A rock deformation stage identification method based on microseismic background noise
By performing correlation processing and analysis on microseismic background noise between stations, and utilizing cross-correlation calculations and sensitive kernel functions, the problem of identifying rock deformation stages was solved, achieving accurate identification of rock deformation stages and improving the accuracy of rockburst prediction.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NORTH CHINA UNIVERSITY OF TECHNOLOGY
- Filing Date
- 2022-12-30
- Publication Date
- 2026-05-19
AI Technical Summary
Existing technologies lack effective methods for identifying precursory features of rock deformation and failure, especially when accurate loading curves cannot be obtained, making it difficult to accurately identify rock deformation stages and resulting in difficulties in predicting and forecasting rockbursts.
By performing correlation processing and analysis between stations on microseismic background noise, and using cross-correlation calculation, dispersion curve extraction, dispersion data correlation statistics and sensitive kernel function, the rock deformation stage can be identified.
It enables precise identification of rock deformation stages, improves the accuracy and early warning capability of rockburst prediction, reduces the impact of external environmental changes, and provides a means of observing the three-dimensional internal deformation evolution.
Smart Images
Figure CN115903040B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of microseismic monitoring and relates to a method for identifying rock deformation stages based on microseismic background noise. Background Technology
[0002] The development of rockburst monitoring and prediction is a crucial aspect of effective rockburst prevention and control research. Since the fundamental cause of rockbursts is rock mass deformation and failure, accurately identifying precursory characteristics of deformation and failure is fundamental to rockburst prediction and forecasting. Based on the spatiotemporal evolution of the rock deformation field and the characteristics of the loading curve, rock deformation and failure can be divided into four stages: the random evolution stage of micro-fractures, the deformation localization stage, the sub-instability stage, and the instability stage. When accurate loading curves are unavailable, utilizing background noise to identify the rock deformation evolution stages provides a better method for predicting and warning of rockbursts. Summary of the Invention
[0003] To address the shortcomings of existing technologies, this invention provides a method for identifying rock deformation stages based on microseismic background noise by performing correlation processing and analysis on the correlation coefficient and sensitive kernel function between stations for background noise.
[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows:
[0005] A method for identifying rock deformation stages based on microseismic background noise includes the following steps:
[0006] (1) Based on the micro-vibration noise data collected by the station sensors, extract the noise waveform data and preprocess the noise waveform data;
[0007] (2) Perform cross-correlation calculations on the processed noise from a certain station and other different stations within the same time period;
[0008] Discretization formula for cross-correlation:
[0009]
[0010] Where x(r) and y(r) are signals from two different stations, N is the number of sampling points, n is the delay sequence, and R is the signal from the station. xy (nΔτ) represents the cross-correlation data between the signals from the two stations;
[0011] (3) Calculate the Green's function from the cross-correlation function, then use the Fk variation method to extract the dispersion curve, and record tx. tx represents the time domain, with the horizontal axis representing time and the vertical axis representing the instantaneous value of the measured parameter. The collected discrete records are as follows:
[0012] f(m,n)=f(x0+mΔx,t+nΔt)(2)
[0013] Where: m = 0, ..., M-1; n = 0, ..., N-1; the number of channels is M, and the number of sampling points is N;
[0014] The fk domain is recorded as F(f,k), and the collected discrete records are as follows:
[0015]
[0016] W = e -2πj (4)
[0017] Where: f = 0, 1, 2, ..., M-1, k = 0, 1, 2, ..., N-1, f represents frequency, k represents wave number, and F(f,k) is the frequency-wave number spectrum of u(t,x);
[0018] Using Relations Transform the fk domain into the fv domain;
[0019] (4) Based on the different degrees of dispersion of dispersion curve data at various stations and markers and the concept of correlation coefficient, a statistical index of dispersion data correlation was introduced:
[0020]
[0021] Where i = 1, 2, 3, ..., n, n is the number of stations, Xn is the dispersion curve data of the stations, and X(t) represents the data of each station after relevant processing.
[0022] The function corr(X,Y) uses the Pearson product moment coefficient, where X and Y represent two sets of related data to be analyzed, which are used to measure the linear correlation between the two sets of data. The values range from [-1, +1], and the two sets of data used are the data matrix of the dispersion curve.
[0023]
[0024] Where cov represents the covariance function, σ X σ Y The standard deviations of X and Y are represented by the correlation coefficients of all stations with time. The time points of the previous deformation stage and the next deformation stage are found by the correlation coefficients of all stations. The group of stations with the highest correlation coefficients is found by the data of different plane stations, so as to carry out the next step of research.
