KR101914657B1 - Method for extracting phase and amplitude information of seismic signal from seismic ambient noise - Google Patents

Abstract

The present invention relates to a method for extracting both the phase information and the amplitude information of seismic signals by using seismic ambient noise. According to the present invention, in order to solve the problems of the prior art, in which no technique for extracting both the phase information and the amplitude information of seismic signals has been provided, provided is a method for extracting the phase and amplitude information of seismic signals by using seismic ambient noise, in which by using three seismic interferometry techniques of deconvolution, coherency and cross correlation, a series of processes for extracting the phase and amplitude information of impulse response functions (IRFs) from seismic ambient noise, which is always recorded in a seismic station regardless of seismic activity, is configured to be performed by a computer or dedicated hardware such that both the phase information and the amplitude information of the seismic signals can be efficiently extracted from the seismic ambient noise with respect to a low earthquake area and the extracted data can be used to analyze and study seismic ground motion. The method for extracting the phase and amplitude information of seismic signals by using seismic ambient noise comprises: a data collecting step; an impulse response function extracting step; a phase and amplitude information calculating step; and a result display step.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for calculating phase information and amplitude information of a seismic signal using an earthquake background noise,

The present invention relates to a method for extracting both phase information and amplitude information of an earthquake signal using seismic ambient noise, and more particularly, to a method for extracting both phase information and amplitude information of an earthquake signal, In order to solve the problem of the prior art in which only the phase information of the seismic signal is limited due to the failure to provide the extracting technique, seismic interferometry technique is used to continuously measure A series of processes for extracting the phase and amplitude information of impulse response functions (IRFs) from the recorded earthquake background noise are performed by a computer or dedicated hardware. And a method for calculating phase and amplitude information.

In addition, the present invention is configured to efficiently extract both the phase and amplitude information of IRFs from the background noise as described above. Thus, even in a low earthquake region, And to a method for calculating the phase and amplitude information of an earthquake signal using the earthquake background noise configured to contribute to the analysis and research of the earthquake.

Recently, natural disasters such as extreme weather are increasing due to climate change such as global warming. In addition, earthquakes are increasing worldwide, including Korea.

In other words, it is important to anticipate and prepare for the occurrence of earthquakes in advance, as earthquakes can cause more damage to lives and property than other natural disasters. In Korea, there is no large-scale earthquake in the past, Earthquake of 5.8 magnitude occurred in Gyeongju on September 12, 2016 and 5.4 magnitude earthquake occurred in Pohang on November 15, 2017, and the earthquake continued until recently. The risk is increasing.

Therefore, in order to prepare for the increasing risk of earthquakes, there have been various studies to detect or predict the occurrence of earthquakes in the past. It has been suggested.

More specifically, as described above, as an example of the prior art for detecting and predicting earthquake occurrence, for example, according to Korean Patent Registration No. 10-1184382, a free space including a horizontal or slope condition The seismic measuring apparatus and the installation method of the apparatus are provided so as to be able to accurately and accurately measure the results of the earthquake measurement on the ground because the stable and reasonable installation and operation are possible on all the surfaces of the condition.

Also, as an example of the prior art for detecting and predicting the occurrence of an earthquake as described above, for example, in Korean Patent Registration No. 10-1523355 and Korean Patent Registration No. 10-1028779, And a P-wave is detected based on the variation in the time-frequency domain and the threshold value in which the background noise is reflected, thereby to reduce the false detection rate due to the background noise. And a method for detecting the seismic early warning and accurate seismic factor analysis is disclosed in Korean Patent Publication No. 10-1520399 by removing the measurement noise and background noise at the time of collecting the seismic data The present invention relates to an apparatus and a method for removing noise of a seismic wave using micro-positional alignment polymerization.

As described above, various techniques for detecting and predicting the occurrence of the earthquake have been proposed. However, the conventional techniques have the following problems.

In other words, the recent magnitude and frequency of earthquake occurrence has increased the necessity of studies on the earthquake occurrence environment and the characteristics of earthquake disasters on the Korean peninsula. However, since the average seismic activity is low in the Korean peninsula region, There was a problem in that it was relatively insufficient.

If earthquake background noise recorded at seismic stations at any time can be used as seismic observation data regardless of seismic activity, it is expected to help secure the base technology for studying seismic propagation characteristics and crustal structure of the Korean peninsula. The prior art contents of the prior art mainly concentrate on improving the accuracy of seismic detection by separating seismic waves and background noise and eliminating noise. However, it is not considered in construction of seismic observation data necessary for analysis of seismic occurrence There was a problem.

In addition, since the seismic signal extracted based on the background noise of the earthquake is influenced by the seasonal change and the spatial distribution characteristic of the background noise source, understanding the regional and temporal distribution characteristics of the background noise source causes the reliability of the extracted seismic signal However, the above-mentioned prior art contents merely focus on removing the background noise, and the meaning of such an earthquake background noise has not been considered.

In addition, the study on the correlation analysis of the background noise of the earthquake has recognized the possibility of applying the seismic data in the area with the small earthquake, and since the mid 2000s, the interest from the identification of the cause of the background noise to the derivation of the velocity structure based on the observation data However, in the past, only the extracted phase information of the seismic signal has been utilized. Furthermore, the technique capable of extracting the amplitude information in addition to the phase information of the seismic signal has also been studied as a seismic wave propagation characteristic including both the perception speed and the damping structure It is still in its infancy, and no standardization work has yet been achieved.

[Prior Art Literature]

1. Korean Registered Patent No. 10-1184382 (2012.09.13.)

2. Korean Patent Registration No. 10-1523355 (May 20, 2015)

3. Korean Patent Registration No. 10-1028779 (April 05, 2011)

4. Korean Patent Registration No. 10-1520399 (Aug.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a technique for extracting both phase information and amplitude information of an earthquake signal, In order to solve the problems of the conventional art, there has been proposed a method of detecting impulse response functions from seismic background noise recorded at seismic stations at all times regardless of seismic activity using seismic interferometry. The present invention is to provide a phase and amplitude information calculation method of an earthquake signal using an earthquake background noise configured to be performed by a computer or dedicated hardware by a series of processes for extracting phase and amplitude information of an IRFs.

