Negative Poisson's ratio performance mixture filling method applied to goaf filling

Abstract

The invention discloses a negative Poisson's ratio performance mixture filling method applied to goaf filling, which comprises the following steps: acquiring three-dimensional form data of a goaf, and constructing a three-dimensional digital model of the goaf; according to the goaf three-dimensional digital model, the volume and distribution density of goaf fractures are obtained, and the negative Poisson's ratio mixture filling amount and a flow turnout layout scheme are determined; obtaining a fluidity parameter of crack filling, and if the fluidity parameter meets a preset threshold value, optimizing a negative Poisson's ratio mixture ratio through numerical simulation to obtain an optimal value of a transverse expansion characteristic; designing parameters of the lattice structure, and generating lattice structure data of the filling body; and through a flow turnout guide algorithm, in combination with the goaf three-dimensional digital model and lattice structure data, obtaining a flow direction path of the negative Poisson's ratio mixture in the fracture, updating a flow turnout layout scheme, and generating an optimized filling scheme. Precise filling of the goaf cracks is achieved, the filling effect is improved, and the stability of the goaf is enhanced.

Description

Technical Field

[0001] The invention belongs to the technical field of mining area filling, and in particular relates to a negative Poisson's ratio performance mixture filling method for filling goaf areas. Background Art

[0002] Goaf management is an important issue faced by mining after mining. Its purpose is to prevent goaf collapse, surface subsidence and related geological disasters. However, current goaf filling technology faces many challenges:

[0003] Difficulty in grasping the three-dimensional shape: Traditional filling methods make it difficult to accurately obtain the three-dimensional shape and crack distribution of the goaf, resulting in uneven filling and inability to effectively support the rock layer at the top of the goaf.

[0004] Filling materials have a difficult time achieving a balance of performance: they must possess good fluidity to flow fully into the cracks, while also forming a high-strength filling body to provide adequate support. Existing materials often struggle to meet both of these requirements.

[0005] Insufficient dynamic monitoring: The goaf will undergo dynamic changes during the filling process, and traditional methods make it difficult to grasp the subsidence situation in real time and adjust the filling strategy in time.

[0006] These problems are interrelated and together constitute the technical difficulties of accurate filling of goaf. In order to solve these problems, it is urgent to propose a filling method of negative Poisson's ratio performance mixture for goaf filling. Summary of the Invention

[0007] In order to solve the above technical problems, the present invention proposes a filling method of a negative Poisson's ratio mixture for filling goaf to solve the problems existing in the above-mentioned prior art.

[0008] To achieve the above object, the present invention provides a method for filling a goaf with a negative Poisson's ratio mixture, comprising the following steps:

[0009] Collecting three-dimensional morphological data of the goaf and constructing a three-dimensional digital model of the goaf, wherein the three-dimensional digital model of the goaf includes crack distribution and spatial characteristics;

[0010] According to the three-dimensional digital model of the goaf, the volume and distribution density of the cracks in the goaf are obtained, and the filling amount of the negative Poisson's ratio mixture and the flow switch layout plan are determined;

[0011] According to the rheological properties of negative Poisson's ratio mixture, the relationship between viscosity and shear rate is analyzed to obtain the fluidity parameters of crack filling;

[0012] If the fluidity parameter meets the preset threshold, the negative Poisson's ratio mixture ratio is optimized through numerical simulation to obtain the optimal value of the lateral expansion characteristic;

[0013] Based on the optimal value of the lateral expansion characteristics, the parameters of the lattice structure are designed to generate the lattice structure data of the filling body;

[0014] By using the flow switch guidance algorithm, combined with the three-dimensional digital model of the goaf and the lattice structure data, the flow path of the negative Poisson's ratio mixture in the cracks is obtained, and the flow switch layout plan is updated.

[0015] Based on the updated flow turnout layout plan, the subsidence displacement changes of the goaf are monitored in real time. According to the real-time monitoring results, the lattice structure data and negative Poisson's ratio mixture ratio are updated to generate an optimized filling plan.

