Intelligent rockburst early warning and protection system based on micro-seismic monitoring and numerical simulation

Abstract

The invention discloses a rockburst intelligent early warning and protection system based on micro-seismic monitoring and numerical simulation, and relates to the technical field of rock mechanics and geotechnical engineering.The system is characterized in that a plurality of sensors are arranged in a tunnel, sensor data are received through a micro-seismic information processing module, noise reduction processing is conducted on the sensor data, the position of a seismic source is positioned, and the seismic source is monitored; generating micro-seismic parameters; the danger alarm module receives the micro-seismic parameters, performs surrounding rock stress prediction through the micro-seismic parameters, compares the predicted surrounding rock stress with a preset threshold value, and performs danger level early warning if the predicted surrounding rock stress is greater than the corresponding threshold value; the micro-seismic information and numerical model display module is used for displaying micro-seismic information and surrounding rock stress; and the prevention module receives the danger level early warning and starts corresponding protection measures. According to the method, accurate prediction of surrounding rock stress of rockburst and positioning of corresponding positions are realized through various micro-seismic parameters.

Description

Technical Field

[0001] The present invention relates to the technical field of rock mechanics and geotechnical engineering, and more specifically to a rockburst intelligent early warning and protection system based on microseismic monitoring and numerical simulation. Background Art

[0002] At present, with the rapid economic development of our country, the scale of infrastructure construction is constantly expanding, the demand for energy is also increasing continuously, and many large-scale water conservancy and hydropower projects have begun to be built, the investment in basic transportation construction has been increased, and mineral resources are gradually developed deeper. The characteristics of "long, large, deep, and clustered" of underground projects such as hydraulic tunnels, traffic tunnels, and mine shafts are becoming more and more obvious, and at the same time, many rock mechanics problems are brought, such as large deformation of surrounding rocks, collapse, and rockburst, etc. These problems seriously restrict the improvement of the construction level of underground projects and pose a great threat to the safety of personnel and equipment. Among them, rockburst is a special form of damage generated during the excavation process in a high-stress environment in brittle rock areas. The strong energy release and surrounding rock damage accompanied by its occurrence pose a serious threat to construction safety and also have a serious impact on the on-site construction progress.

[0003] The rockburst process is complex, occurs quickly, and has a serious degree of damage, making it difficult to predict, and the time is urgent. The microseismic monitoring technology has laid a foundation for solving this scientific problem due to its unique advantages. Scientific researchers use the change of microseismic source parameters to monitor and early warn the rock unloading process. However, using a single parameter for prediction and early warning has a poor prediction accuracy and a large error.

[0004] Therefore, how to improve the prediction accuracy of rockburst is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides a rockburst intelligent early warning and protection system based on microseismic monitoring and numerical simulation, which can accurately predict the surrounding rock stress of rockburst through a variety of microseismic parameters and locate the corresponding position.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A rockburst intelligent early warning and protection system based on microseismic monitoring and numerical simulation, comprising: a microseismic information processing module, a danger alarm module, a microseismic information and numerical model display module, and a prevention module. A number of sensors are arranged inside the tunnel. The microseismic information processing module receives the sensor data, performs noise reduction processing on the sensor data, locates the source position, and generates microseismic parameters; the danger alarm module receives the microseismic parameters, predicts the surrounding rock stress through the microseismic parameters, and compares the predicted surrounding rock stress with a preset threshold. If it is greater than the corresponding threshold, a danger level early warning is issued; the microseismic information and numerical model display module displays the microseismic information and the surrounding rock stress; the prevention module receives the danger level early warning and activates corresponding protection measures.

