Coal mine dynamic disaster drilling blasting optimization method based on submodel stress inheritance

Abstract

The invention provides a coal mine dynamic disaster drilling blasting optimization method based on submodel stress inheritance. Comprising the steps of global model and stress result generation, sub-model cutting and stress inheriting, borehole parameterization information modification, TNT explosive parameter modeling, JWL state equation loading, high-precision structural grid division, model operation to generate a stress nephogram and data extraction, digital twin construction and AI proxy model training and optimization decision. And an optimized drilling parameter generation and parameter feedback and iterative optimization step. According to the method, a dynamic and accurate blasting pressure relief prevention and control scheme is provided by constructing a digital twinborn coupled with a global-local model, extracting blasting stress response, establishing an AI proxy model in combination with a deep neural network and performing dynamic decision and closed-loop optimization on key factors such as blasting drilling arrangement and explosive parameters; the blasting pressure relief effect and the rock burst prevention and control capacity are effectively improved, and safe, efficient and intelligent technical support is provided for rock burst disaster prevention and control.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent prevention and control of coal mine disasters, and in particular to a coal mine dynamic disaster drilling and blasting optimization method based on sub-model stress inheritance. Background Art

[0002] Rock burst, one of the most sudden and devastating dynamic hazards in coal mining, poses a serious threat to mine safety and the lives of workers. Rock burst is particularly frequent, severe, and unpredictable in deep, high-stress, and complex geological structures, where rock mass stresses accumulate and coal and rock masses deform and fail severely. To achieve early regional energy release, localized stress reduction, and lower coal and rock stiffness, drilling and blasting, borehole unloading, and hydraulic fracturing are widely used in current engineering practices.

[0003] Drilling and blasting pressure relief is a highly efficient and rapid energy release method commonly used for proactive prevention and control within impact-prone areas. The principle is to deploy blasting drill holes and implement instantaneous energy release, thereby promoting localized coal and rock destruction and inducing stress redistribution, thereby reducing the probability of impact. However, in actual application, the inventors of this application have discovered through research that existing drilling and blasting pressure relief technology still has the following technical bottlenecks:

[0004] 1. Lack of multi-scale accurate modeling capabilities: Traditional blasting design is often based on simplified two-dimensional models or empirical analogies, which cannot fully reflect the actual geological structure, initial stress state and its evolution process of the mining area. As a result, the blasting parameter design lacks scientificity and specificity, and is prone to secondary disturbance or insufficient pressure relief effect.

[0005] 2. Blasting simulations fail to replicate real-world conditions: In most current numerical simulation methods, the blasting analysis model is disconnected from the global mine mechanical model, making it impossible for the sub-model to accurately replicate the global stress field. Even when local regions are extracted for blasting simulation, boundary stresses are often simplified, ignoring the impact of real-world stress evolution on blasting response. This reduces the simulation's credibility and optimization effectiveness.

[0006] 3. Blasting parameters rely on experience: Key blasting parameters such as drilling layout, explosive loading, detonation sequence, and charge configuration often rely on the experience of on-site personnel or repeated debugging through experiments. They lack a systematic optimization mechanism and cannot effectively adapt to the safety and control needs under different geological conditions.

[0007] 4. Lack of data feedback and intelligent adjustment mechanism: Current blasting operations generally lack effective real-time stress feedback and pressure relief response evaluation mechanisms, making it impossible to dynamically adjust parameters based on numerical simulation results and measured data. This leads to the problem of "passive execution and difficulty in iterative optimization."

[0008] 5. The intelligent optimization mechanism is still imperfect: At present, there is no intelligent optimization system that deeply integrates multi-source simulation data, blasting stress response and deep learning algorithms. There is a lack of agent models that can be used to predict blasting effects and reversely optimize parameters, which seriously restricts the improvement of the quality and efficiency of blasting pressure relief effects.

[0009] In summary, a new intelligent drilling and blasting optimization method is urgently needed. Based on the inheritance of global stress information, a local blasting sub-model is constructed, and high-precision explosion physics modeling (such as the TNT model and the JWL state equation) is integrated with three-dimensional simulation and artificial intelligence algorithms to achieve accurate simulation, dynamic feedback and parameter optimization of the entire blasting decompression process. Summary of the Invention

[0010] In response to the technical problems in existing drilling and blasting pressure relief methods, such as the difficulty in accurately predicting blasting effects, insufficient optimization of pressure relief parameters, and reliance on experience-based configuration, the present invention provides a coal mine dynamic disaster drilling and blasting optimization method based on sub-model stress inheritance. By constructing a digital twin of a coupled global-local model, extracting the blasting stress response, and combining a deep neural network to establish an AI agent model, dynamic decision-making and closed-loop optimization are carried out on key factors such as blasting drilling arrangement and explosive parameters. A dynamic and accurate blasting pressure relief prevention and control solution is provided, which effectively improves the blasting pressure relief effect and impact ground pressure prevention and control capabilities, provides safe, efficient, and intelligent technical support for impact ground pressure disaster prevention and control, and ensures the safety of coal mine mining.

