Rock burst monitoring and early warning method and system based on multi-source information fusion

Abstract

The invention discloses a rock burst monitoring and early warning method and system based on multi-source information fusion. The method comprises the steps of collecting multi-dimensional data of a target mining area; carrying out integration processing on the multi-dimensional data by adopting a Kalman filtering fusion algorithm; generating a three-dimensional geologic structure model of the target mining area by using a geologic structure inversion method; carrying out regional division on the three-dimensional geologic structure model; predicting the rock burst risk level of each region by using the rock burst risk prediction model; and performing graded early warning. According to the system, a data acquisition module is used for acquiring multi-dimensional data of a target mining area; the data fusion module is used for carrying out integration processing on the data; the model construction module is used for generating a three-dimensional geologic structure model; the region division module is used for performing region division on the three-dimensional geological structure model; the risk prediction module is used for performing risk prediction; the grading early warning module is used for carrying out early warning grading; and the alarm is used for executing graded early warning actions. According to the method and the system, accurate monitoring and intelligent early warning of the rock burst can be realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of coal mine rock burst intelligent monitoring, and specifically relates to a rock burst monitoring and early warning method and system based on multi-source information fusion. Background Art

[0002] Rock burst is a common and significant geological disaster during mining operations. Characterized by suddenness and destructive force, rock bursts pose a serious threat to the safety of underground workers and the smooth operation of mining operations. As shallow mineral resources become increasingly depleted, mining operations are gradually extending to deeper areas. Simultaneously, factors such as high ground stress and complex geological structures have significantly increased the frequency and intensity of rock bursts, posing significant challenges to mine safety management.

[0003] Currently, existing rock burst monitoring and early warning methods have many limitations. Traditional monitoring technologies often rely on a single type of sensor, such as stress sensors or microseismic sensors, which can only capture changes in local physical parameters in the mining area, but cannot fully reflect the geological conditions and stress distribution of the entire mining area. At the same time, due to the single data source, the monitoring results are often one-sided, which can easily lead to misjudgments or missed judgments, and thus cannot provide a reliable decision-making basis for rock burst prevention and control. To solve this problem, there is an urgent need to provide a rock burst monitoring and early warning method and system based on multi-source information fusion. Summary of the Invention

[0004] In response to the problems existing in the above-mentioned prior art, the present invention provides a rock burst monitoring and early warning method and system based on multi-source information fusion. The method has a simple implementation process, low implementation cost, and high intelligence. It can realize efficient and accurate monitoring of rock burst, effectively reduce the probability of rock burst occurrence, and ensure safe production operations in mining areas. The system has a simple structure, a high degree of intelligence, and reliable performance. It can realize accurate monitoring and intelligent early warning of rock burst, and effectively ensure safe production operations in coal mines.

[0005] In order to achieve the above object, the present invention provides a rock burst monitoring and early warning method based on multi-source information fusion, comprising the following steps: Step 1: Collect multi-dimensional data of the target mining area, including terrain parameters, geological structure data and rock mass physical property information; Step 2: Process the multi-dimensional data and integrate it using the Kalman filter fusion algorithm to obtain fused data; Step 3: Based on the fused data, use the geological structure inversion method to generate a 3D geological structure model of the target mining area; Step 4: Based on the homogeneity of the rock mass physical property information and the continuity of the geological structure data in the 3D geological structure model, the 3D geological structure model is divided into N independent sub-regions; each independent sub-region includes background feature data, which includes terrain parameters, geological structure data, and rock mass physical property information; Step 5: Collect historical mining area rock burst event data, construct a rock burst risk prediction model based on BP neural network, and use particle swarm optimization to optimize the hyperparameters of the rock burst risk prediction model. The rock burst risk prediction model is trained based on the hyperparameters. Independent sub-regions are used as input data, and risk prediction is performed using the rock burst risk prediction model. The rock burst risk level of each region in the target mining area is output, and the rock burst risk level includes low, medium and high levels. Step 6: Classify warnings according to the rock burst risk level and generate warning reminder suggestions.

[0006] As a preferred embodiment, in step one, the terrain parameters include the surface elevation, slope, slope direction and surface undulation of the mining area; the geological structure data include the fault distribution position, fault strike, fault dip, fold morphology and distribution range; the rock physical property information includes rock density, elastic modulus, Poisson's ratio, uniaxial compressive strength and rock integrity coefficient.

[0007] As a preferred embodiment, in step 2, the process of obtaining fused data is as follows: S21: Eliminate outliers in multidimensional data using the Laida criterion; S22: Using Kalman filter fusion algorithm to integrate data and obtain fused data; S22-1: Establish the system's state equation according to formula (1), and establish the system's observation equation according to formula (2); (1); (2); Where, is the system state at time k, is the state transition matrix, which indicates that the system changes from The dynamic evolution relationship from time to time k, for The system status at the moment, for The system noise driving matrix at time , for System noise at the moment, is the observation vector at time k, is the observation matrix at time k, is the observation noise at time k; S22-2: State prediction; According to formula (3), the state prediction equation is established to The posterior estimate of the moment is used to estimate the state at moment k, and the prior estimate of the moment k is obtained. ; (3); Where, It represents the optimal estimate of the system state vector at time k-1 after combining the observation information at time k-1; S22-3: State update; According to formula (4), the state update equation is established to use the measurement value at time k to correct the estimated value in the prediction stage and obtain the posterior estimated value at time k. ; (4); Where, represents the Kalman gain; S22-4: Execute S22-2 and S22-3 in a loop, continue to fuse the data, and finally obtain fused data.

[0008] In this technical solution, the Laida criterion is used to eliminate outliers and the Kalman filter fusion algorithm is used to integrate data, which effectively reduces data noise and errors, ensures the accuracy of the fused data, and lays a solid data foundation for subsequent model construction and risk prediction.

[0009] As a preferred method, in step 3, the process of using the geological structure inversion method to generate a three-dimensional geological structure model of the target mining area is as follows: S31: Using the fused data as a constraint, the geological structure inversion method based on Bayesian theory is adopted to construct the inversion objective function according to formula (5): ; (5); Where m is the model parameter vector, d is the observation data vector, G(m) is the forward operator, is the initial model parameter vector, C is the covariance matrix of the model parameters; S32: Solving the inversion objective function using the Markov chain Monte Carlo method , obtain the posterior probability distribution of model parameters; S33: Based on the maximum likelihood estimation result of the posterior probability distribution, a three-dimensional grid model of the target mining area is constructed, and the model parameters are assigned to the corresponding grid cells to generate a three-dimensional geological structure model.