[0016] Optionally, collecting three-dimensional morphological data of the goaf and constructing a three-dimensional digital model of the goaf includes:

[0017] Acquire 3D morphological data from the goaf and use stereo microscope scanning technology to obtain original point cloud data;

[0018] For the original point cloud data, if the point cloud density is lower than the preset threshold, the interpolation algorithm is used to supplement the data to obtain uniform point cloud data;

[0019] Through uniform point cloud data, the volume segmentation algorithm is used to extract the crack distribution characteristics and determine the crack spatial location information;

[0020] Based on the spatial location information of the cracks, the spatial characteristic information including the crack distribution is constructed to obtain a preliminary digital model;

[0021] From the preliminary digital model, a surface fitting algorithm is used to generate a smooth spatial surface and obtain a three-dimensional digital model of the goaf.

[0022] Optionally, obtaining the volume and distribution density of cracks in the goaf based on the three-dimensional digital model of the goaf, and determining the filling amount of the negative Poisson's ratio mixture and the flow switch layout plan include:

[0023] The crack distribution data is obtained from the three-dimensional digital model of the goaf, and the crack positioning points are extracted using the volume analysis method to obtain the crack space coordinate set;

[0024] The total volume of the fracture is calculated using the fracture spatial coordinate set. If the total volume exceeds a preset threshold, a segmentation algorithm is used to divide the fracture area and determine the fracture sub-area set.

[0025] According to the set of fracture sub-regions, the distribution density value of each sub-region is calculated to obtain the density distribution matrix;

[0026] Extract high-density areas from the density distribution matrix and generate a mixture filling distribution map using a mapping algorithm;

[0027] Based on the mixture filling distribution map, candidate position points of the mobile turnout are obtained. If the distance between the candidate position points is less than a preset threshold, the position points are merged to obtain the optimized turnout location set;

[0028] Generate turnout connection lines through turnout location sets, adjust connection line distribution using path optimization methods, and determine the mobile turnout layout plan;

[0029] The spatial feature information is extracted from the turnout layout scheme, and the corresponding relationship between the filling amount and the turnout distribution is generated, so as to obtain the final negative Poisson's ratio mixture filling amount and the flow turnout layout scheme.

[0030] Optionally, analyzing the relationship between viscosity and shear rate based on the rheological properties of the negative Poisson's ratio mixture to obtain the fluidity parameters for crack filling includes:

[0031] Based on rheological experimental data, a viscosity-shear rate relationship model of negative Poisson's ratio mixtures was constructed;

[0032] According to the relationship model between viscosity and shear rate, the experimental data were fitted to obtain the curve of viscosity changing with shear rate;

[0033] According to the change curve, key turning points are extracted and viscosity parameter sets are generated;

[0034] If the viscosity value in the viscosity parameter set exceeds the preset fluidity threshold, the fitting parameters are adjusted and the curve is regenerated to obtain an optimized viscosity parameter set;

[0035] Extracting the fluidity parameters that match the crack filling requirements from the optimized viscosity parameter set;

[0036] Using a classification algorithm, the liquidity parameters are divided into high liquidity group and low liquidity group, and the parameter range of the high liquidity group is determined;

[0037] Generate a formula adjustment plan for the mixture through the parameter range of the high fluidity group;

[0038] If the deviation between the adjusted liquidity parameter of the formula and the target liquidity is greater than a preset threshold, the formula is iteratively optimized to obtain the adjusted liquidity parameter;

[0039] According to the adjusted fluidity parameters, the flow behavior of the mixture in the cracks is simulated;

[0040] Use numerical simulation tools to generate flow path distribution maps and determine the flow coverage corresponding to the mobility parameters;

[0041] Extract the mobility parameter scheme that matches the fracture filling requirements from the mobility coverage;

[0042] If the coverage range meets the preset filling ratio, the final fluidity parameter solution is output to obtain the final fluidity parameter of the crack filling.

