JP3836933B2 - Seismic observation method to remove noise - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an earthquake observation method used for earthquake observation, and particularly used for artificial earthquakes used for geological surveys, exploration, active fault investigations, etc., or microearthquake observations for investigating the causes of landslides and seismic activity. It relates to earthquake observation methods.
[0002]
[Prior art]
Conventionally, seismometers have been used to measure the vibrations of artificial earthquakes and microearthquakes caused by dynamite. In that case, one set of seismometer and one sensor is used.
[0003]
[Problems to be solved by the invention]
However, artificial earthquakes are caused by the blasting of dynamite, but even if a large amount of dynamite is used, the vibration level is not so high, and it is easily affected by noise caused by daily vibration such as vehicle running and human walking. It was. Even in the case of natural earthquakes, observations of microseismic vibrations are still susceptible to noise, which hinders accurate observations.
[0004]
Furthermore, in the observation of microearthquakes, we have devised to record the waveform from a few seconds before the trigger operation using a delay circuit in the seismometer, with a vibration above a certain vibration level as a trigger, so that the economy of recording paper etc. However, recording is often triggered by noise as a trigger, which has hindered observation economics.
[0005]
Therefore, the present invention has been made in view of the problems of the prior art, and it is possible to reliably determine noise vibrations among vibrations observed by a seismometer, and to extract only vibrations caused by true earthquakes. Its purpose is to provide an observation method.
[0006]
[Means for Solving the Problems]
In order to achieve the above object, the earthquake observation method according to the first aspect of the present invention decomposes the vibration measured by the seismometer into intensity levels with respect to the frequency component, and sets the vibration level of the predetermined frequency y1 hertz component of the vibration to ay1. When the reference level that is a reference for discriminating whether the vibration is an earthquake vibration or a noise vibration is an decibel, and the vibration level at the vibration frequency fp hertz that gives the vibration peak is ap decibel, the following expression ry1 = ( ay1-an) / (ap-an) × 100 If the ry1 percent exceeds a predetermined value r1 percent, the vibration is determined as noise. At this time, as in the method of claim 2, the predetermined frequency y1 hertz may be any frequency within a range of 15 to 25 hertz, for example, 20 hertz. At this time, as in the method of claim 3, the predetermined value r1 percent may be any value of 55 to 65 percent, for example, 60 percent.
[0007]
In such a method, the ratio of the difference between the vibration level of the predetermined frequency y1 Hertz component of vibration, for example, 20 Hz, and the reference level, and the difference between the vibration level of the vibration peak and the reference level is a predetermined value, for example, 60% Noise can be determined and eliminated based on whether or not the value exceeds.
[0008]
The seismic observation method according to claim 4, wherein the vibration measured by the seismometer is decomposed into intensity levels with respect to frequency components, and a reference level serving as a reference for determining whether the vibration is vibration caused by an earthquake or noise is obtained. an decibel, where the vibration level at the vibration frequency fp hertz that gives the vibration peak is ap decibel, and the predetermined ratio is x percent, ax given by the expression ax = an + (ap−an) × (x / 100) When the bandwidth W in decibels exceeds a predetermined frequency width W1 hertz, the vibration is determined as noise. Here, as described in claim 5, the x percent may be any ratio within a range of 65 to 75 percent, and as described in claim 6, the predetermined frequency width W1. Hertz may be any value between 10 and 20 hertz.
[0009]
In such a method, noise is discriminated based on whether or not the bandwidth W at a vibration level of 70 percent exceeds a predetermined ratio between the vibration reference level and the vibration level at the vibration peak, for example, a vibration level of 70 percent. Can be removed.
[0010]
Seismic observation method according to the invention according to claim 7, the vibration to decompose the vibration measured by seismometers on the intensity level for a frequency component, a predetermined frequency y1 Hz ay1 decibel vibration levels of components of this vibration, the vibration is caused by an earthquake When the reference level serving as a reference for discriminating between vibration and noise is an decibel, and the vibration level at the vibration frequency fp hertz that gives the vibration peak is ap decibel, the expression ry1 = (ay1-an) / (ap-an ) × 100, when the ry1 percent exceeds a predetermined value r1 percent, the vibration is determined as noise, and when the predetermined ratio is x percent, the expression ax = an + (ap−an) × When the bandwidth W hertz in ax decibels given by (x / 100) exceeds a predetermined frequency width W1 Hertz Motion is identified as noise.
