A top coal fracture network reconstruction method based on three-dimensional modeling

Abstract

The application discloses a kind of top coal fracture network reconstruction methods based on three-dimensional modeling, belong to mining information field, including: through field survey and sensor acquisition top coal fracture initial data, and analyze the morphological change characteristics of top coal fracture initial data, obtain the narrowing distribution model of fracture from opening to deep part;Non-uniformity degree is determined based on the narrowing distribution model acquisition geological condition parameter;Three-dimensional dynamic index is extracted from non-uniformity degree, and preliminary space frame is constructed, and the expansion state of fracture in depth direction is obtained;The expansion state of fracture in depth direction is fused with geological condition parameter, and fracture behavior simulation is carried out, and the dynamic adjustment scheme of space network is determined;Based on dynamic adjustment scheme, obtain the optimized three-dimensional model;Fracture behavior prediction value is extracted from the optimized three-dimensional model, and the final top coal fracture network reconstruction result is obtained.The application effectively improves the accuracy and reliability of fracture network analysis under complex geological conditions.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of mining informatization, and particularly relates to a top coal fracture network reconstruction method based on three-dimensional modeling. BACKGROUND

[0002] By accurately simulating the spatial distribution and morphological characteristics of top coal fractures, important basis can be provided for disaster prevention and control and process optimization during mining. However, although some methods have attempted to model the spatial distribution of top coal fractures, most of them ignore the dynamic change characteristics of fractures in three-dimensional space, especially the non-uniformity of morphology and the complex connection relationship between fractures. This neglect leads to deviations between the simulation results and the actual geological environment, making it difficult to truly reflect the fracture behavior and stress distribution of top coal during mining, and thus affecting the accuracy of the prediction.

[0003] The deeper technical difficulty lies in the realistic portrayal of fracture morphology and the overall construction of spatial network. Fractures in three-dimensional space often show a gradually narrowing feature from the opening to the deep part. This morphological change is not only influenced by the geological conditions of the coal seam, but also further affects how fractures intersect and connect with each other in space. If this morphological change cannot be accurately captured, it is difficult to restore the true extension state of fractures at different depths and positions. For example, in actual mining, some fractures may have a larger opening near the roof, but gradually narrow or even close as the depth increases. However, existing techniques often fail to reflect this change, resulting in a large difference between the simulated fracture network and the actual situation, and thus affecting the judgment of top coal stability and mining risk.

[0004] Therefore, how to accurately portray the morphological change characteristics of fractures in three-dimensional space and construct a spatial connection network that conforms to the actual geological environment has become a key problem to be solved. SUMMARY

[0005] To solve the above technical problems, the present application proposes a top coal fracture network reconstruction method based on three-dimensional modeling to solve the problems existing in the prior art.

[0006] To achieve the above purpose, the present application provides a top coal fracture network reconstruction method based on three-dimensional modeling, comprising:

[0007] Collecting initial data of top coal fractures through field investigation and sensors, and analyzing the morphological change characteristics of the initial data of top coal fractures to obtain a narrowing distribution model of fractures from the opening to the deep part;

[0008] Obtaining geological condition parameters based on the narrowing distribution model, and determining the degree of non-uniformity based on the geological condition parameters;

[0009] A three-dimensional dynamic index is extracted from the non-uniformity degree, a preliminary spatial framework is constructed by using a grid division method for the three-dimensional dynamic index, and an extension state of the fissure in the depth direction is obtained;

[0010] The extension state of the fissure in the depth direction is fused with a geological condition parameter to obtain a preliminary network topology;

[0011] A fracture behavior simulation is performed based on the preliminary network topology, and a dynamic adjustment scheme of the spatial network is determined;

[0012] Overall distribution characteristics of the top coal fissure are obtained based on the dynamic adjustment scheme, and an optimized three-dimensional model is obtained based on the overall distribution characteristics;

[0013] Fracture behavior prediction values are extracted from the optimized three-dimensional model, and a final top coal fissure network reconstruction result is obtained by comparing the fracture behavior prediction values with actual geological data.

