A scale-free method for fine-grained monitoring and early warning of dynamic disasters in mines

Abstract

This invention discloses a scale-effect-free method for precise monitoring and dynamic early warning of dynamic disasters in mines, comprising the following steps: setting up a multi-point microseismic monitoring network at the coal mine working face to continuously collect and preprocess microseismic signals; employing an innovative analytical positioning algorithm to accurately analyze the precise location and time information of microseismic sources; combining signal characteristics and positioning data, innovatively applying four-dimensional joint kernel density estimation and confidence interval theory to construct a rockburst risk assessment model and establish hazard level classification standards; and calculating the risk index in real time and dynamically issuing early warnings. This invention significantly improves the accuracy, timeliness, and reliability of dynamic disaster monitoring and early warning in mining areas, providing key technical support for safe coal mine production.

Description

Technical Field

[0001] This invention relates to the field of mine safety technology, specifically to a method for precise monitoring and early warning of mine dynamic disasters without scale effects. Background Technology

[0002] As global mineral resource extraction progresses to deeper levels, increasing mining depth and complex geostress environments, the frequency and intensity of mine dynamic disasters are on the rise. Particularly in deep mines, the "three highs" characteristics—high geostress, high temperature, and high permeability—make the prediction and prevention of dynamic disasters even more challenging. Furthermore, the expansion of mining scale and the increase in mining intensity further exacerbate the severity of this problem.

[0003] In recent years, the rapid development of microseismic monitoring technology and methods has opened up new avenues for the monitoring and early warning of dynamic disasters in mines. This technological innovation enables us to capture microscopic fracture information within rock masses in real time, providing massive data support for mine dynamic disaster early warning systems. The application of microseismic monitoring systems has not only significantly improved the accuracy and coverage of monitoring but also provided valuable real-time data for a deeper understanding of rock mass mechanical behavior and disaster gestation processes. However, despite the significant progress made by microseismic monitoring technology in the field of mine dynamic disaster monitoring and early warning, many challenges remain in practical applications:

[0004] 1. The accuracy of microseismic source location needs improvement: ① Traditional location algorithms have low computational efficiency and often suffer from non-convergence or local convergence during iteration; ② Traditional location methods cannot comprehensively utilize all arrival data at once, and the location accuracy cannot achieve maximum likelihood estimation; ③ Traditional location methods require matrix solving, which may lead to matrix singularity and ill-conditioned situations; ④ Arrival measurement data are severely affected by abnormal signals such as environmental noise; ⑤ The sensor layout of traditional mine area microseismic monitoring networks is not reasonable enough, which may cause sensors to fail to receive signals, or unreasonable sensor geometry may cause serious systematic errors.

[0005] 2. Insufficient adaptability of the model: Traditional early warning models often suffer from scaling effects and lack practicality. Statistical models built based on data from specific mining areas are often difficult to apply directly to other mining areas with different geological conditions.

[0006] 3. The risk assessment model is not comprehensive enough: it does not fully consider the multidimensional characteristics of microseismic events, relies on only a single parameter or simple threshold, fails to make full use of multi-source information, and is difficult to adaptively adjust the early warning threshold based on real-time monitoring data.

[0007] 4. Insufficient real-time performance: Real-time processing and analysis of large-scale data remains a challenge, affecting the timeliness of early warnings.

[0008] 5. Scale effect of early warning models: Traditional models are usually built based on local observation data, making it difficult to accurately reflect the overall behavior of large-scale geological bodies. When extended from small-scale to large-scale, the predictive ability of the model often decreases significantly. Summary of the Invention

[0009] To address the shortcomings of the existing technologies, this invention provides a scale-free method for fine monitoring and dynamic early warning of mine dynamic disasters, thereby solving the technical problems of traditional rockburst monitoring methods in terms of data utilization, model evaluation, and dynamic early warning.