[0011] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0012] The optimization method of drilling and blasting for coal mine dynamic disasters based on sub-model stress inheritance includes the following steps:

[0013] S1. Global model and stress result generation:

[0014] S11. Collect geological structural parameters of the coal mine area and construct a global geomechanical model of the coal mine area. The geological structural parameters include rock layer thickness, elastic modulus E, Poisson's ratio v, density ρ, initial stress tensor σ0, geometric boundary conditions, and structural joint information.

[0015] S12. Use finite element numerical simulation software to perform static or quasi-static analysis on the global geomechanical model, calculate the distribution of the ground stress field, and obtain the global stress result σ global (x,y,z), the stress results of a single node are expressed by the following stress tensor expression:

[0016]

[0017] Among them, σ i is the stress tensor of the i-th element or node; σ xx, σ yy , σ zz is the stress component on the main diagonal, representing the normal stress acting in the x, y, and z directions; τ xy , τ xz , τ yx , τ yz , τ zx , τ zy is the stress component on the non-diagonal line, indicating the shear force;

[0018] σ global (x,y,z)=σ i {σ i |i=1,2,3…,N}

[0019] S2. Submodel cutting and stress inheritance:

[0020] S21. Based on the actual mining area and drilling layout, and based on the global geomechanical model in step S1, a local sub-model area is cut out in the high-risk area for rock burst. The sub-model area covers the drilling impact range and nearby key risk areas.

[0021] S22, the sub-model inherits the boundary stress field of the global geomechanical model as the boundary condition input, and is used to carry out blasting simulation based on the spatial distribution of the original ground stress state. The sub-model stress field σ sub (x,y,z) is guaranteed to be consistent with σ through interpolation and mapping global (x, y, z) are consistent, and dynamic modification of working conditions is allowed;

[0022] S3. Modification of drilling parameter information:

[0023] According to the existing drilling and blasting design information, load the location coordinates (x i ,y i ), aperture Φ i , depth l i , spacing d i The sub-model in step S2 is parametrically modified based on the drilling parameters, including the drilling parameters; the drilling geometry is embedded in the sub-model grid using geometric modeling tools; the model boundary conditions and material properties are modified to simulate the impact of the drilling on ground stress and structural integrity;

[0024] S4. TNT explosive parameter modeling:

[0025] In step S3, an explosive filling model is established inside the borehole of the sub-model. The explosion source area adopts a cylindrical or spherical filling domain. The explosive type is trinitrotoluene (TNT). The explosion parameters include density ρ TNT , initial internal energy E0, detonation velocity D;

[0026] S5, JWL state equation loading:

[0027] The Jones-Wilkins-Lee (JWL) equation of state is defined for the TNT material model to express the pressure-volume relationship of the explosive gas:

[0028]

[0029] Where P is the pressure of the gas generated by the explosion of the explosive; V is the relative volume, that is, the initial volume ratio; E is the internal energy per unit volume; A, B, R1, R2, and ω are TNT material constants;

[0030] S6. High-precision structural meshing:

[0031] For the sub-model adjusted in step S5, structured meshing technology is used to achieve high-precision meshing, and the mesh size is adaptively refined according to the drill hole size and stress gradient area to ensure calculation accuracy;

[0032] S7. Run the model to generate stress cloud diagram and extract data:

[0033] S71, running a blasting simulation on the sub-model after meshing in step S6, and outputting stress response data in the simulation time series;

[0034] S72. Extract the coordinates and stress values ​​of key nodes after blasting to form the following data set:

[0035] D blast ={(x j ,y j ,z j ,σ j )|j=1,2,…,N}

[0036] Among them, (x j ,y j ,z j ) represents the three-dimensional space coordinates of node j, σ j represents the stress value of node j;

[0037] Use cloud maps and post-processing technology to analyze stress concentration and changes in key areas;

[0038] S8. Digital Twin Construction:

[0039] S81, input the point coordinates and stress value I1 in step S7 = {(x j ,y j ,z j ,σ j )};

[0040] Structuring discrete stress data into a three-dimensional matrix through three-dimensional interpolation methods Represents a structured three-dimensional tensor, which represents the stress value of each voxel in a three-dimensional space grid;

[0041] A three-dimensional convolutional neural network is used to extract features and obtain a high-dimensional feature tensor F 3D =3DCNN(M 3D ), 3DCNN represents three-dimensional convolutional neural network;

[0042] Use 3D pooling layer to reduce dimension, reduce feature dimension, and obtain dimensionality reduction features

[0043] The reduced dimensionality features are further extracted through a two-dimensional convolutional neural network to obtain a two-dimensional feature tensor. 2DCNN stands for two-dimensional convolutional neural network;

[0044] S82, input drilling parameter matrix I2 = {(x i ,y i ,Φ i ,l i ,d i )};

[0045] Map the drilling parameter matrix into a two-dimensional matrix format and input it into 2DCNN to extract the feature F hole =2DCNN(I2);

[0046] S83. Feature fusion and matrix generation:

[0047] The feature F of step S81 2D and feature F of step S82 hole Perform feature fusion F fuse =F use (F 2D ,F hole ), F use Represents the feature fusion operation, which is to stack the features and then perform convolution fusion;

[0048] Generate a 2D stress-blast relief matrix f represents the fused output feature tensor;

[0049] Perform a three-dimensional deconvolution operation on the two-dimensional stress-blasting relief matrix to reconstruct the three-dimensional stress-blasting relief matrix Deconv means deconvolution;

[0050] S9. AI agent model training and optimization decision-making:

[0051] S91, using the existing drilling parameterization information and the actual simulation result D obtained in step S7. blast Construct the following AI agent model FAI :

[0052]

[0053] S92. Use supervised learning to train the model by minimizing the prediction error loss function:

[0054]

[0055] in, represents the pressure relief effect index of the i-th sample predicted by the surrogate model; represents the real pressure relief effect index of the i-th sample;

[0056] The three-dimensional stress-blasting relief matrix reconstructed by the digital twin in step S8 As training data for AI agent models;

[0057] S10. Generation of optimized drilling parameters:

[0058] Based on the trained AI agent model and multi-objective optimization algorithm, the optimal drilling and blasting parameter combination is searched.

[0059] S11. Parameter feedback and iterative optimization:

[0060] The optimization parameters generated in step S10 are fed back to step S3, the sub-model working conditions are modified, the blasting simulation is rerun, and steps S3 to S10 are repeated until the model converges, and the optimal drilling and blasting design parameters are finally output.

[0061] Furthermore, the meshing in step S6 utilizes volume units or hexahedron-dominated meshes to ensure simulation stability.

[0062] Compared with the existing technology, the coal mine dynamic disaster drilling and blasting optimization method based on sub-model stress inheritance provided by the present invention accurately simulates the blasting process through the combination of digital twin modeling and JWL state equation and performs dynamic optimization through artificial intelligence algorithm, which has the following significant advantages: 1) Accurate simulation: Based on the TNT blasting model and JWL state equation, it can accurately calculate the blasting load and stress wave propagation, and optimize the blasting unloading effect. 2) Intelligent decision-making: The AI ​​agent model can intelligently adjust the blasting parameters according to the real-time data of the mine, reduce dependence on traditional experience, and improve the scientific nature of the solution. 3) Efficient optimization: Compared with the existing methods, the present invention can realize parameter optimization in multiple scenarios, reduce the probability of rock burst, and improve mine operation safety. Therefore, this method improves the scientific nature and accuracy of blasting unloading, effectively reduces the occurrence of rock burst accidents, and ensures safe and stable mining of the mine. DETAILED DESCRIPTION

[0063] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the present invention is further described below.

[0064] The present invention provides a method for optimizing drilling and blasting for coal mine dynamic disasters based on sub-model stress inheritance, comprising the following steps:

[0065] S1. Global model and stress result generation:

[0066] S11. Collect geological structural parameters of the coal mine area and construct a global geomechanical model of the coal mine area. The geological structural parameters include rock layer thickness, elastic modulus E, Poisson's ratio v, density ρ, initial stress tensor σ0, geometric boundary conditions, and structural joint information;

[0067] S12. Use finite element numerical simulation software to perform static or quasi-static analysis on the global geomechanical model, calculate the distribution of the ground stress field, and obtain the global stress result σ global (x,y,z), the stress results of a single node are expressed by the following stress tensor expression:

[0068]

[0069] Among them, σ i is the stress tensor of the i-th element or node; σ xx , σ yy , σ zz is the stress component on the main diagonal, representing the normal stress acting in the x, y, and z directions; τ xy , τ xz , τ yx , τ yz , τ zx , τ zy is the stress component on the non-diagonal line, indicating the shear force;

[0070] σ global (x,y,z)=σ i {σ i |i=1,2,3…,N}.

