A noise detection method for MEMS seismic intensity meters of an earthquake early warning network

Abstract

The present application relates to the field of earthquake early warning technology, in particular to a noise detection method for MEMS seismic intensity meter of earthquake early warning network, comprising sensitivity coefficient check, vertical polarity determination, directional data missing determination and short-time abnormality detection, compared with the prior art, the present application removes abnormal data noise through sensitivity coefficient comparison, determination of station installation direction and determination of three directional data, and can determine and process short-time abnormal noise, which helps to improve the accuracy of earthquake early warning processing results.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of earthquake early warning, in particular to a noise detection method for MEMS seismic intensity meters of an earthquake early warning network. BACKGROUND

[0002] The seismic intensity meter is the basis of earthquake early warning and intensity rapid reporting, and can provide scientific basis for emergency early warning disposal, rapid disaster judgment, earthquake emergency rescue strategy and action, and engineering rescue repair decision, etc. The device monitors the X, Y and Z three-axis acceleration of earthquake vibration through the MEMS sensor, and transmits the data to the cloud platform or the center server in real time through the 3G / 4G / 5G network.

[0003] During the use of the MEMS seismic intensity meter, it is found that due to the noise interference source near the station where the intensity meter is located, the noise level of the instrument itself is high, and the parameters, directional configuration and other problems, which lead to the failure to accurately detect whether it is a normal earthquake event, thereby affecting the correct handling of earthquake early warning and intensity rapid reporting.

[0004] During the noise analysis of the intensity meter station, it is found that some stations have the following problems:

[0005] Affected by strong noise interference, abnormal parameter configuration leads to noise being multiplied too large or too small, directional configuration error leads to incorrect polarity of three directional, and device problems lead to missing of part of directional data.

[0006] Based on the above reasons, the present application designs a noise detection method for MEMS seismic intensity meters of an earthquake early warning network, which removes abnormal data during data preprocessing, ensures real-time collection of seismic event waveform data, and provides basis for correct handling of earthquake early warning and intensity rapid reporting. SUMMARY

[0007] The purpose of the present application is to overcome the shortcomings of the prior art, provide a noise detection method for MEMS seismic intensity meters of an earthquake early warning network, which removes abnormal data during data preprocessing, ensures real-time collection of seismic event waveform data, and provides basis for correct handling of earthquake early warning and intensity rapid reporting.

[0008] In order to achieve the above purpose, the present application provides a noise detection method for MEMS seismic intensity meters of an earthquake early warning network, which includes sensitivity coefficient inspection, vertical polarity determination, directional data missing determination and short-time abnormality detection.

[0009] The sensitivity coefficient inspection is:

[0010] According to the station sensitivity coefficient stored in the database, the sensitivity coefficient in the received data packet information is compared. If the deviation exceeds 5%, it is considered that the station sensitivity coefficient is wrong, and the corresponding station is marked as station sensitivity abnormality;

[0011] If the sensitivity coefficient value is not included in the data, the root mean square RMS value of the received UD / EW / NS three-direction data is compared with the upper and lower limits of the normal RMS value of the corresponding station equipment. When the station RMS value exceeds the upper limit of the reference value, the corresponding station is identified as station sensitivity abnormality;

[0012] The vertical direction polarity determination is:

[0013] When the installation method is ground installation, it is determined whether the UD direction RMS value is greater than the EW and NS direction RMS values. If it is less than any one of them, the corresponding station is identified as vertical direction polarity abnormality;

[0014] When the installation method is wall-mounted installation, it is determined whether the RMS value of the EW or NS direction is greater than the UD direction. If the RMS values of the EW and NS directions are not greater than the UD direction, the corresponding station is identified as vertical direction polarity abnormality;

[0015] The detection interval of the vertical direction polarity determination includes a specified time interval, the first time the station data is received, the abnormal station is reconnected online, and the center end is manually specified. The station data length is not less than 5 minutes, and in the station non-trigger state;

[0016] The direction data missing determination is:

[0017] When the received three-direction data has individual direction data missing, i.e. all 0 values, the corresponding station is identified as direction data missing abnormality. The detection interval is set as: specified time interval, first time the station data is received, abnormal station is reconnected online, and center end is manually specified. The three-direction data length is not less than 1 minute, and in the station non-trigger state;

[0018] The short-time abnormality detection is:

[0019] The waveform after real-time triggering is determined to prevent abnormal data from entering the subsequent data processing process, and 1s or 0.5s waveform after triggering is detected. The specific steps include:

[0020] S1, calibration abnormality detection: using zero-crossing method, first intercept the original waveform from 9s before triggering to 1s or 0.5s after triggering, calculate the peak acceleration PGA or peak velocity PGV of the wave band, then calculate the intersection point of 0.1×PGA or 0.1×PGV. If the number of intersection points is less than 3 and the waveform turning point is one, it is determined as calibration abnormality. Otherwise, go to the next step;