[0010] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0011] A scale-free method for fine-grained monitoring and dynamic early warning of mine dynamic disasters includes the following steps:

[0012] S1: Multiple high-sensitivity microseismic monitoring sensors are deployed in the monitoring area of ​​the mining area to form a comprehensive three-dimensional monitoring network;

[0013] S2: Real-time acquisition of microseismic signal data through a high-speed data acquisition system to ensure data continuity and integrity;

[0014] S3: Adaptive noise reduction processing is performed on the acquired microseismic signals;

[0015] S4: High-precision three-dimensional positioning of microseismic sources is achieved using an improved analytical positioning model;

[0016] S5: Segment the microseismic energy data according to multi-scale time windows (such as hour, day, week, month) to facilitate multi-level statistical analysis and trend identification;

[0017] S6: Based on the time-frequency characteristics, spatial distribution, and energy release patterns of microseismic signals, and combined with the principles of kernel probability density estimation and probability confidence intervals, a dynamic assessment model for rockburst risk is established.

[0018] S7: Calculates the risk index of rockburst in real time and dynamically classifies disaster early warnings;

[0019] S8: Establish a feedback mechanism for early warning results, continuously optimize and update risk assessment models and early warning systems through historical data analysis, post-disaster investigation, numerical simulation technology and expert knowledge base, and build a comprehensive database of mine dynamic disasters.

[0020] As a preferred embodiment of the present invention, the layout of the omnidirectional three-dimensional monitoring network in step S1 is as follows:

[0021] (1) A regular grid layout is adopted within the mining area to ensure that the sensors cover the entire monitoring area;

[0022] (2) Based on the geological structure of the mining area, sensors are deployed at different depth levels to form a three-dimensional monitoring network;

[0023] (3) Increase sensor density in key areas (such as mining faces, fault zones, and rock strata boundaries);

[0024] (4) The sensors are distributed in a dispersed manner, and the distance between each pair of sensors is maximized to facilitate the subsequent micro-seismic source location.

[0025] As a preferred embodiment of the present invention, the adaptive noise reduction process for the acquired microseismic signal in step S3 is as follows:

[0026] (1) Signal preprocessing:

[0027] The original microseismic signals are sampled and digitized to ensure signal integrity and continuity;

[0028] (2) Empirical Mode Decomposition:

[0029] a) The preprocessed microseismic signal x(t) is decomposed to obtain a series of intrinsic mode functions and a residual term:

[0030]

[0031] Among them, IMF i (t) is the i-th intrinsic modulus function, r n (t) represents the residual term;

[0032] b) Steps for signal decomposition:

[0033] Identify all local extrema of signal x(t);

[0034] The upper and lower envelope lines are generated using cubic spline interpolation.

[0035] Calculate the average value m1(t) of the upper and lower envelopes;

[0036] Subtracting m1(t) from the original signal yields h1(t) = x(t) - m1(t);

[0037] Repeat the above steps until h1(t) satisfies the definition conditions of IMF;

[0038] Define h1(t) as the first IMF, i.e., c1(t) = h1(t);

[0039] Subtract c1(t) from the original signal and repeat the above process to obtain the subsequent IMFs;

[0040] (3) Noise IMF identification:

[0041] a) Calculate the energy density and center frequency of each IMF;

[0042] b) Analyze the spectral characteristics of IMFs to identify high-frequency noise IMFs;

[0043] c) Use correlation analysis or signal-to-noise ratio thresholding to determine the noise IMFs that need to be removed;

[0044] (4) Adaptive threshold denoising:

[0045] The specific steps for identifying the IMF components that represent noise can be taken from the following aspects:

[0046] a) Calculate the energy of each IMF component: Where N is the signal length;

[0047] b) Calculate the average energy:

[0048]

[0049] Where N is the total number of IMF components;

[0050] c) Calculate the standard deviation σ of all IMF energies.

[0051]

[0052] d) Set an adaptive threshold: T = μ - ασ, where α = 2;

[0053] e) IMFs with indices below the threshold T are identified as noise components;

[0054] (5) Signal reconstruction:

[0055] a) Reconstruct the denoised IMF with the retained low-frequency IMF and residual terms to obtain the denoised microseismic signal:

[0056]

[0057] Among them, IMF′ i (t) represents the denoised IMF, and k represents the number of noisy IMFs;

[0058] Through the above steps, empirical mode decomposition technology can be effectively used to reduce noise in mine microseismic signals, improve signal quality, and provide a more reliable data foundation for subsequent microseismic event analysis and early warning.