[0010] This technical solution employs a Bayesian-based geological structure inversion method, combined with fused data as constraints, and solves the inversion objective function using the Markov Chain Monte Carlo method. This method quantifies the uncertainty of model parameters and generates a posterior probability distribution of model parameters that more closely reflects the actual geological conditions. This technology significantly enhances the accuracy of the 3D geological model and enables the constructed 3D geological structure model to clearly represent the geological structure and rock mass physical properties of the mining area, providing a precise spatial framework for subsequent regional delineation.

[0011] As a preferred embodiment, in step 4, the process of dividing the three-dimensional geological structure model into regions is as follows: S41: Using the elastic modulus and uniaxial compressive strength in the rock mass physical property information as the division index, calculate the rock mass physical property difference coefficient between any two points in the three-dimensional geological structure model according to formula (6): ; (6); Where, and are the elastic moduli of the two points, and are the uniaxial compressive strength at two points, and are the maximum values ​​of elastic modulus and uniaxial compressive strength in the mining area, respectively; S42: Setting the difference coefficient threshold , when the difference coefficient of rock mass physical properties at two points When , the two points are judged to be in the same area and divided into the same sub-area; S43: combining the distribution of faults and folds in the geological structure data, dividing the areas separated by faults or with fold morphology differences exceeding a preset difference threshold into different sub-areas; S44: After the initial division is completed by the region growing algorithm, the boundaries are smoothed to obtain N independent sub-regions.

[0012] In this technical solution, regional division is carried out based on the uniformity of rock physical properties and the continuity of geological structure. Combined with the calculation of difference coefficient and geological structure distribution, the independent sub-regions obtained through regional growing algorithm and boundary smoothing are highly representative, ensuring the consistency of geological conditions in each region and providing clear objects for hierarchical monitoring.

[0013] Furthermore, in order to obtain a rock burst risk prediction model with high prediction accuracy, in step five, the historical mining area rock burst event data includes background characteristic data of the event area and specific attributes of the rock burst event; the background characteristic data of the event area includes the terrain parameters, geological structure data and rock physical property information of the area; the specific attributes of the rock burst event include the spatiotemporal information of the event, the impact intensity and the rock burst risk probability value corresponding to the event, and the spatiotemporal information of the event includes the time and location coordinates of the rock burst occurrence.

[0014] As a preferred embodiment, in step five, the process of constructing a rock burst risk prediction model based on a BP neural network is as follows: S151: Constructing a BP neural network basic model, which includes an input layer, multiple hidden layers, and an output layer; the input layer is used to receive background feature data of the independent sub-region; the multiple hidden layers are used to perform nonlinear transformation and feature mapping on the data received by the input layer; and the output layer is used to output the rock burst risk probability value of the independent sub-region; S152: Take the regional characteristic data corresponding to the event occurrence area in the historical mining area rock burst event data as input, take the rock burst risk probability value corresponding to the event as output, train the BP neural network basic model, and save the model parameters that meet the preset accuracy. After training, a rock burst risk prediction model based on the BP neural network is obtained.

[0015] In this technical solution, a risk prediction model based on BP neural network is constructed and trained using historical rock burst event data. This enables the model to effectively capture the complex nonlinear relationship of rock burst and ensure prediction accuracy.

[0016] As a preferred method, in step 5, the process of using the particle swarm optimization algorithm to optimize the hyperparameters of the rock burst risk prediction model is as follows: S251: Hyperparameter combination of learning rate and number of hidden layer neurons based on rock burst risk prediction model; S252: Randomly generate a set of particles, where the position of each particle represents a set of hyperparameters; S253: Use validation set data to evaluate model prediction accuracy as fitness value , as shown in formula (7); (7); Where, To predict the correct number of samples, is the total sample size; S254: According to the fitness value Update the individual optimal position of each particle and the global optimal position of the entire particle swarm, and obtain the updated particle velocity according to formula (8): , according to formula (9) to obtain the updated particle position ; (8); (9); Where, is the velocity of particle i at the kth iteration, w is the inertia weight, and are two different learning factors. and are two random numbers in the interval [0, 1]. is the individual optimal position of particle i up to the kth iteration, is the global optimal position of the entire particle swarm up to the kth iteration, is the position of particle i at the kth iteration; S255: Repeatedly iteratively randomly generate a group of particles and update the individual optimal position of each particle and the global optimal position of the entire particle swarm until the preset number of iterations is reached and the optimal hyperparameter combination is obtained.

[0017] In this technical solution, the particle swarm algorithm is used to optimize the model hyperparameters, avoiding the subjectivity of manual experience selection, so that the model can maintain high prediction accuracy under different geological conditions, accurately output the risk level of each area, and provide a scientific basis for risk prevention and control.

[0018] As a preferred embodiment, in step 6, the process of grading the early warning according to the rock burst risk level and generating early warning reminder suggestions is as follows: When the rock burst risk level is low, a first-level warning is issued, and the alarm is controlled to perform the first-level warning action. At the same time, a reminder message is generated for periodic sampling monitoring of low-risk areas, and displayed in real time on the display; For situations where the rock burst risk level is medium, a second-level warning is issued, and the alarm is controlled to perform the second-level warning action. At the same time, a reminder message is generated to adopt fixed-point continuous monitoring and dynamically adjust the monitoring frequency for the medium-risk area, and the message is displayed in real time on the display. For situations where the impact ground pressure risk level is high, a third-level warning is issued and the alarm is controlled to perform the third-level warning action. At the same time, a reminder message is generated to deploy a full-dimensional three-dimensional monitoring network in the high-risk area and to link the early warning response mechanism, and the message is displayed in real time on the display.

[0019] This technical solution, through the automatic generation of early warning grading and reminder suggestions, not only can relevant personnel be promptly and effectively reminded when an anomaly occurs, but decision-making information can also be intelligently generated for their reference, significantly shortening response time. Furthermore, implementing a graded monitoring strategy based on different risk levels effectively ensures monitoring effectiveness through differentiated monitoring strategies. This also avoids the waste of monitoring resources, allowing for the timely identification of potential risks and the implementation of appropriate measures, significantly reducing the probability of rock bursts and ensuring safe production in the mining area.