[0025] (5) To further investigate the correspondence between the dispersion characteristics of the stations with the highest correlation coefficients and the deformation evolution stages, a sensitive kernel function is introduced:
[0026]
[0027] Where i is the sampling frequency point; k is the order of the dispersion curve; It is the phase velocity of the dispersion curve at the i-th frequency point and the k-th order. To observe the phase velocity of the dispersion curve; m is the order of the dispersion curve; n k It is the number of sampling points for inverting the k-th order dispersion curve; a k It is the dispersion curve weight; the dispersion curve weight is set to 1, and the unit of phase velocity is m / s; Vs represents the wave velocity of the noise.
[0028] By using the dispersion curves of the four stages, sensitive kernel function curves are plotted. The curve values reflect the magnitude and trend of the deformation evolution of the medium fracture, thereby further determining the stage of deformation evolution.
[0029] The preprocessing includes: detrending, mean removal, bandpass filtering, time-domain normalization, and spectral whitening.
[0030] Correlation coefficient and sensitive kernel function index analysis were performed on the dispersion curves at each time point. The station group with the highest correlation coefficient was found through the image. The stage characteristics were identified by the overall trend of the correlation curve and the speckle deformation evolution cloud map, thereby identifying the rock deformation stage. The degree and trend of fracture development were further studied by the changes of sensitive kernel function curves of different layers at different times, so as to achieve the goal of detailed division of different deformation stages of the same medium.
[0031] Beneficial effects of this invention:
[0032] This invention aims to identify the various stages of rock deformation evolution. Using microseismic background noise data, it derives the Green's function and plots dispersion curves through noise preprocessing and cross-correlation calculations. By analyzing the correlation coefficient and sensitive kernel function of the dispersion curves, it identifies the dispersion characteristics of rock deformation evolution at each stage and correlates them with speckle surface deformation contour maps, thus obtaining a method for identifying rock deformation stages based on microseismic background noise. This method utilizes station noise data to reflect the rock deformation evolution, offering fewer limitations and less susceptibility to external environmental changes compared to observing deformation through speckle surface deformation contour maps. While speckle surface deformation contour maps are limited to two-dimensional planar observation of rock deformation evolution, observations based on station data with microseismic background noise can be three-dimensional and internal. Attached Figure Description
[0033] Figure 1 Schematic diagram of speckle size and measuring point arrangement.
[0034] Figure 2 Noise waveform image.
[0035] Figure 3 Graph of cross-correlation function.
[0036] Figure 4 Dispersion curve image during the deformation localization stage.
[0037] Figure 5 Dispersion curve image of the sub-instability stage.
[0038] Figure 6 Correlation coefficient graph.
[0039] Figure 7 Image of sensitive kernel function during the sub-instability stage. Detailed Implementation
[0040] The specific embodiments of the present invention will be further described in detail below with reference to examples.
[0041] Example 1
[0042] A method for identifying rock deformation stages based on microseismic background noise, the specific steps of which are as follows:
[0043] (1) Extract noise waveform data based on the micro-vibration noise data collected by the station sensors. Preprocess the noise waveform data: detrend, mean removal, bandpass filtering, time-domain normalization, and spectral whitening.
[0044] (2) Perform cross-correlation calculations on the processed noise from a certain station and other different stations within the same time period.
[0045] Discretization formula for cross-correlation:
[0046]
[0047] x(r) and y(r) are signals from two different stations, N is the number of sampling points, n is the delay sequence, and R is the signal from the station. xy (nΔτ) represents the cross-correlation data between the two stations.
[0048] (3) The dispersion curve is a curve that represents the relationship between the period and wave speed of the dispersion wave. Extracting the dispersion curve of noise in red sandstone will prepare for subsequent research on its deformation stage.
[0049] The Green's function is obtained by calculating the cross-correlation function (Formula (8)), and then the dispersion curve is extracted using the Fk variation method. tx is recorded. tx represents the time domain, with the horizontal axis representing time and the vertical axis representing the instantaneous value of the measured parameter. The collected discrete records are as follows:
[0050] f(m,n)=f(x0+mΔx,t+nΔt)(9)
[0051] Where: m = 0, ..., M-1; n = 0, ..., N-1; the number of channels is M, and the number of sampling points is N (lowercase m and n represent the variables of the number of channels M and the number of sampling points N);
[0052] The fk domain (fk is a formula representing a frequency beam) is recorded as F(f,k), and the collected discrete records are as follows:
[0053]
[0054] W = e -2πj (11)
[0055] Where: f = 0, 1, 2, ..., M-1, k = 0, 1, 2, ..., N-1, f represents frequency, k represents wavenumber, F(f,k) is the frequency-wavenumber spectrum of u(t,x); u(t,x) represents the time-domain image of the data in the formula at this stage.