[0059] As a preferred embodiment of the present invention, the calculation formula for the improved analytical positioning model in step S4 is as follows:

[0060]

[0061] Where, α1=l1l6l 11 -l1l7l 10 -l2l5l 11 +l2l7l9+l3l5l 10 -l3l6l9

[0062]

[0063] Where, χ (j) The j-th (j = 1, 2, 3, 4) index takes parameters a, b, c, d respectively; a i =2(x i -x1), b i =2(y i -y1), c i =2(z) i -z1), d i =-2v 2 (t i -t1), e i =v 2 (t i 2 -t1 2 )-[(x i 2 -x1 2 )+(y i 2 -y1 2 )+(z i 2 -z1 2 The sensor coordinates are (x) i ,y i ,z i ), i = 1, 2, ..., n, where n is the number of sensors, t i The arrival time data of sensor i is represented, t1 represents the arrival time of the first triggered sensor, v represents the propagation velocity of the medium, (x,y,z) represents the coordinates of the microseismic source to be solved, and t0 represents the time when the microseismic event occurred.

[0064] As a preferred embodiment of the present invention, the dynamic assessment model for rockburst risk in step S6 includes:

[0065] S1. Obtain the four-dimensional joint probability density function for the location of microseismic events.

[0066] A density estimation method is used to fit the spatial location and energy of microseismic events, and the joint probability density function of the microseismic event location-energy is calculated. It can be represented in the following form:

[0067]

[0068] Among them: θ is the source coordinate and energy of the microseismic source, θ = [x, y, z, e] T , the superscript T represents the transpose of a matrix or vector; is the event of the k-th (k = 1, 2,..., m) time series, and m is the number of all events within a certain time period; H is the bandwidth matrix, d is the data dimension, d = 4; σ i is the standard deviation of the i-th vector of θ, med(·) represents taking the median; H ii = 0, i ≠ j;

[0069] S2. Define the confidence interval

[0070] ① Select an appropriate confidence level, such as P = 85%, 95% or 99%, which depends on the mine's acceptance of different levels of risk;

[0071] ② For the above probability distribution, the confidence interval [θ1, θ2] depends on the corresponding confidence level P, and [θ1, θ2] can be calculated by the following formula:

[0072]

[0073] where P(θ1 ≤ X < E) = P(E < X < θ2), and E is the position where the probability density reaches its extreme value;

[0074] S3. Determine the threshold for classifying the hazard level

[0075] ① According to the confidence interval, the microseismic disaster hazard can be divided into multiple levels:

[0076] Low hazard: X < θ1;

[0077] Medium hazard: θ1 ≤ X ≤ θ2;

[0078] High hazard: X > θ2;

[0079] ② The specific threshold selection can consider the following three factors: the energy level that caused disasters in historical data, expert opinions and empirical judgments, and the specific geological conditions and mining methods of the mine.

[0080] To further illustrate the specific process of the microseismic positioning method, the microseismic positioning process is detailed as follows:

[0081] Assume that the sound source position coordinates are (x, y, z), and the sensor coordinates are (x i , y i , z iGiven that the wave velocity of the propagation medium is v, and based on the direct proportionality between propagation distance and travel time, the governing equations of the sound source can be obtained as follows:

[0082]

[0083] Among them, t i This refers to the time when sensor i receives the sound source signal, and t0 refers to the time when the sound source emits the signal. By squaring formula (1), we can obtain:

[0084] v 2 (t i -t0) 2 =(xx) i ) 2 +(yy i ) 2 +(zz i ) 2 (2)

[0085] Subtracting the formula for i=1 from the formula for i>1, we get:

[0086] 2(x i -x1)x+2(y i -y1)y+2(z i -z1)z-2v 2 t0(t i -t1)+v 2 (t i 2 +t1 2 )-[(x i 2 -x1 2 )+(y i 2 -y1 2 )+(z i 2 -z1 2 )]=0 (3)

[0087] Let a i =2(x i -x1), b i =2(y i -y1), c i =2(z) i -z1), d i =-2v 2 (t i -t1), e i =v 2 (t i 2 -t1 2 )-[(x i2 -x1 2 )+(y i 2 -y1 2 )+(z i 2 -z1 2 Then equation (3) can be expressed as:

[0088] a i x+b i y+c i z+d i t0+e i =0 (4)