[0011]
At this time, as defined in claim 8, the predetermined frequency y 1 hertz is any frequency within a range of 15 to 25 hertz, and the predetermined value r1 percent is any of 55 to 65 percent. The x percent is any ratio within a range of 65 to 75 percent, and the predetermined frequency width W1 hertz may be any value between 10 and 20 hertz.
[0012]
In such a case, the determination accuracy is improved because the two conditions are overlapped.
[0013]
Further, as described in claim 9, when the peak frequency fp hertz exceeds a predetermined frequency f1, the vibration is determined as noise, and a vibration level difference ad decibel given by the following expression ad = ap-an. However, when the predetermined value a1 is not exceeded, the vibration may be determined as noise.
[0014]
In this case, since the peak frequency fp and the vibration level difference are added to the determination condition, the determination accuracy is further improved.
[0015]
Further, as in the earthquake observation method according to claim 10, the vibration measured by the seismometer may be measured in each of two directions orthogonal to the horizontal direction and one direction perpendicular to the two directions. good rather than, in this case the sample is also becomes possible to determine the vibration direction based on different data only three times, go up further determination of accuracy.
[0016]
DETAILED DESCRIPTION OF THE INVENTION
The inventor of the present invention has studied characteristics of earthquake and noise vibrations, and found that there is a difference in characteristics between true earthquake vibration and noise vibration. That is, (a) the vibration level of the frequency around 20 Hz is about 60% of the peak vibration level, and the true earthquake and noise vibration can be distinguished, and (b) the frequency of the true earthquake is compared. For example, distant earthquakes such as Chile are around 1 Hertz, domestic earthquake frequencies are around 3 Hertz, generally 3 to 10 Hertz, and noise is a relatively high frequency region such as 10 to 20 Hertz. (C) The bandwidth W hertz at a level of around 70% is relatively narrow in a true earthquake, for example, around 7 hertz, for example, noise is around 17 hertz, and (d) vibration. It was found that true earthquakes and noise can often be distinguished by the difference between the peak value and the reference level. The present invention has been made based on such findings.
[0017]
Hereinafter, embodiments of the present invention will be described. FIG. 8 is a diagram showing an outline of the seismometer system used in the present invention. A signal lead wire 4 connects between the sensor 1 installed at a point where an earthquake is desired to be observed and the seismometer body 2. In the drawing, since the lead wire is long, it is omitted in the middle. The seismometer main body and the computer 3 for processing the measured data are electrically connected. The computer 3 is configured to perform a predetermined determination by performing frequency analysis on the observed vibration and comparing the observed or analyzed numerical value with a set value.
[0018]
The sensor 1 can measure vibrations in a total of three directions including orthogonal coordinates on a plane horizontal to the ground, an east-west direction EW, a north-south direction NS, and a vertical direction UD perpendicular to the ground. It is configured as follows. The amount to be detected is vibration acceleration, but may be vibration speed.
[0019]
The seismometer main body 2 includes a delay circuit, and is configured to detect a vibration as a trigger and record the vibration from a few seconds before the time when the vibration is detected.
[0020]
In the embodiment of the present invention, the vibration measured by the seismometer is subjected to frequency analysis, and the following four items are compared with the set values to determine whether the vibration is caused by a true earthquake or noise. In particular, the discrimination according to whether the vibration level r20 at the predetermined frequency (3) , for example, around 20 Hz, is greater than or less than a set value is the most significant discrimination method.
[0021]
Furthermore, if all the four items are satisfied with the AND condition as a true earthquake, the reliability is very high. The four items that give the conditions for a true earthquake are as follows.
[0022]
(1) The peak frequency is within the set value.
[0023]
(2) The bandwidth of the predetermined vibration level% is within the set value.
[0024]
(3) The vibration level at a predetermined frequency is within a set value.
[0025]
(4) The difference between the peak vibration level value and the reference level value is greater than or equal to the set value.
[0026]
Such vibrations that deviate from the conditions of the items are considered to be due to noise, and can be removed and removed from the recording symmetry, so that the observation accuracy can be improved and the economics of recording can be improved.