[0008] Preferably, the microseismic information processing module includes:

[0009] A data preprocessing module, which obtains the sensor data, combines it with a preset acoustic wave database, extracts the reference noise data under the same geological conditions as the sensor data, then performs the same frame segmentation processing on the two types of data, and extracts the MFCC coefficients of the two types of data; calculates the correlation coefficient between each frame of the sensor data and the reference noise frame, and performs clustering to identify the frame signals without noise and with noise; retains the non-noise frame signals, and performs a difference processing on the noise-containing frame signals and the reference noise frame signals in the frequency domain; restores the non-noise frame signals and the noise-containing frame signals to obtain noise-reduced microseismic data;

[0010] A source position location module, which locates the source position through the noise-reduced microseismic data;

[0011] A data transmission module, which sends the source position to the microseismic information and numerical model display module, processes the noise-reduced microseismic data, and obtains the microseismic parameters and sends them to the danger alarm module.

[0012] Preferably, the data preprocessing module specifically includes:

[0013] Performing the same frame segmentation processing on the sensor data x(t) and the reference noise data z(t), where x(t) is divided into n 1 frames, and z(t) is divided into n 2 frames. The frame segmentation formula is as follows:

[0014] f n =(N - wlen + inc) / inc;

[0015] overlap = wlen - inc;

[0016] where N is the length of the sound data, wlen is the set frame length, inc is the set frame shift, overlap is the frame overlap, and f n is the number of frames into which the data is divided;

[0017] The MFCC coefficients of each frame of the sensor data x(t) and the reference noise data z(t), namely M(z,n 2 ), M(x,n 1 );

[0018] For M(x,k) of each frame signal of x(t), the Pearson coefficient between M(x,k) and all frames M(z,q) of z(t) is calculated respectively, and then the average value is calculated to obtain the correlation magnitude R(k) of each frame signal of x(t) and z(t). The formula is as follows:

[0019]

[0020] where k ∈ (1,n 1 ), q ∈ (1,n 2 ); Con(x,y) is the covariance of variables x and y, and σ x and σ y are the variances of x and y respectively;

[0021] The frame signals without noise and with noise are identified by clustering R(k) using the K_means method; the non-noise frame signals are retained, and the frame signals with noise and the reference noise frame signals are subtracted in the frequency domain; the non-noise frame signals and the frame signals with noise are restored frame by frame to obtain the denoised microseismic data.

[0022] Preferably, the seismic source position positioning module specifically includes:

[0023]

[0024] where a, b, c are the coordinates of the test points, t is the event occurrence time, a i , b i , c i are the coordinates of the i-th sensor, t i is the time when the P wave arrives at the i-th sensor, and v P is the P wave velocity.

[0025] Preferably, the microseismic parameters include: microseismic P wave, microseismic S wave, daily microseismic event number, event energy, dominant frequency information and signal frequency characteristic parameters.

[0026] Preferably, the danger prediction module specifically includes:

[0027] The comprehensive model establishment module selects a classification machine learning model to construct sub-models for each microseismic parameter, establishes a training set through historical microseismic parameters, trains each sub-model to obtain several surrounding rock stress prediction models; analyzes the accuracy of each surrounding rock stress prediction model, calculates the correlation between different models, and the stability of sample data, thereby obtaining the fusion weights of the corresponding surrounding rock stress prediction models; based on the fusion weights, combines all the surrounding rock stress prediction models to construct a comprehensive prediction model;

[0028] The model prediction module predicts the surrounding rock stress for the microseismic parameters to be predicted according to the comprehensive prediction model, and obtains the predicted surrounding rock stress;

[0029] The threshold comparison and warning module compares the surrounding rock stress with a preset threshold and gives a warning corresponding to the danger level.

[0030] Preferably, the obtaining of the fusion weights of the corresponding surrounding rock stress prediction models specifically includes:

[0031] Establish a test set through historical microseismic parameters, analyze the prediction accuracy of each sub-model as the first index of the fusion weight; calculate the parameter correlation degree between different microseismic parameters as the second index of the fusion weight; calculate the data fluctuation coefficient of the microseismic parameters to be predicted as the third index of the fusion weight; add the three weights after normalization to obtain the final fusion weight of each sub-model;

[0032] The importance of the sub-models generated by each microseismic parameter in the fusion model is determined by the fusion weights; multiply the classification results of each sub-model by the weight size and combine them to obtain the final comprehensive prediction model.

