A coal mine rock burst intelligent prevention and control decision system based on multi-source information fusion

Abstract

The present application relates to the technical field of coal mine safety, and discloses a coal mine rock burst intelligent prevention and control decision system based on multi-source information fusion, which comprises the following modules: multi-source data acquisition, wave impedance initial modeling, energy dissipation analysis, prevention and control parameter optimization, while-drilling spatial dynamic reconstruction, hot start local correction, and efficiency evaluation and closed-loop evolution. The system obtains multi-source monitoring data to construct a global initial wave impedance matrix coupled with stress and wave velocity. The target area and dynamic load source are extracted, the residual energy is quantified along the shortest time ray path, and the initial pressure relief borehole parameters are calculated. The mechanical damage variable is calculated by using the while-drilling dynamics parameters to locally update the wave impedance, and the unconstructed borehole is locally searched and compensated. The final residual energy of the whole domain is recalculated to evaluate the efficiency, and the measured microseismic data is used to adjust the basic parameters reversely. The present application improves the accuracy of the rock burst prevention and control decision of deep coal mines through the while-drilling dynamic decision of the rock burst prevention parameters and the long-period adaptive closed-loop calibration.

Description

Technical Field

[0001] This invention relates to the field of coal mine safety technology, specifically to an intelligent decision-making system for the prevention and control of coal mine rockbursts based on multi-source information fusion. Background Technology

[0002] Rockburst is one of the major dynamic hazards faced during deep coal mining. Currently, for the prevention and control of rockburst at the mining face, methods such as microseismic networks and online stress gauges are used to predict dangerous areas, and large-diameter stress relief boreholes are carried out in areas where risks are identified to disrupt the integrity of the coal and rock mass and release localized concentrated stress.

[0003] However, existing anti-seismic decision-making methods have significant limitations in practical applications. At the data processing level, existing monitoring systems analyze microseismic events, static stress, and geological exploration data independently, failing to establish a deep fusion model of heterogeneous monitoring data in three-dimensional space. Conventional analysis of medium energy transfer characteristics often uses static rock mass density and wave velocity background values ​​to calculate wave impedance, ignoring the nonlinear influence of the closure and compaction of rock micro-fractures under deep high-stress environments on actual wave velocity and wave impedance. This leads to deviations between the deduced dynamic stress wave propagation path and energy attenuation law and the actual physical state of deep rock mass.

[0004] During the formulation and execution phase of anti-impact and pressure relief schemes, existing parameter designs are mostly generated once based on preliminary static geological data, and engineering construction is in a one-way execution state. During drilling, dynamic feedback data such as thrust and torque generated when the intelligent drilling rig cuts and breaks rock cannot be extracted to quantify the actual mechanical damage to the rock mass. Due to the lack of an interface to convert physical execution data into geological models, the system cannot use the actual trajectory of completed boreholes to dynamically adjust the parameters of surrounding unconstructed boreholes while drilling. If the recalculation of the entire environmental data is forcibly performed after a single borehole operation, it will generate a huge consumption of computing power and make it difficult to meet the time requirements for real-time guidance of downhole engineering.

[0005] Furthermore, traditional anti-scour and pressure relief engineering acceptance focuses primarily on the statistics of drilling footage and construction quantity, lacking a quantitative assessment of the critical instability energy that the pressure relief area can ultimately withstand and the actual residual energy. This makes it difficult to accurately determine whether a single project meets safety standards. At the same time, during long-term mining, existing models rely on fixed empirical setpoints and fail to use the actual residual energy of newly occurring microseismic events in the pressure relief area to inversely calibrate the energy attenuation coefficient of the system foundation. The lack of a closed-loop evolution mechanism leads to a gradual decrease in the prediction accuracy of the decision-making system as the mining environment changes and time progresses. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides an intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion. This system solves the problem that existing coal mine rockburst prevention decisions rely excessively on static discrete data and are executed in a one-way manner, lacking dynamic correction and long-cycle closed-loop calibration mechanisms based on drilling physical feedback, which leads to rockburst prevention parameters deviating from the true state of the rock mass.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] This invention provides an intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion. The system includes: The system includes a multi-source data acquisition module, a wave impedance initial modeling module, an energy dissipation analysis module, a prevention parameter optimization module, a drilling space dynamic reconstruction module, a hot start local correction module, and an effectiveness evaluation and closed-loop evolution module. The multi-source data acquisition module is used to acquire geological parameters, microseismic event parameters, and online stress monitoring data of the mining face; The initial wave impedance modeling module is used to discretize the mining space into a three-dimensional spatial grid, establish a coupling relationship formula between stress and wave velocity based on monitoring data, and construct a global initial wave impedance matrix. The energy dissipation analysis module is used to extract the static target area and dynamic source from the global initial wave impedance matrix, and to construct an energy decay integral model along the shortest time ray path to quantify the residual energy. The prevention and control parameter optimization module is used to set the pressure relief borehole as a disturbance body, construct the objective function with the minimum total borehole depth and the maximum energy dissipation gradient, and output the three-dimensional spatial parameters of the anti-impact pressure relief borehole. The drilling space dynamic reconstruction module is used to extract drilling attitude and dynamic parameters, calculate rock mass mechanical damage variables, and locally update the global initial wave impedance matrix. The hot-start local correction module is used to extract the envelope mesh around the actual borehole trajectory and perform local gradient search to calculate the parameter compensation amount for the remaining unconstructed boreholes; The performance evaluation and closed-loop evolution module is used to calculate the final residual energy based on the final global wave impedance matrix after drilling and to evaluate the anti-surge performance and update the basic parameters.

[0009] Furthermore, the wave impedance initial modeling module assigns initial values ​​of initial microseismic longitudinal wave velocity and coal and rock mass density to the grid nodes of the three-dimensional spatial grid, extracts the absolute static load stress value at the grid node, and uses the standard value of uniaxial compressive strength of coal and rock mass in the target area to perform dimensionless processing on the absolute static load stress value. The initial three-dimensional wave impedance value of the grid node is calculated by combining the nonlinear strengthening coupling coefficient of static stress on the wave impedance property change, and then the global initial wave impedance matrix is ​​constructed. The above scheme uses measured stress data to correct the initial parameters of the medium, reduces the error caused by the closure of microcracks in the deep high-stress zone on the wave impedance calculation, and improves the accuracy of the medium parameters in the three-dimensional spatial grid.