[0021] S2, pulse or spike anomaly detection: intercept the data segment C1 after 0-0.4s or 0-0.2s of triggering, and the data segment C2 after 0.4s-0.8s or 0.2s-0.4s of triggering, and the amplitude change of the two data segments, that is, the amplitude ratio of the former segment to the latter segment, if the amplitude change ratio is greater than 6, it is determined as pulse or spike anomaly, otherwise, it goes to the next step;

[0022] S3, step anomaly detection: by constructing 5 different PGA or PGV percentage dividing lines, the exceeding probability of each dividing line is calculated, if there is the same exceeding probability, and the characteristic period τ c > 1s, it is determined as step anomaly, otherwise, it goes to subsequent normal processing.

[0023] Sensitivity coefficient check, vertical polarity determination and directional data missing determination are aimed at abnormal configuration of sensitivity coefficient of equipment end, vertical polarity anomaly and sharing data missing.

[0024] Compared with the prior art, the abnormal data noise is removed by the sensitivity coefficient comparison judgment, the station installation direction judgment and the three directional data judgment, and the short-time abnormal noise can be determined and processed, and the processing result accuracy of earthquake early warning and intensity rapid reporting is further improved. DETAILED DESCRIPTION

[0025] The application provides a noise detection method for a MEMS seismic intensity instrument of an earthquake early warning station network, which comprises sensitivity coefficient check, vertical polarity determination, directional data missing determination and short-time anomaly detection.

[0026] The sensitivity coefficient check is:

[0027] According to the station sensitivity coefficient stored in the database and the sensitivity coefficient in the received data packet information, if the deviation exceeds 5%, it is considered that the station sensitivity coefficient is wrong, and the corresponding station is marked as a station sensitivity anomaly;

[0028] If the sensitivity coefficient value is not included in the data, the root mean square RMS value of the UD / EW / NS three directional data received is compared with the upper and lower limits of the normal RMS value of the corresponding station equipment, and when the station RMS value exceeds the upper and lower limits of the reference value, the corresponding station is marked as a station sensitivity anomaly;

[0029] The vertical polarity determination is:

[0030] When the installation mode is ground installation, it is determined whether the UD direction RMS value is greater than the EW and NS direction RMS values, if it is smaller than any one of them, the corresponding station is marked as a vertical polarity anomaly;

[0031] When the installation method is wall-mounted, determine whether the RMS value of the EW or NS direction is greater than that of the UD direction. If the RMS values ​​of both the EW and NS directions are not greater than those of the UD direction, then mark the corresponding station as having vertical polarity anomaly.

[0032] The detection interval for vertical polarity determination includes a specified time interval, the first time the station data is received, the abnormal station coming back online, and manual specification at the central end; the station data length is not less than 5 minutes, and it must be in a state where the station is not triggered;

[0033] Missing data in the distribution direction is determined as follows:

[0034] When some of the received three-way data is missing (i.e., all values ​​are 0), the corresponding station is marked as having missing data. The check interval is set as: a specified time interval, the first time the station data is received, the abnormal station comes back online, and the center manually specifies the check interval. The length of the three-way data is not less than 1 minute, and the check is performed when the station is not triggered.

[0035] Short-term anomaly detection is as follows:

[0036] The waveform after real-time triggering is evaluated to prevent abnormal data from entering subsequent data processing flows. Detection is performed on the waveform at 1 second or 0.5 seconds after triggering. Specific steps include:

[0037] S1, Calibration Anomaly Detection: Using the zero-crossing method, firstly, extract the original waveform from 9 seconds before triggering to 1 second or 0.5 seconds after triggering, calculate the peak acceleration PGA or peak velocity PGV of this band, and then calculate the intersection of 0.1×PGA or 0.1×PGV. If the number of intersection points is less than 3 and there is only 1 waveform inflection point, it is determined to be a calibration anomaly; otherwise, proceed to the next step.

[0038] S2, Pulse or spike anomaly detection: Extract data segment C1 from 0 to 0.4s or 0 to 0.2s after triggering, and data segment C2 from 0.4s to 0.8s or 0.2s to 0.4s after triggering. Calculate the amplitude change between these two data segments, i.e., the amplitude of the first segment is greater than the amplitude of the second segment. If the amplitude change ratio is greater than 6, it is determined to be an anomaly such as a pulse or spike; otherwise, proceed to the next step.

[0039] S3, Step Anomaly Detection: By constructing 5 dividing lines with different PGA or PGV percentages, the exceedance probability of each dividing line is calculated. If there are identical exceedance probabilities and the characteristic period τ... c If the time exceeds 1 second, it is determined to be a step anomaly; otherwise, it proceeds to normal processing.