[0014] Optionally, a process of obtaining a narrowing distribution model of the fissure from an opening to a deep part based on morphological change characteristics of the initial data of the top coal fissure includes:

[0015] Initial data of the top coal fissure are obtained based on field investigation, and morphological change characteristics in the initial data are collected by using a sensor to obtain a width value of an opening part of the fissure;

[0016] The width value and the morphological change characteristics are processed by using a fractal geometry algorithm to obtain a narrowing trend of the fissure from the opening to the deep part;

[0017] Distribution parameters are extracted from the narrowing trend, the distribution parameters are obtained by calculating a width change rate, and a closing point of the deep part of the fissure is determined;

[0018] Based on the closing point and the distribution parameters, a narrowing distribution model is constructed.

[0019] Optionally, an expression of the narrowing distribution model is:

[0020] ;

[0021] In the formula, (x, y, z) is a coordinate of an arbitrary point in space, W is a width, is a probabilistic three-dimensional field, is a standard deviation of the width value, is a core narrowing function, is a probability that the fissure at the space point (x, y, z) is not completely closed and filled.

[0022] Optionally, a process of obtaining a geological condition parameter based on the narrowing distribution model and determining a non-uniformity degree based on the geological condition parameter includes:

[0023] obtaining a geological condition parameter based on the narrowed distribution model, the geological condition parameter comprising: a maximum principal stress, a minimum principal stress, a rock density, and an elastic modulus;

[0024] adopting a deep feature fusion network to perform multi-modal fusion of the geological condition parameter and morphological feature analysis data, enhancing the expression ability of key features through a self-attention mechanism, and inputting the key features into an improved graph convolutional neural network to perform high-dimensional nonlinear classification on the fracture type, outputting a fracture type classification label, and obtaining a fracture type classification result;

[0025] The morphological feature analysis data comprises: a fractal dimension, an average width, and a width variance.

[0026] Based on the fracture type classification result, a closing point position construction data is obtained by integrating dynamic distribution parameters through multi-scale morphological feature analysis.

[0027] The closing point position construction data determines a non-uniformity degree through a non-uniformity quantization model.

[0028] Optionally, the process of obtaining the extension state of the fracture in the depth direction comprises:

[0029] Based on the non-uniformity degree, a spatial gradient calculation method is adopted to extract three-dimensional dynamic indicators.

[0030] For the three-dimensional dynamic indicators, a grid division method is adopted to process, divide uniform grid units, and construct a preliminary spatial framework, to obtain preliminary fracture extension data.

[0031] Based on the preliminary fracture extension data, a depth gradient analysis method is adopted to obtain morphological change parameters in the depth direction.

[0032] A data fusion method is adopted to integrate the morphological change parameters in the depth direction with the preliminary spatial framework, to determine the extension state distribution characteristics.

[0033] Based on the extension state distribution characteristics, a parameter integration method is adopted to process the morphological change parameters in the depth direction, to obtain the extension state of the fracture in the depth direction by merging distribution points.

[0034] Optionally, the process of fusing the extension state of the fracture in the depth direction with the geological condition parameter to obtain a preliminary network topology comprises:

[0035] Based on the extension state of the fracture in the depth direction, connection relationship data is obtained, and the connection relationship data is fused with the geological condition parameter to superimpose a strength value.

[0036] Distribution points in the morphology change analysis are obtained based on the intensity values, and an extended state distribution feature is determined by mapping the coordinates of the points.

[0037] A crack network connection relationship is obtained based on the extended state distribution feature, and a preliminary network topology is obtained based on the crack network connection relationship.

[0038] Optionally, based on the dynamic adjustment scheme, an overall distribution feature of the top coal crack is obtained, and the process of obtaining an optimized three-dimensional model based on the overall distribution feature includes:

[0039] Based on the dynamic adjustment scheme, an overall distribution feature of the top coal crack is obtained.

[0040] The overall distribution feature is processed by using a joint distribution function based on the coupling of the crack density field and the stress field to obtain a quantitative evaluation result.

[0041] Based on the quantitative evaluation result, a distribution uniformity discrimination function is used for processing, and when the discrimination result exceeds a preset threshold, a connection relationship recalculation process based on a graph theory optimization algorithm is activated to obtain updated crack distribution data.

[0042] For the updated crack distribution data, a parameter variation assimilation method is used to fuse and process stress-related parameters and geometric offset values in the dynamic adjustment scheme to obtain optimized connection parameters.