[0043] Optionally, if the fluidity parameter satisfies a preset threshold, optimizing the negative Poisson's ratio mixture ratio through numerical simulation to obtain an optimal value of the lateral expansion characteristic includes:

[0044] If the fluidity parameter meets the preset threshold, the initial proportioning scheme of the negative Poisson's ratio mixture is generated through numerical simulation;

[0045] According to the initial mix ratio, the finite element analysis tool is used to simulate the lateral expansion behavior of the mixture and obtain the expansion behavior data;

[0046] extracting a lateral expansion characteristic parameter from the expansion behavior data, and adjusting the mix ratio scheme if the lateral expansion characteristic parameter deviates from a preset range to obtain an adjusted mix ratio scheme;

[0047] The Monte Carlo algorithm is used to simulate the expansion behavior of the mixture under different working conditions through the adjusted proportion scheme to determine the expansion behavior distribution;

[0048] Extract key distribution features from the expansion behavior distribution. If the key distribution features meet the preset expansion requirements, output the optimized ratio scheme.

[0049] According to the optimized mix ratio scheme, the mechanical response curve of the mixture is generated to obtain the optimized value of lateral expansion.

[0050] Optionally, the step of designing parameters of the lattice structure based on the optimal value of the lateral expansion characteristic and generating lattice structure data of the filling body includes:

[0051] The geometric parameters of the lattice structure are extracted from the optimal values of the lateral expansion characteristics, and the initial lattice model is generated using a meshing tool;

[0052] Based on the initial lattice model, a finite element analysis tool is used to simulate the mechanical response of the lattice structure under loading conditions to obtain mechanical response data;

[0053] Extracting stress distribution parameters from the mechanical response data; if the stress distribution parameters deviate from a preset range, adjusting the geometric parameters to obtain an adjusted lattice model;

[0054] Based on the adjusted lattice model, an optimized layout of the lattice structure is generated by a topology optimization algorithm, and optimized layout data is determined;

[0055] Extract key geometric features from the optimized layout data, use the mesh smoothing tool to process the optimized layout, and obtain smoothed lattice structure data;

[0056] Generate a three-dimensional model data set of the filling body based on the smoothed lattice structure data to obtain filling body data with high mechanical properties;

[0057] The mechanical performance parameters are extracted from the filling volume data. If the performance parameters meet the preset threshold, the final lattice structure data is output.

[0058] Optionally, the flow switch guidance algorithm is combined with the three-dimensional digital model of the goaf and the lattice structure data to obtain the flow path of the negative Poisson's ratio mixture in the cracks and update the flow switch layout plan, including:

[0059] Using the flow switch guidance algorithm, the geometric constraints of the cracks are extracted from the three-dimensional digital model of the goaf and the lattice structure data to obtain the initial distribution data of the crack paths.

[0060] Based on the initial distribution data, a discretization model of the crack area is generated by a meshing tool, and the discretization model data is determined;

[0061] If the point distribution density of the discretized model data deviates from the geometric constraint conditions, the grid division parameters are adjusted to obtain the adjusted discretized model data;

[0062] The adjusted discretized model data is combined with the fluid dynamic parameters through data fusion method to obtain the flow path data of the mixture;

[0063] According to the flow path data of the mixture, the path optimization algorithm is used to process the discontinuous points in the flow path to obtain the final flow path data, and then the flow turnout layout plan is updated.

[0064] Optionally, the method of monitoring the displacement change of the goaf in real time based on the updated flow switch layout plan, updating the lattice structure data and the negative Poisson's ratio mixture ratio according to the real-time monitoring results, and generating an optimized filling plan includes:

[0065] Obtain real-time monitoring data of goaf subsidence displacement, collect displacement change information through the sensor network, and determine the displacement change trend;

[0066] If the displacement change trend exceeds the preset threshold range, the lattice structure analysis tool is used to extract the geometric characteristics of the goaf and obtain the updated lattice structure parameters;

[0067] By using data fusion method, the updated lattice structure parameters are combined with the mixture proportion data to generate the initial filling scheme data;

[0068] The random forest algorithm is used to classify the initial filling scheme data to obtain the classified filling scheme data;

[0069] If the matching degree of the classified filling scheme data does not meet the preset standard, the mixture ratio parameters are adjusted through the grid search tool to generate an optimized filling scheme.