[0089] In fact, in real monitoring environments, due to the influence of many factors such as noise interference and system errors, the above formula (4) cannot be exactly equal to 0, but has a certain deviation, which is called the equation residual, and is specifically expressed as follows:

[0090] a i x+b i y+c i z+d i t0+e i =ε (5)

[0091] The sum of squared residuals of the above governing equations is expressed as follows:

[0092]

[0093] Solving the partial derivatives of the sum of squared residuals with respect to x, y, z, t0, we obtain a system of positive definite equations concerning the source coordinates:

[0094]

[0095] Let the above partial derivatives equal to 0, and rearrange them into the following expression:

[0096]

[0097] in, The j-th (j = 1, 2, 3, 4) index is taken from vectors a, b, c, d respectively.

[0098] By using the elimination method, the unknown t0 in formula (8) is eliminated, and the control equations concerning only the source coordinates x, y, z are obtained as follows:

[0099]

[0100] in:

[0101]

[0102] Solving formula (9) yields the specific values ​​of the source coordinates x, y, and z, as shown in the following expression:

[0103]

[0104] Where, σ1=l1l6l 11 -l1l7l 10 -l2l5l 11 +l2l7l9+l3l5l 10 -l3l6l9.

[0105] This method eliminates the need for combinations of individual probes, allowing the use of all sensors. Furthermore, it employs a least-squares solution approach, which to some extent eliminates the influence of maxima and minima (especially when dealing with a large number of sensors), thus improving computational accuracy.

[0106] Compared with the prior art, the present invention has the following technical advantages:

[0107] 1. Improve the location accuracy of microseismic events: By rationally deploying a multi-directional three-dimensional monitoring network in the monitoring area, the systematic error of location is effectively reduced; environmental noise is effectively reduced and interference from abnormal signals is eliminated; an improved analytical location algorithm is developed, which effectively solves the problem of iterative non-convergence or local convergence; by solving the residual gradient, an approximate maximum likelihood estimation of the microseismic source is achieved; the new method of this invention significantly improves the location accuracy and computational efficiency of microseismic sources, providing reliable basic data for accurately assessing the risk of rockburst.

[0108] 2. Dynamic disaster threshold classification: Using the probability confidence interval theory, a dynamic multi-level threshold for disasters in mining areas is established. The early warning threshold is adaptively adjusted based on real-time monitoring data to accurately classify the disaster level.

[0109] 3. Strong model generalization ability: The model introduces a dynamic assessment model for rockburst risk based on a four-dimensional joint kernel probability density function. It does not rely on specific probability distribution assumptions of the data, nor on fixed data parameter structures, and ignores the scale effect of the data. It has a strong model generalization ability and is suitable for mine disaster early warning at different monitoring scales.

[0110] 4. Full utilization of microseismic data: The risk assessment model comprehensively considers the multidimensional characteristics of microseismic events, including the energy release rate, spatial distribution characteristics, and event time series (i.e., event frequency), which greatly improves the comprehensiveness and accuracy of rockburst risk assessment.

[0111] 5. Good real-time early warning: The real-time processing and analysis of large-scale data has been effectively solved, realizing the real-time acquisition and processing of microseismic signals. It is computationally efficient, easy to use, and can capture the dynamic changes of coal and rock masses in a timely manner, providing time guarantee for rapid response to potential dangers.

[0112] 6. Scale-Free Early Warning Model: This invention employs an effective non-parametric kernel density estimation method that does not rely on pre-defined data distribution assumptions. This allows the model to better adapt to the complex distribution characteristics of geological data at different scales, avoiding potential biases that may arise when traditional parametric methods are applied across scales. Furthermore, the confidence interval theory used is calculated based on the statistical characteristics of the data distribution, rather than relying on absolute values. Therefore, it maintains statistical consistency and stability across different scales, reducing the impact of scale changes.

[0113] 7. The method of this invention improves positioning accuracy and real-time performance through an improved linear positioning algorithm, and establishes a risk assessment model that comprehensively considers "spatiotemporal energy" information through kernel density estimation and probability confidence interval theory, providing strong support for accurate early warning of mine disasters.