[0020] The present invention provides a rock burst monitoring and early warning method based on multi-source information fusion. First, by collecting multi-dimensional information such as terrain parameters, geological structure data and rock physical properties, it breaks through the limitations of traditional single data source and can fully reflect the geological conditions of the mining area. Secondly, by pre-processing the data and fusing the data using the Kalman filter fusion algorithm, the data noise and error can be effectively reduced, and the data reliability is improved by multi-dimensional data fusion. At the same time, the accuracy of the fused data is ensured, laying a solid data foundation for the subsequent accurate construction of the model and accurate prediction of the risk level. Then, with the accurate fused data as the constraint condition, the geological structure inversion method is used to generate a three-dimensional geological structure model of the target mining area, which can clearly and accurately present the geological structure and rock physical property distribution of the mining area, and can fully reflect the geological conditions and stress distribution state of the entire mining area, providing an accurate spatial framework for regional division. Thirdly, based on the uniformity of rock physical properties and the continuity of geological structure, The regional division results in independent sub-regions with strong representativeness, ensuring the consistency of geological conditions in each region and providing clear objects for graded monitoring. This is conducive to reducing the complexity of the subsequent rock burst risk prediction model and saving computing resources in the risk prediction process. Then, by constructing a rock burst risk prediction model based on BP neural network and optimizing the model hyperparameters using particle swarm optimization, the subjectivity of manual experience selection is avoided, so that the model can maintain high prediction accuracy under different geological conditions. The optimized BP neural network is used to predict the risk of the divided independent sub-regions, which not only significantly improves the risk prediction accuracy, but also can simultaneously present the risk situation of each independent sub-region, so that relevant personnel can understand the situation of each independent sub-region simultaneously, and then can simultaneously grasp the overall and local situation of the target mining area, which is conducive to relevant personnel to take a comprehensive consideration when taking disposal measures, and obtain more scientific, reasonable and effective disposal measures. At the same time, it can better optimize resource allocation. Finally, the early warning classification is carried out according to different risk levels, which can timely and effectively remind relevant personnel to take graded treatment measures when an abnormality occurs, providing a scientific basis for risk prevention and control.

[0021] This method has a simple implementation process, low implementation cost and high intelligence. It improves data reliability and enhances the accuracy of three-dimensional geological models by integrating multi-dimensional data. At the same time, it uses the optimized BP neural network for prediction, which significantly improves the risk level prediction ability and can efficiently and accurately output the risk level of each area. Combined with graded early warning, it can provide a reliable decision-making basis for rock burst risk prevention and control. This method can achieve efficient and accurate monitoring of rock burst, effectively reduce the probability of rock burst, and ensure safe production operations in mining areas.

[0022] The present invention also provides a rock burst monitoring and early warning method system based on multi-source information fusion, which is used to implement a rock burst monitoring and early warning method based on multi-source information fusion, including a processor, an alarm, a display and a memory; The processor includes a data acquisition module, a data fusion module, a model building module, a region division module, a risk prediction module, and a graded warning module; The data acquisition module is used to collect multi-dimensional data of the target mining area, the multi-dimensional data including terrain parameters, geological structure data and rock mass physical property information; The data fusion module is used to remove outliers from multi-dimensional data and integrate the multi-dimensional data using a Kalman filter fusion algorithm to obtain fused data. The model building module is used to generate a three-dimensional geological structure model of the target mining area using a geological structure inversion method based on the fused data; The region division module is used to divide the three-dimensional geological structure model into regions based on the homogeneity of the rock mass physical property information and the continuity of the geological structure data in the three-dimensional geological structure model to obtain N independent sub-regions; The risk prediction module is used to use the background feature data of the independent sub-region as input data of the rock burst risk prediction model based on the BP neural network to predict the risk and output the rock burst risk level of each area in the target mining area; The graded warning module is used to classify warnings according to the rock burst risk level and generate warning reminder suggestions; The alarm is used to perform graded warning actions according to the warning classification results; The display is used to display early warning reminder suggestions in real time; The memory is used for storing and reading data by the processor.

[0023] In the present invention, the data acquisition module provides a data receiving channel for the data fusion module, allowing the received multi-dimensional data to be directly input into the data fusion module. The data fusion module automatically performs data denoising and fusion processing, obtaining accurate fused data and providing a reliable data foundation for the subsequent accurate construction of a three-dimensional geological structure model. The model construction module automatically constructs a three-dimensional geological structure model that clearly and accurately represents the geological structure and rock mass physical property distribution of the mining area. The regional division module efficiently divides the target mining area into multiple, highly representative, independent sub-regions, thereby providing clear targets for subsequent risk prediction, reducing the complexity of the risk prediction module, and conserving computing power. The risk prediction module significantly improves the accuracy and efficiency of risk prediction. The hierarchical early warning module allows for the use of a hierarchical early warning system to alert relevant personnel to take targeted measures, thereby shortening response time. The alarm provides timely and effective alerts to relevant personnel when an anomaly occurs. The display allows relevant personnel to easily view early warning information. By setting the memory, historical data can be stored and read in real time.

[0024] The system has a simple structure, high intelligence and reliable performance. It can realize accurate monitoring and intelligent early warning of rock burst, and can effectively ensure safe production operations in coal mines. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a flow chart of the monitoring method of the present invention; Figure 2 It is a flow chart of generating a three-dimensional geological structure model in the present invention; Figure 3 It is a principle block diagram of the monitoring system in the present invention. DETAILED DESCRIPTION

[0026] The present invention will be further described below with reference to the accompanying drawings.

[0027] like Figure 1 As shown, the present invention provides a rock burst monitoring and early warning method based on multi-source information fusion, comprising the following steps: Step 1: Collect multi-dimensional data of the target mining area, including terrain parameters, geological structure data and rock mass physical property information; As a preferred embodiment, the terrain parameters include the surface elevation, slope, slope direction and surface undulation of the mining area; the geological structure data include the fault distribution location, fault direction, fault dip, fold morphology and distribution range; the rock physical property information includes rock density, elastic modulus, Poisson's ratio, uniaxial compressive strength and rock integrity coefficient.

[0028] As an example, terrain parameters are collected using a combination of drone aerial surveys and ground-based total stations. The drones, equipped with high-precision optical cameras and lidar, capture multi-angle, full-coverage aerial images of the mining area along a pre-set route. Photogrammetry technology generates dense point cloud data, which is then processed to extract the mining area's surface elevation, slope, aspect, and surface relief. For complex terrain areas difficult to cover with drones, manual supplementary surveys are performed using total stations to ensure the integrity of terrain parameters.

[0029] Topographic parameters are the basis for constructing the three-dimensional topographic framework of the mining area. They can provide a reference for the spatial positioning of the subsequent geological structure model. At the same time, they affect the stress distribution and rock stability in the mining area and are an important basis for analyzing the environment in which rock burst occurs.