[0056] Using Relations Transform the fk domain into the fv domain.
[0057] (4) Based on the different degrees of dispersion of dispersion curve data at various stations and markers and the concept of correlation coefficient, a statistical index of dispersion data correlation was introduced:
[0058]
[0059] (i = 1, 2, 3, ..., n);
[0060] Where i = 1, 2, 3, ..., n; n is the number of stations; Xn is the dispersion curve data of the stations; X(t) represents the data of each station after relevant processing.
[0061] The function corr(X,Y) uses the Pearson product moment coefficient, where X and Y represent two sets of related data to be analyzed, used to measure the relationship (linear correlation) between the two sets of data, with values ranging from [-1, +1]. The two sets of data used are the data matrix of the dispersion curve.
[0062]
[0063] Where cov represents the covariance function, σ X σ Y The standard deviations of X and Y are represented by the correlation coefficients of all stations with time. The time points of the previous deformation stage and the next deformation stage are found by using data from different planar stations. The group of stations with the highest correlation coefficients is then used to conduct further research.
[0064] (5) To further investigate the correspondence between the dispersion characteristics of the stations with the highest correlation coefficients and the deformation evolution stages, a sensitive kernel function is introduced:
[0065]
[0066] Here, i and k are the sampling frequency points and the order of the dispersion curve, respectively. It is the phase velocity of the dispersion curve at the i-th frequency point and the k-th order. To observe the phase velocity of the dispersion curve, m is the order of the dispersion curve, and n is the phase velocity of the dispersion curve. k It is the number of sampling points for inverting the k-th order dispersion curve, a k is the dispersion curve weight, which is set to 1 here. The phase velocity is in m / s. Vs represents the noise wave velocity.
[0067] By using the dispersion curves of the four stages, sensitive kernel function curves are plotted. The curve values reflect the magnitude and trend of the deformation evolution of the medium fracture, thereby further determining the stage of deformation evolution.
[0068] Correlation coefficient and sensitive kernel function index analysis were performed on the dispersion curves at each time point. The station group with the highest correlation coefficient was found through the image. The stage characteristics were identified by the overall trend of the correlation curve and the speckle deformation evolution cloud map, thereby identifying the rock deformation stage. The degree and trend of fracture development were further studied by the changes of sensitive kernel function curves of different layers at different times, so as to achieve the goal of detailed division of different deformation stages of the same medium.
[0069] Application examples
[0070] (1) The rock specimen used was red sandstone, with dimensions of 100mm × 50mm × 50mm. A smoother side of the rock was uniformly sprayed with black paint, followed by uniformly sprayed white paint dots onto the black surface. The resulting speckle pattern should have high contrast, with random distribution of white paint dots and uniform dot size. The purpose of spraying the speckle pattern was to verify the microseismic signal analysis results.
[0071] (2) A preliminary detection plan was formulated, and microseismic probes were deployed. Microseismic sensors were installed on the rock surface, and a total of 5 stations were set up on two sides. The layout of each measuring point was as follows: 2 stations were located on the left side of the rock speckle surface, and the other 3 stations were located on the opposite side of the rock speckle surface (e.g., Figure 1 As shown in the figure, 1-5 are the station numbers.
[0072] (3) The specimen was subjected to uniaxial compression for destruction. A CCD camera was used to acquire speckle surface images, and micro-vibration noise data was simultaneously acquired and extracted. The noise waveform data was preprocessed (detrending and mean removal, bandpass filtering, time-domain normalization, and spectral whitening) to obtain data suitable for cross-correlation calculation. The noise waveform and cross-correlation function images are shown below. Figure 2 , 3 As shown.
[0073] (4) Calculate the Green's function by calculating the cross-correlation function, and then calculate the dispersion curve by fk transform.
[0074] The deformation evolution was divided into stages based on the speckle deformation evolution cloud map and the press loading curve: In the first and second stages before the peak, the deformation field evolved from random microcrack evolution to localized deformation. Rock deformation evolution analysis revealed that localization initially occurred in the pre-stress peak stage. The characteristics of localized deformation are mainly reflected in two aspects: firstly, the spatial characteristics of localized deformation, i.e., deformation is concentrated on one or more bands, exhibiting a spatial characteristic of deformation concentration; secondly, the numerical characteristics of localized deformation, i.e., the magnitude within the localized deformation band is much greater than the magnitude outside the band. In the third and fourth stages after the peak, the loading curve was used for further division, which visually distinguishes the post-peak sub-instability stage and the instability stage. The dispersion curve images for the localized deformation stage and the sub-instability stage are shown below. Figure 4 , 5 As shown.