[0114] 8. The technical solution proposed in this invention significantly enhances the monitoring accuracy and early warning timeliness of mine dynamic disasters. The method has excellent environmental adaptability and functional expansion potential, providing key technical support for ensuring safe production in mines. Attached Figure Description

[0115] Figure 1 A flowchart illustrating a scale-free method for fine-grained monitoring and dynamic early warning of mine dynamic disasters;

[0116] Figure 2 This is a rendering of a scale-free method for fine monitoring and dynamic early warning of mine dynamic disasters. Detailed Implementation

[0117] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings.

[0118] A scale-free method for fine-grained monitoring and dynamic early warning of mine dynamic disasters, the process of which is as follows: Figure 1 As shown, the specific steps include the following:

[0119] S1: Multiple high-sensitivity microseismic monitoring sensors are deployed in the monitoring area of ​​the mining area to form a comprehensive three-dimensional monitoring network;

[0120] S2: Real-time acquisition of microseismic signal data through a high-speed data acquisition system to ensure data continuity and integrity;

[0121] S3: Adaptive noise reduction processing is performed on the acquired microseismic signals;

[0122] S4: High-precision three-dimensional positioning of microseismic sources is achieved using an improved analytical positioning model;

[0123] S5: Segment the microseismic energy data according to multi-scale time windows (such as hour, day, week, month) to facilitate multi-level statistical analysis and trend identification;

[0124] S6: Based on the time-frequency characteristics, spatial distribution, and energy release patterns of microseismic signals, and combined with the principles of kernel probability density estimation and probability confidence intervals, a dynamic assessment model for rockburst risk is established.

[0125] S7: Calculates the risk index of rockburst in real time and dynamically classifies disaster early warnings;

[0126] S8: Establish a feedback mechanism for early warning results, continuously optimize and update risk assessment models and early warning systems through historical data analysis, post-disaster investigation, numerical simulation technology and expert knowledge base, and build a comprehensive database of mine dynamic disasters.

[0127] The layout of the all-round three-dimensional monitoring network in step S1 is as follows:

[0128] (1) A regular grid layout is adopted within the mining area to ensure that the sensors cover the entire monitoring area;

[0129] (2) Based on the geological structure of the mining area, sensors are deployed at different depth levels to form a three-dimensional monitoring network;

[0130] (3) Increase sensor density in key areas (such as mining faces, fault zones, and rock strata boundaries);

[0131] (4) The sensors are distributed in a dispersed manner, and the distance between each pair of sensors is maximized to facilitate the subsequent micro-seismic source location.

[0132] The adaptive noise reduction process for the acquired microseismic signals in step S3 is as follows:

[0133] (1) Signal preprocessing:

[0134] The original microseismic signals are sampled and digitized to ensure signal integrity and continuity.

[0135] (2) Empirical Mode Decomposition:

[0136] a) The preprocessed microseismic signal x(t) is decomposed to obtain a series of intrinsic mode functions and a residual term:

[0137]

[0138] Among them, IMF i(t) is the i-th intrinsic modulus function, r n (t) represents the residual term.

[0139] b) Steps for signal decomposition:

[0140] Identify all local extrema of signal x(t);

[0141] The upper and lower envelope lines are generated using cubic spline interpolation.

[0142] Calculate the average value m1(t) of the upper and lower envelopes;

[0143] Subtracting m1(t) from the original signal yields h1(t) = x(t) - m1(t);

[0144] Repeat the above steps until h1(t) satisfies the definition conditions of IMF;

[0145] Define h1(t) as the first IMF, i.e., c1(t) = h1(t);

[0146] Subtract c1(t) from the original signal and repeat the above process to obtain the subsequent IMFs.

[0147] (3) Noise IMF identification:

[0148] a) Calculate the energy density and center frequency of each IMF;

[0149] b) Analyze the spectral characteristics of IMFs to identify high-frequency noise IMFs;

[0150] c) Use correlation analysis or signal-to-noise ratio thresholding to determine the noise IMFs that need to be removed.

[0151] (4) Adaptive threshold denoising:

[0152] The specific steps for identifying the IMF components that represent noise can be taken from the following aspects:

[0153] a) Calculate the energy of each IMF component: Where N is the signal length;

[0154] b) Calculate the average energy:

[0155]

[0156] Where N is the total number of IMF components;

[0157] c) Calculate the standard deviation σ of all IMF energies.