[0010] Furthermore, the energy dissipation analysis module calculates the statistical mean and standard deviation of the impedance values ​​of the grid nodes in the global initial wave impedance matrix, and extracts the connected regions that are higher than the judgment threshold as static load target areas. Density clustering algorithm is used to identify spatially dense clusters of microseismic sources and perform weighted calculations to extract them as dynamic load sources; The fast traversal algorithm is used to solve the process function equation to calculate the initial arrival time field distribution, and the reverse tracing is performed along the time gradient descent direction to obtain the shortest time ray path connecting the dynamic source and the static target area; The above scheme solves the time field based on wave theory and clarifies the specific propagation path of the moving carrier in a non-uniform geological body.

[0011] Furthermore, the energy dissipation analysis module discretizes the shortest time ray path into micro-segment path variables, and performs exponential decay spatial integration calculations by combining the medium's basic energy attenuation coefficient with the wave impedance distribution function of the grid nodes along the path to solve for the residual energy value of the dynamic stress wave penetrating the coal and rock mass medium and reaching the boundary of the static load target area.

[0012] Furthermore, the prevention and control parameter optimization module constructs a virtual cylindrical geometric model within a three-dimensional spatial grid based on the configured spatial parameter vector containing coordinates, angles, and depth, extracts grid nodes located within the design radius to form a local disturbance set, and reduces wave impedance values ​​to form an evolution impedance matrix. Re-execute the path integral operation to obtain the updated residual energy, and construct the objective function using the dimensionless value of the total drilling footage and the dimensionless value of the energy dissipation gradient; An optimization algorithm is used to iteratively solve the objective function, extract the global population's historical optimal position, and decode to obtain the three-dimensional spatial parameters of the anti-impact and pressure-relief borehole; The above scheme solves the spatial parameters of the pressure relief borehole by iteratively solving the objective function, thereby reducing the reliance on human experience in the parameter planning process.

[0013] Furthermore, the drilling space dynamic reconstruction module uses the minimum curvature method to transform the drilling dip angle and azimuth angle parameters to obtain the true physical three-dimensional spatial coordinates of the drill bit; Collect thrust and torque parameters to calculate the mechanical specific energy consumed in rock breaking; By using the limit fracture energy constant of the corresponding rock strata in the region to perform normalized mapping calculation on the mechanical energy, a rock mass mechanical damage variable distributed in the interval of zero to one is generated; The above scheme uses drilling measurement parameters to calculate mechanical specific energy and damage variables, realizing the transformation of physical feedback data during construction into model parameters.

[0014] Furthermore, the drilling space dynamic reconstruction module combines rock mass mechanical damage variables and acoustic damage sensitivity adjustment index, uses the local wave impedance dynamic evolution equation based on effective medium theory to calculate the updated impedance value of the damaged grid, and replaces the grid data corresponding to the actual borehole trajectory penetration in the original global initial wave impedance matrix.

[0015] Furthermore, the hot-start local correction module calculates the effective stress relief radius of the plastic yield failure around the borehole based on the Mohr-Coulomb failure criterion, and generates an envelope mesh; Perform a Boolean intersection operation between the envelope mesh and the mesh sequence contained in the shortest time ray path. When there are overlapping nodes, use the spatial parameter vector of the remaining unconstructed boreholes as the initial high-quality solution input to the local search algorithm. Within the optimization and dimensionality reduction constraint domain, the partial derivative gradient is calculated based on the difference in energy increment change. The gradient ascent algorithm is used to perform step-by-step deduction and calculation of the spatial coordinate deflection and depth compensation of adjacent boreholes to be constructed according to the preset search step size. The above scheme determines the overlap between the envelope grid and the energy transfer path, limits the area that needs to be recalculated, reduces the overall computational load of the optimization process, and meets the time requirements for drilling correction.

[0016] Furthermore, the performance evaluation and closed-loop evolution module re-performs reverse tracking and integration calculations based on the final global wave impedance matrix after all pressure relief drilling operations are completed, to obtain the final residual energy value of the moving carrier reaching the boundary of the static target area. The minimum critical energy for rockburst failure in the target area is calculated by combining the elastic strain energy release criterion. When the product of the final residual energy value and the safety reserve coefficient is greater than the minimum critical energy, an anti-impact warning signal is triggered and redirected to the prevention and control parameter optimization module to trigger a supplementary pressure relief drilling planning cycle.

[0017] Furthermore, after the completion of the regional anti-shock and depressurization project, the performance evaluation and closed-loop evolution module extracts the actual residual wave frequency energy of newly occurring microseismic events penetrating the depressurization zone; Calculate the prediction deviation gradient between the actual residual energy and the model-predicted residual energy. When the prediction deviation gradient exceeds the limit, adjust the dielectric energy attenuation coefficient and the nonlinear strengthening coupling coefficient in reverse adaptively. The above scheme adjusts the attenuation coefficient and coupling coefficient in reverse based on measured microseismic data over a long period, providing a closed-loop feedback and correction method for the continuous operation of the system.

[0018] This invention provides an intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion. It has the following beneficial effects: 1. This invention establishes an initial wave impedance matrix with nonlinear coupling between stress and wave velocity, and uses a fast travel algorithm to solve the process function equation to obtain the shortest time ray path of the dynamic carrier wave, transforming multi-source discrete monitoring data into a continuous spatial medium model. The scheme, combined with measured static stress correction of the initial physical parameters of the coal and rock mass, can more objectively reflect the true energy transfer characteristics and dynamic load energy attenuation law of non-uniform rock mass under deep high stress environment.

[0019] 2. This invention uses dynamic parameters obtained from drilling measurements to calculate rock mass mechanical damage variables to update local wave impedance. It also performs local gradient search correction on parameters of unconstructed boreholes within a defined area by judging the intersection of the envelope grid affected by the borehole and the energy propagation path. The scheme dynamically adjusts subsequent prevention and control parameters based on real physical feedback during construction, while avoiding full-domain data recalculation and meeting the real-time requirements of dynamic decision-making for anti-scour and pressure relief parameters during drilling.

[0020] 3. This invention evaluates the effectiveness of a single depressurization project and determines whether to trigger supplementary drilling by comparing the final residual energy obtained from the recalculation of the global wave impedance matrix with the minimum critical energy for impact instability. At the same time, it uses the residual energy prediction deviation of actual microseismic events over a long period to adjust the energy attenuation coefficient and coupling coefficient of the system foundation in reverse. The scheme constructs a closed-loop mechanism that combines short-term construction compliance verification with long-term parameter self-calibration to improve the calculation accuracy and reliability of the system under long-term operating conditions. Attached Figure Description

[0021] Figure 1 This is a system architecture diagram of the present invention; Figure 2 This is a flowchart of the method of the present invention; Figure 3 This is a schematic diagram of the process for optimizing the multi-source data fusion and initial anti-collision scheme of the present invention; Figure 4 This is a schematic diagram of the system architecture for dynamic optimization and performance closed-loop evaluation during drilling according to the present invention. Figure 5 This is a slice diagram of the initial three-dimensional wave impedance distribution of the present invention; Figure 6 This is a three-dimensional scatter plot of the energy conduction shortest time ray path according to the present invention; Figure 7 Contour plots were drawn for the evolution impedance matrix of the drilling space reconstruction in this invention.

