Bayesian method-based rock burst risk dynamic assessment method

Abstract

A rock burst risk dynamic assessment method based on a Bayesian method comprises the steps that a stope working face is divided into M * M regular grids, and each grid serves as an independent assessment unit; collecting geological and mining data, selecting a plurality of influence factors based on historical data and expert knowledge, and quantifying the plurality of influence factors by adopting a static evaluation model; determining weights of a plurality of influence factors by using an analytic hierarchy process, and calculating an initial risk score of each evaluation unit by using a comprehensive index method; normalizing the initial dangerousness score into prior distribution by adopting an S-type function, and constructing an initial dangerousness prior probability field; a power law attenuation model is introduced to calculate influence weights among the evaluation units, and an influence weight matrix is constructed; real-time monitoring data are obtained through a micro-seismic monitoring system, the risk probability of each evaluation unit is updated in real time based on the Bayesian theorem, and dynamic evaluation and evaluation result output are achieved. According to the method, the accuracy and the real-time performance of rock burst risk assessment can be effectively improved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent and safe coal mining technology, specifically a dynamic assessment method for rockburst hazard based on Bayesian methods. Background Technology

[0002] Rockbursts are major dynamic disasters caused by the instantaneous release of elastic potential energy from coal and rock masses during coal mining. Their occurrence is the result of a complex coupling of multiple factors, including geological structure, mining disturbance, and stress environment. To effectively address this threat, researchers have developed various assessment methods, such as the comprehensive index method, the analytic hierarchy process (AHP), and numerical simulation, to quantify static factors and conduct hazard assessments. In recent years, artificial intelligence methods such as deep learning have also been introduced to effectively improve the accuracy of rockburst early warning systems.

[0003] However, existing technologies have significant limitations. Traditional methods largely rely on fixed weighting systems and expert experience, resulting in static evaluations that are difficult to incorporate into real-time monitoring data generated as mining progresses, thus failing to capture the dynamic evolution of hazards. While deep learning models excel at handling complex patterns, their "black box" nature leads to a lack of physical interpretability in the decision-making process, and model updates are difficult, limiting their generalization ability.

[0004] Therefore, there is an urgent need to provide an assessment method that can integrate static geological conditions with dynamic monitoring information, and has clear physical meaning and continuous updating capability, in order to overcome the static and lagging nature of existing technologies and achieve accurate and real-time early warning of rockburst hazards. Summary of the Invention

[0005] To address the problems existing in the prior art, this invention provides a dynamic assessment method for rockburst hazard based on Bayesian methods. This method is simple to implement, has low implementation cost, and high assessment accuracy. It can quantify the influencing factors of rockburst and realize real-time dynamic updates of assessment results, effectively improving the accuracy and real-time performance of rockburst hazard assessment, and providing reliable technical support for safe and efficient coal mining.

[0006] To achieve the above objectives, this invention provides a dynamic assessment method for rockburst hazard based on Bayesian methods, characterized by comprising the following steps: Step 1: Divide the working face of the mining area into a regular grid of M×M, with each grid serving as an independent evaluation unit; Step 2: Collect geological and mining data of the area to be evaluated, select multiple influencing factors based on historical data and expert knowledge, and quantify the multiple influencing factors using a static evaluation model; Step 3: Use the analytic hierarchy process (AHP) to determine the weights of multiple influencing factors, and then use the comprehensive index method to calculate the initial risk score for each assessment unit; Step 4: Use the sigmoid function to normalize the initial hazard score of each assessment unit into a prior distribution, and construct the initial hazard prior probability field for all assessment units. Step 5: Introduce a power-law decay model to calculate the influence weights between evaluation units and construct an influence weight matrix; Step Six: Obtain real-time monitoring data through the microseismic monitoring system, and based on Bayes' theorem, update the hazard probability of each assessment unit in real time by combining the influence weight matrix to realize the dynamic assessment of rockburst hazard and output the assessment results.

[0007] Furthermore, in order to ensure the accuracy of microseismic monitoring while avoiding computational redundancy due to excessively small grid sizes, the working face of the mining area is divided into a regular grid of 30×30 in step one.