[0033] Preferably, the classification machine learning model selects any one of the random forest, gradient boosting tree, and support vector machine classification machine learning models to construct sub-models, and takes the training set of each microseismic parameter as the input of the sub-model to obtain the surrounding rock stress prediction model of each microseismic parameter.

[0034] As can be seen from the above technical solutions, compared with the prior art, the present invention discloses a rockburst intelligent early warning and protection system based on microseismic monitoring and numerical simulation. By combining the microseismic information processing module with the acoustic wave database, precise noise reduction is achieved through technologies such as frame division, feature extraction, and correlation analysis. The source location is accurately calculated using a specific formula, improving the quality of microseismic signals. The danger alarm module constructs sub-models using multiple machine learning models, determines the fusion weights by comprehensively considering the prediction accuracy rate, parameter correlation degree, and data fluctuation coefficient, and establishes a comprehensive prediction model to comprehensively and accurately predict the surrounding rock stress and reduce the risk of misjudgment. The threshold comparison and warning module can quickly compare the predicted stress with the threshold and issue an accurate danger level warning in a timely manner. The prevention module activates corresponding protection measures according to different warning levels to avoid waste of resources and effectively ensure the safety of personnel and equipment. The microseismic information and numerical model display module displays the microseismic information and the surrounding rock stress in real time, and presents the rock mass state intuitively in combination with the numerical model analysis, facilitating relevant personnel to master the situation and prevent dangers in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0036] Figure 1 The system structure diagram provided by the present invention;

[0037] Figure 2 The system operation diagram provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0038] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0039] The embodiments of the present invention disclose a rockburst intelligent early warning and protection system based on microseismic monitoring and numerical simulation, as Figure 1As shown in the figure, it includes: a microseismic information processing module, a danger warning module, a microseismic information and numerical model display module, and a prevention module. A number of sensors are set inside the tunnel. The microseismic information processing module receives the sensor data, performs noise reduction processing on the sensor data, locates the source position, and generates microseismic parameters; the danger warning module receives the microseismic parameters, predicts the surrounding rock stress through the microseismic parameters, and compares the predicted surrounding rock stress with a preset threshold. If it is greater than the corresponding threshold, a danger level warning is issued; the microseismic information and numerical model display module displays the microseismic information and the surrounding rock stress; the prevention module receives the danger level warning and activates corresponding protection measures.

[0040] In a specific embodiment, the microseismic information processing module includes:

[0041] A data preprocessing module, which obtains the sensor data, combines it with a preset acoustic wave database, extracts the reference noise data under the same geological conditions as the sensor data, then performs the same frame division processing on the two types of data, and extracts the MFCC coefficients of the two types of data; calculates the correlation coefficient between each frame of the sensor data and the reference noise frame, and performs clustering to identify the frame signals without noise and with noise; retains the non-noise frame signals, and performs a difference processing on the noisy frame signals and the reference noise frame signals in the frequency domain; restores the non-noise frame signals and the noisy frame signals to obtain the noise-reduced microseismic data;

[0042] A source position location module, which locates the source position through the noise-reduced microseismic data;

[0043] A data transmission module, which sends the source position to the microseismic information and numerical model display module, processes the noise-reduced microseismic data, and obtains the microseismic parameters and sends them to the danger warning module.