[0043] Based on the optimized connection parameters, a physics-based deformation model is used to adjust and process three-dimensional grid node positions and integrate crack distribution probability fields to finally obtain an optimized three-dimensional crack network model.

[0044] The present application provides a computer device, comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to realize the steps of the method.

[0045] The present application provides a computer readable storage medium, which stores computer program instructions, wherein the computer program instructions are executed by a processor to realize the steps of the method.

[0046] The present application provides an electronic device, comprising at least one processor and a memory connected in communication with the at least one processor, wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the steps of the method.

[0047] Compared with the prior art, the present application has the following advantages and technical effects:

[0048] The application discloses a comprehensive analysis and optimization method based on top coal crack network reconstruction, and aims at a business scene problem that top coal crack distribution characteristics are complex, non-uniformity degree is high, and dynamic change is difficult to predict, solves a crack space network reconstruction problem through multidimensional data fusion and intelligent algorithm. BRIEF DESCRIPTION OF DRAWINGS

[0049] The accompanying drawings, which form a part of this application, are included to provide a further understanding of the application and are incorporated in and constitute a part of this application. The embodiments of this application, and of the prior art, will be described and explained with additional specificity and detail by the accompanying drawings.

[0050] Figure 1 A method flowchart of the embodiments of the application. DETAILED DESCRIPTION

[0051] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0052] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a group of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown herein.

[0053] As shown in the accompanying drawings, Figure 1 The embodiments provided in the present embodiment include the following steps:

[0054] S1, collecting initial data of top coal cracks through field investigation and sensors, and analyzing the morphological change characteristics of the initial data of top coal cracks to obtain a narrowing distribution model of cracks from opening to deep part.

[0055] Further, the process of obtaining the narrowing distribution model of the crack from the opening to the deep part by analyzing the morphological change characteristics of the initial data of the top coal crack includes: obtaining the initial data of the top coal crack based on the field survey, collecting the morphological change characteristics in the initial data by using a sensor to obtain the width value of the crack opening part; processing the width value and the morphological change characteristics by using a fractal geometry algorithm to obtain the narrowing trend of the crack from the opening to the deep part; extracting the distribution parameters from the narrowing trend, the distribution parameters are obtained by calculating the width change rate, and the closure point of the deep part of the crack is determined; and constructing the narrowing distribution model based on the closure point and the distribution parameters.

[0056] Further, as a specific implementation of the embodiment, when obtaining the initial data of the top coal crack based on the field survey, an engineering team can be dispatched to enter the mine with a laser scanner and geological drilling equipment to conduct on-site measurement of the top coal seam. The survey process includes marking the crack position, recording the initial crack length and direction, measuring a crack with a length of 5 meters and a direction deviating north, and collecting rock samples for analysis of the initial morphology. These data serve as a basis for subsequent analysis of the stability of the crack. The morphological change characteristics in the initial data are collected by using a sensor to obtain the width value of the crack opening part. An embedded strain sensor is installed at the crack opening to monitor the width change in real time. For example, the initial width is 0.2 meters, and it may expand to 0.3 meters due to stress changes during the coal mining process. This collection method ensures data accuracy by wirelessly transmitting data to a central system, thereby providing reliable input for crack analysis. For the width value, when analyzing the morphological change characteristics in the data by using a fractal geometry algorithm, the fractal geometry algorithm is a mathematical method for describing the complexity of irregular geometric shapes. The input includes the width value and the morphological change characteristics, such as the width sequence and the crack curvature, and the output includes the narrowing trend of the crack from the opening to the deep part.

[0057] In one possible implementation, after determining the narrowing trend, if the narrowing trend exceeds a preset threshold, for example, the trend slope exceeds 0.1 per meter, the distribution parameters are extracted from the narrowing trend. The distribution parameters are obtained by calculating the width change rate in the narrowing trend, for example, the change rate is -0.05, indicating that the width decreases by 0.05 meters per meter of depth, thereby obtaining the closure point of the deep part of the crack, for example, the calculation shows that the closure occurs at a depth of 3 meters. According to the closure point, when constructing the narrowing distribution model, the model integrates the closure point and the distribution parameters to determine the evolution law of the crack in the model, thereby obtaining the narrowing distribution model of the crack from the opening to the deep part.