Another object of the present invention is to efficiently extract both phase and amplitude information of IRFs from an earthquake background noise as described above so that data extracted from an earthquake background noise is also used for a low earthquake region And to provide a method for calculating phase and amplitude information of an earthquake signal using an earthquake background noise configured to contribute to the analysis and study of earthquake ground motions.

According to the present invention, the phase of impulse response functions (IRFs) from seismic ambient noise, which is always recorded in a seismic observation station regardless of an earthquake activity, In the phase and amplitude information calculation method of an earthquake signal using an earthquake background noise for extracting amplitude information, a process of collecting data on an earthquake signal including an earthquake background noise recorded at a seismic observing station installed in each region is performed A data collecting step; The seismic interferometry technique is used to extract impulse response functions (IRFs) between each seismic station for the seismic background noise collected at each seismic station using the seismic signal data collected at the data acquisition step An impulse response function extraction step in which processing is performed; The azimuthal variation according to the position of each station is calculated with respect to the impulse response function IRFs extracted at the extraction step of the impulse response function and the amplitude correction according to the azimuthal angle change is performed, A phase and amplitude information calculation step of calculating phase information and amplitude information reflecting the amplitude correction of the azimuth and azimuth according to the phase and amplitude information; And a result display step in which a process of displaying the result obtained through the data collecting step, the impulse response function extracting step and the phase and amplitude information calculating step is performed through display means including a display, The amplitude and the phase of the earthquake can be calculated by the following equation.

Here, the data collection step may include collecting data on seismic signals including seismic background noise recorded at seismic stations installed in each area, and storing the collected data for each observation station and period to construct a database Is performed.

The extraction of the impulse response function (IRFs) may be performed using a deconvolution method using the following equation.

Where xr is the receiver station, Xs is the source station, v is the background noise rate, ω is the frequency, i and j are the i-th component of the receiving station, is the complex conjugate, ε is the water level defined as 1% of the average spectral power, F ^-1 is the inverse Fourier transform, ≪> represents time domain stacking)

Alternatively, the extracting of the impulse response function may be configured to perform a process of extracting the impulse response functions IRFs through a coherency method using the following equation.

Where Xr is a receiver station, Xs is a source station, b is a 1-bit normalized segment of background noise rate data, ω is frequency, i and j are i Is the complex conjugate, ε is the water level defined as 1% of the average spectral power, F ^-1 is the complex conjugate of the first component and the jth component of the virtual source station, Inverse Fourier transform, and parentheses <> represent time domain stacking, respectively)

Alternatively, the extracting of the impulse response function may be configured to perform a process of extracting the impulse response functions IRFs through a cross correlation method using the following equation.

(여기서, Xr은 수신 관측소(receiver station), Xs는 소스 관측소(source station), v는 배경잡음 속도, ω는 주파수, i 및 j는 각각 수신 관측소의 i번째 성분(component) 및 가상 소스 관측소의 j번째 성분, *는 켤레 복소수 행렬(complex conjugate), F^-1은 역푸리에 변환(inverse Fourier transform), 괄호〈〉는 시간영역 스태킹(time domain stacking)을 각각 나타냄)

Where xr is the receiver station, Xs is the source station, v is the background noise rate, ω is the frequency, i and j are the i-th component of the receiving station, j denotes a j-th component, * denotes a complex conjugate, F ^-1 denotes an inverse Fourier transform, and parentheses denotes time domain stacking)

In addition, the phase and amplitude information calculating step, using the following equation, causal part of the extracted IRFs (causal parts) and the ratio widow (acausal parts) peak ground velocity ratio (peak ground velocity ratio for; PGV _ratio ) Is calculated and an azimuthal variation according to the position of each observation station is calculated.

(Where PGV _causal is the peak surface velocity of the causal parts of the extracted IRFs, and _{PGV acausal} is the peak surface velocity of the acausal parts of the extracted IRFs)

Further, the phase and amplitude information calculation step may include the steps of: (a) obtaining an amplitude value corrected for distance attenuation by multiplying the peak surface velocity of the impulse response function extracted in the impulse response function extraction step by the distance between the observing stations; (b) (C) deriving an amplitude value predicted according to an azimuth angle of the corrected amplitude values, (d) deriving a distance attenuation calculated in (b) above, Calculating an amplitude correction coefficient according to an azimuth angle by calculating a ratio of an amplitude value according to the azimuth calculated in the step (c) to an average of the corrected amplitude value, (e) calculating an amplitude correction coefficient based on the impulse response extracted in the extracting of the impulse response function And the function is multiplied by the amplitude correction coefficient calculated in the above (d), thereby performing processing for correcting the amplitude according to the azimuthal angle change.

Here, the phase and amplitude information calculation step is characterized in that the process of correcting the amplitude according to the azimuthal angle change is performed using the following equation.

은 거리감쇠가 보정된 진폭값의 평균, PGV_noise는 임펄스 응답함수의 피크 지면속도, R은 소스 관측소와 수신 관측소간의 거리를 각각 나타냄) (Where IRF _corrected is the impulse response function after amplitude correction with azimuthal variation, IRF is the impulse response function before amplitude correction with azimuthal variation, y is the amplitude predicted by the azimuth from the source station to the receiving station, value,

Where PGV _noise is the peak ground velocity of the impulse response function, and R is the distance between the source observation station and the receiving station, respectively)

The method may further include a verification step of performing verification by comparing phase and amplitude information obtained through the phase and amplitude information extraction step with actual measured data.

Here, the verification step may include a process of performing quantitative evaluation of the impulse response function by calculating relative event differences (PGV _diff ) between the actual event record and the PGVs of the extracted IRFs using the following equation Is performed.