[0070] The present invention also provides a computer device, comprising a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method.

[0071] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the method when executed by a processor.

[0072] Compared with the prior art, the present invention has the following advantages and technical effects:

[0073] The present invention discloses a method for filling a mixture with negative Poisson's ratio performance applied to goaf filling. The method comprises the following steps: constructing a digital model by collecting three-dimensional morphological data, calculating the volume and distribution density of the cracks to determine the filling amount, analyzing the rheological properties of the mixture to obtain the fluidity parameters, optimizing the negative Poisson's ratio to obtain the best lateral expansion characteristics, designing a lattice structure to generate a high-mechanical performance filling body, calculating the precise flow path using a flow switch guidance algorithm, monitoring the subsidence displacement changes in real time, and dynamically optimizing the filling scheme. The present invention achieves precise filling of cracks in goafs, improves the filling effect, enhances the stability of goafs, and provides a new technical means for goaf management. By combining digital modeling, intelligent algorithms, and real-time monitoring, the present invention can be dynamically adjusted according to the actual situation of the goaf, and has strong adaptability and practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0075] Figure 1 is a flow chart of a method according to an embodiment of the present invention;

[0076] Figure 2 This is a flow chart of optimizing the lateral expansion characteristic parameters according to an embodiment of the present invention;

[0077] Figure 3 This is a flow chart for updating a mobile turnout layout solution according to an embodiment of the present invention. DETAILED DESCRIPTION

[0078] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0079] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0080] Example 1

[0081] like Figure 1 As shown, this embodiment provides a method for filling a mixture with a negative Poisson's ratio performance for filling a goaf, comprising the following steps:

[0082] Collecting three-dimensional morphological data of the goaf and constructing a three-dimensional digital model of the goaf, wherein the three-dimensional digital model of the goaf includes crack distribution and spatial characteristics;

[0083] According to the three-dimensional digital model of the goaf, the volume and distribution density of the cracks in the goaf are obtained, and the filling amount of the negative Poisson's ratio mixture and the flow switch layout plan are determined;

[0084] According to the rheological properties of negative Poisson's ratio mixture, the relationship between viscosity and shear rate is analyzed to obtain the fluidity parameters of crack filling;

[0085] If the fluidity parameter meets the preset threshold, the negative Poisson's ratio mixture ratio is optimized through numerical simulation to obtain the optimal value of the lateral expansion characteristic;

[0086] Based on the optimal value of the lateral expansion characteristics, the parameters of the lattice structure are designed to generate the lattice structure data of the filling body;

[0087] By using the flow switch guidance algorithm, combined with the three-dimensional digital model of the goaf and the lattice structure data, the flow path of the negative Poisson's ratio mixture in the cracks is obtained, and the flow switch layout plan is updated.

[0088] Based on the updated flow turnout layout plan, the subsidence displacement changes of the goaf are monitored in real time. According to the real-time monitoring results, the lattice structure data and negative Poisson's ratio mixture ratio are updated to generate an optimized filling plan.

[0089] As a specific implementation method, the following steps are specifically included:

[0090] S101. By collecting three-dimensional morphological data of the goaf, a three-dimensional digital model of the goaf including crack distribution and spatial characteristics is constructed to generate a three-dimensional morphological feature description.

[0091] Three-dimensional morphological data is acquired from the goaf, and raw point cloud data is generated using stereoscopic microscope scanning technology. For the raw point cloud data, if the point cloud density is below a preset threshold, an interpolation algorithm is used to supplement the data to generate uniform point cloud data. From this uniform point cloud data, a stereoscopic segmentation algorithm is used to extract crack distribution characteristics and determine their spatial location information. Based on this spatial location information, spatial feature information containing crack distribution is constructed to generate a preliminary digital model. From this preliminary digital model, a surface fitting algorithm is used to generate a smooth spatial surface to determine the digital model structure. Morphological feature information is extracted from the digital model structure to generate a three-dimensional morphological description.

[0092] S102. Based on the three-dimensional morphological feature description, calculate the volume and distribution density of the cracks in the goaf, and determine the mixture filling amount and flow switch layout plan.