[0058] Further, the expression of the narrowing distribution model is:

[0059] ;

[0060] In the formula, (x, y, z) is the coordinates of any point in space, W is the width, is the probability of a three-dimensional field, is the standard deviation of the width value, is the core narrowing function, is the probability of a spatial point (x, y, z) being not completely filled by the closure of the fracture.

[0061] ;

[0062] wherein, is the initial width of the opening at position x, whose statistical distribution (such as variance) is related to the fractal dimension D, is the predicted closure depth corresponding to position x, is the base narrowing index, is the dynamic narrowing modulation factor, which is a function of the fractal dimension D and the formation stress σ(z).

[0063] S2, obtain geological condition parameters based on the narrowing distribution model, and determine the non-uniformity degree based on the geological condition parameters.

[0064] Further, the process of obtaining geological condition parameters based on the narrowing distribution model and determining the non-uniformity degree based on the geological condition parameters comprises: obtaining geological condition parameters based on the narrowing distribution model, the geological condition parameters comprising: maximum principal stress, minimum principal stress, rock density, and elastic modulus; using a depth feature fusion network to perform multi-modal fusion of the geological condition parameters and morphological feature analysis data, enhancing the expression ability of key features through a self-attention mechanism, and inputting into an improved graph convolutional neural network to perform high-dimensional nonlinear classification of the fracture type, outputting a fracture type classification label, and obtaining a fracture type classification result; wherein the morphological feature analysis data comprises: fractal dimension, average width, and width variance; based on the fracture type classification result, obtaining closure point position construction data by integrating dynamic distribution parameters through multi-scale morphological feature analysis; and determining the non-uniformity degree through a non-uniformity quantification model based on the closure point position construction data.

[0065] Further, as a specific embodiment of the present embodiment, in the fracture classification based on the graph convolutional neural network, the graph structure construction process comprises: taking the fracture sample as a node, and taking the fused feature vector as the node feature An adjacency matrix is constructed using feature similarity: .

[0066] The GCN classification model adopts a two-layer graph convolutional network: .

[0067] S3, extract a three-dimensional dynamic index from the non-uniformity degree, construct a preliminary spatial framework by using a grid division method for the three-dimensional dynamic index, and obtain the expansion state of the fracture in the depth direction.

[0068] Further, the process of obtaining the extension state of the fissure in the depth direction includes: based on the non-uniformity degree, extracting three-dimensional dynamic indicators by using a spatial gradient calculation method; for the three-dimensional dynamic indicators, processing by using a grid division method, dividing uniform grid cells and constructing a preliminary spatial framework to obtain fissure extension preliminary data; based on the fissure extension preliminary data, obtaining the morphological change parameters in the depth direction by using a depth gradient analysis method; integrating the morphological change parameters in the depth direction with the preliminary spatial framework by using a data fusion method to determine the extension state distribution characteristics; based on the extension state distribution characteristics, processing the morphological change parameters in the depth direction by using a parameter integration method to obtain the extension state of the fissure in the depth direction by merging the distribution points.

[0069] Further, as a specific implementation of the embodiment, the process of extracting three-dimensional dynamic indicators from the non-uniformity degree is realized by processing the scanning data of the coal seam rock sample. First, collect stress distribution data and fissure width change data in the coal seam, which are derived from field information collected by geological exploration equipment such as a borehole camera, and then calculate the change rate of these indicators in the x, y, z spatial coordinates, for example, the gradient of the fissure width in the depth direction as part of the dynamic indicator, and the change rate is calculated by formula such as (width difference / coordinate interval), thereby obtaining three-dimensional dynamic indicators.

[0070] For these three-dimensional dynamic indicators, when constructing a preliminary spatial framework by using a grid division method, the coal seam area is divided into uniform cubic grid cells, each cell size is 1 meter x 1 meter x 1 meter, and each cell is filled with data based on the value of the dynamic indicator, thereby forming a three-dimensional grid model to obtain fissure extension preliminary data.