[0044] In a specific embodiment, the data preprocessing module specifically includes:

[0045] Perform the same frame division processing on the sensor data x(t) and the reference noise data z(t), where x(t) is divided into n 1 frames, and z(t) is divided into n 2 frames. The frame division formula is as follows:

[0046] f n =(N - wlen + inc) / inc;

[0047] overlap = wlen - inc;

[0048] where N is the length of the sound data, wlen is the set frame length, inc is the set frame shift, overlap is the frame overlap, and f n is the number of frames into which the data is divided;

[0049] The MFCC coefficients of each frame of the sensor data x(t) and the reference noise data z(t), namely M(z,n 2 ) and M(x,n 1 );

[0050] For M(x,k) of each frame signal of x(t), the Pearson coefficient between M(x,k) and M(z,q) of all frames of z(t) is calculated respectively, and then the average value is calculated to obtain the correlation magnitude R(k) of each frame signal of x(t) and z(t). The formula is as follows:

[0051]

[0052] where k ∈ (1,n 1 ), q ∈ (1,n 2 ); Con(x,y) is the covariance of variables x and y, and σ x and σ y are the variances of x and y respectively;

[0053] The K_means method is used to cluster and identify the frame signals with and without noise from R(k); the non-noise frame signals are retained, and the difference processing is performed on the noisy frame signals and the reference noise frame signals in the frequency domain; the non-noise frame signals and the noisy frame signals are frame-restored to obtain the denoised microseismic data.

[0054] In a specific embodiment, the seismic source location module specifically includes:

[0055]

[0056] where a, b, c are the coordinates of the test points, t is the event occurrence time, a i , b i , c i are the coordinates of the i-th sensor, t i is the time when the P wave arrives at the i-th sensor, and v P is the P wave velocity.

[0057] In a specific embodiment, the microseismic parameters include: microseismic P wave, microseismic S wave, daily microseismic event number, event energy, main frequency information and signal frequency characteristic parameters.

[0058] In a specific embodiment, the danger prediction module specifically includes:

[0059] The comprehensive model establishment module selects a classification machine learning model to construct sub-models for each microseismic parameter, establishes a training set through historical microseismic parameters, trains each sub-model to obtain several surrounding rock stress prediction models; analyzes the accuracy of each surrounding rock stress prediction model, calculates the correlation between different models, and the stability of sample data, thereby obtaining the fusion weights of the corresponding surrounding rock stress prediction models; based on the fusion weights, combines all surrounding rock stress prediction models to construct a comprehensive prediction model;

[0060] The model prediction module predicts the surrounding rock stress according to the comprehensive prediction model for the microseismic parameters to be predicted, and obtains the predicted surrounding rock stress;

[0061] The threshold comparison and warning module compares the surrounding rock stress with a preset threshold and gives a warning corresponding to the danger level.

[0062] In a specific embodiment, obtaining the fusion weights of the corresponding surrounding rock stress prediction models specifically includes:

[0063] Establish a test set through historical microseismic parameters, analyze the prediction accuracy of each sub-model as the first index of the fusion weight; calculate the parameter correlation degree between different microseismic parameters as the second index of the fusion weight; calculate the data fluctuation coefficient of the microseismic parameters to be predicted as the third index of the fusion weight; add the three weights after normalization to obtain the final fusion weight of each sub-model;

[0064] The fusion weight determines the importance of the sub-models generated by each microseismic parameter in the fusion model; multiply the classification result of each sub-model by the weight size and combine to obtain the final comprehensive prediction model.

[0065] In a specific embodiment, the classification machine learning model selects any one of the random forest, gradient boosting tree, and support vector machine classification machine learning models to construct sub-models, and uses the training set of each microseismic parameter as the input of the sub-model to obtain the surrounding rock stress prediction model of each microseismic parameter.

[0066] In specific embodiment 1, the microseismic information processing module first locally stores the microseismic signal after obtaining the microseismic signal by acquiring microseismic monitoring data, and then uploads it to the cloud through satellite communication; when locally storing, first perform manual processing and adjustment on the microseismic signal, and then use intelligent algorithms for artificial intelligence learning, intelligent noise reduction, intelligent positioning, etc. The artificial intelligence learning is to learn the results of manual processing and adjustment, and automatically process the sensor data after deep learning for a period of time;

[0067] The danger alarm module determines the threshold for three-level early warning through comprehensive analysis of microseismic parameters, numerical simulation results, deep learning, and on-site feedback. The microseismic parameters include the number of daily microseismic events, event energy, dominant frequency information, and other frequency characteristics of the signal. The parameters of the numerical simulation results are mainly the surrounding rock stress. (The specific early warning threshold is determined according to the on-site situation.)