[0022] Among them, 10 is the multi-source data acquisition module; 20 is the wave impedance initial modeling module; 30 is the energy dissipation analysis module; 40 is the prevention parameter optimization module; 50 is the drilling space dynamic reconstruction module; 60 is the hot start local correction module; and 70 is the performance evaluation and closed-loop evolution module. Detailed Implementation

[0023] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] See attached document Figure 1 This invention provides an intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion. The system includes: a multi-source data acquisition module 10, a wave impedance initial modeling module 20, an energy dissipation analysis module 30, a prevention parameter optimization module 40, a drilling space dynamic reconstruction module 50, a hot start local correction module 60, and an effectiveness evaluation and closed-loop evolution module 70.

[0025] The multi-source data acquisition module 10 is used to simultaneously acquire geological parameters, microseismic event parameters, and online stress monitoring data of the mining face and adjacent areas.

[0026] The initial wave impedance modeling module 20 is used to discretize the mining space into a three-dimensional spatial grid. Based on the acquired geological parameters, microseismic event parameters and online stress monitoring data, it establishes a coupling relationship formula between stress and wave velocity and calculates the initial three-dimensional wave impedance value of the grid nodes.

[0027] The energy dissipation analysis module 30 is used to extract the high static load target area and dynamic load source from the initial wave impedance matrix, and construct an energy attenuation integral model along the shortest time propagation path to quantify the residual energy of the dynamic load.

[0028] The prevention and control parameter optimization module 40 is used to set the pressure relief borehole as the minimum disturbance body in the grid region, construct the objective function with the minimum total borehole depth and the maximum energy dissipation gradient, and solve and output the three-dimensional spatial parameters of the anti-impact pressure relief borehole.

[0029] The drilling space dynamic reconstruction module 50 is used to extract the drilling attitude parameters and dynamic parameters of the intelligent drilling rig during the pressure relief drilling construction stage, calculate the true physical coordinates of the drill bit and local mechanical damage variables, and replace the initial three-dimensional wave impedance value of the corresponding spatial grid through acoustic-mechanical coupling reduction calculation.

[0030] The hot-start local correction module 60 is used to extract the influence envelope mesh around the current borehole's true trajectory. It uses the previous calculation result as the initial value to limit the optimization domain and solves the deflection and depth compensation of the remaining unconstructed boreholes through local search.

[0031] The performance evaluation and closed-loop evolution module 70 is used to recalculate the final residual energy of the entire domain after the pressure relief drilling construction, compare it with the critical failure energy to evaluate the anti-scour performance and decide whether to trigger the supplementary drilling plan; and extract the residual energy of the wave frequency of newly generated microseismic events that cross the pressure relief construction area over a long period, calculate the prediction deviation gradient and update the foundation transfer coefficient and coupling coefficient in reverse.

[0032] The aforementioned multi-source data acquisition module 10 acquires the basic data of the downhole heterogeneous monitoring network. The wave impedance initial modeling module 20 and the energy dissipation analysis module 30 transform the discrete monitoring data into a continuous spatial matrix characterizing the mechanical properties of the rock mass. The prevention parameter optimization module 40 outputs initial guidance parameters to the downhole drilling rig execution end based on this spatial matrix. The drilling space dynamic reconstruction module 50 and the hot start local correction module 60 use the physical feedback generated by the drilling rig's execution actions to reverse the intervention decision algorithm. The performance evaluation and closed-loop evolution module 70 performs the final state safety compliance test of a single project and adjusts the system's basic parameters using macroscopic micro-vibration frequency response after leaving the single construction cycle.

[0033] See attached document Figure 2 This invention provides an intelligent decision-making method for preventing and controlling rockbursts in coal mines based on multi-source information fusion. The method includes the following steps: S100 collects geological information, mining information, online stress monitoring values, and microseismic source parameters of the working face and roadway area. The mining area is divided into a three-dimensional spatial grid. Based on the stress-wave velocity nonlinear coupling relationship, the initial three-dimensional wave impedance value of each grid node is calculated, and a global initial wave impedance matrix is ​​constructed.

[0034] S200. Extract the wave impedance anomaly region from the global initial wave impedance matrix as the static load target area, extract the centroid of the microseismic cluster as the dynamic load source, define the ray path of the dynamic carrier from the dynamic load source to the static load target area, and construct a spatial energy dissipation quantification model.

[0035] S300. Set the anti-impact and pressure relief borehole as a set of cylindrical grids with extremely low values ​​in the wave impedance matrix. With the objective function of maximizing the energy dissipation gradient and minimizing the total borehole depth, run the optimization algorithm to solve for the spatial opening coordinates, azimuth angle, inclination angle and hole depth of the initial anti-impact scheme set.

[0036] S400: During drilling operations, the drilling dip angle and azimuth angle parameters are acquired in real time, and the current three-dimensional trajectory coordinates of the borehole are calculated by integration; the thrust and torque dynamic parameters are extracted and converted into mechanical damage variables, and the wave impedance of the real trajectory mesh is reduced using the mechanical damage variables, and the global initial wave impedance matrix is ​​locally updated.

[0037] S500: Extract the spatial decompression influence envelope mesh around the actual trajectory of a single borehole, use the parameter set of the remaining unconstructed boreholes as the initial value for hot-start optimization, and perform local gradient search only on the energy transfer path affected by the envelope mesh update to calculate the compensation parameters for subsequent adjacent boreholes.

[0038] After the completion of the S600 large-diameter pressure relief drilling, the final residual energy value of the static load target area is recalculated and compared with the minimum critical energy for impact instability. If the value is not met, a new round of supplementary pressure relief scheme is triggered. After the pressure relief project in this area is completed, microseismic event data is continuously collected and the actual wave frequency residual energy penetrating the pressure relief area is extracted. The deviation gradient between the actual residual energy and the model-predicted residual energy is calculated, and the attenuation coefficient in the wave impedance matrix construction and reduction model is adjusted in reverse.