[0008] Furthermore, in order to effectively couple the static geological characteristics with the dynamic disturbance factors of mining and ensure the accuracy of the assessment, the geological and mining data include coal seam depth, coal seam thickness, thickness of the hard roof above the coal seam, lithological parameters of the hard roof above the coal seam, distribution of geological structures, and mining intensity.

[0009] Furthermore, in order to integrate geological and mining data of different dimensions and sources into a risk indicator of a unified magnitude, in step two, a static assessment model is used to quantify multiple influencing factors as follows: The quantitative value of the mining depth is calculated according to formula (1). The quantitative value of coal seam thickness is calculated according to formula (2). The thickness of the hard roof above the coal seam is calculated according to formula (3). The quantitative values ​​of the lithological parameters of the hard roof above the coal seam are calculated according to formula (4). The quantitative value of the geological structure distribution is calculated according to formula (5). The quantitative value of the mining intensity is calculated according to formula (6). ; (1); In the formula, For grid The coal seam depth at that location; Minimum coal seam burial depth; This represents the maximum coal seam burial depth. (2); In the formula, For grid Coal seam thickness at the location; This is the minimum thickness of the coal seam; This represents the maximum thickness of the coal seam. (3); In the formula, For grid The thickness of the rigid top plate at that location; (4); In the formula, For grid Uniaxial compressive strength of the top rock; For grid Integrity coefficient of the roof rock mass; (5); In the formula, For grid Total length of the fault within the area; Grid area; For grid The distance from the center to the fault; (6); In the formula, For grid The advancing speed of the working face; For grid Output per unit time at the location; The maximum allowable output; For grid Simultaneous mining area; For grid The total mining area at the site.

[0010] Furthermore, in order to achieve accurate quantification of the initial hazard score, the process of calculating the initial hazard score for each grid in step three is as follows: S31: Calculate and construct the judgment matrix for pairwise comparison of influencing factors according to formula (7). ; (7); S32: Calculate the judgment matrix The feature vectors are used as initial weights; S33: Perform a consistency check. If the consistency ratio CR < 0.1, the check passes; otherwise, adjust the judgment matrix. Until the consistency test is passed; S34: Optimize the initial weights by combining historical rockburst data to determine the final weights; S35: Calculate the first according to formula (8) Initial hazard score for each grid ; (8).

[0011] To facilitate obtaining the probability of the initial occurrence of rockburst, in step four, the probability of the first rockburst is obtained according to formula (9). Prior probabilities of each grid cell ; (9); In the formula, As the reference threshold; This is the steepness parameter.

[0012] Furthermore, in order to obtain a weight matrix that accurately quantifies chain-like risk propagation, the process of constructing the influence weight matrix in step five is as follows: S51: Use the coordinates of the grid center point to represent the evaluation unit; S52: Combining laboratory rock mass acoustic emission test data and field microseismic monitoring data, the range of values ​​for the attenuation coefficient α of elastic waves propagating in rock mass was determined through wave velocity-attenuation experiments under different lithological conditions. S53: Construct the mesh according to formula (10) For the grid The effect of microseismic events on the weakening function ; (10); In the formula, This represents the maximum impact value. For grid and Distance from the center point; S54: Weakening Function Based on the Impact of Microseismic Events Construct according to formula (11) ; (11).

[0013] Furthermore, in order to effectively integrate the physical mechanism of microseismic energy propagation in space with the prior risk field and achieve dynamic probability updates of the response within seconds, the process of updating the hazard probability of each assessment unit in step six is ​​as follows: S61: Based on real-time monitoring data, the likelihood probability is calculated according to formula (12). ; (12); In the formula, This indicates observed evidence of microseismic activity; For grid The total number of microseismic events occurring within the area; For grid Microseismic events occurring within Energy; The number of grid cells; For grid For the grid Influence weight; The energy decay coefficient; S62: Combining the prior probability and the likelihood probability, calculate the risk probability of each assessment unit according to formula (13). ; (13); In the formula, This is the prior probability; For evidence probability; Let be the likelihood probability.

