A dynamic measurement and monitoring system for coal mining cracks

Abstract

The present invention belongs to the technical field of coal mining crack measurement, and specifically discloses a coal mining crack dynamic measurement and monitoring system, which includes a crack signal acquisition module, a crack signal processing module, a crack risk determination module, and a monitoring adjustment feedback terminal. By adopting a distributed high-frequency acoustic emission sensor array and combining a rupture event trigger threshold with time-domain / frequency-domain feature extraction technology, the present invention can capture the tiny energy release signals of rock burst precursors in real time, breaking through the limitation of traditional microseismic monitoring's insufficient sensitivity to tiny signals. By dynamically adjusting the frequency band division weights through wavelet packet decomposition technology, the present invention can effectively distinguish between noise and true crack expansion signals, preventing the tiny energy features from being submerged. At the same time, by dynamically adjusting the sampling frequency, sensor activation ratio, and data transmission frequency, the present invention ensures the integrity and real-time nature of the crack expansion signal, thus making up for the shortcomings of the traditional single data source.

Description

Technical Field

[0001] The present invention belongs to the technical field of coal mining crack measurement, and in particular relates to a coal mining crack dynamic measurement and monitoring system. Background Art

[0002] During coal mining, the dynamic evolution of rock cracks is directly related to underground safety and mining efficiency. Cracks expand in real time as the working face advances, requiring high-frequency and continuous monitoring to capture transient changes in cracks, especially the precise capture and measurement of fine cracks to ensure the safety of coal mining.

[0003] Existing technologies, such as the Chinese invention patent application with application number 202311644095.6, disclose a method for measuring cracks in the shaft wall structure of a coal mine ventilation shaft. The method solves the problem of insufficient accuracy in identifying fine crack deformations caused by noise interference and static measurement errors by setting up multi-stage point cloud data acquisition and filtering processing technology, combined with a gap measurement comparison mechanism between the initial stage and the next stage, and realizes millimeter-level dynamic deformation monitoring of crack defects in the shaft wall structure, thereby improving the safety monitoring response speed and ensuring the stability of the shaft wall structure.

[0004] Regarding the above technical solutions, the current core technology for measuring rock cracks, especially fine cracks, is to use three-dimensional point cloud spatial coordinates to quantify crack deformation, which solves the accuracy problem of static crack size measurement. Obviously, there are still the following problems in the current measurement of fine cracks: 1. Current microseismic monitoring cannot distinguish the tiny energy release characteristics of impact ground pressure precursor signals, resulting in missed warnings.

[0005] 2. The existing equipment relies on frequency settings, and the monitoring data is single, which is insufficient to supplement the crack extension signal and seamlessly obtain more detailed features of fine cracks, thus failing to guarantee the accuracy of fine crack analysis. Summary of the Invention

[0006] In view of this, a dynamic measurement and monitoring system for coal mining cracks is proposed to solve the limitations of existing technologies in preventing false alarms of water vapor, such as poor environmental adaptability, sensitivity to optical contamination, and rough feature analysis.

[0007] The purpose of the present invention can be achieved through the following technical solutions: The present invention provides a dynamic measurement and monitoring system for coal mining cracks, which includes: a crack signal acquisition module, which collects high-frequency acoustic emission signal data of underground coal and rock masses in real time, and the data includes the cumulative frequency characteristics of fracture events, time domain energy change characteristics and frequency domain energy distribution characteristics.

[0008] The crack signal processing module performs wavelet packet decomposition on the signal data to obtain frequency bands and generate signal data of different frequency bands.

[0009] The crack risk determination module performs multi-scale correlation analysis on signal data in different frequency bands to obtain the crack risk probability value.

[0010] The monitoring and adjustment feedback terminal dynamically adjusts the monitoring strategy according to the crack risk probability value, and links the underground disaster prevention and control system to execute risk feedback control instructions.

[0011] Compared with the existing technology, the beneficial effects of the present invention are as follows: (1) The present invention can capture the tiny energy release signal of the precursor of rock burst in real time by adopting a distributed high-frequency acoustic emission sensor array and combining the rupture event trigger threshold with the time domain / frequency domain feature extraction technology, breaking through the limitation of the traditional microseismic monitoring that is not sensitive enough to tiny signals. The wavelet packet decomposition technology is used to dynamically adjust the frequency band division weight. Combining the cumulative frequency characteristics of the rupture event with the average energy ratio of the frequency domain energy, the present invention can effectively distinguish between noise and real crack extension signals, avoid the submergence of tiny energy characteristics, and ensure the timeliness and accuracy of the early warning.

[0012] (2) The present invention dynamically determines the number of wavelet packet decomposition layers by combining the growth rate of fracture events and the average energy ratio in the frequency domain, and allocates the importance of frequency bands by combining the time domain / frequency domain weight coefficients, thereby significantly improving the time-frequency resolution of small crack signals.