[0027]
Moreover, although it can determine whether it is a noise or an earthquake for each item, it can also be used in a superimposed manner. That is, vibrations that deviate from one of the above conditions may be removed as noise. Alternatively, a vibration that satisfies all the above conditions may be selected as a true earthquake.
[0028]
Next, an embodiment of the present invention will be described with reference to FIG. FIG. 1 is a diagram showing the result of frequency analysis of vibration measured by a seismometer, with the horizontal axis representing frequency Hz (Hertz) and the vertical axis representing vibration level dB (decibel). In this embodiment, data of 3 seconds (300 samples) before the time of the trigger vibration and about 7 seconds (723 samples) after the time of the trigger detected by the sensor 1 are taken out, and the high-speed CPU in the computer 3 is extracted. The Fourier transform was performed.
[0029]
In Figure 1, fp frequency Hz for vibrating peaks, ap is the average decibel level of vibration, an, is in vibration level in a frequency component of 40 to 50 Hz at the peak frequency fp Hz, which the reference level To do. a70 dB is a value obtained by adding 70% of the difference between the peak value ap and the reference level an to the reference level, that is, a70 = an + (ap−an) × (70/100), r20 percent is 0% of the reference level an, Level at a frequency component of 20 Hz when the peak value ap is 100%, that is, r20 = (a20−an) / (ap−an) × 100, W Hertz is a frequency bandwidth having a vibration level of a70 or higher, The ad decibel is the difference between the peak level and the reference level, that is, ad = ap−an.
[0030]
Here, the property of each parameter will be described.
[0031]
Peak frequency fp hertz: The peak frequency of earthquakes is most often distributed between 1 and 15 hertz. In order to check whether the measured vibration satisfies this condition, a lower limit flw hertz and an upper limit fhi (f1 of the present invention) hertz are set for the peak frequency.
[0032]
70% bandwidth W Hertz: When frequency analysis of earthquake and noise waveforms is performed, W is often 15 Hz or less in the case of earthquakes, and W is often 10 Hz or more in the case of noises. In order to examine whether or not the measured vibration satisfies this condition, an upper limit Wmax (W1 of the present invention) of W is set.
[0033]
20 Hz level r20: In the case of noise such as cars, r20 often exceeds 60%. In order to check whether or not the measured vibration satisfies this condition, an upper limit r20max decibel (r1 of the present invention) of r20 is set.
[0034]
Difference between peak value ap and reference level an: In order to investigate how many times the vibration level of the measured peak frequency of vibration corresponds to the reference level, a lower limit amin (a1 of the present invention) decibel is set. .
[0035]
Hereinafter, test results which are examples of the present invention will be described and described. In this test, the parameters were set as follows.
[0036]
flw = 1 hertz fhi = 15 hertz Wmax = 15 hertz r20max = 60%
amin = 10 dB (about 3 times the standard level)
The number of samples was 76 (24 earthquakes, 48 noises, 4 unknowns). As a result of the test, 31 cases were judged to be earthquakes, and 45 cases were judged to be noises. That is, 45 out of 48 actual noises were determined to be noises, and the noise discrimination accuracy was 92% in the case of this example, and the effect of noise removal was quite high.
[0037]
In this embodiment, the fact that noise is widely distributed from low to high frequencies (5 to 20 hertz) is utilized, and in particular, the noise and impact of the car generated relatively near the sensor. The discrimination accuracy was high for noise.
[0038]
Specific test results will be described with reference to FIGS.
[0039]
FIG. 2 shows the relationship between the vibration level r20% at a frequency of 20 Hz and the frequency. The horizontal axis is r20%, and the vertical axis is the frequency of samples falling within each% width. In FIG. 2, (a) shows the case of earthquake and (b) shows the case of noise. As can be seen from (a), in the present example, in the case of an earthquake, r20 is below 60%, and the noise exceeds 50%. Therefore, it can be seen that the determination can be made with a considerable probability only under this condition.
[0040]
FIG. 3 shows the relationship between the peak frequency fp and the frequency. The horizontal axis is fp hertz, and the vertical axis is the frequency of samples falling within each frequency width. In FIG. 3, (a) shows the case of earthquake and (b) shows the case of noise. As can be seen from (a), in this example, fp is below 15 hertz for earthquakes, most samples are below 12 hertz, above 3 hertz for noise, and many samples are 9 hertz. You can see that Therefore, it is understood that accurate determination cannot be made only with this condition, but accuracy can be improved by combining with other conditions.