(Where PGV _event is the peak ground velocity of the actual event and PGV _noise is the peak ground velocity of the extracted IRFs)

According to another aspect of the present invention, there is provided an earthquake signal analysis method, comprising: analyzing an earthquake signal using the phase and amplitude information calculation method of the earthquake signal using the earthquake background noise described above.

Further, according to the present invention, there is provided an earthquake analysis system configured to analyze an earthquake signal using the phase and amplitude information calculation method of an earthquake signal using the above-described earthquake background noise.

As described above, according to the present invention, a series of processes for extracting the phase and amplitude information of the IRFs from the seismic background noise, which is always recorded in the seismic observation station regardless of the seismic activity using the seismic interferometry technique, There is provided a method of calculating phase and amplitude information of an earthquake signal using an earthquake background noise configured to be performed by a dedicated hardware so that it is impossible to provide a technique for extracting both phase information and amplitude information of an earthquake signal, It is possible to solve the problems of the prior art in which there is a limit to only use information.

Also, according to the present invention, a phase and amplitude information calculation method of an earthquake signal using an earthquake background noise configured to efficiently extract both phase and amplitude information of IRFs from an earthquake background noise is provided, For the seismic region, we can also contribute to the analysis and study of the seismic ground motion using the data extracted from the background noise of the earthquake.

1 is a schematic view of broadband observation stations applied to the seismic interferometry method of the present invention at the Korea Institute of Geoscience and Mineral Resources (KIGAM) and the Korea Meteorological Administration (KMA) network.
FIG. 2 is a Tk source-type plot showing the results of moment tensor inversion for a mine collapse event occurring at Samcheok at 08:53 am on Jan. 31, 2015. FIG.
FIG. 3 shows a graphical representation of trends in the fall (from September 2015 to November 2015), winter (December 2015 to February 2016), spring (March 2016 to May 2016), and summer (June 2016 And a seasonal change of the source distribution of the background noise from the viewpoint of the PGV _ratio for one year from September 2015 to August 2016.
4 is a graph showing a comparison of PGV _diff before and after azimuth correction.
5 is a graph showing the azimuthal dependence of the background noise source distribution in the Korean peninsula region.
6 is a graph showing the azimuthal change of PGVs after correction for geometrical spreading (or distance attenuation) and the polynomial function regression equation for amplitude correction.
FIG. 7 is a diagram showing a comparison between vertical direction components and radial direction components of IRFs and event records extracted using three types of seismic interferometry techniques of deconvolution, cancellation, and cross correlation.
FIG. 8 is a diagram showing a comparison of Fourier amplitude spectra of IRFs extracted from the event record using the deconvolution, the interference method, and the cross-correlation method, respectively.
FIG. 9 is a diagram showing a comparison of tangential components of IRFs and event records extracted using the deconvolution, the coherence method, and the cross-correlation method, respectively.
FIG. 10 is a graph showing the results of comparing the correlation coefficient, the RMS error and the PGV _diff as quantitative indexes of IRFs extracted using the deconvolution, the interference method and the cross-correlation method, respectively.
FIG. 11 is a table showing average and standard deviation of three criteria applied to each reception station.
FIG. 12 is a diagram showing the result of comparing the waveforms of recorded surface motions with the IRFs obtained from the background noise data using the deconvolution method for the virtual source observation station (MGR) (red) and KNUD (blue).
13 is a graph showing a result of comparing the correlation coefficient, the RMS error and the PGV _diff as quantitative indexes of the IRFs extracted through the source observation station MGR and KNUD.
FIG. 14 is a flowchart schematically showing the overall configuration of a method for calculating phase and amplitude information of an earthquake signal using an earthquake background noise according to an embodiment of the present invention.

Hereinafter, a method of calculating phase and amplitude information of an earthquake signal using the earthquake background noise according to the present invention will be described with reference to the accompanying drawings.

Hereinafter, it is to be noted that the following description is only an embodiment for carrying out the present invention, and the present invention is not limited to the contents of the embodiments described below.

In the following description of the embodiments of the present invention, parts that are the same as or similar to those of the prior art, or which can be easily understood and practiced by a person skilled in the art, It is important to bear in mind that we omit.

Next, with reference to the drawings, the details of the phase and amplitude information calculation method of an earthquake signal using the earthquake background noise according to the present invention will be described.

1 and 2 illustrate a broadband observation station applied to the earthquake interferometry method of the present invention at KIGAM and KMA network, A Tk source-type plot showing the result of moment tensor inversion for a mine collapse event at Samcheok on January 31, 2015 at 08:53 am.

That is, in the present invention described below, through the recording of a mine collapse event (Mw 4.2) occurring in Korea in January 2015, this event has a single downward force at shallow depth It is shown that phase and amplitude information of impulse response functions (IRFs) can be obtained effectively using seismic interferometry.

In the present invention, as described later, three kinds of seismic interferometry techniques of deconvolution, coherency, and cross correlation are applied by applying quantitative metrics. And considering the azimuthal dependency on the source distribution of the ambient noise, seismic ambient noise in the period band of 2 to 10s The best way to extract IRFs between two remote stations using data is presented and if appropriate spectral normalization or whitening schemes are applied during data processing An effective method is presented.

More specifically, it has been proposed that broadband Rayleigh waves and their dispersion characteristics can be extracted from background noise using cross correlation analysis, This is also known as seismic interferometry. Conventionally, a high-resolution tomographic image of California is constructed using short-period surface wave group velocity measurements from the cross-correlation of 1-month background noise data There is a bar.

It is also possible to apply the sum of impulse responses (IRFs) extracted from the background noise of the earthquake to predict strong ground motions, that is, from the deconvolution of the earthquake background noise region to the complex underground structure it is possible to extract reliable phase and amplitude information for IRFs, including the effect of subsurface structures.