[0030] Geological structural data is collected through a combination of geological mapping, geophysical exploration, and drilling. Through field mapping, geologists record the location of faults (marked with longitude and latitude coordinates), fault strike (the direction of the intersection of the fault line with the horizontal plane), fault dip (the downward angle of the fault plane), fold morphology (such as anticlines and synclines), and distribution range. Geophysical exploration methods such as seismic exploration and geological radar are used to detect hidden underground geological structures and determine their extension. Drilling and sampling are conducted in key structural areas to obtain cores to verify and refine geological structural features. Geological structure is a key factor influencing the occurrence of rock bursts. Structures such as faults and folds can cause stress concentrations, and their data can provide direct clues for identifying high-risk areas and are the core content of the structural framework in constructing three-dimensional geological structural models.

[0031] Information on the physical properties of the rock mass is collected through a combination of laboratory and field testing. Rock samples are collected from various locations within the mining area and sent to the laboratory for testing to determine rock density, elastic modulus, Poisson's ratio, and uniaxial compressive strength. On-site testing of the rock mass is performed using an acoustic wave tester, and the rock integrity coefficient is calculated based on the propagation velocity of the acoustic wave.

[0032] Step 2: Process the multi-dimensional data and integrate it using the Kalman filter fusion algorithm to obtain fused data; As a preferred method, the process of obtaining fused data is as follows: S21: Eliminate outliers in multidimensional data using the Laida criterion; S22: Using Kalman filter fusion algorithm to integrate data and obtain fused data; S22-1: Establish the system's state equation according to formula (1), and establish the system's observation equation according to formula (2); (1); (2); Where, is the system state at time k, is the state transition matrix, which indicates that the system changes from The dynamic evolution relationship from time to time k, for The system status at the moment, for The system noise driving matrix at time , for System noise at the moment, is the observation vector at time k, is the observation matrix at time k, is the observation noise at time k; S22-2: State prediction; According to formula (3), the state prediction equation is established to The posterior estimate of the moment is used to estimate the state at moment k, and the prior estimate of the moment k is obtained. ; (3); Where, It represents the predicted value of the system state vector at time k based on all observation information at time k-1 and before. It represents the optimal estimate of the system state vector at time k-1 after combining the observation information at time k-1; S22-3: State update: According to formula (4), the state update equation is established to use the measured value at time k to correct the estimated value in the prediction stage and obtain the posterior estimated value at time k; (4); Where, represents the Kalman gain, Indicates the integration of k-time observation vector After that, the optimal estimate of the system state vector at time k; S22-4: Execute S22-2 and S22-3 in a loop, continue to fuse the data, and finally obtain fused data.

[0033] By using the Laida criterion to eliminate outliers and the Kalman filter fusion algorithm to integrate data, data noise and errors are effectively reduced, the accuracy of the fused data is ensured, and a solid data foundation is laid for subsequent model construction and risk prediction.

[0034] Step 3: Based on the fused data, use the geological structure inversion method to generate a 3D geological structure model of the target mining area; like Figure 2 As shown, as a preferred method, the process of using the geological structure inversion method to generate a three-dimensional geological structure model of the target mining area is as follows: S31: Using the fused data as a constraint, the geological structure inversion method based on Bayesian theory is adopted to construct the inversion objective function according to formula (5): ; (5); Where m is the model parameter vector, d is the observation data vector, G(m) is the forward operator, is the initial model parameter vector, C is the covariance matrix of the model parameters; S32: Solving the inversion objective function using the Markov chain Monte Carlo method , obtain the posterior probability distribution of model parameters; Determine the initial set of model parameters, which are based on existing geological knowledge, regional geological data, and previous exploration data, and serve as the starting point of the Markov chain. Generate new candidate model parameter values ​​according to certain probabilistic rules, taking into account the distribution of the current model parameters and the characteristics of the inversion objective function to ensure that the generated candidate parameters are within a reasonable range.

[0035] The inversion objective function value corresponding to the new candidate parameter is calculated and compared with the objective function value of the current model parameter. If the objective function value of the new candidate parameter is smaller, it means that it better meets the constraints of the fused data, and the candidate parameter is directly accepted as the current model parameter. If the objective function value of the new candidate parameter is larger, it is not directly discarded. Instead, a probability is used to determine whether to accept it. This probability is related to the difference between the objective function values ​​of the two parameters. The smaller the difference, the greater the probability of acceptance. This avoids falling into a local optimal solution and ensures the comprehensiveness of the search.

[0036] The process of generating candidate parameters, calculating the objective function value, and then comparing and judging whether to accept them is repeated repeatedly, forming a Markov chain. During this iterative process, the distribution of model parameters is continuously monitored. When the parameter distribution stabilizes, that is, when the statistical characteristics of the parameters (such as mean and variance) no longer change significantly after multiple iterations, the Markov chain is considered to have reached a state of convergence. At this point, all accepted model parameters during the iterative process are collected to form the posterior probability distribution of the model parameters, which reflects the possible values ​​and probability of each model parameter under the constraints of the fused data.

[0037] S33: Based on the maximum likelihood estimation result of the posterior probability distribution, a three-dimensional grid model of the target mining area is constructed, and the model parameters are assigned to the corresponding grid cells to generate a three-dimensional geological structure model.

[0038] From the posterior probability distribution of the model parameters, find the model parameter combination with the highest probability of occurrence, that is, the maximum likelihood estimation result, which represents the model parameters that are most likely to be close to the actual geological conditions under the current fusion data constraints.

[0039] The size and density of the 3D grid are determined based on the target mining area's topographical scope, geological complexity, and the accuracy requirements for subsequent analysis. For areas with complex geological structures and high interest, the grid is denser to ensure detailed representation of the model; for areas with relatively simple geological conditions, the grid can be appropriately sparse to improve computational efficiency.

[0040] A three-dimensional grid is constructed to cover the entire target mining area, with each grid cell corresponding to a specific spatial location within the mining area. Model parameters derived from maximum likelihood estimation are assigned to the spatial locations of the grid cells. Each grid cell is then assigned geological parameters corresponding to its location, such as the lithology, thickness, porosity, and permeability of the rock formation.

[0041] Professional 3D modeling software integrates and visualizes the grid cells with parameters. Based on the spatial coordinates and parameter information of the grid cells, a three-dimensional model with a sense of depth is constructed, showing the distribution of different rock layers within the mining area, the morphology of geological structures, and the spatial variations of rock mass physical properties, thus obtaining a 3D geological structure model of the target mining area.