[0071] S2. Submodel cutting and stress inheritance:

[0072] S21. Based on the actual mining area and drilling layout, and based on the global geomechanical model in step S1, a local sub-model area is cut out in the high-risk area for rock burst. The sub-model area covers the drilling impact range and nearby key risk areas.

[0073] S22, the sub-model inherits the boundary stress field of the global geomechanical model as the boundary condition input, and is used to carry out blasting simulation based on the spatial distribution of the original ground stress state. The sub-model stress field σ sub (x,y,z) is guaranteed to be consistent with σ through interpolation and mapping global (x, y, z) are consistent, while allowing dynamic modification of working conditions; unlike existing technologies, sub-model working conditions are dynamically modified and inherit stresses to achieve more accurate simulation.

[0074] S3. Modification of drilling parameter information:

[0075] According to the existing drilling and blasting design information, load the location coordinates (x i ,y i ), aperture Φ i , depth l i , spacing d i The sub-model in step S2 is parametrically modified based on the drilling parameters including the following: embed the drilling geometry into the sub-model mesh using geometric modeling tools such as Abaqus; modify the model boundary conditions and material properties to simulate the impact of drilling on ground stress and structural integrity.

[0076] S4. TNT explosive parameter modeling:

[0077] In step S3, an explosive filling model is established inside the borehole of the sub-model. The explosion source area adopts a cylindrical or spherical filling domain. The explosive type is trinitrotoluene (TNT). The explosion parameters include density ρ TNT , initial internal energy E0, detonation velocity D.

[0078] S5, JWL state equation loading:

[0079] The Jones-Wilkins-Lee (JWL) equation of state is defined for the TNT material model to express the pressure-volume relationship of the explosive gas:

[0080]

[0081] Among them, P is the pressure of the gas generated by the explosion of the explosive; V is the relative volume, that is, the initial volume ratio; E is the internal energy per unit volume; A, B, R1, R2, and ω are TNT material constants.

[0082] S6. High-precision structural meshing:

[0083] For the sub-model adjusted in step S5, a structured meshing technique is used to achieve high-precision meshing. The mesh size is adaptively refined based on the drill hole size and stress gradient area to ensure calculation accuracy. Preferably, the meshing is performed using volume elements or hexahedron-dominated meshes to ensure simulation stability.

[0084] S7. Run the model to generate stress cloud diagram and extract data:

[0085] S71, running a blasting simulation on the sub-model after meshing in step S6, and outputting stress response data in the simulation time series;

[0086] S72. Extract the coordinates and stress values ​​of key nodes after blasting to form the following data set:

[0087] D blast ={(x j ,y j ,z j ,σ j )|j=1,2,…,N}

[0088] Among them, (x j ,y j ,z j ) represents the three-dimensional space coordinates of node j, σ j represents the stress value of node j;

[0089] Use cloud maps and post-processing technology to analyze stress concentration and changes in key areas.

[0090] S8. Digital Twin Construction:

[0091] S81, input the point coordinates and stress value I1 in step S7 = {(x j ,y j ,z j ,σ j )};

[0092] Structuring discrete stress data into a three-dimensional matrix through three-dimensional interpolation methods Represents a structured three-dimensional tensor, which represents the stress value of each voxel in a three-dimensional space grid;

[0093] A three-dimensional convolutional neural network is used to extract features and obtain a high-dimensional feature tensor F 3D =3DCNN(M 3D ), 3DCNN represents three-dimensional convolutional neural network;

[0094] Use 3D pooling layer to reduce dimension, reduce feature dimension, and obtain dimensionality reduction features

[0095] The reduced dimensionality features are further extracted through a two-dimensional convolutional neural network to obtain a two-dimensional feature tensor. 2DCNN stands for two-dimensional convolutional neural network;

[0096] Through the above process, step S81 can achieve structural processing of the stress results of the sub-model;

[0097] S82, input drilling parameter matrix I2 = {(x i ,y i ,Φ i ,l i ,d i )};

[0098] Map the drilling parameter matrix into a two-dimensional matrix format and input it into 2DCNN to extract the feature F hole =2DCNN(I2);

