Three-dimensional modeling method for correlating coal permeability with pore and fracture evolution

Abstract

The present application provides a kind of coal body permeability and pore fissure evolution associated three-dimensional modeling method, it is related to coal body structure three-dimensional modeling technical field, the present application is by to each time all voxel gray standardization, classification mark and space structure unit identification, combine voxel level parameter statistics and geometric analysis, can extract porosity, fracture width and fracture connectivity and other structure characteristic parameters, and with the physical field parameter dynamic pairing of synchronous acquisition, not only guarantee the physical accuracy and time sequence response ability of parameter, also realize the accurate assignment of voxel level parameter, so that coal body heterogeneity and dynamic evolution process can be carefully quantified, using voxel scale permeability calculation method, can calculate the permeability value of each voxel according to voxel characteristic parameters and physical field parameters at each time, generate the permeability distribution in three-dimensional space, can fully capture the structural evolution process of coal core sample.

Description

Technical Field

[0001] This invention relates to the field of three-dimensional modeling technology for coal structure, specifically a three-dimensional modeling method relating coal permeability to pore and fracture evolution. Background Technology

[0002] In engineering fields such as coal mining, coalbed methane development, and underground reservoir utilization, coal bodies are constantly subjected to complex and variable physical field environments, including stress disturbances, fluid pressure changes, and heat transfer. The pore and fracture structure inside the coal body not only determines its mechanical stability but also directly affects the gas and liquid seepage process, making it a key factor related to efficient energy development and safe production. With the development of digital core and 3D CT imaging technologies, accurately characterizing the internal structure of coal bodies using high-resolution 3D voxel data has become an important technical approach to reveal the spatial distribution, dynamic evolution, and seepage behavior of pores and fractures in coal bodies.

[0003] Existing technologies for coal structure modeling and permeability analysis typically employ static, single-field conditions using two-dimensional image processing, overall statistical analysis, or single-shot three-dimensional CT scan data. While some technologies can achieve three-dimensional spatial structure visualization, they are mostly limited to single-moment, single-physical-field conditions, lacking the ability to dynamically respond to changes in the coal structure with varying physical fields. Structural parameters such as porosity, fracture width, and connectivity are often assigned values ​​based on regional averages or empirical formulas, failing to reflect spatial heterogeneity, structural details, and dynamic evolution processes. Furthermore, there is a lack of automated, time-series data pairing between physical field parameters and structural features; permeability calculations are mostly based on overall or regional averages, unable to achieve voxel-level dynamic prediction. These shortcomings make it difficult to accurately model and analyze the structural changes and permeability responses of coal under multi-field coupling and dynamic conditions in actual engineering projects.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] (a) Technical problems to be solved

[0006] To address the shortcomings of existing technologies, this invention provides a three-dimensional modeling method for the correlation between coal permeability and pore fracture evolution. By using time-series three-dimensional CT scanning and simultaneous acquisition of multiple physical field parameters, combined with standardized voxel classification, automatic identification of spatial structural units, and assignment of voxel-level feature parameters, dynamic pairing of structural parameters and physical field parameters and voxel-scale permeability distribution modeling are achieved, thus solving the technical problems described in the background art.

[0007] (II) Technical Solution

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A three-dimensional modeling method for the correlation between coal permeability and pore fracture evolution, comprising the following steps:

[0010] Time-series three-dimensional CT scan data of coal core samples under different physical fields were collected to obtain the original three-dimensional voxel dataset containing voxel gray values ​​at each time point, and the physical field parameters corresponding to each time point were recorded synchronously. The physical field parameters include axial stress, pore pressure and temperature.

[0011] The original three-dimensional voxel dataset is subjected to grayscale normalization. Each voxel is classified using a set grayscale threshold to distinguish between the coal body region and the pore and fracture region. The pore and fracture region is marked and identified to obtain the distribution information of pore structure and fracture structure at each time step.

[0012] Based on the distribution information of pore structure, the number of pore voxels at each time moment is counted. The porosity parameter is obtained by the ratio of the number of pore voxels to the total number of voxels in the coal core sample. At the same time, the distribution information of fracture structure is geometrically analyzed to extract fracture width and fracture connectivity parameters. The porosity parameter, fracture width parameter, and fracture connectivity parameter are used as structural characteristic parameters of the coal body.

[0013] The structural feature parameters at each time step are paired with the corresponding physical field parameters and arranged in chronological order. Based on the structural feature parameters at each time step, the porosity feature, fracture width feature, and fracture connectivity feature of each voxel are assigned values ​​according to the voxel classification results, thus forming the voxel feature parameters of each voxel at each time step.

[0014] Based on the physical field parameters corresponding to each moment, and combined with the voxel characteristic parameters of all voxels at that moment, the voxel-scale permeability calculation method is used to generate the permeability distribution of each voxel in the three-dimensional space of the coal body.

[0015] Furthermore, when obtaining the original 3D voxel dataset:

[0016] Set the initial values ​​for axial stress, pore pressure, and temperature, and determine the synchronous step increment for each physical field parameter to ensure that the three physical field parameters increase synchronously during the experiment.

[0017] Axial stress, pore pressure and temperature are increased simultaneously at fixed time intervals. Each increase is recorded as a moment, so that each moment corresponds to a set of synchronously changed physical field parameters.

[0018] At each moment, a three-dimensional CT scan is performed to obtain the spatial coordinates and corresponding gray values ​​of all voxels in the three-dimensional space of the coal core sample, which are then calibrated as the three-dimensional voxel gray value data at that moment.

[0019] Establish a pairing relationship between the three-dimensional voxel grayscale data obtained from each scanning moment and the physical field parameters corresponding to that moment, and establish the correspondence between the moment, physical field parameters and three-dimensional voxel grayscale values;

[0020] The three-dimensional voxel grayscale data and physical field parameters at all times are organized in chronological order to form an original three-dimensional voxel dataset arranged by time. The original three-dimensional voxel dataset is a set of data at each time that contains the spatial coordinates and corresponding grayscale values ​​of all voxels in the coal core sample room, and is paired with the physical field parameters at the corresponding time.

[0021] Furthermore, when performing grayscale normalization on the original 3D voxel dataset, the grayscale values ​​of the 3D voxels at each time step are obtained, the maximum and minimum grayscale values ​​in the 3D voxel grayscale data are determined, and the grayscale value of each voxel is processed according to the minimum-maximum normalization, that is, the minimum grayscale value is subtracted from the current voxel grayscale value, and then divided by the difference between the maximum and minimum values, so that all voxel grayscale values ​​are normalized to between 0 and 1.

[0022] For the 3D voxel grayscale data at multiple time points, the min-max normalization process is performed independently on the 3D voxel grayscale data at each time point to obtain the 3D voxel grayscale data after grayscale normalization at each time point, which is labeled as a 3D voxel dataset.

[0023] When distinguishing between the coal body region and the pore and fracture region, the three-dimensional voxel dataset at each time step is traversed according to the pre-set grayscale threshold that best distinguishes between the coal body and the pore and fracture region. The grayscale value of each voxel is judged. If the grayscale value of the voxel is higher than or equal to the grayscale threshold, the voxel is marked as the coal body region. If the grayscale value of the voxel is lower than the grayscale threshold, the voxel is marked as the pore and fracture region. This distinguishes between the coal body region and the pore and fracture region at each time step.

[0024] Furthermore, the three-dimensional voxel dataset corresponding to each time step is traversed to filter out all voxels classified as pore and fracture regions, and these voxel sets are defined as candidate pore and fracture voxel sets.

[0025] For the candidate pore and fracture voxel set, a three-dimensional spatial connectivity analysis algorithm is used to process them one by one. That is, according to the adjacency relationship of voxels in three-dimensional space, it is determined whether any two pore and fracture voxels are directly connected or indirectly connected. All connected voxels are automatically aggregated into a spatial structural unit, and each spatial structural unit represents an independent pore or fracture.

[0026] For each spatial structural unit, the spatial coordinates of all voxels contained within that unit are statistically analyzed. By comparing the X, Y, and Z coordinates of all voxels within the unit, the maximum span of that unit in the X, Y, and Z directions is obtained.