[0042] By employing a Bayesian-based geological structure inversion method, combined with fused data as constraints and solving the inversion objective function using the Markov Chain Monte Carlo method, we can quantify the uncertainty of model parameters and obtain a posterior probability distribution of model parameters that more closely reflects the actual geological conditions. This technology significantly enhances the accuracy of the 3D geological model and enables the constructed 3D geological structure model to clearly represent the geological structure and rock mass physical characteristics of the mining area, providing a precise spatial framework for subsequent regional delineation.

[0043] Step 4: Based on the homogeneity of the rock mass physical property information and the continuity of the geological structure data in the 3D geological structure model, the 3D geological structure model is divided into N independent sub-regions; each independent sub-region includes background feature data, which includes terrain parameters, geological structure data, and rock mass physical property information; As a preferred method, the process of dividing the three-dimensional geological structure model into regions is as follows: S41: Using the elastic modulus and uniaxial compressive strength in the rock mass physical property information as the division index, calculate the rock mass physical property difference coefficient between any two points in the three-dimensional geological structure model according to formula (6): ; (6); Where, and are the elastic moduli of the two points, and are the uniaxial compressive strength at two points, and are the maximum values ​​of elastic modulus and uniaxial compressive strength in the mining area, respectively; S42: Setting the difference coefficient threshold , when the difference coefficient of rock mass physical properties at two points When , the two points are judged to be in the same area and divided into the same sub-area; S43: combining the distribution of faults and folds in the geological structure data, dividing the areas separated by faults or with fold morphology differences exceeding a preset difference threshold into different sub-areas; The spatial distribution of faults, including their strike, dip, inclination, and extent, is extracted from the 3D geological structure model. 3D modeling software is used to visualize the faults, clarifying their specific location and morphology within the model. For areas completely separated by faults, the presence of faults disrupts the continuity of the rock mass, leading to significant differences in stress state and geological evolution between the two sides of the fault. Therefore, these areas separated by faults are directly divided into different preliminary sub-regions.

[0044] To analyze differences in fold morphology, we first identify indicators to measure fold morphological differences, such as fold axial inclination, hinge strike, and wing inclination. We then calculate the index values ​​for folds in different regions and compare them to a preset difference threshold. When the difference in index values ​​between two adjacent regions exceeds the preset threshold, it indicates significant differences in their fold morphology, reflecting poor tectonic continuity. These two regions are then divided into different preliminary subregions.

[0045] S44: After the initial division is completed by the region growing algorithm, the boundaries are smoothed to obtain N independent sub-regions.

[0046] In the initial partitioning of the region growing algorithm, a point that meets the rock mass physical property difference coefficient requirement and is not separated by geological structures is used as a seed point. Starting from this seed point, adjacent points that meet the requirements are continuously incorporated into the current region until there are no more points to be incorporated, forming multiple initial regions. However, the resulting region boundaries may be rough, with jagged or irregular protrusions and depressions.

[0047] Boundary smoothing is performed using a moving average method. This involves selecting a series of points on the boundary, taking each point as the center, and taking a certain number of neighboring boundary points around it. The coordinates of these neighboring points are averaged, and this average is used as the smoothed coordinates of the point. The boundary is adjusted point by point, smoothing out previously abrupt edges. Furthermore, for some overly sharp corners on the boundary, a curve fitting method is used to fit a smooth curve based on the boundary points near the corners, replacing the original corner segments and making the boundary transition more natural.

[0048] During the smoothing process, we continuously check whether the smoothed boundaries still accurately reflect the actual geological structure, avoiding over-smoothing that may obscure important geological boundary features. After multiple iterations of adjustment, the boundaries are smooth and consistent with geological reality, ultimately obtaining N independent sub-regions with clear boundaries and regular morphology.

[0049] Regional division is carried out based on the homogeneity of rock mass physical properties and the continuity of geological structure. Combined with the calculation of difference coefficient and geological structure distribution, the independent sub-regions obtained through regional growing algorithm and boundary smoothing are highly representative, ensuring the consistency of geological conditions in each region and providing clear objects for hierarchical monitoring.

[0050] Step 5: Collect historical mining area rock burst event data, construct a rock burst risk prediction model based on BP neural network, and use particle swarm optimization to optimize the hyperparameters of the rock burst risk prediction model. The rock burst risk prediction model is trained based on the hyperparameters. Independent sub-regions are used as input data, and risk prediction is performed using the rock burst risk prediction model. The rock burst risk level of each region in the target mining area is output, and the rock burst risk level includes low, medium and high levels. In order to obtain a rock burst risk prediction model with high prediction accuracy, the historical mining area rock burst event data includes background characteristic data of the event area and specific attributes of the rock burst event; the background characteristic data of the event area includes the regional terrain parameters, geological structure data and rock physical property information; the specific attributes of the rock burst event include the spatiotemporal information of the event, the impact intensity and the rock burst risk probability value corresponding to the event, and the spatiotemporal information of the event includes the time and location coordinates of the rock burst occurrence.

[0051] As a preferred method, the process of constructing a rock burst risk prediction model based on BP neural network is as follows: S151: Construct a BP neural network basic model, which includes an input layer, multiple hidden layers, and an output layer. The input layer is used to receive background feature data of independent sub-regions. The multiple hidden layers are used to perform nonlinear transformation and feature mapping on the data received by the input layer. The hidden layers use a Sigmoid activation function to enhance the model's ability to fit nonlinear relationships. The output layer is used to output the rock burst risk probability value of the independent sub-region. The output layer has three nodes, corresponding to low, medium, and high rock burst risk levels, respectively. S152: Take the regional characteristic data corresponding to the event occurrence area in the historical mining area rock burst event data as input, take the rock burst risk probability value corresponding to the event as output, train the BP neural network basic model, and save the model parameters that meet the preset accuracy. After training, a rock burst risk prediction model based on the BP neural network is obtained.

[0052] Constructing a risk prediction model based on BP neural network and using historical rock burst event data for training can enable the model to effectively capture the complex nonlinear relationship of rock burst and ensure prediction accuracy.

[0053] As a preferred method, the process of using particle swarm optimization to optimize the hyperparameters of the rock burst risk prediction model is as follows: S251: Hyperparameter combination of learning rate and number of hidden layer neurons based on rock burst risk prediction model; S252: Randomly generate a set of particles, where the position of each particle represents a set of hyperparameters; S253: Use validation set data to evaluate model prediction accuracy as fitness value , as shown in formula (7); (7); Where, To predict the correct number of samples, is the total sample size; S254: According to the fitness value Update the individual optimal position of each particle and the global optimal position of the entire particle swarm, and obtain the updated particle velocity according to formula (8): , according to formula (9) to obtain the updated particle position ; (8); (9); Where, is the velocity of particle i at the kth iteration, w is the inertia weight, and are two different learning factors. and are two random numbers in the interval [0, 1]. is the individual optimal position of particle i up to the kth iteration, is the global optimal position of the entire particle swarm up to the kth iteration, is the position of particle i at the kth iteration; S255: Repeatedly iteratively randomly generate a group of particles and update the individual optimal position of each particle and the global optimal position of the entire particle swarm until the preset number of iterations is reached and the optimal hyperparameter combination is obtained.