In addition, we apply both depth and focal mechanism corrections to earthquakes and extend these methods to develop a moderate model of Japan, assuming southern California ground motion simulations and point sources. Have been applied to long-term ground motion simulations produced by offshore subduction earthquakes.

Seismic interferometry techniques can be categorized into three types of deconvolution, coherency, and cross correlation. First, deconvolution is used to study seismic amplification in the basin and large It is used to simulate long-term ground motion caused by an earthquake.

In addition, the coherence technique is applied to recover the anelastic attenuation estimates. In order to numerically investigate the recovery of the IRF attenuation, conventionally, the spatial distribution and attenuation of the source are identified, The theoretical basis for calculating the crosscorrelation that represents the quantification of amplitude is presented. The amplification effect of the elastic subsurface structure and the crust and the attenuation of the upper mantle and the building Effects are included.

As described above, the background noise technique is widely used to identify the velocity structure of the earth, in particular, the crustal structure of the Korean peninsula, and the seismic interferometry method has a high-magnitude earthquake activity, , It is particularly useful in low-seismicity regions, but high-density seismic networks are required to record background noise.

Accordingly, in the present invention, a method of efficiently extracting phase and amplitude information of IRFs from background noise data of a low-seismic region through a ground motion excited by a limestone mine collapse event (Mw 4.2) in Korea Here, since the above-mentioned collapse event is regarded as a single impulsive downward force applied to the surface of the earth, in the present invention, in order to simulate the ground motion of the mine collapse event, Three seismic interferometry techniques, deconvolution, coherency and cross correlation, were compared to obtain both phase and amplitude information without applying the instrument calibration. To compensate for the azimuthal dependency, the spatiotemporal distribution of the background noise sources on the Korean Peninsula was investigated.

More specifically, a limestone mine collapse occurred in Samcheok at 08:53 am on Jan. 31, 2015 as a Korean standard time. As shown in FIG. 1, the Korea Institute of Geoscience and Mineral Resources KIGAM) and the Korea Meteorological Administration (KMA), respectively, were recorded for each of the broadband stations.

There was no casualties since the mine had been abandoned a few years before the incident had already occurred and the present inventors have identified five 3-component regional broadband recordings in the frequency range of 0.04 to 0.1 Hz ), And the moment tensor solution is inverted to obtain a negative tensile crack at the depth of several hundred meters, as shown in the Tk source type plot of FIG. 2, And the like.

In other words, when a shallow-depth collapse occurs, a singular vertical force can be approximated by a tensile fault, and this model assumes that the source depth is shallower than the wavelength, It is valid only when recorded by remote observation, and the limestone mine collapse event applied to the present invention satisfies both of these conditions.

Here, from a theoretical point of view, centroid single force inversion may be more effective in understanding the source mechanism, but with the content shown in FIG. 2, the source mechanism of the event is equivalent to a single downward force .

In addition, the field survey revealed that the collapsed area corresponding to a moment magnitude (Mw) of 4.2 was spread over a width of several hundred meters, and the time for mine collapse was estimated to be within a few seconds, Can be modeled as a short-term impulse for the target period band and IRFs extracted using the seismic interferometry can be directly compared to the event record without depth or mechanism correction.

More specifically, the seismic interferometry technique can be applied in three forms according to the correlation analysis method for the background noise data. First, the power spectrum of the source station data is used for normalization Deconvolution of the normalized receiver station data has been applied to extract IRFs from seismic fields in the past.

The second form is a coherency method, which uses 1-bit normalization and spectral whitening of the amplitude spectra on both the source observing station and the receiving observing station.

Here, the 1-bit normalization is generally used for background noise tomography by extracting dispersion curves for the surface wave group velocity, By replacing 1 with 1 and replacing all negative amplitudes by -1. ≪ RTI ID = 0.0 > [0031] < / RTI >

The third form is a method of applying cross correlation without spectral schemes and a comparison of these processing techniques for reliable broadband surface wave dispersion measurement is as follows.

First, the deconvolution method can be expressed as Equation (1) below. This is because the simultaneous data representing the background noise velocity v for two seismic stations located at Xr (reception observatory) and Xs Is performed in the frequency domain omega with simultaneous data segments.

[Equation 1]

In Equation (1), i and j are the i-th component of the receiving station and the j-th component of the virtual source station, respectively, and * denotes a complex conjugate.

In Equation (1), the numerator corresponds to the cross-correlation of the simultaneous background noise data segments in the frequency domain, and is normalized using the power spectrum of the source station.

In order to maintain the stability of deconvolution during the numerical calculation in the above Equation 1, in the present invention, the water level? Defined as 1% of the average spectral power F ^-1 indicates an inverse Fourier transform, parentheses <> indicate time domain stacking, and the time-domain stacking over a long time series increases the signal-to- .

Next, the coherency method includes 1-bit normalization and spectral whitening, as shown in Equation (2) below, And the cross-correlation method.

&Quot; (2) "

In Equation (2), b is a 1-bit normalized segment of the background noise rate data, and the numerator corresponds to the cross-correlation of the 1-bit normalized background noise segment in the frequency domain, and denominator Represents the spectral whitening of the amplitude spectra of the source and the receiving station.

In Equation (2), the water level? Is added to each spectrum similarly to the deconvolution method, and this water level is not mathematically required for the coherent method. However, Equation (1) It was added for consistency.

Next, the cross-correlation method can be expressed by the following equation (3).

&Quot; (3) "

As shown in Equation (3) above, the cross-correlation method does not apply the spectral processing applied to the deconvolution and coherent methods described above and does not apply various assumptions such as, for example, with or without attenuation To extract IRFs.

More specifically, in the present invention, as shown in FIG. 1, IRFs are extracted using a background noise seismic interferometry using an MGR observation station as a virtual source observation station, a broadband observation station of a KIGAM and a KMA network as reception stations , Where earthquake background noise data was collected for one year from September 2015 to August 2016 after the installation of the MGR station.

