An online method for identifying abnormal operation of geophones in a microseismic monitoring system

Abstract

The present application relates to the field of wave detector anomaly detection, in particular to an online discrimination method for abnormal operation of a wave detector in a microseismic monitoring system, comprising the following steps: 1, intercepting stationary observation data; 2, filtering the intercepted observation data; 3, regionally decomposing the filtered observation data; 4, data reconstruction and energy change calculation on the data in step 3; 5, obtaining fitting parameters; 5, obtaining mutation points of parameters changing with time; 6, establishing an anomaly discrimination algorithm to judge the working state of the wave detector, the present application effectively obtains wave detector data information, establishes a wave detector working state automatic discrimination algorithm, can effectively improve the timeliness of judging the working state of the wave detector, ensures the effective work of the microseismic system, further improves the observation data quality, and makes the microseismic event pickup and microseismic positioning result more reliable.

Description

Technical Field

[0001] This invention relates to the field of detector anomaly detection, specifically to an online method for identifying abnormal operation of detectors in a microseismic monitoring system. Background Technology

[0002] Microseismic monitoring systems are mainly deployed in coal mines, where the acquired microseismic event signals are weak and occur at low frequencies. During continuous acquisition, detector malfunctions and human sabotage may occur. Therefore, determining whether the sensors and microseismic systems are functioning properly is crucial. Currently, the normal operation of microseismic systems and detectors is mainly determined by the frequency of microseismic event acquisition and long-term observation. However, this method is highly lagging, especially for detectors deployed in areas where microseismic events are infrequent. Therefore, it is essential to propose a more accurate and lag-free method for determining the detector's operational status. Summary of the Invention

[0003] The purpose of this invention is to provide an online method for identifying abnormal operation of a geophone in a microseismic monitoring system, so that the detection of the geophone is lag-free and the working status of the geophone can be accurately determined, thereby ensuring the working quality of the geophone and solving the existing technical defects and unmet technical requirements.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an online method for identifying abnormal operation of a detector in a microseismic monitoring system, comprising the following steps:

[0005] 1) Extract stable observation data;

[0006] 2) Filter the extracted observation data;

[0007] 3) Perform region decomposition on the filtered observation data;

[0008] 4) Reconstruct the data and calculate the energy changes from the data in step 3);

[0009] 5) Obtain the fitting parameters;

[0010] 6) Obtain the abrupt change points of parameters over time;

[0011] 7) Establish an anomaly detection algorithm to determine the working status of the detector.

[0012] In step 1), for continuously acquired observation data, data of a certain window length is extracted. The stable observation data does not contain microseismic events or other sudden changes.

[0013] This application establishes an anomaly detection method and reflects the working status of the detector through real-time monitoring data. It can accurately judge and check whether the detector is abnormal by means of lag, thereby ensuring the working quality of the detector.

[0014] In step 2), a Butterworth bandpass filter with a frequency range of 10Hz-300Hz is used to filter the data.

[0015] The purpose of selecting a filter in this frequency band in this application is to better preserve waveform information.

[0016] In step 3, the specific method of region decomposition is as follows: wavelet packet decomposition is used to decompose the filtered observation data into regions.

[0017] The frequency domain decomposition method described above is used in this application to effectively analyze all spectral components within a wide bandwidth, exhibiting time-frequency localization characteristics and improving the time-frequency resolution of the signal.

[0018] 1. The number of layers in the region decomposition is 7, and the wavelet packet decomposition method uses wavelet basis functions, specifically Meyer wavelets.

[0019] This application uses this method to improve resolution.

[0020] In step 4), the object of signal reconstruction is the signal within each node frequency band after wavelet packet decomposition;

[0021] The specific method for calculating the energy change is as follows: the energy change is represented by calculating the energy entropy value of each node after wavelet packet decomposition.

[0022] This application uses this calculation to obtain the energy distribution of the observed signal across the entire frequency band.

[0023] In step 5), the method for obtaining the fitting parameters is to obtain the fitting parameters using the distribution fitting method of maximum likelihood estimation.

[0024] The purpose of using the above method to obtain fitting parameters in this application is mainly to extract feature parameters that can represent the working state of the detector. Since the wavelet packet decomposition calculates the energy entropy of each node, which is discrete data, the maximum likelihood estimation method is chosen to fit the discrete data distribution.

[0025] The specific operation for obtaining the fitting parameters is as follows: calculate the energy distribution multiple times within a certain period of time, construct a new energy distribution using the most recent energy distribution and the previous energy distribution, and then calculate the fitting parameter results using the maximum likelihood estimation distribution fitting method.

[0026] This application operates in this way to avoid the isolation of the fitting results.

[0027] In step 6), the observation data is first continuously intercepted using the sliding window method with a window length of 4-6h and an overlap time of 1h. Then, the fitting parameters of the energy distribution are calculated. The method for obtaining the abrupt change points of the parameters over time is the Bayesian online change point detection method.

[0028] The reason for choosing this method for data processing in this application is that this method can be an online detection method that uses probability theory to predict the probability of a new point continuing the previous segment and the probability of a mutation point. Based on the final calculation result, it is determined whether it is a mutation point, which is suitable for online estimation of the detector state.

[0029] 2. In step 7), the method for judging the working state of the detector is as follows: set a threshold for the duration of the mutation point. If the mutation point lasts for a long time and exceeds the threshold, or if no new mutation point is generated, it is judged that the working state of the detector is abnormal. Otherwise, it is considered that the working state of the detector can be restored by itself and is not judged as abnormal.

[0030] The reason why this application judges the detector state based on the characteristics of the mutation point is that the detector is in a stable acquisition state. When the detector state changes, the acquired data will also change.