[0054] The particle swarm algorithm is used to optimize the model hyperparameters, avoiding the subjectivity of manual experience selection, so that the model can maintain high prediction accuracy under different geological conditions, accurately output the risk level of each area, and provide a scientific basis for risk prevention and control.

[0055] Step 6: Classify warnings according to the rock burst risk level and generate warning reminder suggestions.

[0056] As a preferred embodiment, the process of grading warnings according to the rock burst risk level and generating warning reminder suggestions is as follows: for situations where the rock burst risk level is low, a first-level warning is performed, and the alarm is controlled to perform the first-level warning action. At the same time, a reminder message is generated for periodic sampling monitoring of low-level risk areas, and displayed in real time on the display; for situations where the rock burst risk level is medium, a second-level warning is performed, and the alarm is controlled to perform the second-level warning action. At the same time, a reminder message is generated for adopting fixed-point continuous monitoring and dynamically adjusting the monitoring frequency for medium-level risk areas, and displayed in real time on the display; for situations where the rock burst risk level is high, a third-level warning is performed, and the alarm is controlled to perform the third-level warning action. At the same time, a reminder message is generated for deploying a full-dimensional three-dimensional monitoring network for high-risk areas and linking the warning response mechanism, and displayed in real time on the display.

[0057] Low-risk areas indicate areas with no historical record of rock burst events, and according to rock burst risk prediction models, the probability of future rock burst events is extremely low. Furthermore, the rock mass in the area has stable physical properties, simple geological structures, and no significant stress concentration, so mining activities will have minimal impact on the rock mass.

[0058] Medium-risk areas indicate areas where rock burst events have not occurred historically, but risk prediction models indicate a certain probability of future occurrence. The area may have localized differences in rock mass physical properties or complex geological structures, such as small folds without distinct stress concentration zones. Changes in mining practices could lead to potential rock burst risks.

[0059] High-risk areas indicate areas with a history of rock burst events or where risk prediction models assess a high probability of future rock bursts. The rock mass within the region exhibits significant variations in physical properties, complex geological structures, and the presence of significant faults and large folds, which are prone to stress concentration. Mining activities significantly disturb the rock mass, making rock bursts highly likely.

[0060] Periodic sampling monitoring of low-risk areas will formulate a distribution plan for sampling monitoring points based on the scope and terrain characteristics of the low-risk areas to ensure that the monitoring points can evenly cover the entire area.

[0061] Fixed-point continuous monitoring and dynamic frequency adjustment in medium-risk areas. In medium-risk areas, fixed monitoring points are selected at key locations based on geological structure and mining plan. Stress sensors, microseismic sensors and other equipment are installed at each monitoring point to achieve continuous monitoring of parameters such as rock stress and microseismic activity.

[0062] A comprehensive, three-dimensional monitoring network and a coordinated early warning response mechanism are being established for high-risk areas. Multiple monitoring base stations are being installed on the surface, along with high-precision GPS displacement monitoring systems and meteorological monitoring equipment, to monitor surface displacement changes and the impact of meteorological conditions on the rock mass in real time. Stress sensors, displacement sensors, microseismic sensors, and infrared thermometers are densely deployed within the underground mining face and surrounding rock mass to enable comprehensive monitoring of rock mass parameters such as stress, displacement, microseismic activity, and temperature.

[0063] The automatic generation of early warning grading and reminder suggestions not only provides timely and effective reminders to relevant personnel when an anomaly occurs, but also intelligently generates decision-making information for their reference, significantly shortening response time. Furthermore, implementing a graded monitoring strategy based on different risk levels effectively ensures monitoring effectiveness through differentiated monitoring strategies. This also avoids the waste of monitoring resources, allowing for the timely identification of potential risks and the implementation of appropriate measures, significantly reducing the probability of rock bursts and ensuring safe production in the mining area.

[0064] The present invention provides a rock burst monitoring and early warning method based on multi-source information fusion. First, by collecting multi-dimensional information such as terrain parameters, geological structure data and rock physical properties, it breaks through the limitations of traditional single data source and can fully reflect the geological conditions of the mining area. Secondly, by pre-processing the data and fusing the data using the Kalman filter fusion algorithm, the data noise and error can be effectively reduced, and the data reliability is improved by multi-dimensional data fusion. At the same time, the accuracy of the fused data is ensured, laying a solid data foundation for the subsequent accurate construction of the model and accurate prediction of the risk level. Then, with the accurate fused data as the constraint condition, the geological structure inversion method is used to generate a three-dimensional geological structure model of the target mining area, which can clearly and accurately present the geological structure and rock physical property distribution of the mining area, and can fully reflect the geological conditions and stress distribution state of the entire mining area, providing an accurate spatial framework for regional division. Thirdly, based on the uniformity of rock physical properties and the continuity of geological structure, The regional division results in independent sub-regions with strong representativeness, ensuring the consistency of geological conditions in each region and providing clear objects for graded monitoring. This is conducive to reducing the complexity of the subsequent rock burst risk prediction model and saving computing resources in the risk prediction process. Then, by constructing a rock burst risk prediction model based on BP neural network and optimizing the model hyperparameters using particle swarm optimization, the subjectivity of manual experience selection is avoided, so that the model can maintain high prediction accuracy under different geological conditions. The optimized BP neural network is used to predict the risk of the divided independent sub-regions, which not only significantly improves the risk prediction accuracy, but also can simultaneously present the risk situation of each independent sub-region, so that relevant personnel can understand the situation of each independent sub-region simultaneously, and then can simultaneously grasp the overall and local situation of the target mining area, which is conducive to relevant personnel to take a comprehensive consideration when taking disposal measures, and obtain more scientific, reasonable and effective disposal measures. At the same time, it can better optimize resource allocation. Finally, the early warning classification is carried out according to different risk levels, which can timely and effectively remind relevant personnel to take graded treatment measures when an abnormality occurs, providing a scientific basis for risk prevention and control.