[0075] (5) Correlation coefficient and sensitive kernel function index analysis were performed on the dispersion curves at each time point. The station group with the highest correlation coefficient was identified through the images. The stage characteristics were found by corresponding to the overall trend of the correlation curve and the speckle deformation evolution cloud map, thereby identifying the rock deformation stage. The degree and trend of fracture development were further studied by the changes of sensitive kernel function curves of different layers at different times, so as to achieve the goal of detailed classification of different deformation stages of the same medium. Among them, the correlation coefficient and sensitive kernel function images of the sub-instability stage are shown in the figure. Figure 6 , 7 As shown.
Claims
1. A method for identifying rock deformation stages based on microseismic background noise, characterized in that, Includes the following steps: (1) Based on the micro-vibration noise data collected by the station sensors, extract the noise waveform data and preprocess the noise waveform data; (2) Perform cross-correlation calculations on the processed noise from a certain station and other different stations within the same time period; Discretization formula for cross-correlation: Where x(r) and y(r) are signals from two different stations, N is the number of sampling points, n is the delay sequence, and R is the signal from the station. xy (nΔτ) represents the cross-correlation data between the signals from the two stations; (3) Calculate the Green's function from the cross-correlation function, then use the Fk variation method to extract the dispersion curve, and record tx. tx represents the time domain, the horizontal axis represents time, and the vertical axis represents the instantaneous value of the measured parameter. The collected discrete records are as follows: f(m,n)=f(x0+mΔx,t+nΔt) (2) Where: m = 0, ..., M-1; n = 0, ..., N-1; the number of channels is M, and the number of sampling points is N; The fk domain is recorded as F(f,k), and the collected discrete records are as follows: W=e -2πj (4) Where: f = 0, 1, 2, ..., M-1, k = 0, 1, 2, ..., N-1, f represents frequency, k represents wave number, and F(f,k) is the frequency-wave number spectrum of u(t,x); Using Relations Transform the fk domain into the fv domain; (4) Based on the different degrees of dispersion of dispersion curve data at various stations and markers and the concept of correlation coefficient, a statistical index of dispersion data correlation was introduced: Where i = 1, 2, 3, ..., n, n is the number of stations, Xn is the dispersion curve data of the stations; X(t) represents the data of each station after correlation processing; The function corr(X,Y) uses the Pearson product moment coefficient, where X and Y represent two sets of related data to be analyzed, which are used to measure the linear correlation between the two sets of data. The values range from [-1, +1], and the two sets of data used are the data matrix of the dispersion curve. Where cov represents the covariance function, σ X σ Y The standard deviations of X and Y are represented by the correlation coefficients of all stations with time. The time points of the previous deformation stage and the next deformation stage are found by the correlation coefficients of all stations. The group of stations with the highest correlation coefficients is found by the data of different plane stations, so as to carry out the next step of research. (5) To further investigate the correspondence between the dispersion characteristics of the stations with the highest correlation coefficients and the deformation evolution stages, a sensitive kernel function is introduced: Where i is the sampling frequency point; k is the order of the dispersion curve; It is the phase velocity of the dispersion curve at the i-th frequency point and the k-th order. To observe the phase velocity of the dispersion curve; m is the order of the dispersion curve; n k It is the number of sampling points for inverting the k-th order dispersion curve; a k It is the dispersion curve weight; the dispersion curve weight is set to 1, and the unit of phase velocity is m / s; Vs represents the wave velocity of the noise. By using the dispersion curves of the four stages, sensitive kernel function curves are plotted. The curve values reflect the magnitude and trend of the deformation evolution of the medium fracture, thereby further determining the stage of deformation evolution.
2. The method as described in claim 1, characterized in that, The preprocessing includes: detrending, mean removal, bandpass filtering, time-domain normalization, and spectral whitening.
3. The method as described in claim 1, characterized in that, Correlation coefficient and sensitive kernel function index analysis were performed on the dispersion curves at each time point. The station group with the highest correlation coefficient was found through the image. The stage characteristics were identified by the overall trend of the correlation curve and the speckle deformation evolution cloud map, thereby identifying the rock deformation stage. The degree and trend of fracture development were further studied by the changes of sensitive kernel function curves of different layers at different times, so as to achieve the goal of detailed division of different deformation stages of the same medium.