[0013] (3) The present invention realizes multi-dimensional dynamic modeling of the crack propagation process by three-band analysis based on sliding time window and combining the stress distribution data of coal rock mass to construct a coupling matrix. Based on the multi-parameter fusion algorithm of the three-band-stress field coupling matrix, the energy mutation, correlation fluctuation and the spatial distribution of stress concentration area are associated to realize the accurate quantification of the crack risk probability value.

[0014] (4) The present invention ensures the integrity and real-time performance of the crack extension signal by dynamically adjusting the sampling frequency, sensor activation ratio, and data transmission frequency, thus making up for the shortcomings of the traditional single data source, facilitating the acquisition of more detailed features of fine cracks, and ensuring the accuracy of fine crack analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 This is a schematic diagram of the connection of various modules of the system of the present invention.

[0017] Figure 2 It is a schematic diagram of the overall implementation process of the present invention.

[0018] Figure 3 Schematic diagram of the process for confirming the final number of decomposition layers of the present invention. DETAILED DESCRIPTION

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0020] See also Figures 1 to 2 As shown, the present invention provides a coal mining crack dynamic measurement and monitoring system, which includes: a crack signal acquisition module, a crack signal processing module, a crack risk determination module and a monitoring adjustment feedback terminal.

[0021] In the above, the crack signal processing module is connected to the crack signal acquisition module and the crack risk determination module respectively, and the crack risk determination module is connected to the monitoring adjustment feedback terminal.

[0022] The crack signal acquisition module collects high-frequency acoustic emission signal data of the coal and rock mass in the well in real time. The data includes the cumulative frequency characteristics of fracture events, the time domain energy change characteristics and the frequency domain energy distribution characteristics.

[0023] It should be added that the specific installation of the distributed acoustic emission sensor array is as follows: based on the advancement direction of the coal mining face and the distribution of rock burst danger areas, the monitoring area is divided into dynamically updated acoustic emission monitoring sub-areas, and a high-frequency acoustic emission sensor array is deployed in each monitoring sub-area. The sensors cover key positions of the coal wall, roof and goaf boundary.

[0024] For example, 32 distributed acoustic emission sensor nodes are deployed on the surface of the coal and rock mass to be tested, with a node spacing of 5m x 5m and a coverage area of ​​160m x 160m. The sensor operating frequency range is 1kHz-500kHz, and the sampling frequency is dynamically adjustable.

[0025] Specifically, the specific collection process of the high-frequency acoustic emission signal data includes: A1, covering the coal and rock mass area to be measured by a distributed acoustic emission sensor array, and acquiring the time domain waveform of the acoustic emission signal in real time.

[0026] As can be understood, multiple acoustic emission sensors are deployed around the coal and rock mass area to form a distributed sensor array, ensuring comprehensive coverage of the monitoring area. These sensors detect the acoustic emission signals emitted by the coal and rock mass in real time, visually displaying the dynamic characteristics of the signals over time.

[0027] A2. Extracting a time series of rupture events from the time domain waveform based on a preset rupture event trigger threshold, and generating a cumulative frequency feature of the rupture events.

[0028] Understandably, a fracture event trigger threshold is pre-set. When the intensity of the real-time monitored acoustic emission signal exceeds this threshold, a coal-rock fracture event is identified. The occurrence of these fracture events is recorded chronologically to form a fracture event time series. The cumulative number of fracture events in different time periods is counted to obtain a cumulative fracture event frequency characteristic, which reflects the temporal changes in the frequency of coal-rock fracture activity.

[0029] A3. Perform energy integration calculation on the time domain waveform of each rupture event to generate the time domain energy variation feature.

[0030] It can be understood that the specific process of performing energy rate integral calculation on the time domain waveform of each rupture event is: performing integral calculation on the time domain waveform, obtaining the energy release value of a single rupture event, and calculating the gradient of the energy release values ​​of adjacent single rupture events as the time domain energy rate change characteristic, thereby analyzing the energy strength of each rupture event and the change pattern of the overall energy over time.

[0031] A4. Perform fast Fourier transform on the time domain waveform, extract the energy proportion within a preset frequency band, and generate the frequency domain energy distribution feature.

[0032] As can be understood, the Fast Fourier Transform (FFT) converts the time-domain waveform into a frequency-domain representation, revealing the signal's energy distribution across different frequencies. A predetermined frequency range is used to calculate the proportion of energy within that frequency range to the total energy. By analyzing the energy contribution of each frequency range, the frequency-domain energy distribution characteristic is derived, revealing the energy distribution characteristics of the acoustic emission signal along the frequency dimension.

[0033] The crack signal processing module performs wavelet packet decomposition on the signal data to obtain frequency bands, thereby generating signal data of different frequency bands.

[0034] Specifically, the specific processing process of performing wavelet packet decomposition includes: B1, performing filtering, noise reduction and normalization processing on the original high-frequency acoustic emission signal.