[0065] This method has a simple implementation process, low implementation cost and high intelligence. It improves data reliability and enhances the accuracy of three-dimensional geological models by integrating multi-dimensional data. At the same time, it uses the optimized BP neural network for prediction, which significantly improves the risk level prediction ability and can efficiently and accurately output the risk level of each area. Combined with graded early warning, it can provide a reliable decision-making basis for rock burst risk prevention and control. This method can achieve efficient and accurate monitoring of rock burst, effectively reduce the probability of rock burst, and ensure safe production operations in mining areas.

[0066] like Figure 3 As shown, based on the same inventive concept, the present invention also provides a rock burst monitoring and early warning method system based on multi-source information fusion, which is used to implement a rock burst monitoring and early warning method based on multi-source information fusion, including a processor, an alarm, a display and a memory; The processor includes a data acquisition module, a data fusion module, a model building module, a region division module, a risk prediction module, and a graded warning module; The data acquisition module is used to collect multi-dimensional data of the target mining area, the multi-dimensional data including terrain parameters, geological structure data and rock mass physical property information; The data fusion module is used to remove outliers from multi-dimensional data and integrate the multi-dimensional data using a Kalman filter fusion algorithm to obtain fused data. The model building module is used to generate a three-dimensional geological structure model of the target mining area using a geological structure inversion method based on the fused data; The region division module is used to divide the three-dimensional geological structure model into regions based on the homogeneity of the rock mass physical property information and the continuity of the geological structure data in the three-dimensional geological structure model to obtain N independent sub-regions; The risk prediction module is used to use the background feature data of the independent sub-region as input data of the rock burst risk prediction model based on the BP neural network to predict the risk and output the rock burst risk level of each area in the target mining area; The graded warning module is used to classify warnings according to the rock burst risk level and generate warning reminder suggestions; The alarm is used to perform graded warning actions according to the warning classification results; The display is used to display early warning reminder suggestions in real time; The memory is used for storing and reading data by the processor.

[0067] Preferably, the processor is an industrial computer. Preferably, the alarm is an audible and visual alarm that can broadcast preset voice messages. The preset voice message for the first-level warning action is: implement periodic sampling monitoring of low-risk areas; the preset voice message for the second-level warning action is: adopt fixed-point continuous monitoring and dynamically adjust the monitoring frequency for medium-risk areas; and the preset voice message for the third-level warning action is: deploy a full-scale, three-dimensional monitoring network for high-risk areas and link the early warning response mechanism.

[0068] In the present invention, the data acquisition module provides a data receiving channel for the data fusion module, allowing the received multi-dimensional data to be directly input into the data fusion module. The data fusion module automatically performs data denoising and fusion processing, obtaining accurate fused data and providing a reliable data foundation for the subsequent accurate construction of a three-dimensional geological structure model. The model construction module automatically constructs a three-dimensional geological structure model that clearly and accurately represents the geological structure and rock mass physical property distribution of the mining area. The regional division module efficiently divides the target mining area into multiple, highly representative, independent sub-regions, thereby providing clear targets for subsequent risk prediction, reducing the complexity of the risk prediction module, and conserving computing power. The risk prediction module significantly improves the accuracy and efficiency of risk prediction. The hierarchical early warning module allows for the use of a hierarchical early warning system to alert relevant personnel to take targeted measures, thereby shortening response time. The alarm provides timely and effective alerts to relevant personnel when an anomaly occurs. The display allows relevant personnel to easily view early warning information. By setting the memory, historical data can be stored and read in real time.

[0069] The system has a simple structure, high intelligence and reliable performance. It can realize accurate monitoring and intelligent early warning of rock burst, and can effectively ensure safe production operations in coal mines.

Claims

1. A rock burst monitoring and early warning method based on multi-source information fusion, characterized in that: The following steps are involved: Step 1: Collect multi-dimensional data of the target mining area, including terrain parameters, geological structure data and rock mass physical property information; Step 2: Process the multi-dimensional data and integrate it using the Kalman filter fusion algorithm to obtain fused data; Step 3: Based on the fused data, use the geological structure inversion method to generate a 3D geological structure model of the target mining area; Step 4: Based on the homogeneity of the rock mass physical property information and the continuity of the geological structure data in the 3D geological structure model, the 3D geological structure model is divided into N independent sub-regions; each independent sub-region includes background feature data, which includes terrain parameters, geological structure data, and rock mass physical property information; Step 5: Collect historical mining area rock burst event data, construct a rock burst risk prediction model based on BP neural network, and use particle swarm optimization to optimize the hyperparameters of the rock burst risk prediction model. The rock burst risk prediction model is trained based on the hyperparameters. Independent sub-regions are used as input data, and risk prediction is performed using the rock burst risk prediction model. The rock burst risk level of each region in the target mining area is output, and the rock burst risk level includes low, medium and high levels. Step 6: Classify warnings according to the rock burst risk level and generate warning reminder suggestions.

2. The rock burst monitoring and early warning method based on multi-source information fusion according to claim 1 is characterized in that: In step one, the terrain parameters include the surface elevation, slope, slope direction and surface undulation of the mining area; the geological structure data include the fault distribution location, fault direction, fault dip, fold morphology and distribution range; the rock physical property information includes rock density, elastic modulus, Poisson's ratio, uniaxial compressive strength and rock integrity coefficient.

3. The method for monitoring and early warning of rock burst based on multi-source information fusion according to claim 2, characterized in that: In step 2, the process of obtaining fused data is as follows: S21: Eliminate outliers in multidimensional data using the Laida criterion; S22: Using Kalman filter fusion algorithm to integrate data and obtain fused data; S22-1: Establish the system's state equation according to formula (1), and establish the system's observation equation according to formula (2); （1）； （2）； Where, is the system state at time k, is the state transition matrix, which indicates that the system changes from The dynamic evolution relationship from time to time k, for The system status at the moment, for The system noise driving matrix at time , for System noise at the moment, is the observation vector at time k, is the observation matrix at time k, is the observation noise at time k; S22-2: Status prediction; According to formula (3), the state prediction equation is established to The posterior estimate of the moment is used to estimate the state at moment k, and the prior estimate of the moment k is obtained. ; （3）； Where, It represents the optimal estimate of the system state vector at time k-1 after combining the observation information at time k-1; S22-3: State update; According to formula (4), the state update equation is established to use the measurement value at time k to correct the estimated value in the prediction stage and obtain the posterior estimated value at time k. ; （4）； Where, represents the Kalman gain; S22-4: Execute S22-2 and S22-3 in a loop, continue to fuse the data, and finally obtain fused data.

