Waveform Retrieval From The CASS Experiment Based On Time-Varying Narrow Band-pass Filtering Technique

As a signal source in the Precisely and Actively Seismic Monitoring system,the Controlled Accurately Seismic Source(CASS)can stimulate the low-power signal repeat for a long time. At observation stations,the signal is recorded by high-precision Digital seismometer,and extract source signal through different propagation path. This kind of signal is suitable for surveying crustal structures, especially for monitoring the temporal variation of crustal media characteristics.However, With the increase of the propagation distance, useful signal is masked in strong noise as the energy intensity radiated by it is quite small. Strong noise has a great impact on identifying phases and picking accuracy of travel time.In this paper, we firstly introduce the CASS system, features and signal characteristics.In addation, we explain matched filtering method to realize the signal detection and wave transform.Taking into account the characteristics of vibrator sweep curve and signal level is far below the record of the background noise level in the observation station, we propse a new method for time-varying narrow band-pass filter design to achieve precisely filtering effect. and retrieve CASS chrip signal by cross-correlation to identify the major seismic phases traveling through the crust.This paper explain a method for designing digital narrow band-pass filters in the mathematical derivation. Amplitude-frequency and Phase-frequency characteristics of the filter is mapped in this section.In addation,we design zero-phase filter to avoid signal distortion.Finally, The effectiveness and practicability of the filter in suppressing noise by applying it to numerical simulation by programming.In the process of Xinfeng River reservoir experimental data, Three-phase asynchronous servo motor servo motor is limited by the precision,which load deviation between theoretical frequency and the actual frequency. In order to solve the above problems, We propose a sweeping curve correction model to obtain actual sweep curve.Moreover,when we filter the signal in the observation station,we need to estimate delay value in advance.we apply global phase scanning method to obtain delay value of sweep signal through different propagation path.Then,As start of sweep curve using different delay value,we filter the signal using time-varying narrowband filter. At the same time, because of different travel time between different seismic phases,When we retrieve certain impulsive seismic waveforms by using time-varying narrow band-pass filter,which suppresses the other seismic phases. The effectiveness and practicability of the new optimized algorithm in extracting chrip signal,raising the resolution phase and seismic phases separation is proved by applying it to model and actual seismic data.And, retrieving impulsive seismic waveforms at station in 200 km reflectes the capacity of weaker signal detection.