Also, prior to applying the seismic interferometry, successive recordings of background noise data are divided into nonoverlapping time segments of 1 hour, raw data being instrumental responses, The trend and mean are removed from the initial noise data using a Seismic Analysis Code. [0034] [0031] The method of claim 1,

In addition, all time segments are downsampled to 10 samples per second to reduce the computation of the data processing, and segments having spikes greater than 5 times the standard deviation of 1 hour long data to reduce the effects of potential seismic events Is discarded.

Furthermore, the three components of the receiver station background noise data are correlated with the polarity-changed vertical components of the virtual source station data to identify the downward direction of the mine collapse, After the time series of the background noise data is transformed using the FFT, the correlation of each segment in the frequency domain is calculated.

Thereafter, the inverse FFT is applied to the correlated result to accumulate the time series segments over a period of one year in the time domain, and the IRFs (i, j) of the azimuth of the receiving station with respect to the virtual source station (MGR) The horizontal components ZN and ZE of the rotated IRFs are rotated to radial and tangential components and the causal part of the rotated IRFs ) Is considered as a source-to-destination impulse response with a source mechanism consisting of a single downward force, and IRFs extracted using three types of seismic interferometry techniques are used for mine collapse events And compared with observed ground motion data over a time span of 2 to 10 seconds using a fifth-order Butterworth filter.

In order to compare these three seismic interferometry methods, the amplitude of the IRFs should be set to a constant scaling constant for each of the three components to preserve the relative amplitude information of the receiving station, The scaling factor is calculated by the ratio of the average peak amplitude of the IRFs to the ground motion data recorded, which applies to all stations.

Subsequently, to determine the uncertainty of the origin time and focal depth estimates, a constant temporal shift factor is applied to the ground motion data recorded for each receiving station, , And KIGAM's automatic seismic detection system. For the most effective waveform comparison, this time shift was set to 2 seconds according to trial and error method.

Further, in the present invention, the causal parts (positive delay) of the extracted IRFs are the impulse response between the source and the receiving station, and the acausal parts (negative lag) It is assumed that the source of the background noise is not uniformly distributed in the direction of the source, but the distribution of the background source of the background noise in the Korean peninsula is not yet known.

Accordingly, in the present invention, in order to confirm the difference and level of the spatial distribution of the background noise before correcting the amplitude according to the azimuth angle, as shown in the following equation (4), the IRFs extracted using the deconvolution method The azimuthal variation was investigated using the peak ground velocity ratio (PGV _ratio ) according to the azimuth angle.

&Quot; (4) "

The present invention may further comprise a verification step of performing verification by comparing the calculated phase and amplitude information with actual measured data. To this end, a quantitative evaluation of an impulse response function, which is the final result, is performed Lt; / RTI >

That is, the relative amplitude differences in the PGVs between the event record and the extracted IRFs can be calculated as shown in Equation (5) below, and the relative amplitude difference (PGV _diff ) It can be used as an index for quantitative evaluation of the impulse response function.

&Quot; (5) "

3 and 4, FIG. 3 is a graphical representation of the results of the winter (September 2015 to November 2015), winter (December 2015 to February 2016), spring (March 2016 to 2016 And azimuthal and seasonal changes in source distribution of background noise in terms of PGV _ratio for one year from May 2015 to May 2016 and from September 2016 to August 2016 And FIG. 4 is a diagram showing comparison of PGV _diff before and after azimuth correction.

As shown in FIG. 3 and FIG. 4, a significant azimuthal variation and a relatively weak seasonal variation of the background noise source level are observed together, and it is confirmed that the change of the PGV _diff with respect to the azimuth angle after the correction is remarkably reduced.

More specifically, referring to FIG. 5, FIG. 5 is a diagram illustrating the azimuthal dependency of the background noise source distribution in the Korean peninsula region.

In Fig. 5, the causal part (blue) and the non-widow part (red) of IFRs (vertical component) obtained along with the name of the station are displayed on the left side of each panel, and the numbers under the name of the station are shown in the virtual source station MGR And the number on the right side of each waveform is the peak ground velocity ratio (PGV _ratio ) defined in Equation (4), and the positive PGV _ratio value is a value obtained by subtracting the PGV of the causal part from the source- It means stronger than widows, and vice versa.

As shown in FIG. 5, it can be seen that the causal part (blue) generally has a larger PGV when the receiving station is located to the north than the source station. To extract both the amplitude and phase information of the IRFs, It is important to consider the temporal and spatial variation of the distribution.

Accordingly, in the present invention, to accumulate waveforms for one year in order to average seasonal variations, data is generated, and amplitude correction is applied to the azimuth angle to identify changes in the azimuth angle.

6, FIG. 6 illustrates the change of azimuth angle of PGVs after correction for geometrical spreading (or distance attenuation) using 1 /? R (r is the distance between the source observing station and the receiving station) And a polynomial function regression equation for correction.

In FIG. 6, the azimuthal change at the peak ground surface velocity (PGVs) after compensation for the geometric damping effect is shown through a polynomial regression equation (red curve). The three panels are deconvoluted downward to the leftmost, And the polynomial function regression equation required to obtain the correction coefficient by the cross-correlation method is shown on the right side.

More specifically, in order to derive a polynomial regression equation by considering the amplitude value of the PGVs whose distance attenuation has been corrected as a function depending on the azimuth angle, a relation as shown in the following Equation (6) is assumed.

&Quot; (6) "

In Equation (6), PGV _noise is the peak surface velocity of the impulse response function, R is the distance between the source observing station and the receiving observing station, and x is the azimuth of the receiving observing station from the source observing station.

Next, the peak ground speed of the impulse response function is multiplied by the distance between the observing stations to obtain the amplitude value corrected for the distance attenuation, and the relationship between the amplitude values corrected for the distance attenuation and the azimuth angle is calculated using the polynomial function (A), (b), and (c) are polynomial fitted so as to satisfy the following equation (7).