[0041]
FIG. 4 shows the relationship between 70% bandwidth W and frequency. The horizontal axis is W hertz, and the vertical axis is the frequency of samples falling within each frequency width. In FIG. 4, (a) shows the case of earthquake and (b) shows the case of noise. As can be seen from (a), in this example, W is below 15 hertz for earthquakes, many samples are below 12 hertz, above 6 hertz for noise, and many samples are 12 hertz. You can see that Therefore, it can be seen that a certain degree of discrimination is possible only with this condition, and accuracy can be improved by combining with other conditions.
[0042]
FIG. 5 shows the determination result. S014 to s423 in the left column are data numbers corresponding to 76 samples. The central column shows the check result by the person, and the right column shows the determination result by the simulator. In the figure, 0 is determined to be noise, 1 is determined to be an earthquake, and 2 is determined to be unknown. For example, in the data number s113, it is unknown whether the check result by a person is an earthquake or noise, whereas the simulator is determined to be an earthquake. In the data number s405, the check result by a person is noise, whereas the simulator is an earthquake. Although slight differences such as judging are recognized, most of them agree with each other, and the above-mentioned result can be read regarding the high accuracy of the simulator.
[0043]
6 and 7 show the previous four items: (1) peak frequency, (2) bandwidth of predetermined vibration level%, (3) vibration level at a frequency of 20 Hz, (4) peak vibration level value and reference level value. Are the results of determining the three directions EW, NS, and UD, respectively. Comprehensive judgment is a summary of the judgment that the previous judgment is an earthquake under AND conditions. FIG. 7 is a diagram showing a list following FIG. For example, in the data number s406, most of the judgments including the vibration level 3 direction at the frequency 20 Hz with the highest reliability are the judgments of the earthquake, but the level value difference NS is the judgment of the noise, so the comprehensive judgment is noise. As shown in FIG. 5, the result is the same as the human check result.
[0044]
Thus, the possibility of obtaining a correct determination is increased by processing the conditions of the four items with AND.
[0045]
In the above embodiment, the predetermined frequency y1 hertz is set to 20 hertz. However, y1 is one standard, and may be appropriately selected according to the type of earthquake and the type of noise. The rmax (or r1) for determining ry1 may be determined. However, generally, any frequency within the range of 15 to 25 hertz is desirable, and more desirably about 17 to 23 hertz.
[0046]
Further, although the predetermined value r1 percent is 60 percent, an appropriate value may be set according to y1, and it does not stick to 60 percent. Appropriate selection may be made according to the type of earthquake and the type of noise. However, generally 55 to 65 percent is desirable, and more desirably 57 to 63 percent.
[0047]
The predetermined ratio x is set to 70%, but the ratio x is one standard and may be appropriately selected depending on the type of earthquake and the type of noise, and an appropriate Wmax (or W1) may be determined. However, generally, any ratio within the range of 65 to 75% is desirable, and more desirably about 67 to 73%.
[0048]
In addition, although the predetermined frequency width W1 is 15 Hz, W1 is one standard and may be appropriately selected depending on the type of earthquake and the type of noise. However, in general, any frequency within the range of 10 to 20 Hz is desirable, and more desirably about 12 to 18 hertz.
[0049]
In addition, although the predetermined frequency f1 is 15 Hz, f1 is one standard and may be appropriately selected depending on the type of earthquake and the type of noise. However, in general, any frequency within the range of 10 to 20 Hz is desirable, and more desirably about 12 to 18 hertz.
[0050]
The predetermined value a1 is 10 dB, but a1 is one standard and may be appropriately selected depending on the type of earthquake and the type of noise. However, generally, any frequency within the range of 5 to 15 Hz is desirable, and more desirably about 8 to 13 dB.
[0051]
【The invention's effect】
As described above, the present invention decomposes vibrations measured by a seismometer into vibration levels for frequency components, and for a predetermined frequency of the vibrations, a peak frequency, a vibration level% bandwidth, a vibration level, and a peak vibration level. Because it is judged whether the vibration is caused by an earthquake or noise based on the difference between the value and the reference level value, it is possible to reliably determine the noise vibration among the vibrations observed by the seismometer, and to extract only the vibration caused by the true earthquake. .