[0039] In specific implementation, step S100 provided by the present invention includes the following steps: S110 synchronously acquires geological parameters, microseismic event parameters and stress online monitoring data of the mining face and adjacent areas through the data acquisition module 10, and maps the heterogeneous data to a three-dimensional spatial coordinate system. Geological parameters include the thickness and lithology distribution of coal and rock strata obtained from geological boreholes; mining information covers the real-time mining outline of the working face and the topology of surrounding roadways; online stress monitoring data comes from real-time readings transmitted from borehole stress gauges deployed inside coal pillars and solid coal. The microseismic event parameters include the source spatial coordinates, released energy, and regional wave velocity tomography data calculated by the microseismic network. Since the multi-source monitoring data belong to different subsystems and have different spatiotemporal resolutions, the system establishes a unified three-dimensional Cartesian coordinate system. Using the opening of the working face or a specific location in the main transport roadway as the origin of the coordinate system, the multi-source monitoring data are transformed and aligned to the same spatial reference using a spatial transformation matrix. For the source location calculation of the microseismic network and the data parsing of the sensor communication protocol, conventional microseismic location algorithms and industrial standard communication protocols can be used.

[0040] S120 uses the wave impedance modeling module 20 to discretize the mining space into a three-dimensional spatial grid, and assigns independent rock mass density and initial microseismic P-wave velocity to the grid nodes.

[0041] The system sets the discrete step size of the three-dimensional spatial grid according to the accuracy requirements of the anti-impact and pressure relief project and the on-site computing resources, and generates a three-dimensional grid set for the entire domain. In this embodiment, the discrete step size of the three-dimensional spatial grid is set to a range of 0.5m to 2.0m. For any node in the three-dimensional grid, the system calculates and assigns an initial value for the density of the coal and rock mass of the node using a spatial interpolation algorithm based on the geological parameters after the aforementioned transformation and alignment; the system calculates and assigns an initial background value for the propagation velocity of the microseismic longitudinal waves of the node based on the microseismic wave velocity tomography data. In practical implementation, spatial interpolation algorithms can employ either ordinary kriging interpolation or inverse distance weighted interpolation to ensure that the discretized grid nodes can smoothly represent the physical property changes of continuous geological bodies.

[0042] S130, using the wave impedance modeling module 20 based on the coupling relationship formula between stress and wave velocity, calculates the initial three-dimensional wave impedance value of the mesh nodes and constructs the global initial wave impedance matrix.

[0043] In the high-stress environment of deep coal mines, micro-cracks inside the rock mass are closed and compacted under static stress, resulting in the actual stress wave propagation velocity and medium impedance in local areas being higher than the theoretical background value. Conventional impedance calculation models rely solely on the product of density and wave velocity, which is insufficient to accurately reflect the energy transfer characteristics of the medium in high stress concentration zones. The system incorporates stress data to compensate and correct the physical parameters of the grid foundation. The system extracts the absolute static stress values ​​at the grid nodes from online stress monitoring data. The absolute static load stress value was dimensionlessly processed using the uniaxial compressive strength benchmark value of the dominant rock strata. Combined with the stress-wave velocity nonlinear coupling coefficient previously calibrated in the laboratory, the initial three-dimensional wave impedance value of each grid node was calculated using the following formula: ; In the formula: For grid nodes The initial three-dimensional wave impedance value at the location; This refers to the discrete coordinate index variable of the three-dimensional spatial grid along the X-axis. This represents the discrete coordinate index variable of the 3D spatial grid along the Y-axis. This represents the discrete coordinate index variable of the three-dimensional spatial grid along the Z-axis. For grid nodes Initial density distribution values ​​of the coal and rock mass; For grid nodes The initial elastic longitudinal wave propagation velocity of the coal and rock mass; The nonlinear strengthening coupling coefficient of static stress on the change of wave impedance property is denoted as 0.05 to 0.15 in this embodiment. For grid nodes The three-dimensional absolute static stress value borne at the point; The standard value of uniaxial compressive strength of coal and rock mass in the target area.

[0044] The system traverses all three-dimensional spatial grid nodes to complete the above calculations, integrating discrete multi-source monitoring data into a global initial wave impedance matrix that continuously characterizes the energy transfer properties of the mining space medium. The constructed global initial wave impedance matrix will be output and stored, providing a basic spatial data model for subsequent dynamic load energy dissipation path tracing and decompression parameter optimization.

[0045] In specific implementations, step S200 provided by the present invention may include the following sub-steps: S210: Utilize the energy dissipation module 30 to calculate the statistical mean and standard deviation of the impedance values ​​of all grid nodes in the global initial wave impedance matrix. Set the statistical mean plus twice the standard deviation as the high static load judgment threshold, and screen out the set of grid nodes with impedance values ​​greater than the judgment threshold. For multiple spatially unconnected grid node sets, extract the connected domain with the largest envelope volume, calculate the spatial geometric center of the connected domain and mark it as the static load target area. For the set of source spatial coordinate points in the obtained microseismic event parameters, use the DBSCAN density clustering algorithm to identify spatially dense microseismic source clusters. Use the single-release energy value of the microseismic event as the weight value to calculate the three-dimensional weighted average spatial coordinates of all events within the source cluster and mark them as dynamic load sources.

[0046] S220 utilizes the energy dissipation module 30 to construct a propagation path optimization solution rule based on the physical law of stress wave diffraction along the shortest time path when propagating in a non-uniform medium in wave theory. Taking the dynamic load source as the starting node and the static load target area as the target node, the fast travel algorithm is used to solve the equation of motion by combining the local wave velocity values ​​mapped by each grid node in the global initial wave impedance matrix. The first arrival time field distribution of the wavefront from the dynamic load source to all other grid nodes in the matrix is ​​calculated. Based on the calculated first arrival time field, the reverse tracing calculation is performed from the position of the static load target area along the time gradient descent direction to obtain a three-dimensional spatial connection line connecting the dynamic load source and the static load target area with the lowest wave delay characteristics. The three-dimensional spatial connection line and the sequence of physical grids traversed are defined as the shortest time ray path for the dynamic load stress wave to propagate in the current mining space medium.

[0047] S230, using the energy dissipation module 30, constructs an exponentially decaying spatial integral equation along the shortest time ray path. Extracts the sum of energy released by microseismic events within the microseismic source cluster corresponding to the previously identified dynamic load source as the initial energy value. Discretizes the shortest time ray path into continuous micro-segment path variables. Combined with the impedance distribution function of each grid node along the path in the global initial wave impedance matrix, performs path integration to calculate the residual energy value corresponding to the dynamic stress wave penetrating the non-uniform medium of the coal and rock mass and reaching the boundary of the static target area. The calculation formula satisfies: ; In the formula: This represents the residual energy value of the dynamic stress wave reaching the static target area. This represents the initial energy released when the dynamic load source is triggered. It is a natural exponential function with the natural constant as its base; The starting point of the wavefront driven source Target point in static load area The set of shortest time ray paths in space; The dielectric-based energy attenuation coefficient is defined as 0.01 to 0.05 in this embodiment. Three-dimensional continuous spatial coordinates Wave impedance distribution function at point; The X-axis represents the continuous spatial coordinates of a three-dimensional Cartesian coordinate system. The Y-axis represents the continuous spatial coordinate value in a three-dimensional Cartesian coordinate system. The Z-axis represents the continuous spatial coordinate value in a three-dimensional Cartesian coordinate system. For the propagation path of the ray The differential length variable of the direction.