[0014] As a preferred option, in step six, S62, the evidence probability is obtained. The process is as follows: S62-1: Starting from the initial parameter values, generate candidate samples through the proposal distribution; S62-2: Calculate the contact probability according to formula (14) ; (14); In the formula, and These are the likelihood probabilities of the candidate sample and the current sample, respectively. and These are the prior probabilities of the candidate sample and the current sample, respectively; and These are the proposal distributions of the candidate samples and the current sample, respectively. S62-3: Based on acceptance probability Decide whether to accept candidate samples; S62-4: After the pre-burning period, the collected sample sequences approximately follow a posterior distribution. ; S62-5: Calculate the probability of evidence using a weighted average of likelihood values ​​from the posterior samples or an importance sampling method. .

[0015] To facilitate intuitive observation of the hazardous conditions in different assessment units, and to facilitate automatic alarm reminders when limits are exceeded, in step six, during the output of rockburst hazard assessment results, the real-time hazard probability map of each assessment unit is iteratively updated, and a dynamic hazard map is output using visualization tools. At the same time, a hazard alarm threshold is set, and an early warning action is triggered when the hazard probability exceeds the hazard alarm threshold.

[0016] This invention provides a dynamic assessment method for rockburst hazard based on Bayesian updates. First, the working face of the mining area is divided into grids, facilitating precise location of risk areas and avoiding spatial ambiguity inherent in traditional methods, thus improving spatial resolution. Next, combining geological and mining data, multiple influencing factors are selected based on historical data and expert knowledge. This efficiently and accurately identifies the truly important influencing factors for the assessment objective. Then, a static model is used to objectively quantify these influencing factors, ensuring consistent assessment standards across all assessment units. This provides an objective data foundation for subsequent accurate assessments, facilitating data-driven decision-making. Furthermore, the analytic hierarchy process (AHP) is used to determine the weights of each influencing factor, balancing expert subjective experience with objective data, thereby reducing the bias of single methods. Simultaneously, the comprehensive index method effectively integrates multi-dimensional indicators, adapting to the risk characteristics of different assessment units. The combination of AHP and the comprehensive index method offers the advantages of scientific decision-making and dynamic adaptation, making it suitable for complex risk assessment scenarios and ensuring the accuracy of the initial hazard score. Furthermore, an S-shaped function is used to normalize the initial hazard scores of the assessment units into a prior distribution, which can reasonably map unbounded scores into probability values. This provides standardized input for Bayesian updates, risk quantification calculations, and information fusion, facilitating subsequent risk assessment processes. Then, a power-law decay model is introduced to accurately reflect the attenuation of mine hazards with distance. Through physically driven spatial weight allocation, it addresses the technical bottlenecks of neighborhood relationship distortion and neglect of hazard transmission in gridded risk assessments. Simultaneously, the obtained influence weight matrix quantifies the spatial correlation strength between assessment units and enables efficient calculation of assessment results through matrix operations in subsequent processes, providing a mathematical foundation for the construction of a dynamic risk field. Finally, based on Bayesian inference theory, real-time microseismic monitoring data is used as a likelihood function. Combined with the initial prior probability field, the posterior probabilities of each assessment unit are dynamically updated in real time. Through a multi-source coupling update process, continuous rolling assessment of impact hazard is achieved, significantly reducing the risk identification cycle from hours to seconds, enabling real-time risk perception and early warning capabilities. In assessing the probability of hazard, the incorporation of an influence weight matrix effectively calculates the risk increment of adjacent assessment units, facilitating the quantification of chain propagation and significantly improving assessment accuracy. This invention integrates static geological mining conditions with dynamic monitoring information within a probabilistic framework, overcoming the static limitations of traditional assessment methods and the "black box" defects of deep learning models. This significantly improves the timeliness, accuracy, and interpretability of rockburst hazard identification, making rockburst hazard assessment more precise and scientific.

[0017] This method is simple to implement, low in cost, and highly accurate in assessment. It can quantify the influencing factors of rockburst and realize real-time dynamic updates of assessment results, which can effectively improve the accuracy and real-time performance of rockburst hazard assessment and provide reliable technical support for safe and efficient coal mining. Attached Figure Description

[0018] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0019] The invention will now be further described with reference to the accompanying drawings.

