed and implemented a digital resonant filter with an infinite impulse response, tailored to eliminate frequency characteristics specific to the DOBS site. I further enhanced the implementation with a multi-resonant filter composed of several second-order filters connected in series. The frequency parameters for each filter were determined using the power spectral density and the sos2tf function.
The results of the analysis show that the basic resonant filter already significantly improves the identification of secondary waves, while the multi-resonant filter further reduces the influence of resonance. In certain cases, the best identification of the onset of secondary waves is achieved through a combination of both filters. In this way, I improved the accuracy of arrival time estimation of secondary waves across all three components of the seismic record. The results demonstrate that with appropriate filtering, it is possible to successfully compensate for the deficiencies caused by improper installation of the seismic station.
|