[0048] In specific implementation, step S300 provided by the present invention includes the following sub-steps: S310, using the parameter optimization module 40, sets the initial anti-scour scheme set to be planned, which includes multiple large-diameter pressure relief boreholes. For each borehole, the three-dimensional coordinates of the borehole opening, azimuth angle, dip angle, and borehole depth are configured as spatial parameter vectors. The search range of the three-dimensional coordinates of the borehole opening is limited to the spatial grid plane where the working face cut and the surface of the adjacent roadway are located to meet the physical boundary conditions of the actual drilling rig layout. Based on the spatial parameter vector, a virtual cylindrical geometric model representing the actual borehole is constructed inside the three-dimensional spatial grid. For each grid node in the global initial wave impedance matrix, the vertical Euclidean distance from the spatial coordinates to the central axis of the corresponding virtual cylindrical geometric model is calculated. Grid nodes with a vertical Euclidean distance less than the borehole design radius are selected to form a local disturbance set. The wave impedance values ​​of all grid nodes in the local disturbance set are forcibly replaced with one percent of the original values ​​to objectively reflect the actual medium energy transfer state after the rock mass is drilled and hollowed out.

[0049] S320: Using the parameter optimization module 40, the residual energy of the static target area calculated without drilling disturbance is extracted as the reference energy. The extremely low wave impedance value of the local disturbance set is substituted into the global initial wave impedance matrix to form the evolved impedance matrix. Based on the evolved impedance matrix, ray path optimization and path integration are re-executed to obtain the updated residual energy after drilling. The difference between the reference energy and the updated residual energy is calculated as the energy dissipation gradient. The hole depth parameters of all holes in the initial anti-impact scheme set are summarized to obtain the total drilling footage. Linear mapping calculation is performed based on the maximum and minimum values ​​of the parameters in the current iteration population. The total drilling footage and energy dissipation gradient are transformed into dimensionless values ​​distributed in the interval between zero and one. The objective function is constructed by combining the normalized energy dissipation gradient and the total drilling footage. The calculation formula of the objective function satisfies: ; In the formula: The minimum fitness value output by the objective function; The weighting constant for engineering cost control is 0.3 to 0.5 in this embodiment. The sorting index number variable for the individual pressure relief borehole; The total number of large-diameter pressure relief boreholes included in the initial anti-impact scheme set; For the first Hole depth parameters corresponding to each borehole; The dimensionless value of total borehole footage after extreme value mapping normalization; As the anti-impact and pressure relief efficiency weight constant, in this embodiment the value range of the anti-impact and pressure relief efficiency weight constant is 0.5 to 0.7, and the sum of the engineering cost control weight constant and the anti-impact and pressure relief efficiency weight constant is equal to 1; The reference residual energy value is obtained based on the global initial wave impedance matrix. The updated residual energy value is obtained based on the evolution impedance matrix calculation; This is the dimensionless value of the energy dissipation gradient after extreme value mapping normalization.

[0050] S330, the parameter optimization module 40 uses the particle swarm optimization algorithm to iteratively solve the objective function. The combination and encoding of all the borehole space parameter vectors contained in the initial anti-impact scheme set are encoded into the position vector of a single particle in the algorithm cluster. The search boundary conditions of the corresponding parameters are set according to the travel limit of the drilling equipment on site. The fitness value generated after substituting the current position vector of each particle into the objective function is calculated. According to the fitness value, the historical best position of the individual particle and the historical best position of the global population are continuously updated according to the generation until the number of iterations reaches the preset maximum generation. In this embodiment, the preset maximum generation ranges from 100 to 500 times. The historical best position of the global population at the end of the iteration is extracted and mathematically decoded to obtain the final three-dimensional space parameters of the anti-impact and pressure relief borehole, corresponding to the speed and position update rules and parameter boundary mapping of the particle swarm optimization algorithm.

[0051] See attached document Figure 4 In specific implementations, step S400 provided by the present invention may include the following sub-steps: S410 utilizes the spatial reconstruction module 50 to acquire the current borehole depth and corresponding drilling dip and azimuth parameters in real time through the drilling measurement system deployed on the downhole intelligent drilling rig. Based on the objective physical fact that the drill bit's trajectory deviates due to the rock strata dip and its own gravity, the minimum curvature method is used to convert the dip and azimuth obtained from two adjacent measuring points into spatial tangent vectors. The micro-bending of the drill bit between adjacent measuring points is fitted into a smooth three-dimensional spatial arc, and differential and integral operations are performed along the borehole depth direction to iteratively solve and obtain the true physical three-dimensional spatial coordinates of the drill bit at the current moment. The true borehole spatial curve fitted by discrete coordinate points is output and mapped back to the global initial wave impedance matrix to locate the sequence of three-dimensional spatial grid nodes actually penetrated by the drill bit.

[0052] S420 utilizes the spatial reconstruction module 50 to collect dynamic operating parameters such as thrust, torque, drill rod rotation speed, and actual drilling speed of the intelligent drilling rig in real time during the rock-breaking process of the drill bit. It combines the drill bit cross-sectional area constant to calculate the sum of the thrust work and the rotational torque work and divides it by the rock-breaking volume to obtain the mechanical specific energy consumed by the drill bit to break a unit volume of coal and rock mass at the current coordinate point. It extracts the limit fracture specific energy constant of the corresponding rock strata obtained during the geological exploration stage of the working face. Based on the continuous damage mechanics theory, it divides the above mechanical specific energy by the limit fracture specific energy constant to perform normalization mapping calculation, generating a rock mass mechanical damage variable that objectively reflects the degree of expansion of micro-cracks in the rock mass at the grid node. Through algorithm-level numerical truncation processing, the value range of the rock mass mechanical damage variable is strictly limited to the interval between 0 and 1.