[0040] Sensitivity coefficient check, vertical polarity determination, and missing directional data determination are for situations where the device-side sensitivity coefficient configuration is abnormal, the vertical polarity is abnormal, and the shared data is missing.

[0041] The above are only preferred embodiments of the present application, and are used to help understand the method and its core idea of the present application. The protection scope of the present application is not limited to the above-mentioned embodiments. Any technical solution falling within the idea of the present application belongs to the protection scope of the present application. It should be noted that, for ordinary skilled persons in the art, some improvements and refinements without departing from the principle of the present application are also considered to be within the protection scope of the present application.

[0042] The present application solves the problem that the real seismic event data cannot be accurately detected due to the noise interference source near the station where the intensity instrument is located, the high noise level of the instrument itself, the parameter configuration, and the like, thereby affecting the correct processing of the earthquake warning and intensity rapid reporting. The abnormal data is removed during the data preprocessing, the real-time collection of the seismic event waveform data is ensured, and the basis for the normal processing of the earthquake warning and intensity rapid reporting is provided.

Claims

1. A noise detection method for MEMS seismic intensity meters in earthquake early warning networks, characterized in that, This includes sensitivity coefficient checking, vertical polarity determination, directional data missing determination, and short-term anomaly detection; The sensitivity coefficient check is as follows: The sensitivity coefficients of the stations stored in the database are compared with the sensitivity coefficients in the received data packets. If the deviation exceeds 5%, the station sensitivity coefficient is considered to be incorrect, and the corresponding station is marked as having abnormal sensitivity. If the sensitivity coefficient value is not included in the data, the root mean square RMS value of the received UD / EW / NS triaxial data is compared with the upper and lower limits of the normal RMS value of the corresponding station equipment. When the station RMS value exceeds the upper and lower limits of the reference value, the corresponding station is marked as having abnormal station sensitivity. The vertical polarity determination is as follows: When the installation method is ground installation, it is determined whether the RMS value in the UD direction is greater than the RMS value in the EW and NS directions. If it is less than either of them, the corresponding station is marked as having vertical polarity anomaly. When the installation method is wall-mounted installation, it is determined whether the RMS value in the EW or NS direction is greater than the UD direction. If the RMS values ​​in both the EW and NS directions are not greater than the UD direction, the corresponding station is marked as having vertical polarity anomaly. The detection interval for determining vertical polarity includes a specified time interval, the first receipt of station data, the re-connection of an abnormal station, and manual specification at the central end; the station data length is not less than 5 minutes, and it is in the case of a station not being triggered. The determination of missing directional data is as follows: When some of the received three-way data is missing (i.e., all values ​​are 0), the corresponding station is marked as having missing data. The check interval is set as: a specified time interval, the first time the station data is received, the abnormal station comes back online, and the center manually specifies the interval. The length of the three-way data is not less than 1 minute, and the station is not triggered. The short-term anomaly detection is as follows: The waveform after real-time triggering is evaluated to prevent abnormal data from entering subsequent data processing flows. Detection is performed on the waveform at 1 second or 0.5 seconds after triggering. Specific steps include: S1, Calibration Anomaly Detection: Using the zero-crossing method, firstly, extract the original waveform from 9 seconds before triggering to 1 second or 0.5 seconds after triggering, calculate the peak acceleration PGA or peak velocity PGV of this band, and then calculate the intersection of 0.1×PGA or 0.1×PGV. If the number of intersection points is less than 3 and there is only 1 waveform inflection point, it is determined to be a calibration anomaly; otherwise, proceed to the next step. S2, Pulse or spike anomaly detection: Extract data segment C1 from 0 to 0.4s or 0 to 0.2s after triggering, and data segment C2 from 0.4s to 0.8s or 0.2s to 0.4s after triggering. Calculate the amplitude change between these two data segments, i.e., the amplitude of the first segment is greater than the amplitude of the second segment. If the amplitude change ratio is greater than 6, it is determined to be an anomaly such as a pulse or spike; otherwise, proceed to the next step. S3, Step Anomaly Detection: By constructing 5 dividing lines with different PGA or PGV percentages, the exceedance probability of each dividing line is calculated. If there are identical exceedance probabilities and the characteristic period τ... c If the time exceeds 1 second, it is determined to be a step anomaly; otherwise, it proceeds to normal processing.

2. The noise detection method for MEMS seismic intensity meters in earthquake early warning networks according to claim 1, characterized in that, The sensitivity coefficient check, vertical polarity determination, and directional data missing determination are for situations where the device-side sensitivity coefficient configuration is abnormal, the vertical polarity is abnormal, and the directional data is missing.