&Quot; (7) "

)과 위의 다항함수 회귀식으로부터 방위각(x)에 따라 예측되는 진폭값(y)의 비율을 계산하면 그 비율(

)이 방위각에 따른 진폭보정계수에 해당한다. Thereafter, the distance attenuation is averaged over the corrected amplitude value (

) And the ratio of the amplitude value (y) predicted according to the azimuth angle (x) from the above polynomial function regression equation,

) Corresponds to the amplitude correction coefficient according to the azimuth angle.

Therefore, an equation for multiplying the impulse response function by the amplitude correction coefficient and correcting the amplitude according to the azimuth angle change is as follows: " (8) "

&Quot; (8) "

는 거리감쇠가 보정된 진폭값의 평균, y는 다항함수 회귀식으로부터 예측되는 진폭값, IRF_corrected는 방위각 변화에 따른 진폭 보정 이후의 임펄스 응답함수를 각각 나타낸다. In Equation (8), IRF denotes an impulse response function before the amplitude correction according to the azimuthal angle change,

Y is the amplitude value predicted from the polynomial regression equation, and IRF _corrected is the impulse response function after amplitude correction according to the azimuthal variation, respectively.

)을 계산하며(도 6의 검정색 가로 직선으로 표시), (3) [수학식 6] 및 [수학식 7]에 나타낸 바와 같이 하여, (1)에서 거리감쇠가 보정된 진폭(PGV*sqrt(r)) 값들의 방위각(x)에 따라 예측되는 진폭값(y)을 다항함수 회귀식으로 도출(polynomial fitting) 하고(도 6의 빨간색 곡선으로 표시), (4) (2)의 평균값에 대한 (3)의 다항함수값의 비율(

)로부터 방위각에 따른 진폭보정계수를 계산하며, (5) [수학식 8]에 나타낸 바와 같이 하여, 추출된 임펄스 응답함수에 (4)의 진폭보정계수를 곱하는 것에 의해 방위각 변화에 따른 진폭보정이 수행될 수 있다. That is, as shown in Fig. 6 and the above-mentioned equations (6) to (8), the method of correcting the amplitude according to the azimuthal angle change is as follows. (1) The peak ground velocity of the extracted impulse response function (PGV * sqrt (r), indicated by the blue circle in FIG. 6), (2) the amplitude at which the distance attenuation is corrected (PGV * sqrt (r)) Average of values

(3) The amplitude attenuation-corrected amplitude PGV * sqrt (1) is calculated as shown in Equations (6) and (7) (2) is polynomial fitting (indicated by a red curve in FIG. 6), and the amplitude value (y) predicted according to the azimuth angle (x) The ratio of the polynomial function value of (3)

(5) By multiplying the extracted impulse response function by the amplitude correction coefficient of (4) as shown in [Expression 8], the amplitude correction according to the azimuth angle change is calculated .

The azimuthal amplitude correction factor is suitable for the data using the second-order polynomial, minimizing the inconsistency between the polynomial and the data. That is, as shown in FIG. 4, It can be seen that the average value is decreased with the correction even if the value of PGV _diff increases for several stations and the large change of PGV _diff with respect to the azimuth angle is also decreased.

Following the azimuthal correction, in the present invention, the waveforms of the IRFs obtained using the ground motion recorded from the mine collapse event and the three types of seismic interferometry are compared and the vertical and radial components are shown in FIG.

7, FIG. 7 shows the vertical and radial directions of IRFs (red) and event records (gray) extracted using the three techniques of seismic interferometry, namely, deconvolution, Are compared with each other.

7, the vertical direction component is shown on the upper side and the radiation direction component is shown on the lower side, and the numbers on the right side of each panel are calculated between two waveforms within a time window defined by two blue lines on each waveform .

In FIG. 7, it can be seen that both the deconvolution and the coherent method successfully represent the relative amplitude of the ground motions as compared to the recorded ground motions, wherein the coherent method ignores the amplitude of the background noise data Lt; RTI ID = 0.0 > 1-bit < / RTI > normalization and spectral whitening.

That is, for waveform fitting of phase and amplitude information, the IRFs obtained using the deconvolution method are slightly better than the results obtained using the cancellation method and provide more detailed results, but the cross- The result is unstable due to the absence of spectral treatment such as whitening.

More specifically, the peak amplitudes of IRFs extracted using the cross-correlation method for the KSA and HKU receiving stations are larger than those of the other receiving stations, and therefore, spectral normalization scheme.

8, FIG. 8 is a graph comparing the Fourier amplitude spectra of the IRFs and the event record (gray) extracted using the deconvolution (red), the interference method (green), and the cross correlation (yellow) Fig.

8, the gray line represents the amplitude spectrum of each receiving station, the black line represents the average amplitude spectrum thereof, the frequency band used is 0.1 to 0.5 Hz, and the period band is 2 ~ 10 seconds, and is shown in shaded background.

As shown in FIG. 8, it can be seen that the waveform obtained by the deconvolution method shows a generally higher frequency component, and the average Fourier amplitude spectrum obtained using the deconvolution method is similar to the event recording.

9, FIG. 9 is a diagram showing the tangential components of the IRFs (red) and the event records (gray) extracted using the deconvolution, the crosstalk method and the cross-correlation method, respectively, 9, the number on the right side of each panel is a correlation coefficient calculated between two waveforms within a time window defined by two blue lines in each waveform.

7, the vertical direction and radial direction components of Rayleigh waves generated by the mine collapse event are satisfactorily obtained, while the amplitudes of the tangential components are the same as those shown in FIG. 9 As such, it is relatively weak due to the single downward force mechanism of the event.

Subsequently, the phase and amplitude information obtained by the above-mentioned three methods are subjected to correlation coefficient (CC), root-mean-square (RMS) error, and peak ground speed difference between the IRFs and the event record _diff ) are metered quantitatively according to the three metrics.