[0027] For each spatial structural unit, compare its maximum span values ​​in three spatial directions. If the maximum span of the spatial structural unit in one direction is greater than or equal to at least twice the average of the maximum spans in the other two directions, then mark the spatial structural unit as a fractured structural unit; otherwise, mark the spatial structural unit as a porous structural unit.

[0028] The labeling results of each spatial structural unit are associated with the spatial coordinates of each voxel and the number of voxels inside it. The spatial structural units labeled as pore structural units and crack structural units are classified as pore structure distribution and crack structure distribution, respectively.

[0029] The classification and spatial distribution results of the pore structure distribution and fracture structure distribution of all spatial structural units are organized in chronological order to form the distribution information of pore structure and fracture structure at each time point.

[0030] Furthermore, when obtaining porosity parameters, based on the distribution information of pore structure at each time point, the number of voxels inside all marked as pore structure units at that time point is counted, the total number of voxels in the scanning area of ​​the coal core sample is counted, and the porosity parameters at that time point are obtained by dividing the number of voxels inside the pore structure units at that time point by the total number of voxels.

[0031] At each moment, for each fracture structure element, the three-dimensional spatial coordinate information of all fracture voxels in the fracture structure element is extracted. Using principal component analysis, the principal axis direction of the coordinates of all voxels in the fracture structure element is calculated as the principal extension direction of the fracture in the fracture structure element.

[0032] Along the main axis, at preset intervals, several two-dimensional plane sections orthogonal to the main axis are cut. On each section, voxels belonging to the section are identified, and the maximum Euclidean distance of the voxels on the section plane is calculated as the crack width value of the section.

[0033] The width values ​​measured on all cross sections are statistically analyzed and averaged to obtain the representative crack width value of the crack structure unit. The above steps are repeated for all crack structure units at each time point to obtain the representative crack width value of each crack structure unit. The representative crack width values ​​of all crack structure units at each time point are obtained, and the average value is calculated and used as the crack width parameter at that time point.

[0034] At each time step, based on all identified fracture structural elements and their spatial coordinate data, a spatial relationship network between fracture structural elements is constructed. Specifically, it is determined whether the nearest distance between any two fracture structural elements is less than a preset threshold. If it is less than the preset threshold, it is determined that the two fracture structural elements have a spatial relationship.

[0035] Based on the spatial relationship network, graph theory is used to identify all sets of fracture structure units with spatial relationships. Each set is a connected cluster. For each connected cluster, the total number of fracture voxels contained in it is counted.

[0036] The proportion of the number of voxels in the largest connected cluster to the total number of voxels in the fracture is determined as the fracture connectivity parameter at that moment.

[0037] Furthermore, the voxel feature parameters include porosity features, fracture width features, and fracture connectivity features. The logic for constituting the voxel feature parameters of each voxel at each time step is as follows:

[0038] For voxels within the coal body region, the porosity, fracture width, and fracture connectivity features are all assigned a value of zero. For voxels within the pore structure, the fracture width and fracture connectivity features are both assigned a value of zero, and the porosity feature is assigned the porosity parameter at that time. For voxels within the fracture structure, the porosity feature is assigned a value of zero, and the fracture width and fracture connectivity features are assigned the fracture width and fracture connectivity parameters at that time.

[0039] Furthermore, the logic for generating the permeability distribution of each voxel within the three-dimensional space of the coal body is as follows:

[0040] At each moment, for each voxel of the coal core sample, the physical field parameters corresponding to that moment are retrieved, and physical field parameters such as axial stress, pore pressure, and temperature are paired with the voxel. The voxel characteristic parameters of all voxels at that moment are retrieved simultaneously.

[0041] For each voxel, the permeability is calculated by taking the physical field parameters and voxel characteristic parameters as inputs according to the voxel-scale permeability calculation method. The permeability calculation method quantitatively analyzes the permeability performance of the voxel based on the voxel's porosity parameters, fracture width parameters, and fracture connectivity parameters, combined with the current physical field parameters, to obtain the permeability value of the voxel at that moment.

[0042] Arrange the permeability values ​​of all voxels in spatial coordinate order to form the three-dimensional spatial permeability distribution data at that moment, so that the specific position of each voxel in the three-dimensional coordinate system corresponds one-to-one with its corresponding permeability value.

[0043] For all time points, the process of constructing three-dimensional spatial permeability distribution data is repeated sequentially. The three-dimensional spatial permeability distribution data at each time point are stored in chronological order to form a time-series three-dimensional spatial permeability distribution dataset, generating the permeability distribution of each voxel in the three-dimensional space of the coal body.

[0044] Furthermore, the specific formula for calculating penetration rate is as follows:

[0045]

[0046] in, Indicates the first At the moment, the first The permeability value of individual elements, This is the benchmark coefficient for permeability. , and These are the weighted indices for porosity, fracture width, and fracture connectivity, respectively. , and These are the influence coefficients of axial stress, pore pressure, and temperature, respectively. , and The first At the moment, the first Porosity characteristics, fracture width characteristics, and fracture connectivity characteristics of individual elements. , and The first Axial stress, pore pressure, and temperature at each moment.

[0047] (III) Beneficial Effects

[0048] This invention performs grayscale standardization, classification and labeling, and spatial structural unit identification on all voxels at each moment. Combined with voxel-level parameter statistics and geometric analysis, it can automatically extract structural feature parameters such as porosity, fracture width, and fracture connectivity. These parameters are then dynamically paired with synchronously acquired physical field parameters. This not only ensures the physical accuracy and temporal response capability of the parameters but also achieves precise assignment of voxel-level parameters, enabling detailed quantification of coal body heterogeneity and dynamic evolution processes.

[0049] The voxel-scale permeability calculation method can calculate the permeability value of each voxel at any given time based on voxel characteristic parameters and physical field parameters, generating the permeability distribution in three-dimensional space. This method breaks through the limitations of regional averages or overall statistics, and can comprehensively capture the structural evolution process of coal core samples under different stress, pressure and temperature conditions, effectively revealing the distribution changes of the coal body, pores and fractures in three-dimensional space. Attached Figure Description

[0050] Figure 1 This is a schematic diagram of the overall method flow of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0052] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0053] Example:

[0054] Please see Figure 1 The present invention provides a technical solution:

[0055] A three-dimensional modeling method for the correlation between coal permeability and pore fracture evolution, comprising the following steps:

[0056] S1: Collect time-series three-dimensional CT scan data of coal core samples under different physical fields to obtain the original three-dimensional voxel dataset containing voxel gray values ​​at each time point, and simultaneously record the physical field parameters corresponding to each time point, including axial stress, pore pressure and temperature.

[0057] Unlike traditional single-physical-field or static scanning techniques, this embodiment simultaneously controls three physical fields—axial stress, pore pressure, and temperature—during sample testing, enabling dynamic response acquisition under multi-field coupling. This allows for a true reflection of the structural evolution of coal under complex actual working conditions. By employing time-series 3D CT scanning, it not only acquires the spatial structure under different working conditions but also captures the entire process of structural evolution. Unlike existing technologies that only provide single-moment or single-state data, time-series acquisition provides a solid data foundation for dynamic modeling, evolution analysis, and structural prediction.

[0058] By utilizing high-resolution 3D CT scans, detailed voxel grayscale distributions at each moment are obtained, enabling coal structure analysis to be upgraded from an overall scale to a fine-grained spatial distribution at the voxel level. This is far superior to traditional 2D images or coarse partitioning methods. At each moment, physical field parameters are precisely paired with 3D voxel data, ensuring the consistency of physical field and structural state during subsequent modeling and analysis.

[0059] Step S1 provides a dynamic and realistic three-dimensional data foundation for subsequent core steps such as pore and fracture identification, voxel classification, and structural parameter extraction. The synchronous acquisition and pairing of physical field parameters enables subsequent permeability calculations to accurately respond to changes in different physical fields, realizing the dynamic correlation and accurate prediction of coal permeability and structural evolution.