[0099] Through the above process, step S82 can realize the extraction of drilling parameter features;

[0100] S83. Feature fusion and matrix generation:

[0101] The feature F of step S81 2D and feature F of step S82 hole Perform feature fusion F fuse =F use (F 2D ,F hole ), F use Represents the feature fusion operation, which is to stack the features and then perform convolution fusion;

[0102] Generate a 2D stress-blast relief matrix f represents the fused output feature tensor;

[0103] Perform a three-dimensional deconvolution operation on the two-dimensional stress-blasting relief matrix to reconstruct the three-dimensional stress-blasting relief matrix Deconv means deconvolution.

[0104] S9. AI agent model training and optimization decision-making:

[0105] S91, using the existing drilling parameterization information and the actual simulation result D obtained in step S7. blast Construct the following AI agent model F AI :

[0106]

[0107] S92. Use supervised learning to train the model by minimizing the prediction error loss function:

[0108]

[0109] in, represents the pressure relief effect index of the i-th sample predicted by the surrogate model; represents the real pressure relief effect index of the i-th sample;

[0110] The three-dimensional stress-blasting relief matrix reconstructed by the digital twin in step S8 As training data for AI agent models.

[0111] S10. Generation of optimized drilling parameters:

[0112] Based on the trained AI agent model and multi-objective optimization algorithms (such as genetic algorithm, Bayesian optimization, etc.), search for the optimal drilling and blasting parameter combination

[0113] S11. Parameter feedback and iterative optimization:

[0114] The optimization parameters generated in step S10 are fed back to step S3, the sub-model working conditions are modified, the blasting simulation is rerun, and steps S3 to S10 are repeated (in this iterative cycle) until the model converges and the optimal drilling and blasting design parameters are finally output.

[0115] Compared with the existing technology, the coal mine dynamic disaster drilling and blasting optimization method based on sub-model stress inheritance provided by the present invention accurately simulates the blasting process through the combination of digital twin modeling and JWL state equation and performs dynamic optimization through artificial intelligence algorithm, which has the following significant advantages: 1) Accurate simulation: Based on the TNT blasting model and JWL state equation, it can accurately calculate the blasting load and stress wave propagation, and optimize the blasting unloading effect. 2) Intelligent decision-making: The AI ​​agent model can intelligently adjust the blasting parameters according to the real-time data of the mine, reduce dependence on traditional experience, and improve the scientific nature of the solution. 3) Efficient optimization: Compared with the existing methods, the present invention can realize parameter optimization in multiple scenarios, reduce the probability of rock burst, and improve mine operation safety. Therefore, this method improves the scientific nature and accuracy of blasting unloading, effectively reduces the occurrence of rock burst accidents, and ensures safe and stable mining of the mine.