[0020] like Figure 1 As shown, this invention provides a dynamic assessment method for rockburst hazard based on Bayesian methods, characterized by comprising the following steps: Step 1: Divide the working face of the mining area into a regular grid of M×M, with each grid serving as an independent evaluation unit; Step 2: Collect geological and mining data of the area to be evaluated, select multiple influencing factors based on historical data and expert knowledge, and quantify the multiple influencing factors using a static evaluation model; Step 3: Use the analytic hierarchy process (AHP) to determine the weights of multiple influencing factors, and then use the comprehensive index method to calculate the initial risk score for each assessment unit; Step 4: Use the sigmoid function to normalize the initial hazard score of each assessment unit into a prior distribution, and construct the initial hazard prior probability field for all assessment units. Step 5: Introduce a power-law decay model to calculate the influence weights between evaluation units and construct an influence weight matrix; Step Six: Obtain real-time monitoring data through the microseismic monitoring system, and based on Bayes' theorem, update the hazard probability of each assessment unit in real time by combining the influence weight matrix to realize the dynamic assessment of rockburst hazard and output the assessment results.

[0021] In order to ensure the accuracy of microseismic monitoring while avoiding computational redundancy caused by excessively small grid size, the working face of the mining area is divided into a regular grid of 30×30 in step one.

[0022] In order to effectively couple the static geological characteristics with the dynamic disturbance factors of mining and ensure the accuracy of the assessment, the geological and mining data include coal seam depth, coal seam thickness, thickness of the hard roof above the coal seam, lithological parameters of the hard roof above the coal seam, distribution of geological structures, and mining intensity.

[0023] In order to integrate geological and mining data of different dimensions and sources into a risk indicator of a unified magnitude, in step two, a static assessment model is used to quantify multiple influencing factors as follows: The quantitative value of the mining depth is calculated according to formula (1). The quantitative value of coal seam thickness is calculated according to formula (2). The thickness of the hard roof above the coal seam is calculated according to formula (3). The quantitative values ​​of the lithological parameters of the hard roof above the coal seam are calculated according to formula (4). The quantitative value of the geological structure distribution is calculated according to formula (5). The quantitative value of the mining intensity is calculated according to formula (6). ; (1); In the formula, For grid The coal seam depth at that location; Minimum coal seam burial depth; This represents the maximum coal seam burial depth. (2); In the formula, For grid Coal seam thickness at the location; This is the minimum thickness of the coal seam; This represents the maximum thickness of the coal seam. (3); In the formula, For grid The thickness of the rigid top plate at that location; (4); In the formula, For grid Uniaxial compressive strength of the top rock; For grid Integrity coefficient of the roof rock mass; (5); In the formula, For grid Total length of the fault within the area; Grid area; For grid The distance from the center to the fault; (6); In the formula, For grid The advancing speed of the working face; For grid Output per unit time at the location; The maximum allowable output; For grid Simultaneous mining area; For grid The total mining area at the site.

[0024] To achieve accurate quantification of the initial hazard score, the process of calculating the initial hazard score for each grid in step three is as follows: S31: Calculate and construct the judgment matrix for pairwise comparison of influencing factors according to formula (7). ; (7); S32: Calculate the judgment matrix The feature vectors are used as initial weights; S33: Perform a consistency check. If the consistency ratio CR < 0.1, the check passes; otherwise, adjust the judgment matrix. Until the consistency test is passed; S34: Optimize the initial weights by combining historical rockburst data to determine the final weights; S35: Calculate the first according to formula (8) Initial hazard score for each grid ; (8).

[0025] To facilitate obtaining the probability of the initial occurrence of rockburst, in step four, the probability of the first rockburst is obtained according to formula (9). Prior probabilities of each grid cell ; (9); In the formula, This serves as the baseline threshold, corresponding to a risk threshold with a probability of 0.5. The steepness parameter is used to control the sensitivity of the transition curve to ensure that the probability changes smoothly within the (0,1) interval.