[0060] In actual engineering, coal bodies often undergo the combined effects of multiple physical fields. The collected data can reflect the real evolution of the coal body structure under multi-field coupling, enhancing the applicability and accuracy of the model. Voxel gray values ​​can quantitatively reflect the microstructural information such as internal density and pore / fracture distribution of the material, providing a high-resolution data source for subsequent voxel classification, structural parameter extraction, and spatial distribution modeling. Physical field parameters are the direct driving factors of coal body structure evolution and permeability changes. The one-to-one pairing of physical fields and structural data ensures the accuracy of physical field response during modeling, realizing the modeling of the dynamic evolution of coal body structure and permeability.

[0061] In this embodiment, when obtaining the original three-dimensional voxel dataset:

[0062] Set the initial values ​​of axial stress, pore pressure and temperature, and determine the synchronous step increment of each physical field parameter to ensure that the three physical field parameters increase synchronously during the experiment. In practice, on a triaxial experimental device or a dedicated multiphysics loading system, set the initial values ​​of axial stress, pore pressure and temperature, and preset the step increment of each physical field parameter to ensure that the three parameters increase synchronously during the experiment.

[0063] At fixed time intervals, axial stress, pore pressure and temperature are increased simultaneously. Each increase is recorded as a moment, so that each moment corresponds to a set of synchronously changed physical field parameters. According to the fixed time intervals preset in the experiment, three-dimensional CT scans are automatically or manually triggered to perform tomographic imaging on coal core samples, and the spatial coordinates and gray values ​​of all voxels at that moment are collected. The gray values ​​can be directly output by the CT equipment, representing voxel density or material microstructure information.

[0064] At each moment, a three-dimensional CT scan is performed to obtain the spatial coordinates and corresponding gray values ​​of all voxels in the three-dimensional space of the coal core sample, which are then calibrated as the three-dimensional voxel gray value data at that moment.

[0065] The grayscale data of the three-dimensional voxels obtained at each scanning moment are paired with the physical field parameters corresponding to that moment. The correspondence between time, physical field parameters and grayscale values ​​of three-dimensional voxels is established. After each scan, the physical field parameters and grayscale data of three-dimensional voxels at that moment are paired and stored. The one-to-one correspondence between time, physical field parameters and grayscale values ​​of three-dimensional voxels is established. All the collected data are organized in chronological order to form a complete original three-dimensional voxel dataset.

[0066] The three-dimensional voxel grayscale data and physical field parameters at all times are organized in chronological order to form an original three-dimensional voxel dataset arranged by time. The original three-dimensional voxel dataset is a set of data at each time that contains the spatial coordinates and corresponding grayscale values ​​of all voxels in the coal core sample room, and is paired with the physical field parameters at the corresponding time.

[0067] Existing technologies often only control a single physical field or cannot achieve synchronous increments, resulting in a lack of comprehensiveness in structural evolution data. This embodiment ensures synchronous increments of multiple physical field parameters, which can truly reflect the dynamic response of the coal body under complex working conditions, improve the representativeness and engineering value of the data, and achieve high temporal resolution data acquisition through time-sharing scanning and data pairing. Compared with traditional methods that only compare the initial / final state, it can capture the gradual changes in structural details, which is convenient for subsequent evolution modeling and dynamic analysis. Three-dimensional CT voxel scanning enables spatial structural analysis to achieve voxel-level resolution, which is superior to two-dimensional slicing or macroscopic partitioning methods, laying the foundation for fine structure identification and parameter extraction.

[0068] Setting initial axial stress, pore pressure, and temperature to represent the actual stress, fluid pressure, and thermal environment of the coal body is consistent with engineering practice, ensuring that experimental data can be used in various practical scenarios and improving the applicability of the scheme. Simultaneously increasing the physical field parameters reflects the dynamic evolution of the coal body structure under the coupling of multiple fields, making it easier to obtain full-process data of the structure changing with the physical field and enhancing the model's predictive ability.

[0069] The initial value of axial stress is set based on the actual geological depth of the coal core sample, the engineering background of the mining area, or the experimental purpose, with reference to formation pressure, well depth stress, or relevant standards. For example, it can be set close to the original stress of the coal seam or according to the safe range of the experimental equipment. The initial value of pore pressure is set based on the internal fluid pressure of the coal body or the seepage scenario to be simulated, combined with field measurements or relevant literature data. The initial value of temperature is set based on the natural temperature conditions of the coal body environment or the actual needs of special working conditions.

[0070] The step increment is generally set based on the experimental design objectives and the requirements of actual engineering conditions. For example, to simulate the evolution of coal during mining activities or gas injection environments, the step increment should reflect the actual rate of change of stress, pressure, or temperature. Considering the loading accuracy, adjustability, and safety limits of the experimental equipment, the step increment should not exceed the equipment's allowable range to ensure the controllability of the experimental process and the safety of the sample. The time interval needs to be coordinated with the stepping of physical field parameters to ensure that the physical field parameters have stabilized to the next set value at each data acquisition. The time interval can be set based on the response time of the loading equipment and the expected rate of change of the coal structure.

[0071] S2: Perform grayscale normalization on the original three-dimensional voxel dataset, classify each voxel using a set grayscale threshold, distinguish the coal body region from the pore and fracture region, mark and identify the pore and fracture region, and obtain the distribution information of pore structure and fracture structure at each time step.

[0072] Step S2 involves grayscale standardization of the original three-dimensional voxel dataset and classification of each voxel using a set grayscale threshold. This process achieves accurate differentiation between the coal body region and the pore and fracture region. This process lays the foundation for subsequent extraction of structural feature parameters and spatial distribution modeling. Standardization eliminates grayscale deviations between different CT scan batches, ensuring data consistency and comparability. Setting a grayscale threshold as the classification basis makes the voxel classification process more objective, improving the efficiency and accuracy of structural identification.

[0073] After adopting voxel classification, the pore and fracture regions are marked and identified, so that the pore structure and fracture structure at each time point can be effectively extracted and distinguished in three-dimensional space. This not only provides the necessary data source for the subsequent calculation of porosity parameters, fracture width parameters and fracture connectivity parameters, but also enables the coal body structure evolution process to be visualized and quantified at the voxel level.

[0074] Gray-scale normalization eliminates instrument errors and batch variations in the original data, ensuring the scientific rigor of subsequent analyses. Gray-scale thresholding reduces subjective errors caused by human intervention in voxel classification. Marking and identifying pore and fracture regions directly improves the accuracy and completeness of spatial structure extraction, enabling clear segmentation of complex coal body structures and providing a data foundation for the extraction and evolution analysis of dynamic structural characteristic parameters.

[0075] Furthermore, the grayscale threshold was determined through statistical analysis and comparative experiments of voxel grayscale data obtained from three-dimensional CT scans of coal core samples. First, representative sample areas were selected, and typical locations of the coal body, pores, and fractures were identified, with grayscale values ​​recorded for these areas. Subsequently, frequency statistics were performed on all voxel grayscale values, and a grayscale distribution histogram was plotted. By observing the peaks and valleys of the grayscale distribution, the grayscale boundary points between different structures were determined.

[0076] In this embodiment, when performing grayscale normalization on the original three-dimensional voxel dataset, the grayscale value data of the three-dimensional voxels at each time step is obtained, the maximum and minimum grayscale values ​​in the three-dimensional voxel grayscale value data are determined, and the grayscale value of each voxel is processed according to the minimum-maximum normalization, that is, the minimum grayscale value is subtracted from the current voxel grayscale value, and then divided by the difference between the maximum and minimum values, so that all voxel grayscale values ​​are normalized to between 0 and 1.

[0077] For the 3D voxel grayscale data at multiple time points, min-max normalization is performed independently on the 3D voxel grayscale data at each time point to obtain the 3D voxel grayscale data after grayscale normalization at each time point, which is then labeled as a 3D voxel dataset.