[0031] Compared with the prior art, the beneficial effects of the present invention are: the present application effectively acquires the data information collected by the detector and establishes an automatic judgment algorithm for the working status of the detector, which can effectively improve the timeliness of judging the working status of the detector, ensure the effective operation of the microseismic system, further improve the quality of observation data, and make the results of microseismic event picking and microseismic location more reliable. Attached Figure Description

[0032] Figure 1 This is a flowchart of the method of the present invention; Detailed Implementation

[0033] The following will refer to the appendices in the embodiments of the present invention. Figure 1 The technical solutions in the embodiments of the present invention are clearly and completely described herein. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0034] Please see Figure 1 Embodiments of the present invention

[0035] Example:

[0036] The specific operation method and data processing method of this application are as follows:

[0037] 1. For continuously acquired observation data, select data with a 1-hour window length. This data does not include microseismic events and other sudden changes, and only considers observation data in a stable state.

[0038] 2. The intercepted observation data is filtered. To better preserve waveform information, a Butterworth bandpass filter with a frequency range of 300Hz is selected.

[0039] 3. Wavelet packet decomposition was used to perform frequency domain decomposition on the filtered observation data. To improve resolution, the number of layers was selected as 7, and the wavelet basis function was Meyer wavelet.

[0040] 4. Reconstruct the signals within the frequency bands of each node after wavelet packet decomposition, and use energy entropy to calculate the energy change of each node after wavelet packet decomposition, thereby obtaining the energy distribution of the observed signal in the entire frequency band.

[0041] 5. Obtain fitting parameters using the maximum likelihood estimation distribution fitting method. To avoid the isolation of the fitting results, select observation data with a duration of 4-6 hours to calculate the energy distribution within a window of 1 hour multiple times. Superimpose the energy distributions several times to construct a new energy distribution. Then, use the maximum likelihood estimation distribution fitting method to calculate the fitting parameters of the new energy distribution.

[0042] 6. Using the sliding window method, continuously extract window length data, and obtain the fitting parameter results of the energy distribution through the above steps. Then, use the Bayesian online change point detection method to obtain the abrupt change points of the parameters over time.

[0043] 7. An anomaly detection algorithm is established. This application considers the duration of the mutation point and sets a threshold for the duration of the mutation point. If the mutation point lasts for a long time and exceeds the threshold, or if no new mutation point is generated, it is determined that the working state of the detector is abnormal. Otherwise, it is considered that the working state of the detector can be restored by itself and is not considered as an anomaly.

[0044] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the scope of the invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0045] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. An online method for identifying abnormal operation of a detector in a microseismic monitoring system, comprising the following steps: 1) Extract stable observation data; 2) Filter the extracted observation data; 3) Perform frequency domain decomposition on the filtered observation data; 4) Reconstruct the data and calculate the energy changes from the data in step 3); 5) Obtain the fitting parameters; 6) Obtain the abrupt change points of parameters over time; 7) Establish an anomaly detection algorithm to determine the working status of the detector; In step 5), the method for obtaining the fitting parameters is: to obtain the fitting parameters using the distribution fitting method of maximum likelihood estimation; The specific operation for obtaining the fitting parameters is as follows: select observation data with a duration of 4-6 hours, calculate the energy distribution within a window of 1 hour multiple times, superimpose the energy distributions several times to construct a new energy distribution, and then use the maximum likelihood estimation distribution fitting method to calculate the fitting parameters of the new energy distribution.

2. The online method for identifying abnormal operation of a detector in a microseismic monitoring system according to claim 1, characterized in that, In step 1), for continuously collected observation data, data with a 1-hour window length is extracted. The stationary observation data does not include microseismic events, electromagnetic fields, working conditions, or human activities.

3. The online method for identifying abnormal operation of a detector in a microseismic monitoring system according to claim 1, characterized in that, In step 2), a Butterworth bandpass filter with a frequency range of 10Hz-300Hz is used to filter the data.

4. The online method for identifying abnormal operation of a detector in a microseismic monitoring system according to claim 1, characterized in that, In step 3, the frequency domain decomposition is specifically performed by using wavelet packet decomposition to decompose the filtered observation data in the frequency domain.

5. The online method for identifying abnormal operation of a detector in a microseismic monitoring system according to claim 4, characterized in that, The frequency domain decomposition has 7 layers, and the wavelet packet decomposition method uses wavelet basis functions, specifically Meyer wavelets.

6. The online method for identifying abnormal operation of a detector in a microseismic monitoring system according to claim 1, characterized in that, In step 4), the object of data reconstruction is the signal within the frequency band of each node after wavelet packet decomposition; The specific method for calculating the energy change is as follows: the energy change is represented by the energy entropy value of each node after wavelet packet decomposition.

7. The online method for identifying abnormal operation of a detector in a microseismic monitoring system according to claim 1, characterized in that, In step 6), the observation data is first continuously intercepted using the sliding window method with a window length of 4-6h and an overlap time of 1h, and then the fitting parameters of the energy distribution are calculated. The method for obtaining the abrupt change points of the parameters over time is the Bayesian online change point detection method.

8. The online method for identifying abnormal operation of a detector in a microseismic monitoring system according to claim 1, characterized in that, In step 7), the method for judging the working state of the detector is as follows: a threshold for the duration of the mutation point is set. If the mutation point lasts for a long time and exceeds the threshold, or if no new mutation point is generated, the working state of the detector is judged to be abnormal. Otherwise, it is considered that the working state of the detector can be restored by itself and is not judged as abnormal.

Images