[0071] In one possible implementation, the morphological change parameters in the depth direction are obtained from the fissure extension preliminary data, which are realized by analyzing the depth gradient in the data, for example, gradient calculation is performed on the indicator values along the z axis in the grid model, and the change trend of the fissure width or density is identified, such as the gradient increasing from 50 meters to 100 meters in depth indicating accelerated morphological change, thereby obtaining the morphological change parameters.

[0072] By integrating the morphological change parameters with the preliminary spatial framework, the extension state distribution characteristics are determined. The morphological change parameters are mapped into the grid cells, the comprehensive distribution value of each cell is calculated, such as the weighted sum of the average gradient and the dynamic indicator, a distribution characteristic map is formed, and a high expansion risk area is highlighted. The extension state distribution characteristics are judged, if the characteristic value such as the average distribution density exceeds the preset threshold value 0.5, the morphological change parameters in the depth direction are integrated, the integration result is obtained by merging the distribution points, for example, the parameter values of adjacent points are clustered and merged to generate a unified expansion state model, and finally the extension state of the fissure in the depth direction is obtained.

[0073] S4, the extension state of the fracture in the depth direction is fused with the geological condition parameters to obtain a preliminary network topology.

[0074] Further, the process of fusing the extension state of the fracture in the depth direction with the geological condition parameters to obtain the preliminary network topology includes: obtaining connection relationship data based on the extension state of the fracture in the depth direction, fusing the connection relationship data with the geological condition parameters, and superimposing and calculating the intensity value; obtaining the distribution points in the morphological change analysis based on the intensity value, and determining the extension state distribution characteristics through coordinate mapping of the points; obtaining the fracture network connection relationship based on the extension state distribution characteristics, and obtaining the preliminary network topology based on the fracture network connection relationship.

[0075] Further, as a specific embodiment of the present embodiment, the connection relationship data is obtained from the extension state extraction, and the connection relationship data is superimposed and calculated with the geological parameters through the geological condition fusion to obtain the intensity value, wherein the geological parameters include the rock density and the stress distribution, the intensity value is obtained through a weighted average formula, and the formula is S=(a x D+b x T) / (a+b), wherein S represents the intensity value, D represents the rock density, T represents the stress distribution, and a and b are weight coefficients.

[0076] The distribution points in the morphological change analysis are obtained based on the intensity value, and the extension state distribution characteristics are determined through coordinate mapping of the points. When the distribution points in the morphological change analysis are obtained, the coordinate points of the fracture morphology in each grid are recorded, such as point (x=10, y=15, z=20), and then the distribution mode of these points in the three-dimensional space is calculated through coordinate mapping of the points, so as to determine the extension state distribution characteristics. The fracture network connection model is constructed through a connection intensity matrix, and the matrix is an nxn array, wherein n is the number of points, and each element represents the connection weight between two points, such as the weight value calculated based on the distance and the intensity. In this way, the extension network model of the fracture in the depth direction is finally obtained.

[0077] S5, based on the preliminary network topology, a fracture behavior simulation is performed to determine a dynamic adjustment scheme of the spatial network.

[0078] Further, the fracture behavior data is obtained from the preliminary network topology, and a morphological change index is calculated by adjusting variable values through multiple cycles using an iterative optimization process, wherein the variable values are iteratively updated based on the initial state of the data to obtain updated morphological change data. For the morphological change data, spatial network adjustment attributes are fused, including node position offset and connection elastic coefficient, and if the change index exceeds a preset threshold, a dynamic scheme determination is triggered to obtain network adjustment parameters. According to the network adjustment parameters, a topology behavior simulation and a simulation update mechanism are integrated, wherein the topology behavior simulation predicts fracture points through grid mapping, and the simulation update mechanism uses an incremental adjustment method to gradually build an adjustment framework through data iterative fusion to determine a dynamic generation path of the scheme. The dynamic generation path of the scheme is obtained, combined with fracture simulation analysis and an iterative optimization process, wherein the fracture simulation analysis evaluates the fracture propagation mode, and the iterative optimization process refines the path variables to update the overall network structure, and a dynamic adjustment scheme of the spatial network is obtained.