[0078] The original 3D voxel dataset is subjected to grayscale normalization. First, the grayscale values ​​of all voxels at each time step are extracted individually. The maximum and minimum grayscale values ​​in the dataset are obtained by traversing the dataset. The purpose is to obtain the grayscale distribution range of the overall internal structure of the sample under CT scan conditions at that time step. Then, the grayscale value of each voxel is normalized according to the minimum-maximum normalization method. That is, the minimum grayscale value is subtracted from the current voxel's grayscale value, and then divided by the difference between the maximum and minimum grayscale values. The grayscale values ​​of all voxels are uniformly mapped to the range of 0 to 1. The above normalization process is performed independently for data from multiple time steps, and finally the normalized 3D voxel grayscale value dataset at each time step is obtained.

[0079] During CT imaging, the absolute values ​​of grayscale may fluctuate at different scanning times or in different batches due to various factors such as equipment conditions, sample density, scanning parameters, and external environment. Directly classifying or analyzing the original grayscale values ​​can easily introduce systematic errors. Standardization can eliminate instrument drift and background differences between different batches and time points, making the voxel grayscale data obtained at different times and in different physical fields comparable and consistent, thereby ensuring the objectivity of subsequent voxel classification, structure identification, and parameter statistics.

[0080] The maximum and minimum gray values ​​required for grayscale normalization are obtained directly by traversing and comparing the grayscale dataset of 3D voxels at each time step, without the need for external parameters. Traditional voxel data processing often ignores the issue of grayscale consistency between different scanning times and different batches, leading to significant deviations in subsequent structural segmentation and parameter analysis results.

[0081] Furthermore, when distinguishing between the coal body region and the pore and fracture region, based on the pre-set grayscale threshold that best distinguishes between the coal body and the pore and fracture, the three-dimensional voxel dataset at each time moment is traversed, and the grayscale value of each voxel is judged. If the grayscale value of the voxel is higher than or equal to the grayscale threshold, the voxel is marked as the coal body region. If the grayscale value of the voxel is lower than the grayscale threshold, the voxel is marked as the pore and fracture region, thus distinguishing between the coal body region and the pore and fracture region at each time moment.

[0082] The step of distinguishing between the coal body and pore / fracture regions is achieved by setting a grayscale threshold and then performing a voxel-by-voxel traversal judgment on the standardized three-dimensional voxel dataset. Specifically, firstly, based on the sample's grayscale distribution characteristics, experimental experience, or automated algorithms, a grayscale threshold that can best distinguish between the coal body and pores / fractures is selected. Then, each voxel is traversed, and its grayscale value is compared with the threshold. If it is higher than or equal to the threshold, the voxel is marked as a coal body region; if it is lower than the threshold, it is marked as a pore / fracture region. This step is usually completed automatically by data processing programs or image segmentation algorithms. The acquisition of parameters depends on prior grayscale statistical analysis, experimental calibration, and segmentation evaluation to ensure the scientific validity and accuracy of the classification results. This embodiment divides the complex spatial structure of coal into two major categories—body and pores / fractures—with voxel-level precision, significantly improving the efficiency and objectivity of structure recognition.

[0083] In this embodiment, the three-dimensional voxel dataset corresponding to each time step is traversed to filter out all voxels classified as pore and fracture regions, and these voxel sets are defined as candidate pore and fracture voxel sets.

[0084] For the candidate pore and fracture voxel set, a three-dimensional spatial connectivity analysis algorithm is used to process them one by one. That is, according to the adjacency relationship of voxels in three-dimensional space, it is determined whether any two pore and fracture voxels are directly connected or indirectly connected. All connected voxels are automatically aggregated into a spatial structural unit, and each spatial structural unit represents an independent pore or fracture.

[0085] For each spatial structural unit, the spatial coordinates of all voxels contained within that unit are statistically analyzed. By comparing the X, Y, and Z coordinates of all voxels within the unit, the maximum span of that unit in the X, Y, and Z directions is obtained.

[0086] For each spatial structural unit, compare its maximum span values ​​in three spatial directions. If the maximum span of the spatial structural unit in one direction is greater than or equal to at least twice the average of the maximum spans in the other two directions, then mark the spatial structural unit as a fractured structural unit; otherwise, mark the spatial structural unit as a porous structural unit.

[0087] The labeling results of each spatial structural unit are associated with the spatial coordinates of each voxel and the number of voxels inside it. The spatial structural units labeled as pore structural units and crack structural units are classified as pore structure distribution and crack structure distribution, respectively.

[0088] The classification and spatial distribution results of the pore structure distribution and fracture structure distribution of all spatial structural units are organized in chronological order to form the distribution information of pore structure and fracture structure at each time point.

[0089] In this embodiment, the three-dimensional voxel dataset corresponding to each time step is traversed, and all voxels classified as pore and fracture regions are selected. These voxel sets are defined as candidate pore and fracture voxel sets, which are the foundation for achieving fine identification of coal seam spatial structure. Subsequently, the candidate voxel sets are processed using a three-dimensional spatial connectivity analysis algorithm. Based on the adjacency relationship of voxels in three-dimensional space, it is automatically determined whether any two pore and fracture voxels are directly or indirectly connected, thereby aggregating all connected voxels into a spatial structural unit, realizing the spatial segmentation of independent pores and fractures. This process is usually automatically completed by image processing algorithms or spatial data structure analysis software. Internal parameters such as voxel coordinates and adjacency rules are all derived from three-dimensional scanning data and voxel mesh structure, requiring no external manual intervention.

[0090] For each spatial structural unit, the spatial coordinates of all voxels contained therein are counted. By calculating the maximum and minimum values ​​of voxel coordinates in the X, Y, and Z directions, the maximum span of the spatial structural unit in each direction is obtained. The maximum spans are then compared. If the maximum span in one direction is greater than or equal to at least twice the average of the maximum spans in the other two directions, the structural unit is marked as a fractured structural unit; otherwise, it is marked as a porous structural unit. This classification method can effectively distinguish the geometric characteristics of spatial structural units and determine the properties of complex three-dimensional structures.

[0091] By associating the labeling results of each spatial structural unit with the spatial coordinates and quantity of its internal voxels, the structural unit attributes, spatial location, and voxel details are combined. Subsequently, the spatial structural units labeled as pores and fractures are classified into pore structure distribution and fracture structure distribution, respectively, forming the spatial structure distribution information of the coal sample at each time step. By organizing the data at all times, the evolution process of the pore and fracture structure of the coal can be dynamically tracked, providing input data for subsequent structural parameter statistics, voxel feature assignment, and permeability modeling.

[0092] Furthermore, a three-dimensional spatial connectivity analysis algorithm is used to process the set of pore and fracture voxels one by one. In specific implementation, a three-dimensional voxel model is first established for all voxels that have been marked as pore and fracture regions. By traversing the voxel set, each unclassified pore and fracture voxel is selected as the starting point. A breadth-first search algorithm is used to retrieve all voxels that are directly adjacent or indirectly connected to the current voxel. The adjacency relationship is usually set according to the 26-neighborhood, 18-neighborhood or 6-neighborhood rules to ensure the physical reality of voxel connectivity in three-dimensional space.

[0093] Each voxel directly connected to the current connected cluster is added to the same spatial structural unit, and the search continues to expand until all interconnected voxels are classified into the same structural unit. This process continues until all pore and fracture voxels are classified into a spatial structural unit or an isolated unit. Each independent voxel connected cluster represents an independent pore or fracture structural unit. The three-dimensional spatial connectivity analysis algorithm is used to segment and classify voxel data into structural units. It is a well-known basic technology in the fields of image processing and spatial data analysis, and has been widely used in many fields such as CT image segmentation, medical image analysis, geological modeling, and materials science.

[0094] S3: Based on the distribution information of pore structure, the number of pore voxels at each time moment is counted. The porosity parameter is obtained by the ratio of the number of pore voxels to the total number of voxels in the coal core sample. At the same time, the distribution information of fracture structure is geometrically analyzed to extract fracture width and fracture connectivity parameters. The porosity parameter, fracture width parameter, and fracture connectivity parameter are used as structural characteristic parameters of the coal body.

[0095] Based on the distribution information of pore structure, the number of pore voxels at each moment is counted, and the porosity parameter is obtained by the ratio of the number of pore voxels to the total number of voxels in the coal core sample. This is the basis for realizing the spatial quantification and parameterization of coal structure. This method can accurately reflect the actual proportion of pore space in coal and avoid the errors caused by traditional two-dimensional or low-resolution estimation.