[0035] B2. Determine the final number of decomposition layers based on the cumulative frequency characteristics of rupture events and the frequency domain energy distribution characteristics.

[0036] B3. Based on the set wavelet basis function, the processed high-frequency acoustic emission signal is subjected to frequency band decomposition according to the confirmed final decomposition layer number to generate time-frequency components of each sub-band.

[0037] B4. Allocate frequency band weights based on the time domain energy variation characteristics and the frequency domain energy distribution characteristics, and divide the sub-frequency band into a low frequency band, a mid frequency band, and a high frequency band based on the weight allocation result.

[0038] It should be noted that the frequency band division based on the weighted allocation results can be performed according to a weighted allocation table. For example, the sub-bands are divided into the following: 1) Low frequency band (0-50kHz): weight value ≤ 0.5, representing the macrostructural instability characteristics of the coal and rock mass. 2) Mid-frequency band (50-150kHz): weight value 0.5-0.8, representing the transition zone of fracture propagation. 3) High frequency band (150-300kHz): weight value ≥ 0.8, representing the transient propagation characteristics of microcracks.

[0039] It should be noted that when the high-frequency band weight value is ≥2.0 for three consecutive times, its upper limit can be extended to enhance the microcrack resolution. For example, the upper limit can be extended to 350 kHz.

[0040] Further, see Figure 3 As shown, the specific confirmation process of the final decomposition layer number in step B2 includes: B21, calculating the growth rate of rupture events per unit time based on the cumulative frequency characteristics of rupture events.

[0041] B22. If the growth rate exceeds the set first growth rate threshold, the upper limit value of the preset decomposition layer range will be used as the basic decomposition layer number; if it exceeds the second growth rate threshold and is lower than or equal to the first growth rate threshold, the median of the preset decomposition layer range will be used as the basic decomposition layer number; if it is lower than or equal to the second growth rate threshold, the lower limit value of the preset decomposition layer range will be used as the basic decomposition layer number.

[0042] B23. Extract the energy-concentrated frequency band from the frequency domain energy distribution characteristics and calculate the average energy proportion of the energy-concentrated frequency band.

[0043] B24. If the energy concentration frequency band is located in the preset risk frequency band and the average energy proportion exceeds the preset threshold, the sum of the set compensation decomposition layers and the basic decomposition layers will be used as the final decomposition layers. If the energy concentration frequency band is not located in the preset risk frequency band or the average energy proportion does not exceed the preset threshold, the basic decomposition layers will be used as the final decomposition layers.

[0044] Understandably, calculating the growth rate of fracture events per unit time can reflect the intensity of coal and rock fracture activity. By comparing the growth rate with different thresholds to determine the number of basic decomposition layers, the level of wavelet packet decomposition can be flexibly adjusted based on the intensity of coal and rock fracture activity. More intense fracture activity indicates more complex internal changes within the coal and rock mass, requiring a higher number of decomposition layers for more detailed signal analysis to capture more fracture-related information.

[0045] It can be understood that by extracting the energy-concentrated frequency band from the frequency domain energy distribution characteristics and calculating its average energy ratio, we can understand the frequency concentration of the energy of the acoustic emission signal. Comparing the energy-concentrated frequency band with the preset risk frequency band, and comparing the average energy ratio with the preset threshold, can be used to determine whether the current acoustic emission signal is in a state that may indicate a dangerous coal rock mass. If the energy-concentrated frequency band is in the preset risk frequency band and the average energy ratio exceeds the preset threshold, it means that the coal rock mass may be in a high-risk state. At this time, increasing the number of compensatory decomposition layers can further improve the accuracy of signal analysis, so as to more accurately assess the state of the coal rock mass. Conversely, the basic decomposition layer number is used to ensure that sufficient information can be obtained while avoiding the waste of computing resources and increased analysis complexity caused by excessive decomposition.

[0046] It's important to note that the number of decomposition levels determines the degree of signal segmentation achieved by wavelet packet decomposition. A higher number of levels allows for more detailed signal decomposition, enabling more precise analysis of signal characteristics across frequency bands and capturing subtler variations. For complex, non-stationary signals like coal-rock acoustic emission signals, different fracture states may exhibit unique characteristics across frequency bands. The appropriate number of decomposition levels helps accurately identify these characteristics, leading to a better understanding of the physical processes within the coal-rock mass.

[0047] It should be further explained that when the growth rate exceeds the set first growth rate threshold, it is the high-frequency growth stage. The maximum number of decomposition layers is used to enhance the resolution of the high-frequency band and capture the transient characteristics of microcracks. When the growth rate exceeds the second growth rate threshold and is lower than or equal to the first growth rate threshold, it is the steady-state stage. The median of the preset decomposition layer range is used as the basic decomposition layer to balance the computational efficiency and feature extraction accuracy. When the growth rate is lower than or equal to the second growth rate threshold, it is the low-frequency attenuation stage. The minimum number of decomposition layers is used to reduce redundant calculations.