[Brief description of the drawings]
FIG. 1 is a diagram showing a typical example of frequency analysis used in the present invention.
FIG. 2 is a diagram showing the relationship between the level% at 20 Hz and the frequency for the sample of the embodiment of the present invention.
FIG. 3 is a diagram showing the relationship between the peak frequency and the frequency for the sample of the embodiment of the present invention.
FIG. 4 is a diagram showing the relationship between 70% bandwidth and frequency for a sample of an example of the present invention.
FIG. 5 is a diagram showing a comprehensive determination result of an example of the present invention.
FIG. 6 is a diagram showing an individual determination result and a comprehensive determination result of a sample indicated by each data number of the present invention.
FIG. 7 is a diagram showing a continuation of FIG. 6;
FIG. 8 is a diagram showing an outline of a seismometer system used in the present invention.
[Explanation of symbols]
1 Sensor 2 Seismometer body 3 Computer

Claims (10)

The vibration measured by the seismometer is decomposed into intensity levels with respect to the frequency component, the vibration level of the predetermined frequency y1 hertz component of the vibration is ay1 decibels, and a reference for determining whether the vibration is an earthquake vibration or a noise vibration When the level is an decibel, and the vibration level at the vibration frequency fp hertz that gives the vibration peak is ap decibel, the following expression ry1 = (ay1-an) / (ap-an) × 100
The seismic observation method for discriminating the vibration as noise when the ry1 percent given in (1) exceeds a predetermined value r1 percent. The earthquake observation method according to claim 1, wherein the predetermined frequency y1 hertz is any frequency within a range of 15 to 25 hertz. The earthquake observation method according to claim 1 or 2, wherein the predetermined value r1 percent is any value of 55 to 65 percent. The vibration measured by the seismometer is decomposed into intensity levels with respect to the frequency component, the reference level serving as a reference for discriminating whether the vibration is vibration caused by an earthquake or noise is an decibel, and the vibration frequency fp hertz that gives the peak of the vibration Where the vibration level is ap decibels and the predetermined ratio is x percent, the following equation: ax = an + (ap−an) × (x / 100)
A seismic observation method for discriminating vibration as noise when the bandwidth W hertz in ax decibels exceeds a predetermined frequency width W1 hertz. The earthquake observation method according to claim 4, wherein the predetermined ratio x percent is any value within a range of 65 to 75 percent. The earthquake observation method according to claim 4 or 5, wherein the predetermined frequency width W1 hertz is any value of 10 to 20 hertz. The vibration measured by the seismometer is decomposed into intensity levels with respect to the frequency component, the vibration level of the predetermined frequency y1 hertz component of the vibration is ay1 decibels, and a reference for determining whether the vibration is an earthquake vibration or a noise vibration When the level is an decibel and the vibration level at the vibration frequency fp hertz at which the vibration peak is given is ap decibel, the following equation ry1 = (ay1-an) / (ap-an) × 100
When the ry1 percent given by (1) exceeds a predetermined value r1 percent, the vibration is determined as noise, and
When the predetermined ratio is x percent, when the bandwidth W hertz in ax decibels given by the following equation ax = an + (ap−an) × (x / 100) exceeds the predetermined frequency width W1 hertz An earthquake observation method that distinguishes vibration from noise. The predetermined frequency y 1 hertz is any frequency within the range of 15 to 25 hertz, the predetermined value r1 percent is any value of 55 to 65 percent, and the x percent is 65 to 75. The earthquake observation method according to claim 7, wherein the ratio is any ratio in a range of percent, and the predetermined frequency width W1 hertz is any value of 10 to 20 hertz. Further, when the peak frequency fp hertz exceeds a predetermined frequency f1, the vibration is determined as noise, and the following expression ad = ap−an
The earthquake observation method according to any one of claims 1 to 8, wherein when the vibration level difference ad decibel given by (1) does not exceed a predetermined value a1, the vibration is determined as noise. The earthquake observation method according to any one of claims 1 to 9, wherein the vibration measured by a seismometer is measured in each of two directions orthogonal to a horizontal direction and one direction perpendicular to the two directions.

Images