[0096] By performing geometric analysis on the distribution information of fracture structures and extracting fracture width and connectivity parameters, the complex spatial structure of the coal body is further transformed into quantifiable characteristic indicators. The fracture width parameter reflects the scale characteristics of the fluid channels inside the coal body through spatial geometric statistics, while the connectivity parameter characterizes the connectivity of the fracture network and has a direct impact on the permeability, mechanical response, and fluid migration of the coal body. Compared with existing technologies, traditional methods often only make rough estimates of the overall porosity or fracture density, lacking the data foundation of dynamic three-dimensional spatial distribution, and making it difficult to capture the dynamic evolution and heterogeneity details of the structure. This embodiment can accurately extract structural characteristic parameters at every moment and at every spatial location through voxel-level distribution analysis.

[0097] Porosity, fracture width, and fracture connectivity parameters not only constitute the structural index system of coal bodies, but also serve as the basic inputs for subsequent voxel-level feature assignment, permeability calculation, and evolution analysis. Porosity parameters quantify spatial distribution, fracture width parameters provide a true reflection of structural scale, and connectivity parameters ensure the physical interpretability of connectivity and fluid migration capabilities.

[0098] In this embodiment, when obtaining the porosity parameter, based on the distribution information of the pore structure at each time point, the number of voxels inside all marked as pore structure units at that time point is counted, the total number of voxels in the scanning area of ​​the coal core sample is counted, and the porosity parameter at that time point is obtained by dividing the number of voxels inside the pore structure unit at that time point by the total number of voxels.

[0099] The method for obtaining porosity parameters involves, at each time step, first counting the number of voxels marked as pore structure units based on the results of three-dimensional voxel classification and spatial structure unit identification. This statistical process involves traversing the voxel dataset and accumulating the voxel counts of each voxel belonging to a pore structure unit to obtain the total number of voxels in the pore space inside the coal body at that time step. Simultaneously, the total number of voxels in the scanning area is counted, i.e., the total number of voxels in the three-dimensional voxel grid of the entire coal core sample at that time step is obtained. By dividing the number of voxels in the pore structure unit by the total number of voxels, the porosity parameter at that time step is directly calculated, thereby quantifying the spatial proportion of the microstructure of the coal body.

[0100] The reason for adopting this voxel-level statistical method is that CT scans and 3D voxel modeling can achieve high-resolution spatial structure identification, avoiding the errors and subjectivity in porosity calculations caused by traditional 2D slicing or empirical estimation. Voxel statistics are not only highly automated, but can also dynamically update porosity parameters with changes in time and physical fields, fully reflecting the evolution of coal body structure. All parameters are obtained from the 3D voxel dataset and spatial structure unit labeling results.

[0101] Furthermore, at each moment, for each fracture structure unit, the three-dimensional spatial coordinate information of all fracture voxels within the fracture structure unit is extracted, and the principal component analysis method is used to calculate the principal axis direction of the coordinates of all voxels within the fracture structure unit, which is taken as the principal extension direction of the fracture in the fracture structure unit.

[0102] Along the main axis, at preset intervals, several two-dimensional plane sections orthogonal to the main axis are cut. On each section, voxels belonging to the section are identified, and the maximum Euclidean distance of the voxels on the section plane is calculated as the crack width value of the section.

[0103] The width values ​​measured on all cross sections are statistically analyzed and averaged to obtain the representative crack width value of the crack structure element. The above steps are repeated for all crack structure elements at each time point to obtain the representative crack width value of each crack structure element. The representative crack width values ​​of all crack structure elements at each time point are obtained, and the average value is calculated and used as the crack width parameter at that time point.

[0104] For each fracture structure unit, the three-dimensional spatial coordinate information of all fracture voxels within the unit is first extracted. Principal component analysis is then used to perform eigenvalue decomposition on these coordinate data to obtain the direction of maximum variance in spatial distribution, which is the principal axis direction of the fracture structure unit. Principal component analysis can automatically identify the main extension direction of the fracture spatial distribution without subjective assumptions and is applicable to fracture structures of various shapes in complex three-dimensional space.

[0105] Along the main axis, two-dimensional plane sections orthogonal to the main axis are cut at preset intervals. On each section, all fracture voxels belonging to the section are identified by determining whether the voxel coordinates fall within the section range. Then, on each section, the maximum Euclidean distance between all fracture voxels is calculated, and this distance is used as the fracture width value of the section. This method can effectively reflect the true width variation of the fracture at different locations and avoid local errors caused by a single section or subjective estimation.

[0106] The width values ​​measured on all cross sections are summarized and statistically analyzed, and their average value is taken as the representative crack width value of the crack structure unit. The above steps are repeated for all crack structure units at each time point to obtain the representative width value of all crack structure units. Finally, the average of these width values ​​is taken as the crack width parameter at that time point. This method ensures the representativeness of the parameter extraction and the balance of spatial distribution, and can truly reflect the overall width characteristics of different crack structures.

[0107] Three-dimensional cracks often exhibit complex extension and deformation in space. Traditional two-dimensional slicing or single-point measurement cannot accurately describe their overall width characteristics. Principal component analysis can adaptively identify the principal axes in space, and cross-sectional width measurement can systematically reflect the width changes along the principal axis, improving the scientific rigor and automation of parameter extraction. All parameters are directly derived from voxel space coordinates and structural unit division results, ensuring the objectivity of the processing flow.

[0108] At each time step, based on all identified fracture structural elements and their spatial coordinate data, a spatial relationship network between fracture structural elements is constructed. Specifically, it is determined whether the nearest distance between any two fracture structural elements is less than a preset threshold. If it is less than the preset threshold, it is determined that the two fracture structural elements have a spatial relationship.

[0109] Based on the spatial relationship network, graph theory is used to identify all sets of fracture structure units with spatial relationships. Each set is a connected cluster. For each connected cluster, the total number of fracture voxels contained in it is counted.

[0110] The proportion of the number of voxels in the largest connected cluster to the total number of voxels in the fracture is determined as the fracture connectivity parameter at that moment.

[0111] At each time step, the spatial coordinate data of all identified fracture structure units are first extracted, and the geometric center or boundary voxel coordinates of each structure unit are used as representative points. Then, for any two fracture structure units, the Euclidean distance between their nearest voxel pairs is calculated to determine whether the distance is less than a preset spatial relationship threshold. If the distance is less than the threshold, the two fracture structure units are considered to be spatially related, that is, there is a potential connection or fluid migration path. All fracture structure units with spatial relationships are connected in the data structure to form a spatial relationship network between fracture structure units.

[0112] Furthermore, based on the spatial relationship network, graph theory methods are used to connect and group all fracture structural units. Through connected component analysis, structural units with spatial relationships are aggregated into a connected cluster. Each connected cluster represents a fracture network that may actually form in space. For each connected cluster, the total number of fracture voxels contained within it is counted, which can reflect the spatial scale and fluid channel capacity of the fracture network in the coal body. Finally, the number of voxels in the largest connected cluster is selected, and its proportion to the total number of fracture voxels is calculated as the fracture connectivity parameter at that moment.

[0113] This spatial relationship network and connectivity cluster analysis method is adopted because the distribution of fractures within the coal body is complex. Identifying individual fracture structural units cannot reflect the connectivity and fluid migration capacity of the overall network. By using spatial distance discrimination and graph theory grouping, potential flow channels and structural continuity can be automatically identified, providing a physical basis for permeability modeling. All parameters are directly derived from voxel coordinates, fracture structural unit grouping, and spatial relationship criteria.

[0114] The setting of the preset threshold is usually selected by combining the actual fracture width distribution of the coal body, the fluid migration scale, and the experimental resolution. A reasonable spatial distance threshold that can reflect the actual fluid migration capacity and exclude the influence of isolated fractures can be set by analyzing the distance distribution between fracture structural units through pre-experiment analysis or by referring to the physical connection scale of the coal body fracture network in existing literature.