[0026] In order to obtain a weight matrix that accurately quantifies chain-like risk propagation, the process of constructing the influence weight matrix in step five is as follows: S51: Use the coordinates of the grid center point to represent the evaluation unit; S52: Since elastic waves exhibit power-law attenuation characteristics when propagating in rock media, by combining laboratory rock mass acoustic emission test data and field microseismic monitoring data, and through wave velocity-attenuation experiments under different lithological conditions, the range of values ​​for the attenuation coefficient α of elastic waves propagating in rock media is determined. As an optimal choice, the range of values ​​for α is 1.5 to 2.5. S53: The attenuation model is based on a power function of the distance between grid center points, and the grid is constructed according to formula (10). For the grid The effect of microseismic events on the weakening function ; (10); In the formula, This represents the maximum impact value. For grid and Distance from the center point; S54: Weakening Function Based on the Impact of Microseismic Events Construct according to formula (11) ; (11).

[0027] In order to effectively integrate the physical mechanism of microseismic energy propagation in space with the prior risk field and achieve dynamic probability updates of the response in seconds, the process of updating the hazard probability of each assessment unit in step six is ​​as follows: S61: Based on real-time monitoring data, the likelihood probability is calculated according to formula (12). ; (12); In the formula, This indicates observed evidence of microseismic activity; For grid The total number of microseismic events occurring within the area; For grid Microseismic events occurring within Energy; The number of grid cells; For grid For the grid Influence weight; The energy decay coefficient; S62: Combining the prior probability and the likelihood probability, calculate the risk probability of each assessment unit according to formula (13). ; (13); In the formula, This is the prior probability; For evidence probability; Let be the likelihood probability.

[0028] As a preferred embodiment, in step six, S62, due to the evidence probability... The calculation involves complex integrals, which are approximated by sampling directly from the posterior distribution using the MCMC method. Specifically, the Metropolis-Hastings algorithm is employed to obtain the evidence probability. The process is as follows: S62-1: Starting from the initial parameter values, generate candidate samples through the proposal distribution; S62-2: Calculate the contact probability according to formula (14) ; (14); In the formula, and These are the likelihood probabilities of the candidate sample and the current sample, respectively. and These are the prior probabilities of the candidate sample and the current sample, respectively; and These are the proposal distributions of the candidate samples and the current sample, respectively. S62-3: Based on acceptance probability Decide whether to accept candidate samples; S62-4: After the pre-burning period, the collected sample sequences approximately follow a posterior distribution. ; S62-5: Calculate the probability of evidence using a weighted average of likelihood values ​​from the posterior samples or an importance sampling method. .

[0029] To facilitate intuitive observation of the hazardous conditions in different assessment units, and to facilitate automatic alarm reminders when limits are exceeded, in step six, during the output of rockburst hazard assessment results, the real-time hazard probability map of each assessment unit is iteratively updated, and a dynamic hazard map is output using visualization tools. At the same time, a hazard alarm threshold is set, and an early warning action is triggered when the hazard probability exceeds the hazard alarm threshold.