[0079] Further, as a specific embodiment of the present embodiment, the preliminary network topology represents a fracture distribution map in an underground rock layer, and fracture behavior data is extracted from the topology by a data acquisition device such as a seismic wave detector, which includes initial length and direction vector of the fracture. When using an iterative optimization process, a number of cycles is set, such as adjusting variable values through multiple cycles, wherein the variable values are iteratively updated based on the initial state of the data, for example, the initial state of the fracture length is 10 meters, and the stress variable is 5 MPa. In the first iteration, the length is adjusted to 11 meters and the stress is adjusted to 6 MPa according to the feedback formula, and after several cycles, the morphological change index is calculated, such as the fracture expansion rate reaching 20%, thereby obtaining the updated morphological change data.

[0080] In one possible implementation, the step of fusing spatial network adjustment attributes for morphological change data is achieved by integrating node position offset and connection elastic coefficient. According to the network adjustment parameters, the topology behavior simulation and the simulation update mechanism are integrated, wherein the topology behavior simulation can predict fracture points through grid mapping, for example, the network is divided into 10x10 grid units, and each unit maps the position of the potential fracture point. The simulation update mechanism uses an incremental adjustment method to gradually build an adjustment framework in data iterative fusion, such as starting from the initial parameters, adding a correction value of 0.1 each time, and determining the dynamic generation path of the scheme after fusion, which may be an optimized curve from the starting point to the ending point, representing the simulation trajectory of fracture propagation.

[0081] In one possible implementation, when the acquisition scheme dynamically generates the path and combines the fracture simulation analysis and the iterative optimization process, the fracture simulation analysis evaluates the fracture propagation mode, such as linear propagation or branch propagation, and the iterative optimization process refines the path variable, for example, adjusts the path curvature from the initial 0.5 to 0.3, updates the overall network structure, and finally obtains the dynamic adjustment scheme of the spatial network.

[0082] S6, based on the dynamic adjustment scheme, obtaining the overall distribution characteristics of the top coal fracture, and based on the overall distribution characteristics, obtaining the optimized three-dimensional model.

[0083] Further, the process of obtaining the overall distribution characteristics of the top coal fracture based on the dynamic adjustment scheme and obtaining the optimized three-dimensional model based on the overall distribution characteristics includes: based on the dynamic adjustment scheme, obtaining the overall distribution characteristics of the top coal fracture; using a joint distribution function based on the coupling of the fracture density field and the stress field to process the overall distribution characteristics to obtain a quantitative evaluation result; based on the quantitative evaluation result, using a distribution uniformity discrimination function for processing, when the discrimination result exceeds a preset threshold, activating a connection relationship recalculation process based on a graph theory optimization algorithm to obtain updated fracture distribution data; for the updated fracture distribution data, using a parameter variational assimilation method to fuse the stress-related parameters and the geometric offset values in the dynamic adjustment scheme to obtain optimized connection parameters; based on the optimized connection parameters, using a physics-based deformation model to adjust the three-dimensional grid node positions and integrate the fracture distribution probability field to finally obtain the optimized three-dimensional fracture network model.

[0084] S7, extracting the fracture behavior prediction value from the optimized three-dimensional model, and obtaining the final top coal fracture network reconstruction result by comparing the fracture behavior prediction value and the actual geological data.

[0085] Further, the fracture prediction extraction is obtained from the three-dimensional model construction, the stress field simulation is fused through the geological data comparison, wherein the geological data comparison extracts the matching features from the actual geological data and superimposes them with the fracture prediction extraction, the stress field simulation calculates the internal stress distribution by using the finite element method and integrates it into the comparison result to obtain the preliminary network reconstruction generation. For the preliminary network reconstruction generation, the distribution characteristic acquisition integrates the connection parameter optimization, wherein the distribution characteristic acquisition calculates the fracture distribution uniformity from the preliminary network reconstruction generation and combines it with the connection parameter optimization to determine whether the overall distribution deviation exceeds the preset threshold to adjust the fracture density evaluation, update the distribution data by recalculating the density value, and determine the updated verification module application. According to the updated verification module application, the matching value of the fracture prediction extraction and the geological data comparison is extracted, wherein the matching value is obtained by calculating the similarity index between the two, and the final top coal fracture network reconstruction result is obtained.