4. The method for monitoring and early warning of rock burst based on multi-source information fusion according to claim 3 is characterized in that: In step 3, the process of using the geological structure inversion method to generate a 3D geological structure model of the target mining area is as follows: S31: Using the fused data as a constraint, the geological structure inversion method based on Bayesian theory is adopted to construct the inversion objective function according to formula (5): ; （5）； Where m is the model parameter vector, d is the observation data vector, G(m) is the forward operator, is the initial model parameter vector, C is the covariance matrix of the model parameters; S32: Solving the inversion objective function using the Markov chain Monte Carlo method , obtain the posterior probability distribution of model parameters; S33: Based on the maximum likelihood estimation result of the posterior probability distribution, a three-dimensional grid model of the target mining area is constructed, and the model parameters are assigned to the corresponding grid cells to generate a three-dimensional geological structure model.

5. The method for monitoring and early warning of rock burst based on multi-source information fusion according to claim 4 is characterized in that: In step 4, the process of regional division of the 3D geological structure model is as follows: S41: Using the elastic modulus and uniaxial compressive strength in the rock mass physical property information as the division index, calculate the rock mass physical property difference coefficient between any two points in the three-dimensional geological structure model according to formula (6): ; （6）； Where, and are the elastic moduli of the two points, and are the uniaxial compressive strength at two points, and are the maximum values ​​of elastic modulus and uniaxial compressive strength in the mining area, respectively; S42: Setting the difference coefficient threshold , when the difference coefficient of rock mass physical properties at two points When , the two points are judged to be in the same area and divided into the same sub-area; S43: combining the distribution of faults and folds in the geological structure data, dividing the areas separated by faults or with fold morphology differences exceeding a preset difference threshold into different sub-areas; S44: After the initial division is completed by the region growing algorithm, the boundaries are smoothed to obtain N independent sub-regions.

6. The rock burst monitoring and early warning method based on multi-source information fusion according to claim 5 is characterized in that: In step five, the historical mining area rock burst event data includes background characteristic data of the event area and specific attributes of the rock burst event; the background characteristic data of the event area includes the terrain parameters, geological structure data and rock physical property information of the area; the specific attributes of the rock burst event include the spatiotemporal information of the event, the impact intensity and the rock burst risk probability value corresponding to the event, and the spatiotemporal information of the event includes the time and location coordinates of the rock burst occurrence.

7. The method for monitoring and early warning of rock burst based on multi-source information fusion according to claim 6, characterized in that: In step five, the process of constructing a rock burst risk prediction model based on BP neural network is as follows: S151: Constructing a BP neural network basic model, which includes an input layer, multiple hidden layers, and an output layer; the input layer is used to receive background feature data of the independent sub-region; the multiple hidden layers are used to perform nonlinear transformation and feature mapping on the data received by the input layer; and the output layer is used to output the rock burst risk probability value of the independent sub-region; S152: Take the regional characteristic data corresponding to the event occurrence area in the historical mining area rock burst event data as input, take the rock burst risk probability value corresponding to the event as output, train the BP neural network basic model, and save the model parameters that meet the preset accuracy. After training, a rock burst risk prediction model based on the BP neural network is obtained.

8. The method for monitoring and early warning of rock burst based on multi-source information fusion according to claim 7, characterized in that: In step 5, the process of using particle swarm optimization to optimize the hyperparameters of the rock burst risk prediction model is as follows: S251: Hyperparameter combination of learning rate and number of hidden layer neurons based on rock burst risk prediction model; S252: Randomly generate a set of particles, where the position of each particle represents a set of hyperparameters; S253: Use validation set data to evaluate model prediction accuracy as fitness value , as shown in formula (7); （7）； Where, To predict the correct number of samples, is the total sample size; S254: According to the fitness value Update the individual optimal position of each particle and the global optimal position of the entire particle swarm, and obtain the updated particle velocity according to formula (8): , according to formula (9) to obtain the updated particle position ; （8）； （9）； Where, is the velocity of particle i at the kth iteration, w is the inertia weight, and are two different learning factors. and are two random numbers in the interval [0, 1]. is the individual optimal position of particle i up to the kth iteration, is the global optimal position of the entire particle swarm up to the kth iteration, is the position of particle i at the kth iteration; S255: Repeatedly iteratively randomly generate a group of particles and update the individual optimal position of each particle and the global optimal position of the entire particle swarm until the preset number of iterations is reached and the optimal hyperparameter combination is obtained.

9. The method for monitoring and early warning of rock burst based on multi-source information fusion according to claim 8, characterized in that: In step 6, the process of grading warnings based on rock burst risk levels and generating warning reminders is as follows: When the rock burst risk level is low, a first-level warning is issued, and the alarm is controlled to perform the first-level warning action. At the same time, a reminder message is generated for periodic sampling monitoring of low-risk areas, and displayed in real time on the display; For situations where the rock burst risk level is medium, a second-level warning is issued, and the alarm is controlled to perform the second-level warning action. At the same time, a reminder message is generated to adopt fixed-point continuous monitoring and dynamically adjust the monitoring frequency for the medium-risk area, and the message is displayed in real time on the display. For situations where the impact ground pressure risk level is high, a third-level warning is issued and the alarm is controlled to perform the third-level warning action. At the same time, a reminder message is generated to deploy a full-dimensional three-dimensional monitoring network in the high-risk area and to link the early warning response mechanism, and the message is displayed in real time on the display.

10. A rock burst monitoring and early warning method system based on multi-source information fusion, used to implement the rock burst monitoring and early warning method based on multi-source information fusion according to any one of claims 1 to 9, characterized in that: Includes processor, alarm, display and memory; The processor includes a data acquisition module, a data fusion module, a model building module, a region division module, a risk prediction module, and a graded warning module; The data acquisition module is used to collect multi-dimensional data of the target mining area, the multi-dimensional data including terrain parameters, geological structure data and rock mass physical property information; The data fusion module is used to remove outliers from multi-dimensional data and integrate the multi-dimensional data using a Kalman filter fusion algorithm to obtain fused data. The model building module is used to generate a three-dimensional geological structure model of the target mining area using a geological structure inversion method based on the fused data; The region division module is used to divide the three-dimensional geological structure model into regions based on the homogeneity of the rock mass physical property information and the continuity of the geological structure data in the three-dimensional geological structure model to obtain N independent sub-regions; The risk prediction module is used to use the background feature data of the independent sub-region as input data of the rock burst risk prediction model based on the BP neural network to predict the risk and output the rock burst risk level of each area in the target mining area; The graded warning module is used to classify warnings according to the rock burst risk level and generate warning reminder suggestions; The alarm is used to perform graded warning actions according to the warning classification results; The display is used to display early warning reminder suggestions in real time; The memory is used for storing and reading data by the processor.

Images