[0053] S430, using the spatial reconstruction module 50, local acoustic impedance property reduction is performed on the grid nodes traversed by the actual trajectory based on the calculated rock mass mechanical damage variables. The initial three-dimensional acoustic impedance values ​​of the corresponding penetrating grid nodes in the original global initial acoustic impedance matrix are extracted. Combining the nonlinear coupling physical law of rock mass acoustic response characteristics and mechanical damage state, the effective medium theory is used to establish a local acoustic impedance dynamic evolution equation. The updated impedance value of the damaged grid is calculated and used to locally cover and replace the corresponding grid data in the original global initial acoustic impedance matrix. The calculation formula of the local acoustic impedance dynamic evolution equation satisfies: ; In the formula: Based on the local physical property reduction of the actual borehole trajectory, the corresponding three-dimensional continuous spatial coordinates Updated impedance value at the location; For corresponding three-dimensional continuous space coordinates The initial three-dimensional wave impedance distribution function value at the location; The X-axis represents the continuous spatial coordinates of a three-dimensional Cartesian coordinate system. The Y-axis represents the continuous spatial coordinate value in a three-dimensional Cartesian coordinate system. The Z-axis represents the continuous spatial coordinate value in a three-dimensional Cartesian coordinate system. For corresponding three-dimensional continuous space coordinates Rock mass mechanical damage variables at the location; To objectively reflect the degree of energy dissipation during the propagation of stress waves in a medium with microcracks, the acoustic damage sensitivity adjustment index in this embodiment is obtained by calibration based on the acoustic wave penetration attenuation experiment of the target coal and rock mass, and the conventional value range is between 1.5 and 3.0.

[0054] In specific implementations, step S500 provided by the present invention may include the following sub-steps: S510: Using the local correction module 60 to extract the real borehole space curve output by the spatial reconstruction module 50 as the reference center axis, the three-dimensional absolute static load stress value of the corresponding grid node in the global initial wave impedance matrix is ​​extracted and combined with the obtained internal friction angle and cohesion parameters of the rock mass in the target area. Based on the Mohr-Coulomb failure criterion, the elastic-plastic mechanical equilibrium equation of the surrounding rock is established, and the effective pressure relief influence radius of the plastic yield failure around the borehole is calculated. With the real borehole space curve as the axis and the effective pressure relief influence radius as the radial boundary, an irregular cylindrical envelope geometry is generated in the three-dimensional Cartesian coordinate system. All three-dimensional grid nodes in the global initial wave impedance matrix are traversed and the topological inclusion relationship with the envelope geometry is determined by the spatial distance. The three-dimensional spatial grid sequence that falls completely into the envelope geometry is extracted to form the spatial pressure relief influence envelope grid.

[0055] S520, using the local correction module 60, perform a Boolean intersection operation on the extracted spatial decompression influence envelope grid and the grid sequence contained in the shortest time ray path constructed in step S200. When the operation result determines that there are overlapping nodes in the two types of grid sets, extract the spatial parameter vector of the currently remaining unconstructed boreholes in the initial anti-scour scheme set as the initial high-quality solution for local iterative optimization. The initial high-quality solution is input into the local search algorithm to perform a hot start operation. Combining the actual movable physical travel range of the downhole drilling rig and the construction specifications for the spacing between adjacent pressure relief boreholes, the spatial boundary of the optimization calculation operation is strictly limited to the reachable domain three-dimensional grid around the envelope grid. When the calculation results determine that there are no overlapping nodes in the two types of mesh sets, it is determined that the current drilling construction has not caused topological changes in the energy conduction path. The local hot start operation is skipped and the original initial anti-scour scheme output in step S300 is directly used to guide the subsequent construction.

[0056] S530 utilizes the local correction module 60 to construct a local objective function aimed at maximizing the energy dissipation increment of the target area after the envelope grid update. Within the dimensionality-reduced domain, a gradient ascent algorithm combined with the finite difference principle is used to perform local search calculations. Before performing gradient calculations, the coordinate, angle, and depth parameters in the spatial parameter vector are mapped to a dimensionless space through maximum and minimum values. The partial derivative gradient of the objective function with respect to the spatial parameter vector is approximated based on the energy increment change difference caused by applying small parameter perturbations in each dimension. The partial derivative gradient direction is used for step-by-step derivation at a preset search step size until the iterative convergence condition is met. In this embodiment, the preset search step size ranges from 0.01 to 0.1. The converged dimensionless parameter vector is then de-normalized and decoded to calculate the spatial coordinate deflection and depth compensation of the actual physical dimensions required for adjacent boreholes and sends this information to the downhole intelligent drilling rig in real time. The local iterative update formula for the parameters of the corresponding adjacent boreholes satisfies: ; In the formula: For the first The spatial parameter vector of adjacent boreholes to be constructed, calculated in the next iteration; For the first The spatial parameter vector of adjacent boreholes to be constructed, calculated in the next iteration; The current iteration number is a discrete variable for the local optimization search algorithm; The discrete variable representing the next iteration step of the local optimization search algorithm; The preset iteration update step size is the gradient search algorithm. In this embodiment, the value range of the preset iteration update step size is 0.01 to 0.1. The energy dissipation increment is obtained by recalculating the path integral based on the updated local wave impedance; For the energy dissipation increment to the first The gradient of the partial derivative of the local rate of change of the spatial parameter vector in the next iteration.

[0057] In specific implementations, step S600 provided by the present invention may include the following sub-steps: S610 utilizes the performance evaluation and closed-loop evolution module 70 to extract and merge the final global wave impedance matrix after all large-diameter pressure relief boreholes in the initial anti-impact scheme set have been constructed and all local wave impedances have been updated. Based on the objective physical law of dynamic stress wave diffracting along the shortest path in the non-uniform fractured medium after being subjected to multi-hole joint disturbance, the fast travel algorithm is used again to solve the equation and recalculate the initial arrival time field from the dynamic source to the static target area. Based on the regenerated initial arrival time field, the latest shortest time ray path is obtained by reverse tracing and spatial exponential decay integral operation is performed along the path. The final residual energy value of the dynamic stress wave penetrating the composite damage area composed of all pressure relief boreholes and reaching the boundary of the static target area is calculated. The global recalculation process objectively reflects the changes in the spatial offset and attenuation characteristics of the energy dissipation path caused by the cumulative effect of multiple borehole constructions.

[0058] S620 utilizes the performance evaluation and closed-loop evolution module 70 to extract the elastic modulus parameters and absolute static stress values ​​of the coal and rock strata in the static load target area of ​​the target working face. Combined with the elastic strain energy release criterion, the elastic strain energy density of the coal and rock mass is calculated, and spatial integration is performed within the three-dimensional volume of the target area. The minimum critical energy required to induce rockburst damage in the induced area is calculated through inversion. A safety reserve coefficient conforming to the safety management level of the target mining area is extracted. The product of the calculated final residual energy value and the safety reserve coefficient is numerically compared with the minimum critical energy to establish a closed-loop feedback rockburst prevention performance evaluation criterion. The specific comparison relationship expression of the rockburst prevention performance evaluation criterion satisfies: ; In the formula: The final residual energy value is calculated by integrating the final global wave impedance matrix along the latest shortest time ray path. To meet the safety management level of the target mining area, the safety reserve coefficient in this embodiment ranges from 1.2 to 1.5. This refers to the minimum critical energy required for dynamic instability and failure of coal and rock strata in a statically loaded target area, obtained by inversion integral calculation based on the elastic strain energy release criterion.