[0086] Further, as a specific implementation of the present embodiment, firstly, matching features are extracted from actual geological data, such as rock sample data collected in the coal roof stratum, including fracture spacing and direction information, which are superimposed with fracture prediction extraction, which is a preliminary estimate obtained by analyzing potential fracture points in the three-dimensional model. Stress field simulation uses finite element method to calculate internal stress distribution, such as using ANSYS software to build finite element grid model of the coal roof, inputting geological data to simulate stress changes in the mining process, such as shear stress and tensile stress distribution, and then integrating these stress values into the comparison results to obtain the preliminary network reconstruction generation.

[0087] In one possible implementation, for preliminary network reconstruction generation, distribution feature acquisition is used to integrate connection parameter optimization. The fracture distribution uniformity is calculated from the preliminary network reconstruction generation, which can be achieved by calculating the coefficient of variation of fracture density in different regions, such as dividing the grid in the x, y, z coordinate system of the coal roof model, calculating the average value of the number of fractures in each grid, and adjusting the fracture density evaluation if the coefficient of variation exceeds the preset threshold of 0.5. The specific process includes recalculating the density value, such as estimating the density of unsampled areas from existing data points using the Kriging interpolation method, and determining the updated verification module application after updating the distribution data. According to the updated verification module application, the matching value of the fracture prediction extraction and the geological data comparison is extracted, wherein the matching value is obtained by calculating the similarity index between the two, such as comparing the feature vectors using the cosine similarity formula, and obtaining the final coal roof fracture network reconstruction result.

[0088] In one possible implementation, when calculating the fracture distribution uniformity, the distribution feature acquisition can be combined with the connection parameter optimization, such as adjusting the virtual connection weight between the fractures, and dynamically modifying the parameters according to the density deviation, if the deviation exceeds the threshold value such as 10%, the density is re-evaluated by Monte Carlo simulation, and the verification module is applied after updating the data.

[0089] The similarity index calculation of the extracted matching value can include multi-dimensional analysis, such as structural similarity and statistical similarity, comparing the position deviation of the predicted fracture path and the actual geological fault in the coal roof fracture network to obtain the final result.

[0090] The present application provides a computer device, comprising a memory, a processor and a computer program stored on the memory, the processor executing the computer program to implement the steps of the method.

[0091] The present application provides a computer readable storage medium having computer program instructions stored thereon, the computer program instructions being executed by a processor to implement the steps of the method.

[0092] The application provides an electronic device, comprising: at least one processor; and a memory connected with the at least one processor in communication; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to perform the steps of the method.

[0093] The above merely describes the preferred embodiments of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for reconstructing the top coal fracture network based on three-dimensional modeling, characterized in that, Includes the following steps: Initial data of top coal fractures were collected through on-site surveys and sensors, and the morphological change characteristics of the initial data of top coal fractures were analyzed to obtain a narrowing distribution model of fractures from the opening to the depth. Geological condition parameters are obtained based on the narrowing distribution model, and the degree of non-uniformity is determined based on the geological condition parameters. Three-dimensional dynamic indicators are extracted from the degree of non-uniformity. A preliminary spatial framework is constructed based on the three-dimensional dynamic indicators using a mesh generation method to obtain the expansion state of the crack in the depth direction. By integrating the depth-direction expansion state of the fractures with geological condition parameters, a preliminary network topology is obtained; Based on the preliminary network topology, a rupture behavior simulation is performed to determine a dynamic adjustment scheme for the spatial network; The overall distribution characteristics of top coal fractures are obtained based on the dynamic adjustment scheme, and an optimized three-dimensional model is obtained based on the overall distribution characteristics. The predicted values ​​of fracture behavior are extracted from the optimized 3D model, and the final top coal fracture network reconstruction results are obtained by comparing the predicted values ​​of fracture behavior with the actual geological data. The process of analyzing the morphological change characteristics of the initial data of the top coal fractures to obtain the narrowing distribution model of the fractures from the opening to the depth includes: Based on the initial data of top coal fractures obtained from on-site surveys, sensors are used to collect the morphological change characteristics in the initial data to obtain the width value of the fracture opening. The width value and morphological change characteristics are processed using a fractal geometry algorithm to obtain the narrowing trend of the crack from the opening to the depth. Distribution parameters are extracted from the narrowing trend, and these distribution parameters are obtained by calculating the width change rate to determine the closure point in the depth of the fracture. Based on the closed points and distribution parameters, a narrowing distribution model is constructed; The expression for the narrowing distribution model is: ； In the formula, (x, y, z) are the coordinates of any point in space, and W is the width. It is a probabilistic three-dimensional field. The standard deviation of the width value. For the core narrowing function, This represents the probability that the crack at the spatial point (x, y, z) is not completely closed and filled.