[0093] Crack distribution data is obtained from three-dimensional morphological features, and a volumetric analysis method is used to extract crack location points to obtain a set of crack spatial coordinates. The total crack volume is calculated using the set of crack spatial coordinates. If the total volume exceeds a preset threshold, a segmentation algorithm is used to demarcate the crack region and determine a set of crack subregions. Based on the set of crack subregions, the distribution density values of each subregion are calculated to obtain a density distribution matrix. High-density areas are extracted from the density distribution matrix, and a mapping algorithm is used to generate a mixture filling distribution map to determine the filling amount range. Based on the filling distribution map, candidate location points for the flow turnout are obtained. If the distance between candidate location points is less than a preset threshold, the location points are merged to obtain an optimized turnout location set. From the turnout location set, turnout connecting lines are generated. Path optimization methods are used to adjust the connection line distribution and determine the flow turnout layout plan. Spatial feature information is extracted from the turnout layout plan, and a correspondence between the filling amount and the turnout distribution is generated to obtain the final mixture filling and turnout layout pairing set.

[0094] When obtaining fracture distribution data from 3D morphological features, it is feasible to extract fracture location points through volumetric analysis. Based on the principle of geometric projection, volumetric analysis projects 3D point cloud data onto multiple planes to identify fracture boundaries. For example, in goaf point cloud data, the projection planes can be set at 0.3-meter intervals to generate a series of 2D projections. The pixel coordinates of the fracture edges can then be extracted and back-projected into 3D space to obtain the spatial coordinate set of the fracture.

[0095] Preferably, noise points can be eliminated by clustering method to ensure coordinate accuracy.

[0096] In one possible implementation, a voxel grid method can be used to calculate the total fracture volume based on a set of fracture spatial coordinates. This method divides the three-dimensional space into small cubes and counts the number of voxels containing fractures. For example, the space can be divided into voxels with a side length of 0.2 meters. The voxels containing fracture coordinates are marked and their volumes are accumulated.

[0097] If the total volume exceeds a preset threshold, such as 500 cubic meters, a segmentation algorithm can be triggered. It should be noted that the segmentation algorithm can use a graph-cut-based region partitioning method to divide the fracture coordinate set into multiple subregions. For example, using a minimum cut algorithm, the fracture can be divided into three subregion sets, each containing an independent fracture cluster.

[0098] Specifically, when calculating the distribution density of a set of fracture subregions, a density distribution matrix can be generated using kernel density estimation. Kernel density estimation calculates the density of points within a neighborhood centered on a coordinate point. For example, if the neighborhood radius is set to 1 meter, the point density of each subregion is counted to generate a 10×10 density matrix.

[0099] Preferably, when extracting high-density areas from the density distribution matrix, a density threshold of 0.8 can be set to screen out high-density grid cells. For example, after extracting five high-density areas, an interpolation mapping algorithm can be used to map the density values to the mixture filling volume to generate a filling distribution map with a filling volume range of 50 to 200 cubic meters.

[0100] In one embodiment, when obtaining candidate locations for a mobile turnout based on a fill distribution map, hotspot analysis can be used to identify the center of a high fill area. For example, a hotspot radius of 2 meters is set to select 10 candidate locations.

[0101] It should be noted that if the distance between candidate points is less than a preset threshold, such as 1.5 meters, a merging algorithm can be used to cluster close points into one. For example, after merging, six optimized switch positioning points are obtained. Switch connecting lines can be generated based on a minimum spanning tree algorithm, connecting the positioning points to form an initial layout. For example, the total length of the connecting lines is approximately 45 meters. When adjusting the distribution of the connecting lines, path optimization methods such as simulated annealing can be used to optimize the connectivity between the switches. After adjustment, the total length of the connecting lines is shortened to 40 meters, improving layout efficiency.

[0102] When extracting spatial feature information from the turnout layout, we can calculate the turnout spacing and angle distribution. For example, the average spacing is 8 meters, and the angle distribution is between 30 and 60 degrees.