[0115] In this embodiment, principal component analysis is used to determine the principal extension direction (i.e., principal axis direction) of each fracture structural unit, so as to achieve objective and automatic extraction of the spatial geometric features of the fracture. The specific implementation process is as follows:

[0116] First, for each fracture structural unit, the three-dimensional spatial coordinate data of all fracture voxels within that structural unit are extracted. The coordinates (X, Y, Z) of these voxels are treated as a point cloud dataset. Then, principal component analysis is performed on this point cloud dataset, that is, the mean coordinates of all points are calculated, and the mean is subtracted from the coordinates of each voxel to form a decentralized three-dimensional coordinate matrix. The covariance matrix of this coordinate matrix is ​​calculated to reflect the distribution variance and correlation of voxels in three-dimensional space.

[0117] Eigenvalue decomposition of the covariance matrix yields three eigenvalues ​​and their corresponding eigenvectors. The magnitude of the eigenvalues ​​reflects the variance of the point cloud distribution in each direction, while the eigenvectors provide the principal directions. The eigenvector with the largest eigenvalue is selected as the principal component direction, defined as the principal axis direction of the fracture structural unit. In practical applications, this principal axis direction represents the maximum extension direction of the fracture spatial distribution, objectively reflecting the overall spatial orientation and principal geometric characteristics of the fracture.

[0118] Once the main axis direction is obtained, two-dimensional cross-sections orthogonal to the main axis can be cut along that direction at set intervals to achieve subsequent width measurement and structural quantification.

[0119] The core advantage of principal component analysis (PCA) lies in its adaptability to complex, curved, or irregular fracture structures. Compared with traditional methods that estimate fracture direction through manual assumptions or two-dimensional projections, PCA can accurately and objectively capture the true principal axis direction of the spatial structure distribution, improving the consistency of geometric parameter extraction.

[0120] S4: Pair the structural feature parameters at each time step with the corresponding physical field parameters and arrange them in chronological order. Based on the structural feature parameters at each time step, assign values ​​to the porosity feature, fracture width feature, and fracture connectivity feature of each voxel according to the voxel classification results to form the voxel feature parameters of each voxel at each time step.

[0121] Step S4 pairs the structural feature parameters at each moment with the corresponding physical field parameters and organizes them in chronological order, which enables a high degree of coupling between structural parameters and physical field data. This processing method allows subsequent analysis to accurately reflect the dynamic evolution of the coal structure under different physical field conditions, improving the temporal integrity and physical correlation of the data. Compared with existing technologies that only have static parameters or a single data stream, this temporal pairing method can provide a solid data foundation for permeability modeling, structural evolution tracking, and multi-field coupling analysis.

[0122] Based on the structural feature parameters at each moment, the porosity, fracture width, and fracture connectivity features of each voxel are assigned values ​​according to the voxel classification results. This achieves precise allocation of structural parameters at the voxel level, which can reflect the heterogeneity of the coal body's spatial structure and make each physical feature correspond one-to-one with the voxel spatial distribution. This not only improves the resolution of 3D modeling but also provides high-quality input for subsequent voxel-level permeability calculation and spatial distribution analysis.

[0123] The temporal pairing of structural feature parameters and physical field parameters provides a complete data chain for dynamic modeling and engineering analysis. Voxel-level parameter assignment ensures the complete expression of model spatial resolution and physical details. The pairing of structural parameters and physical fields enhances temporal modeling capabilities. Parameter assignment based on voxel classification strengthens the characterization of spatial heterogeneity. The temporal sequence arrangement provides data support for long-period structural evolution analysis.

[0124] In this embodiment, the voxel feature parameters include porosity features, fracture width features, and fracture connectivity features. The logic for constituting the voxel feature parameters of each voxel at each time step is as follows:

[0125] For voxels within the coal body region, the porosity, fracture width, and fracture connectivity features are all assigned a value of zero. For voxels within the pore structure, the fracture width and fracture connectivity features are both assigned a value of zero, and the porosity feature is assigned the porosity parameter at that time. For voxels within the fracture structure, the porosity feature is assigned a value of zero, and the fracture width and fracture connectivity features are assigned the fracture width and fracture connectivity parameters at that time.

[0126] In this embodiment, the assignment of voxel feature parameters is performed sequentially based on the spatial classification results of the voxels. Specifically, each voxel is first classified into its spatial structure type: coal body region, pore structure region, or fracture structure region. For voxels classified as coal body regions, their porosity, fracture width, and fracture connectivity features are all set to zero during the assignment process. This is because coal body regions do not physically possess pore or fracture attributes, therefore these parameters do not participate in subsequent permeability modeling and structural analysis. For voxels belonging to pore structure regions, their fracture width and fracture connectivity features are assigned zero, while the porosity feature is assigned the porosity parameter obtained statistically at that moment. This process ensures that each pore voxel only reflects the structural characteristics of the pore space and is not affected by fracture parameters. For voxels belonging to fracture structure regions, the porosity feature is assigned zero, while the fracture width and fracture connectivity features are assigned the fracture width and fracture connectivity parameters obtained at the current moment, respectively. This assignment method ensures that the physical meaning of the voxel parameters is completely consistent with the spatial structural properties. The internal structure of the coal body is highly heterogeneous, and the assignment of voxel-level parameters can fully reflect the spatial distribution and physical properties of the microstructure. All parameters are obtained from the aforementioned three-dimensional structural analysis, geometric statistics, and connectivity discrimination.

[0127] Traditional methods often use overall averaging or regional averaging to assign values, which fails to reflect the spatial heterogeneity and microstructural variations at the voxel level. This approach, through voxel parameter assignment based on classification, not only improves the spatial resolution of 3D modeling but also ensures the consistency and precision of the model's physical properties. Each voxel obtains unique and accurate structural parameters based on its spatial assignment, providing high-quality input for subsequent permeability calculations, structural evolution analysis, and physical field modeling.

[0128] S5: Based on the physical field parameters corresponding to each moment, and combined with the voxel characteristic parameters of all voxels at that moment, the voxel-scale permeability calculation method is used to generate the permeability distribution of each voxel in the three-dimensional space of the coal body.

[0129] Step S5 combines the physical field parameters corresponding to each moment with the voxel characteristic parameters of all voxels and uses a voxel-scale permeability calculation method to accurately generate the permeability distribution of each voxel in the three-dimensional space of the coal body. The advantage of this technical approach is that it can provide a fine-grained description of the spatial heterogeneity of the coal body structure and the physical field response. Compared with the common overall mean or regional average permeability modeling methods in existing technologies, this step can dynamically capture the influence of structural and physical field changes on permeability at the voxel level, and realize the coupled modeling of microstructure and macroscopic seepage behavior.

[0130] By inputting parameters at the voxel scale and using the permeability calculation formula, independent permeability values ​​can be assigned to voxels of different spatial locations and structural types, fully reflecting the differences in permeability of pores, fractures, and bulk space within the coal body under the influence of physical fields.

[0131] The synergistic input of voxel characteristic parameters and physical field parameters enables the permeability distribution to truly reflect the coal body structure evolution and the linkage effect of physical fields, providing a high-quality data foundation for 3D modeling and evolution analysis. The voxel-scale calculation method ensures the refinement and automation of the model, and the temporal response of physical field parameters enhances the dynamic modeling capability.

[0132] In this embodiment, the logic for generating the permeability distribution of each voxel in the three-dimensional space of the coal body is as follows:

[0133] At each moment, for each voxel of the coal core sample, the physical field parameters corresponding to that moment are retrieved, and physical field parameters such as axial stress, pore pressure, and temperature are paired with the voxel. The voxel characteristic parameters of all voxels at that moment are retrieved simultaneously.

[0134] For each voxel, the permeability is calculated by taking the physical field parameters and voxel characteristic parameters as inputs according to the voxel-scale permeability calculation method. The permeability calculation method quantitatively analyzes the permeability performance of the voxel based on the voxel's porosity parameters, fracture width parameters, and fracture connectivity parameters, combined with the current physical field parameters, to obtain the permeability value of the voxel at that moment.