2. The method for reconstructing the top coal fracture network based on three-dimensional modeling according to claim 1, characterized in that, The process of obtaining geological condition parameters based on the narrowing distribution model and determining the degree of non-uniformity based on the geological condition parameters includes: Geological condition parameters are obtained based on the narrowing distribution model, including: maximum principal stress, minimum principal stress, rock layer density, and elastic modulus. A deep feature fusion network is used to perform multimodal fusion of the geological condition parameters and morphological feature analysis data. The expressive power of key features is enhanced through a self-attention mechanism. The data is then input into an improved graph convolutional neural network to perform high-dimensional nonlinear classification of fracture types and output fracture type classification labels to obtain fracture type classification results. The morphological feature analysis data includes: fractal dimension, average width, and width variance; Based on the fracture type classification results, the data for constructing closure points is obtained by integrating dynamic distribution parameters through multi-scale morphological feature analysis. The degree of non-uniformity in the data for constructing closed points is determined using a non-uniformity quantification model.

3. The method for reconstructing the top coal fracture network based on three-dimensional modeling according to claim 1, characterized in that, The process of obtaining the propagation state of the crack in the depth direction includes: Based on the degree of non-uniformity, a spatial gradient calculation method is used to extract three-dimensional dynamic indicators; For the aforementioned three-dimensional dynamic indicators, a mesh generation method is used to process them, dividing them into uniform mesh cells and constructing a preliminary spatial framework to obtain preliminary data on crack propagation. Based on the preliminary data of the crack propagation, the morphological change parameters in the depth direction are obtained by using the depth gradient analysis method; The morphological change parameters in the depth direction are integrated with the preliminary spatial framework using a data fusion method to determine the extended state distribution characteristics; Based on the aforementioned expansion state distribution characteristics, a parameter integration method is used to process the morphological change parameters in the depth direction, and the expansion state of the crack in the depth direction is obtained by merging the distribution points.

4. The method for reconstructing the top coal fracture network based on three-dimensional modeling according to claim 1, characterized in that, The process of fracturing the depth-direction expansion state of the fractures with geological condition parameters to obtain a preliminary network topology includes: Based on the expansion state of the fracture in the depth direction, the connectivity data is obtained, and the connectivity data is fused with geological condition parameters to calculate the intensity value. Based on the intensity value, the distribution points in the morphological change analysis are obtained, and the distribution characteristics of the extended state are determined by the coordinate mapping of the points. Based on the extended state distribution characteristics, the gap network connection relationship is obtained, and based on the gap network connection relationship, a preliminary network topology is obtained.

5. The method for reconstructing the top coal fracture network based on three-dimensional modeling according to claim 1, characterized in that, The process of obtaining the overall distribution characteristics of top coal fractures based on the dynamic adjustment scheme, and obtaining the optimized three-dimensional model based on the overall distribution characteristics, includes: Based on the aforementioned dynamic adjustment scheme, the overall distribution characteristics of top coal fractures are obtained; The overall distribution characteristics are processed using a joint distribution function based on the coupling of the fracture density field and the stress field to obtain quantitative evaluation results; Based on the quantitative evaluation results, a distribution uniformity discrimination function is used for processing. When the discrimination result exceeds a preset threshold, the connection relationship recalculation process based on graph theory optimization algorithm is activated to obtain updated fracture distribution data. For the updated crack distribution data, the parameter variational assimilation method is used to fuse the stress-related parameters with the geometric offset values ​​in the dynamic adjustment scheme to obtain the optimized connection parameters. Based on the optimized connection parameters, a physics-based deformation model is used to adjust the positions of the three-dimensional mesh nodes, and the crack distribution probability field is integrated to finally obtain the optimized three-dimensional crack network model.

6. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method of claim 1.

7. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the steps of the method of claim 1.

8. An electronic device, characterized in that, The electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the steps of the method of claim 1.

Images