[0048] In a specific embodiment, the number of compensation decomposition layers may be set to 2. By adding 2 layers to the basic decomposition layer number, the frequency band can be specifically refined.

[0049] Exemplarily, the db4 wavelet basis function is used, and the preset decomposition layer range is 4-8 layers. When the rupture event growth rate exceeds or is equal to 10% / min, 8 layers are selected as the basic decomposition layer number. When the rupture event growth rate exceeds 4% / min and is lower than or equal to 10% / min, 6 layers are selected as the basic decomposition layer number. When the rupture event growth rate is lower than or equal to 4% / min, 4 layers are selected as the basic decomposition layer number. If the average energy proportion of the frequency domain energy is greater than 70% and is located in the 0-50kHz risk band, an additional 2 layers of compensatory decomposition are added. When the basic decomposition layer number is 8 layers, the final decomposition layer number is 10 layers. When the basic decomposition layer number is 6 layers, the final decomposition layer number is 8 layers. When the basic decomposition layer number is 4 layers, the final decomposition layer number is 6 layers.

[0050] It should also be added that the energy-concentrated frequency band refers to a frequency band whose energy proportion exceeds a preset energy proportion threshold, for example, a frequency band whose energy proportion exceeds 40%.

[0051] The embodiment of the present invention adopts a distributed high-frequency acoustic emission sensor array and combines the rupture event trigger threshold with time domain / frequency domain feature extraction technology to capture the tiny energy release signals of impact rock pressure precursors in real time, breaking through the limitation of traditional microseismic monitoring's insufficient sensitivity to tiny signals. It also dynamically adjusts the frequency band division weights through wavelet packet decomposition technology. Combined with the cumulative frequency characteristics of rupture events and the average energy ratio of frequency domain energy, it can effectively distinguish between noise and real crack extension signals, avoid the submersion of tiny energy characteristics, and thus ensure the timeliness and accuracy of early warning.

[0052] Furthermore, the decomposition of the high-frequency acoustic emission signal by the wavelet basis function in step B3 belongs to the existing technical means, and the specific decomposition process will not be described here in detail.

[0053] Furthermore, the specific allocation process of allocating frequency band weights in step B4 includes: B41, acquiring the time domain energy data of the high-frequency acoustic emission signal in real time, calculating the instantaneous gradient value of the time domain energy, and generating a time domain weight coefficient based on the instantaneous gradient value of the time domain energy.

[0054] B42. Extract the energy proportion of each frequency band from the frequency domain energy distribution characteristics, identify the energy concentrated frequency bands, and generate frequency domain weight coefficients based on the average energy proportion of the energy concentrated frequency bands.

[0055] B43. The time domain weight coefficient and the frequency domain weight coefficient of each frequency band are integrated according to a preset rule to generate a comprehensive weight coefficient of each frequency band.

[0056] It should be added that the instantaneous gradient value of the time-domain energy can be obtained by calculating the first-order derivative of the time-domain signal energy, and the gradient value reflects the rate of change of energy per unit time.

[0057] The generation of the time-domain weight coefficient based on the instantaneous gradient value of the time-domain energy is primarily performed using a time-domain allocation rule. For example, the weight increases by 15% for every 10% the gradient value exceeds a preset threshold. For example, if the preset threshold is 5 dB / ms, the weight increases by 22.5% when the gradient is 7.5 dB / ms. To ensure a consistent numerical range for subsequent analysis, the upper limit of the time-domain weight coefficient is set to 1, and the lower limit is set to 0.

[0058] Similarly, the frequency domain weight coefficient generated based on the energy proportion of the energy-concentrated frequency band is also mainly executed through the frequency domain allocation rule. For example, for every time the energy proportion exceeds 1 times the standard deviation of the historical mean, the corresponding frequency band weight increases by 10%. For example, if the historical mean is 25%, if the current capacity proportion is 35%, the corresponding frequency band weight increases by 20%.

[0059] It should also be added that the fusion according to the preset rules can specifically adopt the weighted fusion or product method. When weighted fusion is adopted, the specific distribution weight values ​​of the time domain weight coefficient and the frequency domain weight coefficient can be adjusted according to the specific situation.

[0060] It is important to note that the specific fusion rules need to be determined based on the purpose of analyzing the acoustic emission signal and experience. Through this fusion method, the information in the time domain and frequency domain are combined to generate a comprehensive weight coefficient for each frequency band. This allows for a more comprehensive consideration of the characteristics of the signal in different domains, providing a more accurate weight basis for subsequent further analysis and processing of the signal.

[0061] The embodiment of the present invention dynamically determines the number of wavelet packet decomposition layers by combining the growth rate of fracture events and the average energy ratio of frequency domain energy, and allocates the importance of frequency bands by combining time domain / frequency domain weight coefficients, thereby significantly improving the time-frequency resolution of small crack signals.