[0030] This invention provides a dynamic assessment method for rockburst hazard based on Bayesian updates. First, the working face of the mining area is divided into grids, facilitating precise location of risk areas and avoiding spatial ambiguity inherent in traditional methods, thus improving spatial resolution. Next, combining geological and mining data, multiple influencing factors are selected based on historical data and expert knowledge. This efficiently and accurately identifies the truly important influencing factors for the assessment objective. Then, a static model is used to objectively quantify these influencing factors, ensuring consistent assessment standards across all assessment units. This provides an objective data foundation for subsequent accurate assessments, facilitating data-driven decision-making. Furthermore, the analytic hierarchy process (AHP) is used to determine the weights of each influencing factor, balancing expert subjective experience with objective data, thereby reducing the bias of single methods. Simultaneously, the comprehensive index method effectively integrates multi-dimensional indicators, adapting to the risk characteristics of different assessment units. The combination of AHP and the comprehensive index method offers the advantages of scientific decision-making and dynamic adaptation, making it suitable for complex risk assessment scenarios and ensuring the accuracy of the initial hazard score. Furthermore, an S-shaped function is used to normalize the initial hazard scores of the assessment units into a prior distribution, which can reasonably map unbounded scores into probability values. This provides standardized input for Bayesian updates, risk quantification calculations, and information fusion, facilitating subsequent risk assessment processes. Then, a power-law decay model is introduced to accurately reflect the attenuation of mine hazards with distance. Through physically driven spatial weight allocation, it addresses the technical bottlenecks of neighborhood relationship distortion and neglect of hazard transmission in gridded risk assessments. Simultaneously, the obtained influence weight matrix quantifies the spatial correlation strength between assessment units and enables efficient calculation of assessment results through matrix operations in subsequent processes, providing a mathematical foundation for the construction of a dynamic risk field. Finally, based on Bayesian inference theory, real-time microseismic monitoring data is used as a likelihood function. Combined with the initial prior probability field, the posterior probabilities of each assessment unit are dynamically updated in real time. Through a multi-source coupling update process, continuous rolling assessment of impact hazard is achieved, significantly reducing the risk identification cycle from hours to seconds, enabling real-time risk perception and early warning capabilities. In assessing the probability of hazard, the incorporation of an influence weight matrix effectively calculates the risk increment of adjacent assessment units, facilitating the quantification of chain propagation and significantly improving assessment accuracy. This invention integrates static geological mining conditions with dynamic monitoring information within a probabilistic framework, overcoming the static limitations of traditional assessment methods and the "black box" defects of deep learning models. This significantly improves the timeliness, accuracy, and interpretability of rockburst hazard identification, making rockburst hazard assessment more precise and scientific.

[0031] This method is simple to implement, low in cost, and highly accurate in assessment. It can quantify the influencing factors of rockburst and realize real-time dynamic updates of assessment results, which can effectively improve the accuracy and real-time performance of rockburst hazard assessment and provide reliable technical support for safe and efficient coal mining.

Claims

1. A method for dynamic assessment of rockburst hazard based on Bayesian method, characterized in that, Includes the following steps: Step 1: Divide the working face of the mining area into a regular grid of M×M, with each grid serving as an independent evaluation unit; Step 2: Collect geological and mining data of the area to be evaluated, select multiple influencing factors based on historical data and expert knowledge, and quantify the multiple influencing factors using a static evaluation model; Step 3: Use the analytic hierarchy process (AHP) to determine the weights of multiple influencing factors, and then use the comprehensive index method to calculate the initial risk score for each assessment unit; Step 4: Use the sigmoid function to normalize the initial hazard score of each assessment unit into a prior distribution, and construct the initial hazard prior probability field for all assessment units. Step 5: Introduce a power-law decay model to calculate the influence weights between evaluation units and construct an influence weight matrix; Step Six: Obtain real-time monitoring data through the microseismic monitoring system, and based on Bayes' theorem, update the hazard probability of each assessment unit in real time by combining the influence weight matrix to realize the dynamic assessment of rockburst hazard and output the assessment results.

2. The rock burst hazard dynamic assessment method based on the Bayesian method according to claim 1, characterized in that, In step one, the working face of the mining area is divided into a regular grid of 30×30.

3. The rock burst hazard dynamic assessment method based on Bayesian method according to claim 1, characterized in that, Geological and mining data include coal seam depth, coal seam thickness, thickness of the hard roof above the coal seam, lithological parameters of the hard roof above the coal seam, distribution of geological structures, and mining intensity.

4. The rock burst hazard dynamic assessment method based on Bayesian method according to claim 1, characterized in that, In step two, the static evaluation model is used to quantify multiple influencing factors as follows: The quantitative value of the mining depth is calculated according to formula (1) The quantitative value of the coal seam thickness is calculated according to formula (2) The quantitative value of the hard roof thickness above the coal seam is calculated according to formula (3) The quantitative value of the hard roof lithology parameter above the coal seam is calculated according to formula (4) The quantitative value of the geological structure distribution is calculated according to formula (5) The quantitative value of the mining intensity is calculated according to formula (6) ; （1）； wherein is the grid depth of the coal seam at the location; is the minimum coal seam depth; is the maximum coal seam depth; （2）； wherein is the grid thickness of the coal seam at the location; is the minimum thickness of the coal seam; is the maximum thickness of the coal seam; （3）； In the formula, is the thickness of the hard roof at the grid point; （4）； In the formula, is a grid The uniaxial compressive strength of the roof rock at the position; is a grid The integrity coefficient of the roof rock mass at the position; （5）； wherein is the grid total length of faults within the bin; grid area; is the grid distance from the center to the fault; （6）； wherein is the grid is the rate of advance of the working face at the point; is the grid is the production per unit of time at the point; is the maximum allowable production; is the grid is the simultaneous area of extraction at the point; is the grid is the total area of extraction at the point.