[0103] As you can see, when generating a correspondence between fill volume and turnout distribution, a pairing set can be generated through association analysis. For example, an area with a fill volume of 100 cubic meters corresponds to three turnouts, spaced 7 meters apart. This pairing set provides precise guidance for subsequent fill construction and optimizes resource allocation.

[0104] S103. Based on the rheological properties of the mixture, analyze the relationship between its viscosity and shear rate to obtain fluidity parameters suitable for crack filling.

[0105] A model for the relationship between the mixture's viscosity and shear rate was constructed using rheological experimental data. A fluid dynamics model was used to fit the experimental data, generating a viscosity versus shear rate curve. Key turning points were extracted from this curve to generate a viscosity parameter set. If the viscosity value in the parameter set exceeded the preset fluidity threshold, the fitting parameters were adjusted and the curve was regenerated to obtain an optimized viscosity parameter set. From this optimized viscosity parameter set, fluidity parameters that matched the crack filling requirements were extracted. A classification algorithm was used to categorize the parameters into high-fluidity and low-fluidity groups, and the parameter range for the high-fluidity group was determined. Based on the parameter range for the high-fluidity group, a formula adjustment plan for the mixture was generated. If the deviation between the adjusted fluidity parameters and the target fluidity was greater than a preset threshold, the formula was iteratively optimized to obtain the adjusted fluidity parameters. Based on the adjusted fluidity parameters, the flow behavior of the mixture in cracks was simulated. A numerical simulation tool was used to generate a flow path distribution map and determine the flow coverage range corresponding to the fluidity parameters. From this flow coverage range, a fluidity parameter solution that matched the crack filling requirements was extracted. If the coverage range met the preset filling ratio, the final fluidity parameter solution was output, resulting in the fluidity parameters suitable for crack filling.

[0106] When modeling the relationship between mixture viscosity and shear rate, rheological experiments can be used to obtain basic data. These experiments typically use a rotational rheometer to measure the viscosity of the mixture at different shear rates. Suppose the experiment is set up with shear rates ranging from 0.1 to 100 s^-1, and the corresponding viscosity values are recorded. For example, the viscosity at a low shear rate is 500 Pa·s, and at a high shear rate it drops to 50 Pa·s. This data reflects the shear-thinning properties of the mixture and provides a basis for subsequent modeling.

[0107] In one possible implementation, when fitting experimental data with a fluid dynamics model, a power-law model can be selected to describe the relationship between viscosity and shear rate. The power-law model assumes that viscosity varies nonlinearly with shear rate and is therefore suitable for characterizing the flow behavior of goaf-filled mixtures.

[0108] The fitting process optimizes parameters using the least squares method to generate a viscosity curve. For example, a curve shows that the viscosity is 200 Pa·s at a shear rate of 10 s^-1 and drops to 80 Pa·s at 50 s^-1. Key turning points can be identified by changes in the slope of the curve. For example, the point where the rate of viscosity decreases significantly slows is designated as the turning point, and the corresponding parameters, such as 100 Pa·s and 20 s^-1, are extracted.

[0109] It should be noted that if the viscosity parameter exceeds the fluidity threshold, for example, the preset threshold of 150 Pa·s, the fitting parameters can be adjusted to regenerate the curve. Adjustments can include changing the power law exponent or adding experimental data points, such as adding a measurement at a medium shear rate of 30 s^-1. The optimized curve may show a stable viscosity around 120 Pa·s, meeting the fluidity requirements for crack filling.

[0110] Specifically, when extracting fluidity parameters suitable for fracture filling, one should focus on the balance between viscosity and flow rate. Low viscosity facilitates flow but may reduce filling stability. Assuming a fracture width of 0.5 meters, the optimized parameter set selects a viscosity range of 100 to 130 Pa·s. Using classification algorithms such as K-means clustering, the parameters are divided into a high-fluidity group of 80 to 110 Pa·s and a low-fluidity group of 120 to 150 Pa·s. The high-fluidity group is more suitable for rapidly filling narrow fractures.