[0062] The crack risk determination module performs multi-scale correlation analysis on signal data of different frequency bands to obtain a crack risk probability value.

[0063] Specifically, the specific analysis process of the crack risk probability value includes: C1, analyzing the signal data of the high-frequency band through a sliding time window, calculating the energy mutation gradient and waveform complexity as high-frequency band analysis features.

[0064] C2. Perform wavelet coherence analysis on the signal data in the mid-frequency band, calculate the energy fluctuation correlation between adjacent sensor nodes, and use it as the mid-frequency band analysis feature.

[0065] C3. Based on the signal data in the low-frequency band, calculate its energy excess persistence index and use it as the low-frequency band analysis feature.

[0066] C4. Combining the stress distribution data of coal rock mass with the three-frequency band analysis characteristics, a three-frequency band-stress field coupling matrix is ​​constructed, and multi-parameter fusion of the coupling matrix is ​​performed through dynamic weight allocation to generate a risk probability value.

[0067] The embodiment of the present invention realizes multi-dimensional dynamic modeling of the crack propagation process by three-band analysis based on a sliding time window and constructing a coupling matrix in combination with coal rock stress distribution data. The multi-parameter fusion algorithm based on the three-band-stress field coupling matrix associates energy mutations, correlation fluctuations and the spatial distribution of stress concentration areas, thereby realizing accurate quantification of the crack risk probability value.

[0068] Furthermore, the specific calculation process of the high-frequency band analysis feature in step C1 is as follows: C11, calculating the energy value of the high-frequency acoustic emission signal data in each window, and performing differential calculation on the energy values ​​of adjacent windows to obtain an energy mutation gradient.

[0069] C12. Cover the signal waveform in each window in a grid, count the minimum number of grids required to cover the waveform, calculate the fractal dimension based on the minimum number of grids using the Higuchi algorithm, and use it as the waveform complexity.

[0070] It should be added that the energy value is obtained by accumulating the squares of the signal amplitudes in each window, using a common calculation method, and its specific calculation formula is not shown here.

[0071] It should be added that when analyzing the high-frequency acoustic emission signal data in the high-frequency band through a sliding time window, the window parameters are also set. For example, a motion detection window with a length of 10ms can be set, and the window can be advanced every 5ms. At the same time, the Hanning window function can be used for signal interception to reduce spectrum leakage.

[0072] It should also be added that the calculation formula for the energy mutation gradient is: , Indicates the The first window and the The energy mutation gradient corresponding to the window is Indicates the The time domain energy value of the window, Indicates the The time domain energy value of the window, Indicates the window time interval.

[0073] It can be understood that the Higuchi algorithm is a relatively mature algorithm currently available, and its specific calculation formula and calculation process will no longer be displayed. The fractal dimension calculated by the Higuchi algorithm can be used as a quantitative indicator of the complexity of the signal waveform, which is used to describe the complexity changes of high-frequency acoustic emission signals in different windows, and then analyze the characteristic changes of acoustic emission signals of coal rock masses under different states.

[0074] Furthermore, the specific calculation process of the mid-frequency band analysis feature in step C2 is as follows: C21, performing continuous wavelet transform on adjacent sensor signals respectively to obtain wavelet coefficients.

[0075] C22. Calculate the cross wavelet power spectrum of adjacent sensor signals based on the wavelet coefficients, and then calculate the wavelet coherence coefficient.

[0076] C23. Calculate the mean of the wavelet coherence coefficients of adjacent sensor signals in each time window, and use the calculation result as the corresponding energy fluctuation correlation degree.

[0077] It should be added that wavelet transformation is an existing common data processing method, that is, the specific process of obtaining wavelet coefficients through wavelet transformation will not be repeated.

[0078] For example, the wavelet coefficients of adjacent sensor signals are recorded as and , the cross wavelet power spectrum is recorded as , , yes The cross wavelet power spectrum reflects the common variation characteristics of the two signals on the time-frequency plane, and its value indicates the similarity and energy distribution of the two signals on the corresponding time and frequency scales.

[0079] It should also be added that in order to more accurately measure the correlation between two signals at different frequency scales, the wavelet coherence coefficient is calculated based on the cross wavelet power spectrum, which is denoted as , , It represents a smoothing operator, which usually uses weighted average in time and scale to achieve smoothing operation to reduce the influence of noise and fluctuation. The value of wavelet coherence coefficient is between 0 and 1. The closer it is to 1, the stronger the correlation between the two signals in the corresponding time and frequency scales, and the closer it is to 0, the weaker the correlation.

[0080] Understandably, the final mean calculation yields a value that comprehensively reflects the degree of correlation between the energy fluctuations of adjacent sensor signals within a given time window. Different time windows correspond to different observation periods. By analyzing the correlation of energy fluctuations within these different time windows, we can understand how the correlation between adjacent sensor signals changes over time, and thus infer the energy transfer and interaction characteristics between different locations within the coal and rock mass.