[0059] S630, when the performance evaluation and closed-loop evolution module 70 meets the condition of being less than or equal to the above comparison judgment result, the final global wave impedance matrix and the latest shortest time ray path are extracted and a three-dimensional visualization cloud map is generated using a volume rendering algorithm for archiving and display. When the judgment result does not meet the condition, the system generates an anti-impact warning signal and sends it to the control terminal. At the same time, the final global wave impedance matrix is ​​used as the input environment and redirected to the parameter optimization module 40 to trigger a new round of supplementary pressure relief drilling scheme planning loop. When the number of supplementary loops reaches the preset maximum number of interventions, the automatic optimization is terminated and manual evaluation intervention is forcibly triggered. In this embodiment, the preset maximum number of interventions is set to a range of 3 to 5 times.

[0060] S640: After the completion of the anti-shock and pressure relief project, the performance evaluation and closed-loop evolution module 70 continuously collects the actual residual energy of newly occurring microseismic events passing through the composite damage area as measured values. It calculates the prediction deviation gradient between the measured values ​​and the predicted residual energy values ​​calculated based on the final global wave impedance matrix. When the prediction deviation gradient is greater than the preset tolerance threshold, the medium foundation energy attenuation coefficient in step S230 and the nonlinear strengthening coupling coefficient in step S130 are adjusted in the reverse direction along the gradient descent direction with the aforementioned preset search step size, so as to realize the adaptive iterative closed loop of the long-period system foundation parameters.

[0061] To aid in understanding the technical solution of this invention, the following provides an embodiment of an intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion.

[0062] In this application embodiment, the burial depth of the coal mining face is set to 920 meters. Due to the superposition of support pressure from the surrounding historical goaf areas, a high static load stress concentration zone is formed inside the solid coal. The frequency of high-energy dynamic load events in the early stage of the field microseismic network is 3 times per day. Conventional anti-scour measures use large-diameter boreholes with a linear spacing of 3 meters to relieve pressure. After the system of the present invention is deployed to the mine, a closed-loop intelligent prevention and control experiment is carried out.

[0063] The experimental verification steps include the following: By connecting to the mine microseismic network and the online stress monitoring subsystem, the coal mining face is divided into a three-dimensional spatial grid with a step length of 1.0m. The system constructs a global initial wave impedance matrix based on the initial stress data and lithological distribution.

[0064] The system extracts the clusters of microseismic events as dynamic load sources, marks the high wave impedance anomaly areas as static load target areas, and reverse-tracks to obtain the shortest time ray path connecting the dynamic load source and the static load target area.

[0065] A control group and an experimental group were established. The control group was guided by a layout scheme with a spacing of 3 meters, while the experimental group used the intelligent decision-making scheme of this invention to carry out local optimization and output the opening coordinates, inclination angle and hole depth of the pressure relief borehole to guide the construction of the intelligent drilling rig on site.

[0066] During the drilling process, the experimental group collected the drilling rig torque and thrust in real time and converted them into mechanical specific energy. They calculated the mechanical damage variables of the rock mass and locally reduced the grid wave impedance. When the actual borehole trajectory was deflected, causing the wave impedance update to fail to cover the shortest time ray path, the system triggered a hot start local correction and generated the compensation angle and depth of the subsequent boreholes, which were then sent to the drilling rig.

[0067] After the pressure relief drilling is completed, the final residual energy of the static load target area is recalculated, and the physical microseismic response data and stress gauge readings within 30 days after completion are extracted.

[0068] The experimental verification data comparison results are shown in the table below: Table 1. Comparison of decompression construction parameters and energy evolution:

[0069] Table 2 Comparison of measured data on on-site monitoring effects within 30 days after implementation of the plan:

[0070] From Table 1, Table 2 and appendix Figure 5 Appendix Figure 6 Appendix Figure 7 As can be seen, compared with the traditional spacing scheme, the intelligent decision-making scheme of this invention reduces the total drilling footage from 561.4m to 342.7m. The intelligent decision-making scheme of this invention configures 5 local dynamic corrections while drilling, which reduces the final residual energy of the target area after completion from 47319.2J to 18504.8J. Within 30 days after implementation, the frequency of events with energy greater than 10000J is reduced from 17 times to 2 times. The maximum single micro-vibration energy release is 14932.6J, and the average stress reduction of the coal pillar in the target area reaches 7.84MPa. The average stress reduction of the coal pillar in the target area of ​​the traditional spacing scheme is 2.51MPa. The three-dimensional spatial parameters of the borehole output by this invention change the state of the energy transfer medium and reduce the frequency of micro-vibration events and the micro-vibration energy release parameters.

[0071] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A smart decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion, characterized in that, include: The multi-source data acquisition module is used to acquire geological parameters, microseismic event parameters, and online stress monitoring data of the mining face; The initial wave impedance modeling module is used to discretize the mining space into a three-dimensional spatial grid, establish the coupling relationship between stress and wave velocity based on the geological parameters, microseismic event parameters and online stress monitoring data, and construct a global initial wave impedance matrix. The energy dissipation analysis module is used to extract the static load target area from the global initial wave impedance matrix, extract the dynamic load source in combination with the microseismic event parameters, and construct an energy attenuation integral model along the shortest time ray path connecting the dynamic load source and the static load target area to quantify the residual energy. The prevention and control parameter optimization module is used to set the pressure relief borehole as a disturbance body, call the energy attenuation integral model to calculate the energy change after disturbance, construct the objective function with the minimum total borehole footage and the maximum energy dissipation gradient, and iteratively optimize and output the three-dimensional spatial parameters of the planned anti-impact pressure relief borehole. The drilling space dynamic reconstruction module is used to extract the drilling attitude and dynamic parameters, calculate the rock mass mechanical damage variables, generate the real drilling trajectory, and locally update the global initial wave impedance matrix based on the rock mass mechanical damage variables to form the drilling evolution wave impedance matrix. The hot-start local correction module is used to extract the envelope mesh around the real borehole trajectory, perform local gradient search in combination with the drilling evolution impedance matrix, and calculate the parameter compensation amount of the remaining unconstructed boreholes to correct the subsequent three-dimensional spatial parameters. The performance evaluation and closed-loop evolution module is used to obtain the final global wave impedance matrix after all pressure relief drilling is completed, call the energy attenuation integral model to calculate the final residual energy and perform anti-impact performance evaluation, and update the basic parameters in the wave impedance initial modeling module and energy dissipation analysis module based on the evaluation results.