[0135] Arrange the permeability values ​​of all voxels in spatial coordinate order to form the three-dimensional spatial permeability distribution data at that moment, so that the specific position of each voxel in the three-dimensional coordinate system corresponds one-to-one with its corresponding permeability value.

[0136] For all time points, the process of constructing three-dimensional spatial permeability distribution data is repeated sequentially. The three-dimensional spatial permeability distribution data at each time point are stored in chronological order to form a time-series three-dimensional spatial permeability distribution dataset, generating the permeability distribution of each voxel in the three-dimensional space of the coal body.

[0137] In the specific implementation of the process of generating the permeability distribution of each voxel in the three-dimensional space of the coal body, firstly, at each moment, for all voxels of the coal core sample, the corresponding physical field parameters at that moment are retrieved, including axial stress, pore pressure, and temperature, etc., and these physical field parameters are paired with each voxel. At the same time, the voxel characteristic parameters of all voxels at that moment are retrieved simultaneously, including porosity parameters, fracture width parameters, and fracture connectivity parameters. Using the physical field parameters and voxel characteristic parameters as input, the permeability is calculated for each voxel according to the voxel-scale permeability calculation method. This process ensures that the permeability value of each voxel fully considers its spatial structural properties and physical field response, and realizes the organic combination of the microstructure and macroscopic seepage characteristics of the coal body.

[0138] After the calculation is completed, the permeability values ​​of all voxels are arranged in spatial coordinate order to construct the three-dimensional spatial permeability distribution data at that time. This ensures that the specific position of each voxel in the three-dimensional coordinate system corresponds one-to-one with its corresponding permeability value. This data construction process is repeated for all time points, and the three-dimensional spatial permeability distribution data at each time point are stored sequentially in chronological order. Finally, a time-series three-dimensional spatial permeability distribution dataset is formed. This can comprehensively reflect the permeability changes of coal in different physical fields and structural evolution processes, providing solid data support for dynamic analysis and process prediction.

[0139] Coal structure and permeability are inherently highly heterogeneous and dynamically evolving. Only voxel-level distribution calculations can accurately capture spatial differences and temporal changes. All calculation parameters, including voxel spatial coordinates, structural feature parameters, and physical field parameters, are generated from previous 3D modeling, spatial structure identification, and physical field control experiments. No subjective adjustments or external inputs are required, ensuring the scientific validity of the data.

[0140] Furthermore, the specific formula for calculating penetration rate is as follows:

[0141]

[0142] in, Indicates the first At the moment, the first The permeability value of individual elements, This is the benchmark coefficient for permeability. , and These are the weighted indices for porosity, fracture width, and fracture connectivity, respectively. , and These are the influence coefficients of axial stress, pore pressure, and temperature, respectively. , and The first At the moment, the first Porosity characteristics, fracture width characteristics, and fracture connectivity characteristics of individual elements. , and The first Axial stress, pore pressure, and temperature at each moment.

[0143] The calculation of voxel-scale permeability adopts a parameterized exponential-power function coupling formula. In practice, for each voxel, at each time step, its voxel characteristic parameters, including porosity characteristics, fracture width characteristics, and fracture connectivity characteristics, are automatically retrieved, and the physical field parameters at that time step, namely axial stress, pore pressure, and temperature, are simultaneously retrieved. By substituting these parameters into the formula, the permeability value of the voxel at the corresponding time step can be calculated.

[0144] The reason for adopting this formula structure is that coal permeability is affected by the coupling of various structural parameters and physical field factors. Porosity, fracture width, and fracture connectivity reflect the microstructural characteristics of the space in which the voxel is located. Permeability is positively correlated with these structural factors, and their nonlinear contributions are often reflected by power exponents. Physical field parameters such as stress, pressure, and temperature can regulate permeability by affecting the pore and fracture state and fluid dynamics. Using exponential correction terms can better describe the suppression or enhancement effects of these external fields on the permeability system. This configuration is compatible with the engineering practice of the coupling of coal structure and physical fields, and also has good parameter scalability and numerical stability.

[0145] The baseline coefficient for permeability can be initially set based on experimental results or numerical simulations. Weighting index. , and The quantitative contributions of porosity, fracture width, and connectivity to permeability were controlled separately. These contributions were obtained through regression analysis or optimization fitting, referencing theoretical models and actual test data. The influence coefficients were then calculated. , and The sensitivity of permeability to external physical fields is adjusted, and its value can be determined through physical experiments, historical data statistics, or numerical calibration. All weighting parameters can be flexibly adjusted according to the coal material type, experimental scenario, and data sample.

[0146] Compared with existing technologies, traditional permeability models often use global regional parameters, simplify structural assumptions, or ignore the dynamic influence of physical fields, making it difficult to reflect the true permeability behavior of coal bodies at the microscale and under dynamic conditions. This embodiment realizes dynamic response modeling of microstructural heterogeneity and physical field coupling effects through voxel-level parameterization formulas. In the overall modeling process, this formula achieves deep coupling between structural characteristic parameters and physical field parameters. The power function of structural parameters reflects the contribution of structural heterogeneity, the exponential correction of the physical field reflects the dynamic effect of the external environment, and the parameter adjustment mechanism ensures the flexibility and applicability of the scheme.

[0147] This embodiment, by performing grayscale normalization, classification, and spatial structural unit identification on all voxels at each moment, combined with voxel-level parameter statistics and geometric analysis, can automatically extract structural feature parameters such as porosity, fracture width, and fracture connectivity. These parameters are then dynamically paired with synchronously acquired physical field parameters. This not only ensures the physical accuracy and temporal response capability of the parameters but also achieves precise assignment of voxel-level parameters, enabling detailed quantification of coal body heterogeneity and dynamic evolution. Employing a voxel-scale permeability calculation method, it can calculate the permeability value of each voxel at each moment based on voxel feature parameters and physical field parameters, generating a three-dimensional permeability distribution. This overcomes the limitations of regional averages or overall statistics, comprehensively capturing the structural evolution process of coal core samples under different stress, pressure, and temperature conditions, effectively revealing the distribution changes of the coal body, pores, and fractures in three-dimensional space.

[0148] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0149] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0150] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0151] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a division of some logical functions, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0152] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0153] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A three-dimensional modeling method for the correlation between coal permeability and pore fracture evolution, characterized by the following steps: include: Time-series three-dimensional CT scan data of coal core samples under different physical fields were collected to obtain the original three-dimensional voxel dataset containing voxel gray values ​​at each time point, and the physical field parameters corresponding to each time point were recorded synchronously. The physical field parameters include axial stress, pore pressure and temperature. The original three-dimensional voxel dataset is subjected to grayscale normalization. Each voxel is classified using a set grayscale threshold to distinguish between the coal body region and the pore and fracture region. The pore and fracture region is marked and identified to obtain the distribution information of pore structure and fracture structure at each time step. Based on the distribution information of pore structure, the number of pore voxels at each time moment is counted. The porosity parameter is obtained by the ratio of the number of pore voxels to the total number of voxels in the coal core sample. At the same time, the distribution information of fracture structure is geometrically analyzed to extract fracture width and fracture connectivity parameters. The porosity parameter, fracture width parameter, and fracture connectivity parameter are used as structural characteristic parameters of the coal body. The structural feature parameters at each time step are paired with the corresponding physical field parameters and arranged in chronological order. Based on the structural feature parameters at each time step, the porosity feature, fracture width feature, and fracture connectivity feature of each voxel are assigned values ​​according to the voxel classification results, thus forming the voxel feature parameters of each voxel at each time step. Based on the physical field parameters corresponding to each moment, and combined with the voxel characteristic parameters of all voxels at that moment, the voxel-scale permeability calculation method is used to generate the permeability distribution of each voxel in the three-dimensional space of the coal body. The voxel feature parameters include porosity features, fracture width features, and fracture connectivity features. The logic for constituting the voxel feature parameters of each voxel at each time step is as follows: For voxels within the coal body region, the porosity, fracture width, and fracture connectivity features are all assigned a value of zero. For voxels within the pore structure, the fracture width and fracture connectivity features are both assigned a value of zero, and the porosity feature is assigned the porosity parameter at that time. For voxels within the fracture structure, the porosity feature is assigned a value of zero, and the fracture width and fracture connectivity features are assigned the fracture width and fracture connectivity parameters at that time. The logic for generating the permeability distribution of each voxel in the three-dimensional space of the coal body is as follows: At each moment, for each voxel of the coal core sample, the physical field parameters corresponding to that moment are retrieved, and physical field parameters such as axial stress, pore pressure, and temperature are paired with the voxel. The voxel characteristic parameters of all voxels at that moment are retrieved simultaneously. For each voxel, the permeability is calculated by taking the physical field parameters and voxel characteristic parameters as inputs according to the voxel-scale permeability calculation method. The permeability calculation method quantitatively analyzes the permeability performance of the voxel based on the voxel's porosity parameters, fracture width parameters, and fracture connectivity parameters, combined with the current physical field parameters, to obtain the permeability value of the voxel at that moment. Arrange the permeability values ​​of all voxels in spatial coordinate order to form the three-dimensional spatial permeability distribution data at that moment, so that the specific position of each voxel in the three-dimensional coordinate system corresponds one-to-one with its corresponding permeability value. For all time points, the process of constructing three-dimensional spatial permeability distribution data is repeated sequentially. The three-dimensional spatial permeability distribution data at each time point are stored sequentially in chronological order to form a time-series three-dimensional spatial permeability distribution dataset, which generates the permeability distribution of each voxel in the three-dimensional space of the coal body. The specific formula for calculating penetration rate is as follows: in, Indicates the first At the moment, the first The permeability value of individual elements, This is the benchmark coefficient for penetration rate. , and These are the weighted indices for porosity, fracture width, and fracture connectivity, respectively. , and These are the influence coefficients of axial stress, pore pressure, and temperature, respectively. , and The first At the moment, the first Porosity characteristics, fracture width characteristics, and fracture connectivity characteristics of individual elements. , and The first Axial stress, pore pressure, and temperature at each moment.