[0081] Furthermore, the specific calculation process of the energy excess duration index in step C3 is as follows: C31, dividing the signal data in the low-frequency band into continuous monitoring time windows, calculating the energy value in each monitoring window and performing normalization processing.

[0082] C32. Count the number of consecutive monitoring time windows in which the normalized energy value exceeds the preset threshold, and record it as the number of consecutive over-limit monitoring time windows.

[0083] C33. Calculate the average of the normalized energy values ​​in each monitoring time window that exceed the preset threshold value to obtain an average over-limit normalized energy value.

[0084] C34. The energy overlimit duration index is calculated by combining the number of continuous overlimit monitoring time windows and the average overlimit normalized energy value.

[0085] Understandably, in order to facilitate comparison and analysis of energy values ​​between different monitoring time windows, the calculated energy values ​​need to be normalized. The normalization method can be to divide the energy value in each monitoring time window by the sum of the energy values ​​of all monitoring time windows, or by a pre-set reference energy value.

[0086] For example, the energy excess duration index can be calculated by a weighted formula, such as ,in, is the energy excess duration index, and It is the weight coefficient corresponding to the number of continuous over-limit monitoring time windows and the average over-limit normalized energy value, which is determined according to the actual situation and the importance attached to the two factors. The value range of the weight coefficient is usually between 0 and 1, and , is the number of continuous over-limit monitoring time windows, is the total number of monitoring time windows, is the average excess normalized energy value, is the preset threshold corresponding to the energy value, is a natural constant.

[0087] in, This section reflects the contribution of the proportion of the number of consecutive over-limit monitoring time windows to the energy over-limit persistence index. The larger the proportion, the longer the energy over-limit has persisted, and the greater the positive contribution to the index. This part reflects the impact of the average excess normalized energy value relative to the preset threshold on the index. Beyond The more, The closer the value of is to 1, the greater the contribution to the index.

[0088] Understandably, this formula comprehensively reflects the persistence of low-frequency signal energy exceeding the limit. It integrates two key factors—the temporal continuity of energy exceeding the limit and the degree of energy exceeding the limit—into a single formula, avoiding the one-sidedness of considering either factor alone.

[0089] Furthermore, in step C4, the stress distribution data of the coal and rock mass is monitored by stress sensors, and the specific implementation process of step C4 is supplemented as follows: C41. The stress values ​​of different positions of the coal and rock mass in the well are obtained through the stress sensor network to form a three-dimensional stress field distribution matrix. , ,in, Indicates spatial location, represents the stress component,

[0090] C42. Normalize the original stress data to eliminate the dimension effect and obtain the normalized stress matrix .

[0091] C43, let waveform complexity be , and then the energy gradient is changed and waveform complexity Arranged into high-frequency feature vectors , , the energy fluctuation correlation is recorded as , and sorted into mid-frequency feature vectors , , the energy excess duration index Arranged into low-frequency feature vectors , .

[0092] C44, spatially match the acoustic sensor position with the stress sensor position and establish a corresponding relationship. For each spatial position , construct the local coupling vector , Respectively The three-band feature vector corresponding to the spatial position, For the The normalized stress vector corresponding to the spatial position.

[0093] C44. Combine the local coupling vectors of all spatial positions into a three-band-stress field coupling matrix , , is the matrix transpose operator, which facilitates the subsequent unified processing and analysis of the three-band analysis characteristics and stress field data of multiple monitoring points. is the number of monitoring points, the matrix The dimension is , is the number of stress components, usually .

[0094] C45, coupling matrix Each row of is weighted fusion to obtain the comprehensive eigenvalue , , ~ are the weights of high frequency band, medium frequency band, low frequency band and stress field characteristics, and satisfy .

[0095] C46, select Sigmoid function as the risk mapping function, and convert the comprehensive eigenvalue The risk probability value corresponding to each spatial position is obtained by inputting it into the risk mapping function. Combined with the impact weight corresponding to each spatial position, the comprehensive risk probability value is obtained by weighted summation as the final risk probability value.

[0096] It should be noted that the weights of high-frequency, mid-frequency, low-frequency, and stress field features can be objectively determined by calculating the information entropy of each feature. The lower the information entropy, the greater the amount of information provided by the feature, and the higher the weight. Initial weights can also be set based on domain expert knowledge and adjusted based on real-time data. For example, in the early stages of coal and rock mass failure, the low-frequency energy excess persistence index is more important, while at the moment of failure, the waveform complexity changes more significantly in the high-frequency band. These are not specifically defined here; they can be selected based on actual needs.

[0097] The monitoring adjustment feedback terminal dynamically adjusts the monitoring strategy according to the crack risk probability value, and links the underground disaster prevention and control system to execute the risk feedback control instruction.

[0098] Specifically, the specific adjustment process of the dynamic adjustment monitoring strategy includes: D1, matching the crack risk probability value with the pre-set crack risk probability value intervals corresponding to each crack risk level to obtain a matching crack risk level.