2. The intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion as described in claim 1, characterized in that, The wave impedance initial modeling module is specifically used for: Based on the geological parameters and microseismic event parameters, the grid nodes of the three-dimensional spatial grid are assigned initial values ​​for microseismic P-wave velocity and coal-rock mass density. The absolute static load stress value at the grid node is extracted from the online stress monitoring data, and the absolute static load stress value is dimensionless by using the standard value of the uniaxial compressive strength of the coal and rock mass in the target area. The initial three-dimensional wave impedance value of the grid node is calculated by combining the preset nonlinear enhanced coupling coefficient of the effect of static load stress on the wave impedance property, and then the global initial wave impedance matrix is ​​constructed.

3. The intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion as described in claim 1, characterized in that, The energy dissipation analysis module is specifically used for: Calculate the statistical mean and standard deviation of the impedance values ​​of the grid nodes in the global initial wave impedance matrix, and extract the connected regions that are higher than the judgment threshold as the static load target area; Density clustering algorithm is used to identify spatially dense clusters of microseismic sources in the microseismic event parameters and to extract them as dynamic load sources through weighted calculation; The fast traversal algorithm is used to solve the process function equation to calculate the initial arrival time field distribution, and reverse tracing is performed along the time gradient descent direction to obtain the shortest time ray path connecting the dynamic source and the static target area.

4. The intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion as described in claim 1, characterized in that, The construction of an energy decay integral model along the shortest time ray path connecting the dynamic source and the static target region to quantify the residual energy specifically includes: The shortest time ray path is discretized into micro-segment path variables. Then, an exponential decay spatial integral operation is performed by combining the preset medium-based energy attenuation coefficient and the wave impedance distribution function of the grid nodes along the line to solve for the residual energy value of the dynamic stress wave penetrating the coal and rock medium and reaching the boundary of the static load target area.

5. The intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion as described in claim 1, characterized in that, The prevention and control parameter optimization module is specifically used for: Based on the configured spatial parameter vector containing coordinates, angles and depth, a virtual cylindrical geometric model is constructed inside the three-dimensional spatial grid as the disturbance body. The grid nodes located within the design radius are extracted to form a local disturbance set, and the wave impedance value in the global initial wave impedance matrix is ​​reduced to form an evolved impedance matrix. The evolution impedance matrix is ​​input into the energy decay integral model to re-execute the path integral operation to obtain the updated residual energy, and the objective function is constructed using the dimensionless value of the total drilling footage and the dimensionless value of the energy dissipation gradient. An optimization algorithm is used to iteratively solve the objective function, extract the global population's historical optimal position, and decode to obtain the three-dimensional spatial parameters of the planned anti-scour and pressure-relief borehole.

6. The intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion according to claim 1, characterized in that, The drilling-while-drilling spatial dynamic reconstruction module is specifically used for: calculating rock mass mechanical damage variables and generating the actual borehole trajectory. The drilling dip angle and azimuth angle parameters are extracted from the drilling attitude and dynamic parameters, and the minimum curvature method is used to transform them to obtain the real physical three-dimensional space coordinates of the drill bit. The real drilling trajectory is then fitted to generate the real drilling trajectory. Thrust and torque parameters are extracted from the drilling attitude and dynamic parameters to calculate the mechanical specific energy consumed in rock breaking; The mechanical specific energy is normalized and mapped using the regional corresponding rock strata limit fracture specific energy constant to generate the rock mass mechanical damage variable distributed in the interval of zero to one.

7. The intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion according to claim 6, characterized in that, The drilling-while-space dynamic reconstruction module is specifically used for: When locally updating the global initial wave impedance matrix, the drilling-while-space dynamic reconstruction module is used for: Combining the rock mass mechanical damage variables with the preset acoustic damage sensitivity adjustment index, the updated impedance value of the damaged grid is calculated using the local wave impedance dynamic evolution equation based on the effective medium theory, and the grid data corresponding to the actual borehole trajectory penetration in the original global initial wave impedance matrix is ​​replaced to form the drilling evolution wave impedance matrix.

8. The intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion as described in claim 1, characterized in that, The hot-start local correction module is specifically used for: The effective stress relief radius around the actual borehole trajectory that causes plastic yield failure is calculated based on the Mohr-Coulomb failure criterion, and the envelope mesh is generated accordingly. Perform a Boolean intersection operation between the envelope grid and the grid sequence contained in the shortest time ray path generated by the energy dissipation analysis module. When there are overlapping nodes, use the spatial parameter vector of the remaining unconstructed boreholes as the initial high-quality solution input to the local search algorithm. Within the optimization and dimensionality reduction constraint domain, the partial derivative gradient is calculated based on the energy increment change difference brought about by the drilling evolution wave impedance matrix. The gradient ascent algorithm is used to perform step-by-step deduction according to the preset search step size to solve the spatial coordinate deflection and depth compensation of adjacent boreholes to be constructed.

9. The intelligent decision-making system for preventing and controlling rockbursts in coal mines based on multi-source information fusion as described in claim 1, characterized in that, The performance evaluation and closed-loop evolution module is specifically used for: When performing anti-impact performance evaluation, it is used for: Based on the final global impedance matrix, reverse tracking and integration are performed again to obtain the final residual energy value of the moving carrier at the boundary of the static target area. The minimum critical energy for rockburst failure in the target area is calculated by combining the elastic strain energy release criterion. When the product of the final residual energy value and the safety reserve coefficient is greater than the minimum critical energy, an anti-impact warning signal is triggered and redirected to the prevention and control parameter optimization module to trigger a supplementary pressure relief drilling planning cycle.

10. The intelligent decision-making system for preventing and controlling coal mine rockbursts based on multi-source information fusion according to claim 9, characterized in that, The performance evaluation and closed-loop evolution module is specifically used for updating basic parameters as follows: After the completion of the regional anti-scour and depressurization project, the actual residual wave frequency energy of newly occurring microseismic events penetrating the depressurization zone is extracted; The prediction deviation gradient between the actual residual energy and the final residual energy predicted by the model is calculated. When the prediction deviation gradient exceeds the limit, the dielectric energy attenuation coefficient and the nonlinear strengthening coupling coefficient are adjusted in reverse adaptively.

Images