[0068] The microseismic information and numerical model display module establishes a numerical calculation model based on detailed on-site geological data and CAD construction drawings, and simulates the on-site excavation progress. It displays the continuously changing state of the surrounding rock stress and can also display the damage state of the surrounding rock, and real-time displays the numerical simulation results, microseismic event distribution, on-site excavation conditions, and intelligent early warning analysis results at the local end.

[0069] The prevention module corresponds different prevention measures to different early warning levels. When an early warning is issued on-site, corresponding defense measures, including three-level defense measures, will also be proposed.

[0070] Furthermore, the artificial intelligence learning at the local end identifies and denoises the microseismic signal to obtain the effective waveform of the microseismic signal, then picks up the P-wave and S-wave of the microseismic signal and conducts intelligent positioning of the microseismic event, and makes feedback corrections based on the actual on-site situation. The positioning accuracy needs to meet the standards.

[0071] Furthermore, the existing technology generally gives an early warning when a single parameter value is abnormal, resulting in a large early warning error. This intelligent early warning protection system does not give an early warning due to the abnormality of a certain parameter. It comprehensively analyzes the stability of the surrounding rock through multiple microseismic parameters and numerical simulation results to determine the threshold. When this comprehensive value exceeds the threshold, an early warning is issued. The system has good stability and high accuracy.

[0072] Furthermore, the acoustic-optic early warning system and the satellite early warning system give early warnings simultaneously. Since rock bursts occur suddenly, the time difference can be eliminated as much as possible to ensure on-site safety. The acoustic-optic early warning system can generally notify front-line workers in a timely manner, and the satellite early warning system can notify leaders at all levels. At the same time, the on-site situation can be grasped and corresponding responses can be made in a timely manner.

[0073] Furthermore, the intelligent early warning system includes a three-level early warning system and corresponding protection measures, which can enable on-site construction personnel to take corresponding measures in the first time, ensuring the safety of personnel and equipment to the greatest extent, and improving construction efficiency and economic benefits.

[0074] Furthermore, in the microseismic information display module, the numerical simulation operation results are displayed through 3D space software, which can show the stress distribution characteristics and potential dangerous areas of the tunnel during the excavation process, and this is not available in other systems; it can also display the microseismic event monitoring results; display the surrounding rock stability warning level information of comprehensive analysis; display the on-site construction situation information; enable the relevant construction personnel to grasp any on-site construction dynamics at any time, ensure on-site safety; and can also reduce the illegal operations of on-site personnel.

[0075] In the specific embodiment 2, as Figure 2 shown, it includes:

[0076] The first step is to use the microseismic monitoring system to monitor the excavation process of the tunnel in real time, record and obtain the waveform information of rock microfractures in real time; collect the on-site engineering geological data.

[0077] The second step is to first manually identify the microseismic waveforms, and then perform artificial intelligence learning to improve the accuracy of waveform recognition and give feedback.

[0078] The third step is that the waveforms identified by intelligence contain some noise information, which affects the determination of subsequent parameters. Introduce and improve the time-frequency analysis technology, select a suitable intelligent noise reduction method, which can not only retain the original signal characteristics but also achieve the purpose of maximum noise reduction.

[0079] The fourth step is to then pick up the P-wave and S-wave of the microseismic signal and perform intelligent positioning of the microseismic event; compare the positioning result with the on-site situation. If the error exceeds 10m, improve the intelligent noise reduction technology and re-noise the waveform until the positioning error is less than 10m and meets the specification requirements.