[0099] D2. Based on the matched crack risk level, extract the set sampling frequency, sensor node activation ratio and data transmission frequency under the corresponding level from the crack risk sampling setting table, and combine them to generate a dynamically adjusted monitoring strategy.

[0100] In a specific embodiment, for example, at a certain moment, the calculated crack risk probability value is 0.6. This crack risk probability value is matched with the pre-set crack risk probability value intervals corresponding to each crack risk level. Assuming that the pre-set probability value intervals for the low risk level are [0, 0.3), the medium risk level is [0.3, 0.7), and the high risk level is [0.7, 1], the matching shows that the corresponding crack risk level at this moment is medium risk.

[0101] Based on the medium risk level obtained from the matching, the corresponding sampling frequency, sensor node activation ratio, and data transmission frequency are extracted from the crack risk sampling setting table. For example, the crack risk sampling setting table specifies a sampling frequency of every 5 minutes, a sensor node activation ratio of 60%, and a data transmission frequency of every half hour for a medium risk level.

[0102] These extracted parameters are combined to generate a dynamically adjusted monitoring strategy, which monitors the underground coal and rock mass by sampling signals every five minutes, activating 60% of the sensor nodes, and transmitting data every half hour.

[0103] Based on the generated, dynamically adjusted monitoring strategy and the current crack risk level, combined with pre-set risk response rules, risk feedback control instructions are generated. For example, for a medium risk level, the pre-set rule is to activate disaster warnings in selected areas and moderately increase ventilation. The generated instructions include sending activation signals to specific areas underground, such as audible and visual alarms within a certain range of the suspected crack location, and simultaneously sending instructions to the ventilation control system to increase the ventilation equipment power in that area by a certain percentage, such as 20%.

[0104] Upon receiving a command, the underground disaster prevention and control system first interprets it. For commands to activate the audible and visual alarms, an electrical signal is sent to the corresponding audible and visual alarms in the corresponding area, causing them to flash and sound alarms, alerting nearby personnel to potential crack risks. For commands to increase ventilation power, the ventilation control system's actuator adjusts the input voltage or current to the ventilation motor, increasing power by 20%, increasing air circulation and reducing the risk of harmful gas accumulation.

[0105] The embodiments of the present invention ensure the integrity and real-time performance of crack extension signals by dynamically adjusting the sampling frequency, sensor activation ratio, and data transmission frequency, thereby overcoming the shortcomings of traditional single data sources, facilitating the acquisition of more detailed features of fine cracks, and ensuring the accuracy of fine crack analysis.

[0106] The above-mentioned preset parameters are set by technicians in this field according to actual conditions.

[0107] The above embodiments may be implemented in whole or in part through software, hardware, firmware or any other combination. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0108] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0109] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0110] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0111] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A coal mining crack dynamic measurement and monitoring system, characterized in that: The system includes: The fracture signal acquisition module collects high-frequency acoustic emission signal data of underground coal and rock masses in real time. The data includes the cumulative frequency characteristics of fracture events, time domain energy change characteristics, and frequency domain energy distribution characteristics. a crack signal processing module, performing wavelet packet decomposition on the signal data to obtain frequency bands and generate signal data of different frequency bands; The crack risk determination module performs multi-scale correlation analysis on signal data of different frequency bands to obtain the crack risk probability value; The monitoring and adjustment feedback terminal dynamically adjusts the monitoring strategy according to the crack risk probability value, and links the underground disaster prevention and control system to execute risk feedback control instructions; The specific processing process of performing wavelet packet decomposition includes: The final number of decomposition layers is determined based on the cumulative frequency characteristics of rupture events and the frequency domain energy distribution characteristics; The specific confirmation process of the final number of decomposition layers includes: Calculate the growth rate of rupture events per unit time based on the cumulative frequency characteristics of rupture events; If the growth rate exceeds the set first growth rate threshold, the upper limit of the preset decomposition layer interval is used as the basic decomposition layer number; if it exceeds the second growth rate threshold and is lower than or equal to the first growth rate threshold, the median of the preset decomposition layer interval is used as the basic decomposition layer number; if it is lower than or equal to the second growth rate threshold, the lower limit of the preset decomposition layer interval is used as the basic decomposition layer number; Extract the energy concentrated frequency band from the frequency domain energy distribution characteristics and calculate the average energy proportion of the energy concentrated frequency band; If the energy concentration frequency band is located in the preset risk frequency band and the average energy proportion exceeds the preset threshold, the sum of the set compensation decomposition layers and the basic decomposition layers will be used as the final decomposition layers. If the energy concentration frequency band is not located in the preset risk frequency band or the average energy proportion does not exceed the preset threshold, the basic decomposition layers will be used as the final decomposition layers.