2. The three-dimensional modeling method for the correlation between coal permeability and pore fracture evolution according to claim 1, characterized in that, When obtaining the original 3D voxel dataset: Set the initial values ​​for axial stress, pore pressure, and temperature, and determine the synchronous step increment for each physical field parameter to ensure that the three physical field parameters increase synchronously during the experiment. Axial stress, pore pressure and temperature are increased simultaneously at fixed time intervals. Each increase is recorded as a moment, so that each moment corresponds to a set of synchronously changed physical field parameters. At each moment, a three-dimensional CT scan is performed to obtain the spatial coordinates and corresponding gray values ​​of all voxels in the three-dimensional space of the coal core sample, which are then calibrated as the three-dimensional voxel gray value data at that moment. Establish a pairing relationship between the three-dimensional voxel grayscale data obtained from each scanning moment and the physical field parameters corresponding to that moment, and establish the correspondence between the moment, physical field parameters and three-dimensional voxel grayscale values; The three-dimensional voxel grayscale data and physical field parameters at all times are organized in chronological order to form an original three-dimensional voxel dataset arranged by time. The original three-dimensional voxel dataset is a set of data at each time that contains the spatial coordinates and corresponding grayscale values ​​of all voxels in the coal core sample room, and is paired with the physical field parameters at the corresponding time.

3. The three-dimensional modeling method for the correlation between coal permeability and pore fracture evolution according to claim 2, characterized in that, When performing grayscale normalization on the original 3D voxel dataset, the grayscale values ​​of the 3D voxels at each time step are obtained, the maximum and minimum grayscale values ​​in the 3D voxel grayscale data are determined, and the grayscale value of each voxel is processed according to the minimum-maximum normalization, that is, the minimum grayscale value is subtracted from the current voxel grayscale value, and then divided by the difference between the maximum and minimum values, so that all voxel grayscale values ​​are normalized to between 0 and 1. For the 3D voxel grayscale data at multiple time points, the min-max normalization process is performed independently on the 3D voxel grayscale data at each time point to obtain the 3D voxel grayscale data after grayscale normalization at each time point, which is labeled as a 3D voxel dataset. When distinguishing between the coal body region and the pore and fracture region, the three-dimensional voxel dataset at each time step is traversed according to the pre-set grayscale threshold that best distinguishes between the coal body and the pore and fracture region. The grayscale value of each voxel is judged. If the grayscale value of the voxel is higher than or equal to the grayscale threshold, the voxel is marked as the coal body region. If the grayscale value of the voxel is lower than the grayscale threshold, the voxel is marked as the pore and fracture region. This distinguishes between the coal body region and the pore and fracture region at each time step.

4. The three-dimensional modeling method for the correlation between coal permeability and pore fracture evolution according to claim 3, characterized in that, By iterating through the three-dimensional voxel dataset corresponding to each time step, all voxels classified as pore and fracture regions are selected, and these voxel sets are defined as candidate pore and fracture voxel sets. For the candidate pore and fracture voxel set, a three-dimensional spatial connectivity analysis algorithm is used to process them one by one. That is, according to the adjacency relationship of voxels in three-dimensional space, it is determined whether any two pore and fracture voxels are directly connected or indirectly connected. All connected voxels are automatically aggregated into a spatial structural unit, and each spatial structural unit represents an independent pore or fracture. For each spatial structural unit, the spatial coordinates of all voxels contained within that unit are statistically analyzed. By comparing the X, Y, and Z coordinates of all voxels within the unit, the maximum span of that unit in the X, Y, and Z directions is obtained. For each spatial structural unit, compare its maximum span values ​​in three spatial directions. If the maximum span of the spatial structural unit in one direction is greater than or equal to at least twice the average of the maximum spans in the other two directions, then mark the spatial structural unit as a fractured structural unit; otherwise, mark the spatial structural unit as a porous structural unit. The labeling results of each spatial structural unit are associated with the spatial coordinates of each voxel and the number of voxels inside it. The spatial structural units labeled as pore structural units and crack structural units are classified as pore structure distribution and crack structure distribution, respectively. The classification and spatial distribution results of the pore structure distribution and fracture structure distribution of all spatial structural units are organized in chronological order to form the distribution information of pore structure and fracture structure at each time point.

5. A three-dimensional modeling method for the correlation between coal permeability and pore fracture evolution according to claim 4, characterized in that, When obtaining porosity parameters, based on the distribution information of pore structure at each time point, the number of voxels inside all marked as pore structure units at that time point is counted, the total number of voxels in the scanning area of ​​the coal core sample is counted, and the porosity parameters at that time point are obtained by dividing the number of voxels inside the pore structure units at that time point by the total number of voxels. At each moment, for each fracture structure element, the three-dimensional spatial coordinate information of all fracture voxels in the fracture structure element is extracted. Using principal component analysis, the principal axis direction of the coordinates of all voxels in the fracture structure element is calculated as the principal extension direction of the fracture in the fracture structure element. Along the main axis, at preset intervals, several two-dimensional plane sections orthogonal to the main axis are cut. On each section, voxels belonging to the section are identified, and the maximum Euclidean distance of the voxels on the section plane is calculated as the crack width value of the section. The width values ​​measured on all cross sections are statistically analyzed and averaged to obtain the representative crack width value of the crack structure unit. The above steps are repeated for all crack structure units at each time point to obtain the representative crack width value of each crack structure unit. The representative crack width values ​​of all crack structure units at each time point are obtained, and the average value is calculated and used as the crack width parameter at that time point. At each time step, based on all identified fracture structural elements and their spatial coordinate data, a spatial relationship network between fracture structural elements is constructed. Specifically, it is determined whether the nearest distance between any two fracture structural elements is less than a preset threshold. If it is less than the preset threshold, it is determined that the two fracture structural elements have a spatial relationship. Based on the spatial relationship network, graph theory is used to identify all sets of fracture structure units with spatial relationships. Each set is a connected cluster. For each connected cluster, the total number of fracture voxels contained in it is counted. The proportion of the number of voxels in the largest connected cluster to the total number of voxels in the fracture is determined as the fracture connectivity parameter at that moment.

Images