[0080] The fifth step is to use intelligent algorithms to extract and read the waveform parameters, and at the same time import the microseismic parameters (number of events, energy) into the numerical model. Among them, establishing the numerical model is the prior art.

[0081] The sixth step is to perform intelligent analysis on the numerical simulation results. The system can display the stress state of the tunnel surrounding rock and the surrounding rock damage characteristics, automatically delineate the on-site surrounding rock damage area and potential rockburst-prone areas, compare and feedback the numerical simulation results with the on-site situation, and gradually correct the numerical model;

[0082] The seventh step is the intelligent warning system, which comprehensively analyzes the stability of the surrounding rock through multiple microseismic parameters and numerical simulation results, and comprehensively determines the state of the surrounding rock. That is, the surrounding rock is stable, the surrounding rock is at the first-level warning, the surrounding rock is at the second-level warning, the surrounding rock is at the third-level warning, and implement the three-level warning system for hierarchical management.

[0083] The eighth step is that at the beginning, the warning level needs to be compared and feedback with the on-site situation until it is consistent with the on-site situation.

[0084] Step 9: The early warning system is divided into an acoustic-optical early warning system, which includes blue lights, yellow lights, and red lights respectively; and a satellite communication system, which has high reliability and includes telephone, text message, and large-screen early warning respectively. The intelligent early warning protection system of the present invention can reach the personnel related to the construction simultaneously in the shortest time.

[0085] Step 10: The intelligent early warning system simultaneously proposes corresponding protection measures, and the relevant personnel at all levels who receive the early warning reminder immediately take corresponding measures.

[0086] Step 11: The general control center and the on-site large screen display the on-site excavation status, construction progress, surrounding rock stress distribution and damage status, microseismic monitoring status, and surrounding rock stability status in real time.

[0087] In this specification, each embodiment is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple. For the relevant parts, reference can be made to the description in the method section.

[0088] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intelligent rockburst early warning and protection system based on microseismic monitoring and numerical simulation, characterized in that: include: A microseismic information processing module, a danger alarm module, a microseismic information and numerical model display module, and a prevention module are provided inside the tunnel. A plurality of sensors are arranged inside the tunnel. The microseismic information processing module receives sensor data, performs noise reduction processing on the sensor data, locates the earthquake source, and generates microseismic parameters. The danger alarm module receives the microseismic parameters, predicts the surrounding rock stress through the microseismic parameters, and compares the predicted surrounding rock stress with a preset threshold. If the predicted surrounding rock stress is greater than the corresponding threshold, a danger level warning is issued. The microseismic information and numerical model display module displays the microseismic information and surrounding rock stress; The prevention module receives the danger level warning and initiates corresponding protective measures.

2. The intelligent rockburst early warning and protection system based on microseismic monitoring and numerical simulation according to claim 1 is characterized in that: The microseismic information processing module includes: The data preprocessing module obtains the sensor data and combines it with the preset acoustic wave database, extracts the reference noise data under the same geological conditions as the sensor data, and then performs the same frame processing on the two data, and extracts the MFCC coefficients of the two data; calculates the correlation coefficient between each frame of the sensor data and the reference noise frame, and performs clustering to identify the noise-free and noise-containing frame signals; retains the non-noise frame signal, and performs difference processing on the noisy frame signal and the reference noise frame signal in the frequency domain; performs frame restoration on the non-noise frame signal and the noisy frame signal to obtain the noise-reduced microseismic data; A seismic source location positioning module, which locates the seismic source position through the noise-reduced microseismic data; The data transmission module sends the earthquake source position to the microseismic information and numerical model display module, processes the noise-reduced microseismic data, obtains microseismic parameters and sends them to the danger alarm module.