2. A coal mining crack dynamic measurement and monitoring system according to claim 1, characterized in that: The specific collection process of the high-frequency acoustic emission signal data includes: The distributed acoustic emission sensor array covers the coal and rock mass area to be tested, and the time domain waveform of the acoustic emission signal is obtained in real time; extracting a time series of rupture events from the time domain waveform based on a preset rupture event trigger threshold, and generating a cumulative frequency feature of the rupture events; Performing energy integration calculation on the time domain waveform of each rupture event to generate the time domain energy variation feature; Performing a fast Fourier transform on the time domain waveform, extracting the energy proportion within a preset frequency band, and generating the frequency domain energy distribution feature.

3. The coal mining crack dynamic measurement and monitoring system according to claim 1, characterized in that: The specific processing process of performing wavelet packet decomposition includes: The original high-frequency acoustic emission signal is filtered, denoised and normalized respectively; Decomposing the processed high-frequency acoustic emission signal by frequency band according to the determined final decomposition layer number based on the set wavelet basis function to generate time-frequency components of each sub-band; Frequency band weights are allocated based on time domain energy variation characteristics and frequency domain energy distribution characteristics, and based on the weight allocation result, the sub-frequency band is divided into a low frequency band, a mid frequency band, and a high frequency band.

4. A coal mining crack dynamic measurement and monitoring system according to claim 3, characterized in that: The specific allocation process of allocating frequency band weights includes: Acquiring time-domain energy data of high-frequency acoustic emission signals in real time, calculating a time-domain energy instantaneous gradient value, and generating a time-domain weight coefficient according to the time-domain energy instantaneous gradient value; Extract the energy proportion of each frequency band from the frequency domain energy distribution characteristics, identify the energy concentrated frequency band, and generate the frequency domain weight coefficient based on the average energy proportion of the energy concentrated frequency band; The time domain weight coefficient and frequency domain weight coefficient of each frequency band are fused according to preset rules to generate a comprehensive weight coefficient of each frequency band.

5. The coal mining crack dynamic measurement and monitoring system according to claim 1, characterized in that: The specific analysis process of the crack risk probability value includes: Analyze the signal data in the high-frequency band through a sliding time window, calculate the energy mutation gradient and waveform complexity as the high-frequency band analysis features; Perform wavelet coherence analysis on the signal data in the mid-frequency band, calculate the energy fluctuation correlation between adjacent sensor nodes, and use it as the mid-frequency band analysis feature; Based on the signal data in the low-frequency band, the energy excess persistence index is calculated and used as the low-frequency band analysis feature; Combining the stress distribution data of coal rock mass with the three-frequency band analysis characteristics, a three-frequency band-stress field coupling matrix is ​​constructed. The multi-parameter fusion of the coupling matrix is ​​performed through dynamic weight allocation to generate the risk probability value.

6. A coal mining crack dynamic measurement and monitoring system according to claim 5, characterized in that: The specific calculation process of the high-frequency band analysis feature is as follows: The energy value of the high-frequency acoustic emission signal data in each window is calculated, and the energy values ​​of adjacent windows are differentially calculated to obtain the energy mutation gradient; The signal waveform in each window is covered in a grid, and the minimum number of grids required to cover the waveform is counted. The fractal dimension is calculated based on the minimum number of grids using the Higuchi algorithm and is used as the waveform complexity.

7. A coal mining crack dynamic measurement and monitoring system according to claim 5, characterized in that: The specific calculation process of the mid-frequency band analysis feature is as follows: Perform continuous wavelet transform on adjacent sensor signals to obtain wavelet coefficients; Calculate the cross wavelet power spectrum of adjacent sensor signals based on the wavelet coefficients, and then calculate the wavelet coherence coefficient; The wavelet coherence coefficients of adjacent sensor signals in each time window are averaged and the calculation results are used as the corresponding energy fluctuation correlation.

8. The coal mining crack dynamic measurement and monitoring system according to claim 5, characterized in that: The specific calculation process of the energy excess duration index is as follows: Divide the low-frequency signal data into continuous monitoring time windows, calculate the energy value in each monitoring window and perform normalization processing; The number of consecutive monitoring time windows in which the normalized energy value exceeds the preset threshold is counted and recorded as the number of consecutive over-limit monitoring time windows; The normalized energy values ​​in each monitoring time window that exceed the preset threshold are averaged to obtain the average over-limit normalized energy value; The energy overlimit duration index is calculated by combining the number of continuous overlimit monitoring time windows and the average overlimit normalized energy value.

9. The coal mining crack dynamic measurement and monitoring system according to claim 1, characterized in that: The specific adjustment process of the dynamic adjustment monitoring strategy includes: Matching the crack risk probability value with the pre-set crack risk probability value interval corresponding to each crack risk level to obtain a matching crack risk level; Based on the matched crack risk level, the set sampling frequency, sensor node activation ratio and data transmission frequency under the corresponding level are extracted from the crack risk sampling setting table, and combined to generate a dynamically adjusted monitoring strategy.

Images