[0111] In one embodiment, when generating a recipe adjustment plan, the mix components can be adjusted based on the high-flowability group parameters. For example, increasing the water-cement ratio from 0.4 to 0.5 can reduce the viscosity to 90 Pa·s. If the adjusted flowability deviates by more than 10%, iterative optimization can be performed, such as adding a small amount of water-reducing agent to further adjust the viscosity to 85 Pa·s. This adjustment ensures efficient flow of the mix through cracks.

[0112] When simulating the flow behavior of a mixture, computational fluid dynamics software can be used to generate a flow path distribution diagram. Assuming a fracture model of 10 meters long and 0.3 meters wide, the simulation shows that a high-flowing mixture covers 90% of the fracture area, with the path concentrated in the center. A low-flowing mixture covers only 70%, with the path skewed to one side. When the coverage area meets an 85% fill ratio, the final parameter solution is output, such as a viscosity of 95 Pa·s and a shear rate of 15 s^-1.

[0113] As you can understand, the flow path distribution map generated depends on the boundary conditions set. For example, with the inlet pressure set to 0.2 MPa, simulation results show that high-flowability mixtures provide uniform coverage, while low-flowability mixtures tend to clog narrow areas. The final solution selected high-flowability parameters to ensure filling efficiency and uniformity.

[0114] S104. If the fluidity parameter meets a preset threshold, the negative Poisson's ratio mixture ratio is optimized through numerical simulation to obtain an optimal value of the lateral expansion characteristic.

[0115] like Figure 2As shown, if the fluidity parameter meets the preset threshold, the initial mix ratio of the negative Poisson's ratio mixture is generated through numerical simulation to obtain the mix ratio. Based on the mix ratio, the finite element analysis tool is used to simulate the lateral expansion behavior of the mixture to obtain expansion behavior data. The lateral expansion characteristic parameters are extracted from the expansion behavior data. If the characteristic parameters deviate from the preset range, the mix ratio is adjusted to obtain an adjusted mix ratio. Based on the adjusted mix ratio, the Monte Carlo algorithm is used to simulate the expansion behavior of the mixture under different working conditions to determine the expansion behavior distribution. The key distribution characteristics are extracted from the expansion behavior distribution. If the characteristic values meet the preset expansion requirements, the optimized mix ratio is output. Based on the optimized mix ratio, the mechanical response curve of the mixture is generated to obtain the optimized value of the lateral expansion.

[0116] It is feasible to numerically simulate and optimize the lateral expansion characteristics of negative Poisson's ratio mixtures using finite element analysis software after the fluidity parameters meet a preset threshold. Assuming an initial mix ratio of cement, sand, and fiber of 1:2:0.1 by mass, the simulation boundary conditions are set as zero displacement at the fixed end and a lateral pressure of 0.1 MPa applied to the loading end. Through iterative calculations, the lateral expansion coefficients are recorded for different mix ratios. For example, when the fiber content increases to 0.15, the lateral expansion coefficient increases from -0.2 to -0.3.

[0117] A genetic algorithm was used to optimize the mix ratio, with the objective function being to maximize the lateral expansion coefficient and the constraint being that the compressive strength of the mixture must be at least 30 MPa. During the optimization process, the algorithm generated multiple candidate mix ratios, such as a ratio of 1:1.8:0.12 cement:sand:fiber, which yielded a calculated lateral expansion coefficient of -0.28 and a compressive strength of 32 MPa.

[0118] Further analysis found that when the fiber content exceeds 0.15, the increase in the lateral expansion coefficient slows down, while the compressive strength decreases significantly. Therefore, the optimization result selects a fiber content of 0.13, a lateral expansion coefficient of -0.25, and a compressive strength of 35MPa.

[0119] To verify the stability of the optimization results, a Monte Carlo simulation was conducted, randomly generating 100 mix parameter combinations and calculating the distribution of their lateral expansion coefficients and compressive strengths. The results showed that the lateral expansion coefficient of the optimized mix fluctuated between -0.24 and -0.26, and the compressive strength ranged from 33 to 37 MPa, meeting the project requirements. The final optimized mix ratio was 1:1.85:0.13 cement:sand:fiber, with a lateral expansion coefficient of -0.25 and a compressive strength of 35 MPa, providing a reliable basis for the design of negative Poisson's ratio mixtures.