3. The intelligent rockburst early warning and protection system based on microseismic monitoring and numerical simulation according to claim 2 is characterized in that: The data preprocessing module specifically includes: The same frame division process is performed on the sensor data x(t) and the reference noise data z(t), where x(t) is divided into n1 frames and z(t) is divided into n2 frames. The frame division formula is as follows: f n =(N-wlen+inc) / inc; overlap=wlen-inc; Among them, N is the length of the sound data, wlen is the set frame length, inc is the set frame shift, overlap is the frame overlap, f n The number of frames the data is divided into; MFCC coefficients of each frame of sensor data x(t) and reference noise data z(t), i.e., M(z,n2) and M(x,n1); For each frame signal M(x,k) of x(t), the Pearson coefficient with all frames M(z,q) of z(t) is calculated respectively, and then the average value is calculated to obtain the correlation size R(k) between each frame signal of x(t) and z(t). The formula is as follows: Among them, k∈(1,n1), q∈(1,n2); Con(x,y) is the covariance of variables x and y, σ x and σ y are the variances of x and y respectively; The K_means method is used to cluster R(k) to identify the noise-free and noisy frame signals; the non-noise frame signals are retained, and the noisy frame signals are subtracted from the reference noise frame signals in the frequency domain; the non-noise frame signals and the noisy frame signals are restored to obtain the denoised microseismic data.

4. The rockburst intelligent early warning and protection system based on microseismic monitoring and numerical simulation according to claim 2 is characterized in that: The earthquake source location positioning module specifically includes: Among them, a, b, c are the coordinates of the test points, t is the time when the event occurs, a i ,b i ,c i is the coordinate of the i-th sensor, t i is the time when the P wave reaches the i-th sensor, v P is the P-wave velocity.

5. The intelligent rockburst early warning and protection system based on microseismic monitoring and numerical simulation according to claim 2 is characterized in that: The microseismic parameters include: microseismic P waves, microseismic S waves, number of microseismic events per day, event energy, main frequency information and signal frequency characteristic parameters.

6. The rockburst intelligent early warning and protection system based on microseismic monitoring and numerical simulation according to claim 5 is characterized in that: The danger prediction module specifically includes: The comprehensive model building module selects the classification machine learning model to construct a sub-model for each microseismic parameter, establishes a training set through historical microseismic parameters, trains each sub-model, and obtains several surrounding rock stress prediction models; performs accuracy analysis on each surrounding rock stress prediction model, and calculates the correlation between different models and the stability of sample data, thereby obtaining the fusion weight of the corresponding surrounding rock stress prediction model; based on the fusion weight, combines all surrounding rock stress prediction models to build a comprehensive prediction model; A model prediction module predicts surrounding rock stress for the microseismic parameters to be predicted according to the comprehensive prediction model to obtain predicted surrounding rock stress; The threshold comparison alarm module compares the surrounding rock stress with a preset threshold and gives a corresponding danger level warning.

7. The intelligent rockburst early warning and protection system based on microseismic monitoring and numerical simulation according to claim 6 is characterized in that: The fusion weight of the corresponding surrounding rock stress prediction model is obtained specifically including: A test set was established through historical microseismic parameters, and the prediction accuracy of each sub-model was analyzed as the first indicator of the fusion weight; the parameter correlation between different microseismic parameters was calculated as the second indicator of the fusion weight; the data fluctuation coefficient of the microseismic parameters to be predicted was calculated as the third indicator of the fusion weight; the three weights were normalized and added to obtain the final fusion weight of each sub-model; The importance of the sub-model generated by each microseismic parameter in the fusion model is determined by the fusion weight; the classification result of each sub-model is multiplied by the weight, and the final comprehensive prediction model is obtained by combining them.

8. The rockburst intelligent early warning and protection system based on microseismic monitoring and numerical simulation according to claim 6 is characterized in that: The classification machine learning model selects any one of the random forest, gradient boosting tree, and support vector machine classification machine learning models to construct a sub-model, and uses the training set of each microseismic parameter as the input of the sub-model to obtain a surrounding rock stress prediction model for each microseismic parameter.

Images