[0158]

[0159] d) Set an adaptive threshold: T = μ - ασ, where α = 2;

[0160] e) IMFs with indices below the threshold T are identified as noise components.

[0161] (5) Signal reconstruction:

[0162] a) Reconstruct the denoised IMF with the retained low-frequency IMF and residual terms to obtain the denoised microseismic signal:

[0163]

[0164] Among them, IMF′ i (t) represents the denoised IMF, and k represents the number of noisy IMFs.

[0165] Through the above steps, empirical mode decomposition technology can be effectively used to reduce noise in mine microseismic signals, improve signal quality, and provide a more reliable data foundation for subsequent microseismic event analysis and early warning.

[0166] The calculation process of the improved source inversion model in step S4 is as follows:

[0167]

[0168] Where, α1=l1l6l 11 -l1l7l 10 -l2l5l 11 +l2l7l9+l3l5l 10 -l3l6l9

[0169]

[0170] Where, χ (j) The j-th index (j = 1, 2, 3, 4) takes parameters a, b, c, and d respectively. i =2(x i -x1), b i =2(y i -y1), c i =2(z) i -z1), d i =-2v 2 (t i -t1), e i =v 2 (t i 2 -t1 2 )-[(x i 2 -x1 2 )+(yi 2 -y1 2 )+(z i 2 -z1 2 The sensor coordinates are (x) i ,y i ,z i ), i = 1, 2, ..., n, where n is the number of sensors, t i The arrival time data of sensor i is represented, t1 represents the arrival time of the first triggered sensor, v represents the propagation velocity of the medium, (x,y,z) represents the coordinates of the microseismic source to be solved, and t0 represents the time when the microseismic event occurred.

[0171] The dynamic assessment model for rockburst risk in step S6 includes:

[0172] S1. Obtain the four-dimensional joint probability density function for the location of microseismic events.

[0173] A density estimation method is used to fit the spatial location and energy of microseismic events, and the joint probability density function of the microseismic event location-energy is calculated. It can be represented in the following form:

[0174]

[0175] in: θ represents the source coordinates and energy of the microseismic source, θ = [x, y, z, e]. T The superscript T indicates the transpose of a matrix or vector; Let be the event of the k-th (k = 1, 2, ..., m) time series, where m is the number of all events within a certain time period; H is the bandwidth matrix. d represents the data dimension, d = 4; σ i Let be the standard deviation of the i-th vector of θ. med(·) represents the median; H ii =0, i≠j.

[0176] S2. Define the confidence interval

[0177] ① Choose an appropriate confidence level, such as P = 85%, 95%, or 99%, depending on the mine's tolerance for different levels of risk;

[0178] ② For the above probability distribution, the confidence interval [θ1, θ2] depends on the corresponding confidence level P, and [θ1, θ2] can be calculated by the following formula:

[0179]

[0180] Among them, P(θ1≤X<E) = P(E<X<θ2), where E is the position where the probability density reaches its extreme value.

[0181] S3. Determine the threshold for hazard level classification

[0182] ① According to the confidence interval, the microseismic disaster hazard can be divided into multiple levels:

[0183] Low hazard: X<θ1;

[0184] Medium hazard: θ1≤X≤θ2;

[0185] High hazard: X>θ2.

[0186] ② For the specific threshold selection, the following three factors can be considered: the energy level that caused disasters in historical data, expert opinions and empirical judgments, and the specific geological conditions and mining methods of the mine.

[0187] Specific implementation case 1:

[0188] In a microseismic monitoring system of a mine with dimensions of 1000 meters × 1000 meters × 1000 meters, 16 sensors are deployed at key positions in the system. The coordinates of these sensors cover all central areas of the monitoring space, ensuring comprehensive coverage. The coordinates of these sensors are (0,0,0), (1000,0,0), (1000,1000,0), (0,1000,0), (0,0,1000), (1000,0,1000), (1000,1000,1000), (0,1000,1000), (500,0,500), (1000,500,500), (500,1000,500), (0,500,500), (500,500,0), (500,500,1000), (0,0,500), (1000,1000,500) (meters).