5. The rock burst hazard dynamic assessment method based on Bayesian method according to claim 1, characterized in that, In step three, the process of calculating the initial hazard score for each grid is as follows: S31: Calculate the judgment matrix of pairwise comparison of the construction influence factors according to formula (7) ; （7）； S32: Calculate the judgment matrix as the initial weight vector; S33: consistency check is performed, if the consistency ratio CR < 0.1, the check is passed, otherwise the judgment matrix is adjusted until the consistency check is passed; S34: Optimize the initial weights by combining historical rockburst data to determine the final weights; S35: Calculate the first according to formula (8) Initial hazard score for each grid ; （8）。 6. The method for dynamic assessment of rockburst hazard based on Bayesian method according to claim 1, characterized in that, In step four, the first step is obtained according to formula (9). Prior probabilities of each grid cell ; （9）； In the formula, As the reference threshold; This is the steepness parameter.

7. The method for dynamic assessment of rockburst hazard based on Bayesian method according to claim 1, characterized in that, In step five, the process of constructing the influence weight matrix is ​​as follows: S51: Use the coordinates of the grid center point to represent the evaluation unit; S52: Combining laboratory rock mass acoustic emission test data and field microseismic monitoring data, the range of values ​​for the attenuation coefficient α of elastic waves propagating in rock mass was determined through wave velocity-attenuation experiments under different lithological conditions. S53: Construct the mesh according to formula (10) For the grid The effect of microseismic events on the weakening function ; （10）； In the formula, This represents the maximum impact value. For grid and Distance from the center point; S54: Weakening Function Based on the Impact of Microseismic Events Construct according to formula (11) ; （11）。 8. The method for dynamic assessment of rockburst hazard based on Bayesian method according to claim 1, characterized in that, In step six, the process of updating the hazard probability of each assessment unit is as follows: S61: Based on real-time monitoring data, the likelihood probability is calculated according to formula (12). ; （12）； In the formula, This indicates observed evidence of microseismic activity; For grid The total number of microseismic events occurring within the area; For grid Microseismic events occurring within Energy; The number of grid cells; For grid For the grid Influence weight; The energy decay coefficient; S62: Combining the prior probability and the likelihood probability, calculate the risk probability of each assessment unit according to formula (13). ; （13）； In the formula, This is the prior probability; For evidence probability; Let be the likelihood probability.

9. The method for dynamic assessment of rockburst hazard based on Bayesian method according to claim 1, characterized in that, In step six, S62, the probability of evidence is obtained. The process is as follows: S62-1: Starting from the initial parameter values, generate candidate samples through the proposal distribution; S62-2: Calculate the contact probability according to formula (14) ; （14）； In the formula, and These are the likelihood probabilities of the candidate sample and the current sample, respectively. and These are the prior probabilities of the candidate sample and the current sample, respectively; and These are the proposal distributions of the candidate samples and the current sample, respectively. S62-3: Based on acceptance probability Decide whether to accept candidate samples; S62-4: After the pre-burning period, the collected sample sequences approximately follow a posterior distribution. ; S62-5: Calculate the probability of evidence using a weighted average of likelihood values ​​from the posterior samples or an importance sampling method. .

10. The method for dynamic assessment of rockburst hazard based on Bayesian method according to claim 1, characterized in that, In step six, during the output of the rockburst hazard assessment results, the real-time hazard probability map of each assessment unit is iteratively updated, and a dynamic hazard map is output using visualization tools. At the same time, a hazard alarm threshold is set, and an early warning action is triggered when the hazard probability exceeds the hazard alarm threshold.

Images