[0116] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the purpose and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A drilling and blasting optimization method for coal mine dynamic disasters based on sub-model stress inheritance, characterized by: The following steps are involved: S1. Global model and stress result generation: S11. Collect geological structural parameters of the coal mine area and construct a global geomechanical model of the coal mine area. The geological structural parameters include rock layer thickness, elastic modulus E, Poisson's ratio v, density ρ, initial stress tensor σ0, geometric boundary conditions, and structural joint information. S12. Use finite element numerical simulation software to perform static or quasi-static analysis on the global geomechanical model, calculate the distribution of the ground stress field, and obtain the global stress result σ global (x,y,z), the stress results of a single node are expressed by the following stress tensor expression: Among them, σ i is the stress tensor of the i-th element or node; σ xx , σ yy , σ zz is the stress component on the main diagonal, representing the normal stress acting in the x, y, and z directions; τ xy , τ xz , τ yx , τ yz , τ zx , τ zy is the stress component on the non-diagonal line, indicating the shear force; s global (x,y,z)=σ i {s} i |i=1,2,3…,N} S2. Submodel cutting and stress inheritance: S21. Based on the actual mining area and drilling layout, and based on the global geomechanical model in step S1, a local sub-model area is cut out in the high-risk area for rock burst. The sub-model area covers the drilling impact range and nearby key risk areas. S22, the sub-model inherits the boundary stress field of the global geomechanical model as the boundary condition input, and is used to carry out blasting simulation based on the spatial distribution of the original ground stress state. The sub-model stress field σ sub (x,y,z) is guaranteed to be consistent with σ through interpolation and mapping global (x, y, z) are consistent, and dynamic modification of working conditions is allowed; S3. Modification of drilling parameter information: According to the existing drilling and blasting design information, load the location coordinates (x i ,y i ), aperture Φ i , depth l i , spacing d i The sub-model in step S2 is parametrically modified based on the drilling parameters, including the drilling parameters; the drilling geometry is embedded in the sub-model grid using geometric modeling tools; the model boundary conditions and material properties are modified to simulate the impact of the drilling on ground stress and structural integrity; S4. TNT explosive parameter modeling: In step S3, an explosive filling model is established inside the borehole of the sub-model. The explosion source area adopts a cylindrical or spherical filling domain. The explosive type is trinitrotoluene (TNT). The explosion parameters include density ρ TNT , initial internal energy E0, detonation velocity D; S5, JWL state equation loading: The Jones-Wilkins-Lee (JWL) equation of state is defined for the TNT material model to express the pressure-volume relationship of the explosive gas: Where P is the pressure of the gas generated by the explosion of the explosive; V is the relative volume, that is, the initial volume ratio; E is the internal energy per unit volume; A, B, R1, R2, and ω are TNT material constants; S6. High-precision structural meshing: For the sub-model adjusted in step S5, structured meshing technology is used to achieve high-precision meshing, and the mesh size is adaptively refined according to the drill hole size and stress gradient area to ensure calculation accuracy; S7. Run the model to generate stress cloud diagram and extract data: S71, running a blasting simulation on the sub-model after meshing in step S6, and outputting stress response data in the simulation time series; S72. Extract the coordinates and stress values ​​of key nodes after blasting to form the following data set: D blast {(x j ,y j ,z j ,σ j )|j=1,2,…,N} Among them, (x j ,y j ,z j ) represents the three-dimensional space coordinates of node j, σ j represents the stress value of node j; Use cloud maps and post-processing technology to analyze stress concentration and changes in key areas; S8. Digital Twin Construction: S81, input the point coordinates and stress value I1 in step S7 = {(x j ,y j ,z j ,σ j )}; Structuring discrete stress data into a three-dimensional matrix through three-dimensional interpolation methods Represents a structured three-dimensional tensor, which represents the stress value of each voxel in a three-dimensional space grid; A three-dimensional convolutional neural network is used to extract features and obtain a high-dimensional feature tensor F 3D =3DCNN(M 3D ), 3DCNN represents three-dimensional convolutional neural network; Use 3D pooling layer to reduce dimension, reduce feature dimension, and obtain dimensionality reduction features The reduced dimensionality features are further extracted through a two-dimensional convolutional neural network to obtain a two-dimensional feature tensor. 2DCNN stands for two-dimensional convolutional neural network; S82, input drilling parameter matrix I2 = {(x i ,y i ,Φ i ,l i ,d i )}; Map the drilling parameter matrix into a two-dimensional matrix format and input it into 2DCNN to extract the feature F hole =2DCNN(I2); S83. Feature fusion and matrix generation: The feature F of step S81 2D and feature F of step S82 hole Perform feature fusion F fuse =F use (F 2D ,F hole ), F use Represents the feature fusion operation, which is to stack the features and then perform convolution fusion; Generate a 2D stress-blast relief matrix f represents the fused output feature tensor; Perform a three-dimensional deconvolution operation on the two-dimensional stress-blasting relief matrix to reconstruct the three-dimensional stress-blasting relief matrix Deconv means deconvolution; S9. AI agent model training and optimization decision-making: S91, using the existing drilling parameterization information and the actual simulation result D obtained in step S7. blast Construct the following AI agent model F AI : S92. Use supervised learning to train the model by minimizing the prediction error loss function: in, represents the pressure relief effect index of the i-th sample predicted by the surrogate model; represents the real pressure relief effect index of the i-th sample; The three-dimensional stress-blasting relief matrix reconstructed by the digital twin in step S8 As training data for AI agent models; S10. Generation of optimized drilling parameters: Based on the trained AI agent model and multi-objective optimization algorithm, the optimal drilling and blasting parameter combination is searched. S11. Parameter feedback and iterative optimization: The optimization parameters generated in step S10 are fed back to step S3, the sub-model working conditions are modified, the blasting simulation is rerun, and steps S3 to S10 are repeated until the model converges, and the optimal drilling and blasting design parameters are finally output.

2. The method for optimizing drilling and blasting for coal mine dynamic disasters based on sub-model stress inheritance according to claim 1 is characterized in that: In step S6, the meshing is performed using volume elements or hexahedron-dominated meshes to ensure simulation stability.

Images