[0189] To verify the effectiveness of the method of the present invention, a microseismic event with a true hypocenter coordinate of (248.36, 833.78, 16.37) (meters) and a wave speed of 5000 m / s is set. Through simulation calculations, the arrival time data of the signals received by each sensor are obtained. To simulate the noise interference in the actual environment, random errors with a standard deviation of 0.3 milliseconds are added to these data.

[0190] Using this set of noisy arrival data, the source coordinates calculated by the method of this invention are (248.54, 833.87, 15.27) meters, with a positioning error of (0.18, 0.09, -1.1) meters and an absolute distance error of only 1.11 meters. In comparison, the positioning errors of existing methods such as LLSUV (wave velocity-free linear least squares method), NIUV (wave velocity-free non-iterative method), USBM (method proposed by the U.S. Bureau of Mines), and CAS (comprehensive analytical method) are 1.86 meters, 1.67 meters, 3.13 meters, and 16.69 meters, respectively.

[0191] The positioning accuracy of this method is improved by 40.2%, 33.4%, 64.5%, and 93.3% compared to these existing methods, respectively. Furthermore, this method can accurately calculate the triggering time of microseismic events with a calculation error of 1.83 × 10⁻⁶. -7 The error is negligible.

[0192] The effect of the scale-effect-free fine monitoring and dynamic early warning method for mine dynamic disasters of the present invention is as follows: Figure 2 As shown, these results demonstrate that the method of the present invention has significant advantages in terms of microseismic positioning accuracy and time determination, providing more reliable technical support for real-time monitoring and prevention of mine disasters.

[0193] Furthermore, the energy release of microseismic events can be determined by the envelope area of ​​the microseismic signal waveform (represented by voltage level data). Then, statistical analysis is performed on the spatiotemporal energy information of the acquired microseismic events to achieve accurate classification and early warning of mine dynamic disasters.

[0194] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A scale-free method for fine-grained monitoring and early warning of dynamic disasters in mines, characterized in that, Includes the following steps: S1: Multiple high-sensitivity microseismic monitoring sensors are deployed in the monitoring area of ​​the mining area to form a comprehensive three-dimensional monitoring network; S2: Real-time acquisition of microseismic signal data through a high-speed data acquisition system to ensure data continuity and integrity; S3: Adaptive noise reduction processing is performed on the acquired microseismic signals; S4: High-precision three-dimensional positioning of microseismic sources is achieved using an improved analytical positioning model; S5: Segment the microseismic energy data into multi-scale time windows to enable multi-level statistical analysis and trend identification; S6: Based on the time-frequency characteristics, spatial distribution, and energy release patterns of microseismic signals, and combined with the principles of kernel probability density estimation and probability confidence intervals, a dynamic assessment model for rockburst risk is established. S7: Calculates the risk index of rockburst in real time and dynamically classifies disaster early warnings; S8: Establish a feedback mechanism for early warning results, continuously optimize and update risk assessment models and early warning systems through historical data analysis, post-disaster investigation, numerical simulation technology and expert knowledge base, and build a comprehensive database of mine dynamic disasters.

2. The method for fine monitoring and early warning of mine dynamic disasters without scale effects according to claim 1, characterized in that, The layout of the all-round three-dimensional monitoring network in step S1 is as follows: (1) A regular grid layout is adopted within the mining area to ensure that the sensors cover the entire monitoring area; (2) Based on the geological structure of the mining area, sensors are deployed at different depth levels to form a three-dimensional monitoring network; (3) Increase sensor density in key areas; (4) The sensors are distributed in a dispersed manner, and the distance between each pair of sensors is maximized to facilitate the subsequent micro-seismic source location.