10, a correlation coefficient, an effective value error, and a PGVdiff (i) are quantitative indices of IRFs extracted by using the deconvolution (red), the interference method (green) Respectively. As shown in Fig.

In FIG. 10, correlation coefficients, effective value errors, and PGV _diff are shown from the upper side in the vertical direction and the radial direction components, respectively. These reference values are obtained from the waveforms in FIG. 7 between two waveforms within a time window defined by a blue line (Dotted line) and after (solid line), respectively.

That is, to calculate the above criteria, a finite time window (blue line in FIG. 7) is applied to each receiving station for more efficient phase and amplitude comparison, And the arrival time of the propagation velocity of 2.0 to 5.0 km / s for the receiving station within 100 km was determined for the wave velocity of 2.5 to 4.0 km / s.

As a result of comparison, both CC and RMS errors are considered to provide efficient criteria for evaluating the accuracy of phase information obtained using seismic interferometry techniques, and RMS error and PGV _diff are both suitable for evaluating relative amplitude information .

Among the three seismic interferometry techniques, the deconvolution method showed the largest CC value (FIG. 10A) and the smallest RMS value (FIG. 10B) for both vertical and radial components, and PGV _diff showed the CC and RMS error criterion , But it can be seen that the deconvolution method has the best overall result.

That is, referring to FIG. 11, FIG. 11 is a table showing average and standard deviation of three criteria applied to each reception station.

From these results, it can be seen that the deconvolution method is the best method to obtain both the phase and amplitude information of IRFs from the background noise among the three seismic interferometry techniques, and the interference method also shows a compatible result , The quantitative comparison of the tangential components is meaningless because the amplitude of the extracted waveform is relatively small.

In addition, recently, in order to reliably obtain the amplitude of IRFs from the Kanto Basin in Japan, a comparison of different seismic interferometry techniques has been presented. As a result, , Deconvolution techniques have been shown to be the most appropriate method for obtaining IRFs from seismic background noise regions, but the results are limited to the analysis of vertical components, whereas in the present invention, the results for both vertical and horizontal components And more clearly demonstrates that IRFs can be obtained from seismic background noise regions without correction of occurrence depth or source mechanism using data from shallow depth mine collapse events.

In the present invention, the availability of an adjacent source station located about 5 km away from the epicenter was investigated. The single broadband observation station (KNUD) was already installed before the mine collapse event, and MGR A deconvolution method was used to extract IRFs from KNUD, and the results are shown in FIGS. 12 and 13. FIG.

12 and 13, FIG. 12 shows the IRFs obtained from the background noise data using the deconvolution method for the virtual source observatory (MGR) (red) and KNUD (blue) In FIG. 12, the ground motion obtained for KNUD is shifted along the time axis with respect to the distance to the KNUD and the MGR at each receiving station and the azimuth difference, respectively.

FIG. 13 is a diagram showing a result of comparing the correlation coefficient, effective value error and PGV _diff of IRFs extracted through the source observation station MGR and KNUD.

12 and 13, it is possible to confirm that it is more efficient to utilize a source observing station (i.e., MGR) close to the epicenter if possible, but it is also possible to use a source source (source) within a period range of 2 to 10 seconds. a distance of 5 km may be allowed to obtain both the amplitude and phase information of the wave.

Therefore, it is possible to implement a method of calculating the phase and amplitude information of an earthquake signal using the earthquake background noise according to the present invention based on the above description. That is, referring to FIG. 14, 5 is a flowchart schematically showing the overall configuration of a phase and amplitude information calculation method of an earthquake signal using background noise.

14, the phase and amplitude information calculation method of an earthquake signal using an earthquake background noise according to an embodiment of the present invention is roughly divided into an earthquake signal including an earthquake background noise recorded at a seismic station installed in each area, (S10) for constructing a database by collecting data on the earthquake observing station and signal processing including pre-processing of the seismic background noise collected for each seismic observing station, and performs a deconvolution method, an incoherent method, (S20) for extracting impulse response functions (IRFs) between seismic stations using at least one seismic interferometry technique among the three types of seismic interferometry, A phase and amplitude information calculating step (S30) of calculating phase and amplitude information in which the amplitude correction is performed by performing amplitude correction according to the azimuth angle, and And a result display step (S40) of displaying the result obtained through each of the steps through a display means such as a display, can be configured to be performed by a computer or dedicated hardware.

1, the above-described database building step (S10) collects seismic signal data including seismic background noise from KIGAM and KMA from regular seismic stations operated in the Korean peninsula region , Preprocessing processes such as removing trends and averages from the collected seismic signal data and correcting device characteristics, and then arranging and storing the preprocessed data according to each observation station and period to construct a database.

In addition, the impulse response function extracting step S20 may include three methods, i.e., a deconvolution method, a coherent method, and a cross-correlation method, as described above with reference to Equations (1) (IRFs) between each station by applying at least one of the seismic interferometry methods, preferably the deconvolution method.

In addition, the phase and amplitude information extraction step S30 may be performed as described above with reference to Equations (4) to (8) and FIG. 6, taking into consideration the azimuthal angle change depending on the position of each observation station And to extract phase and amplitude information reflecting the amplitude correction.

Further, the above-described method may further comprise a verification step of performing verification by comparing the extracted phase and amplitude information with actual measured data. To this end, it is preferable that the method described with reference to Equation (5) As well as to perform a quantitative evaluation of the final result, the impulse response function.

Therefore, the method of calculating the phase and amplitude information of the seismic signal using the earthquake background noise according to the present invention can be implemented as described above.

In addition, according to the present invention, it is possible to analyze the seismic signal using the phase and amplitude information calculation method of the seismic signal using the seismic background noise according to the present invention, The present invention can provide an earthquake signal analysis method and an earthquake analysis system configured to enable analysis of signals.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the present invention as defined by the following claims and their equivalents. I will work.