3. The method for fine monitoring and early warning of mine dynamic disasters without scale effects according to claim 1, characterized in that, The adaptive noise reduction process for the acquired microseismic signals in step S3 is as follows: (1) Signal preprocessing: The original microseismic signals are sampled and digitized to ensure signal integrity and continuity; (2) Empirical Mode Decomposition: a) The preprocessed microseismic signal x(t) is decomposed to obtain a series of intrinsic mode functions and a residual term: Among them, IMF i (t) is the i-th intrinsic modulus function, r n (t) represents the residual term; b) Steps for signal decomposition: Identify all local extrema of signal x(t); The upper and lower envelope lines are generated using cubic spline interpolation. Calculate the average value m1(t) of the upper and lower envelopes; Subtracting m1(t) from the original signal yields h1(t) = x(t) - m1(t); Repeat the above steps until h1(t) satisfies the definition conditions of IMF; Define h1(t) as the first IMF, i.e., c1(t) = h1(t); Subtract c1(t) from the original signal and repeat the above process to obtain the subsequent IMFs; (3) Noise IMF identification: a) Calculate the energy density and center frequency of each IMF; b) Analyze the spectral characteristics of IMFs to identify high-frequency noise IMFs; c) Use correlation analysis or signal-to-noise ratio thresholding to determine the noise IMFs that need to be removed; (4) Adaptive threshold denoising: The specific steps for identifying the IMF components that represent noise can be taken from the following aspects: a) Calculate the energy of each IMF component: b) Calculate the average energy: Where N is the total number of IMF components; c) Calculate the standard deviation σ of all IMF energies. d) Set an adaptive threshold: T = μ - ασ, where α = 2; e) IMFs with indices below the threshold T are identified as noise components; (5) Signal reconstruction: a) Reconstruct the denoised IMFs, the retained low-frequency IMFs, and the residual term to obtain the denoised microseismic signal: Among them, IMF′ i (t) represents the denoised IMF, and k represents the number of noisy IMFs; Through the above steps, the empirical mode decomposition technology can be effectively utilized to achieve noise reduction of the microseismic signals in the mining area, improve the signal quality, and provide a more reliable data basis for subsequent microseismic event analysis and early warning.

4. The method for fine monitoring and early warning of mine dynamic disasters without scale effects according to claim 1, characterized in that, The calculation formula of the improved analytical positioning model in step S4 is as follows: Where, α1=l1l6l 11 -l1l7l 10 -l2l5l 11 +l2l7l9+l3l5l 10 -l3l6l9 Where, χ (j) The j-th (j = 1, 2, 3, 4) index takes parameters a, b, c, d respectively; a i =2(x i -x1), b i =2(y i -y1), c i =2(z) i -z1), d i =-2v 2 (t i -t1), e i =v 2 (t i 2 -t1 2 )-[(x i 2 -x1 2 )+(y i 2 -y1 2 )+(z i 2 -z1 2 The sensor coordinates are (x) i ,y i ,z i ), i = 1, 2, ..., n, where n is the number of sensors, t i The arrival time data of sensor i is represented, t1 represents the arrival time of the first triggered sensor, v represents the propagation velocity of the medium, (x,y,z) represents the coordinates of the microseismic source to be solved, and t0 represents the time when the microseismic event occurred.

5. The method for fine monitoring and early warning of mine dynamic disasters without scale effects according to claim 1, characterized in that, The dynamic assessment model for rock burst risk in step S6 includes: S1. Obtain the four-dimensional joint probability density function of the microseismic event location A density estimation method is used to fit the spatial location and energy of microseismic events, and the joint probability density function of the microseismic event location-energy is calculated. It can be represented in the following form: in: θ represents the source coordinates and energy of the microseismic source, θ = [x, y, z, e]. T The superscript T indicates the transpose of a matrix or vector; Let be the event of the k-th (k = 1, 2, ..., m) time series, where m is the number of all events within a certain time period; H is the bandwidth matrix. d represents the data dimension, d = 4; σ i Let be the standard deviation of the i-th vector of θ. med(·) represents the median; H ii =0, i≠j; S2. Define the confidence interval ① Select an appropriate confidence level, which depends on the mine's acceptance of different levels of risk; ② For the above probability distribution, the confidence interval [θ1, θ2] depends on the corresponding confidence level P, and [θ1, θ2] can be calculated by the following formula: where P(θ1 ≤ X < E) = P(E < X < θ2), and E is the location where the probability density reaches its extreme value; S3. Determine the threshold for classifying the hazard level ① According to the confidence interval, the microseismic disaster hazard can be classified into multiple levels: Low hazard: X < θ1; Medium hazard: θ1 ≤ X ≤ θ2; High hazard: X > θ2; ② The specific threshold selection considers the following three factors: the energy level that caused disasters in historical data, expert opinions and empirical judgments, and the specific geological conditions and mining methods of the mine.

Images