Claims (12)

An earthquake constructed to extract the phase and amplitude information of impulse response functions (IRFs) from seismic ambient noise, which is always recorded at seismic stations irrespective of seismic activity. A method for calculating phase and amplitude information of a signal,
A data collecting step of collecting data on an earthquake signal including an earthquake background noise recorded at a seismic observing station installed in each area;
The seismic interferometry technique is used to extract impulse response functions (IRFs) between each seismic station for the seismic background noise collected at each seismic station using the seismic signal data collected at the data acquisition step An impulse response function extraction step in which processing is performed;
The azimuthal variation according to the position of each station is calculated with respect to the impulse response function IRFs extracted at the extraction step of the impulse response function and the amplitude correction according to the azimuthal angle change is performed, A phase and amplitude information calculation step of calculating phase information and amplitude information reflecting the amplitude correction of the azimuth and azimuth according to the phase and amplitude information; And
A processing step including a result display step in which a process of displaying the result obtained through the data collection step, the impulse response function extraction step and the phase and amplitude information calculation step is performed through a display means including a display, Wherein the phase and amplitude information of the seismic signal are generated by hardware of the earthquake background noise.
The method according to claim 1,
Wherein the data collection step comprises:
A process for collecting data on an earthquake signal including an earthquake background noise recorded at an earthquake observing station installed in each area and storing the collected data for each observation station and period to construct a database is performed A method for calculating phase and amplitude information of seismic signals using earthquake background noise.
The method according to claim 1,
Wherein the extracting of the impulse response function comprises:
And a process of extracting the impulse response functions (IRFs) using a deconvolution method is performed using the following equation: < EMI ID = 1.0 > Calculation method.

Where xr is the receiver station, Xs is the source station, v is the background noise rate, ω is the frequency, i and j are the i-th component of the receiving station, is the complex conjugate, ε is the water level defined as 1% of the average spectral power, F ^-1 is the inverse Fourier transform, ≪> represents time domain stacking)
The method according to claim 1,
Wherein the extracting of the impulse response function comprises:
Wherein the processing of extracting the impulse response functions (IRFs) through an coherency method is performed using the following equation: < EMI ID = Way.

Where Xr is a receiver station, Xs is a source station, b is a 1-bit normalized segment of background noise rate data, ω is frequency, i and j are i Is the complex conjugate, ε is the water level defined as 1% of the average spectral power, F ^-1 is the complex conjugate of the first component and the jth component of the virtual source station, Inverse Fourier transform, and parentheses <> represent time domain stacking, respectively)
The method according to claim 1,
Wherein the extracting of the impulse response function comprises:
And a process of extracting the impulse response functions (IRFs) through a cross correlation method is performed using the following equation: < EMI ID = Way.

Where xr is the receiver station, Xs is the source station, v is the background noise rate, ω is the frequency, i and j are the i-th component of the receiving station, j denotes a j-th component, * denotes a complex conjugate, F ^-1 denotes an inverse Fourier transform, and parentheses denotes time domain stacking)
The method according to claim 1,
Wherein the phase and amplitude information calculation step comprises:
Using the following equation, the peak when the speed ratio of the causal portion (causal parts) of IRFs the ratio widow (acausal parts) derived from the impulse response function extraction step; by calculating (peak ground velocity ratio PGV _ratio) And calculating an azimuthal variation according to a position of each of the observing stations is performed.

(Where PGV _causal is the peak surface velocity of the causal parts of the extracted IRFs, and _{PGV acausal} is the peak surface velocity of the acausal parts of the extracted IRFs)
The method according to claim 6,
Wherein the phase and amplitude information calculation step comprises:
(a) multiplying the peak ground speed of the impulse response function extracted in the step of extracting the impulse response function by the distance between the observing stations to obtain an amplitude value corrected by the distance attenuation,
(b) calculating an average of the amplitude values with the attenuation corrected,
(c) polynomial fitting the amplitude value predicted according to the azimuth angle of the amplitude values with the attenuation corrected,
(d) calculating an amplitude correction coefficient according to the azimuth angle by calculating a ratio of the amplitude value according to the azimuth calculated in (c) to the average of the amplitude attenuation-corrected amplitude value calculated in (b)
(e) processing for correcting the amplitude according to the azimuthal angle change is performed by multiplying the impulse response function extracted at the step of extracting the impulse response function by the amplitude correction coefficient calculated at (d) A method for calculating phase and amplitude information of an earthquake signal using noise.
8. The method of claim 7,
Wherein the phase and amplitude information calculation step comprises:
Wherein the process of correcting the amplitude according to the azimuthal angle change is performed using the following equation: < EMI ID = 17.0 >

(Where IRF _corrected is the impulse response function after amplitude correction with azimuthal variation, IRF is the impulse response function before amplitude correction with azimuthal variation, y is the amplitude predicted by the azimuth from the source station to the receiving station, value,

Where PGV _noise is the peak ground velocity of the impulse response function, and R is the distance between the source observation station and the receiving station, respectively)
The method according to claim 1,
The method comprises:
And a verification step of performing verification by comparing the phase and amplitude information calculated through the phase and amplitude information extraction step with actual measured data. Method of calculating phase and amplitude information.
10. The method of claim 9,
Wherein the verifying step comprises:
The process of performing the quantitative evaluation of the impulse response function is performed by calculating the relative amplitude differences (PGV _diff ) between the actual event record and the PGVs of the extracted IRFs using the following equation A method for calculating the phase and amplitude information of an earthquake signal using an earthquake background noise.

(Where PGV _event is the peak ground velocity of the actual event and PGV _noise is the peak ground velocity of the extracted IRFs)
A method for analyzing an earthquake signal using the method of calculating phase and amplitude information of an earthquake signal using the earthquake background noise according to any one of claims 1 to 10.
An earthquake analysis system for analyzing an earthquake signal using the phase and amplitude information calculation method of an earthquake signal using the earthquake background noise according to any one of claims 1 to 10.

Images