[0120] S105. Design parameters of the lattice structure based on the optimal value of the lateral expansion characteristic, and generate filler structure data with high mechanical properties.

[0121] The geometric parameters of the lattice structure are extracted from the optimal values of the lateral expansion characteristics. An initial lattice model is generated using a meshing tool to obtain lattice model data. Based on the lattice model data, a finite element analysis tool is used to simulate the mechanical response of the lattice structure under loading conditions to obtain mechanical response data. Stress distribution parameters are extracted from the mechanical response data. If the stress distribution parameters deviate from the preset range, the geometric parameters are adjusted to obtain adjusted lattice model data. Based on the adjusted lattice model data, an optimized layout of the lattice structure is generated using a topology optimization algorithm to determine the optimized layout data. Key geometric features are extracted from the optimized layout data, and the optimized layout is processed using a mesh smoothing tool to obtain smoothed lattice structure data. Based on the smoothed lattice structure data, a three-dimensional model dataset of the infill is generated to obtain infill data with high mechanical properties. Mechanical performance parameters are extracted from the infill data. If the performance parameters meet the preset threshold, the final lattice structure dataset is output.

[0122] It is feasible that in the design of lattice structures of negative Poisson's ratio materials, extracting the geometric parameters of the lateral expansion characteristics is a key step. For example, the lateral expansion characteristics reflect the degree of lateral deformation of the material when subjected to force, which is usually related to the geometric shape of the lattice unit. Assuming that the optimal lateral expansion coefficient is -0.25 measured through experiments, the wall thickness of the lattice unit can be extracted to be 1mm and the inclination angle is 30 degrees. These parameters provide the basis for subsequent modeling, ensuring that the lattice structure can exhibit the expected characteristics when under pressure. When using the meshing tool to generate the initial lattice model.

[0123] Specifically, specialized software can be used to discretize geometric parameters. For example, the lattice elements can be divided into quadrilateral grids with a grid size of 0.5 mm to balance computational accuracy and efficiency.

[0124] It should be noted that meshing must ensure the continuity of element boundaries to avoid stress concentration. After the model is generated, the lattice model data is output, including node coordinates and element connection information.

[0125] In a possible implementation, the mechanical response of the lattice structure is simulated by a finite element analysis tool.

[0126] Preferably, the loading condition is set to apply a pressure of 0.2 MPa in the longitudinal direction, and the boundary condition is fixed at one end. The simulation results show that the stress is concentrated at the nodes of the lattice unit, and the maximum stress is about 10 MPa. The extracted mechanical response data includes stress distribution and displacement field, which provide a basis for subsequent optimization. After extracting the stress distribution parameters from the mechanical response data, if it is found that the maximum stress exceeds the preset range of 8 MPa, the geometric parameters need to be adjusted. For example, the wall thickness can be increased to 1.2 mm and the inclination angle can be reduced to 25 degrees to reduce stress concentration. The adjusted lattice model data is re-entered into the finite element tool for verification to ensure uniform stress distribution.

[0127] When generating the optimized layout of the lattice structure through the topology optimization algorithm, the algorithm aims to minimize the stress peak, and the constraint is that the material volume ratio does not exceed 30%.

[0128] In one embodiment, the optimization results show that the wall thickness of some cells is reduced to 0.8 mm, and the layout is more uniform. The extracted optimized layout data includes the adjusted node positions and cell shapes.

[0129] In one embodiment, a mesh smoothing tool is used to optimize the layout. For example, surface fitting techniques are used to eliminate sharp corners and generate smoothed lattice structure data. After smoothing, the transition between lattice elements is more natural, which helps reduce stress concentration areas.

[0130] When generating a 3D model dataset of the infill volume from the smoothed data, the lattice structure can be expanded into a 3D solid model using stereomicroscopy. For example, the infill volume is set to 10 cm × 10 cm × 5 cm and contains periodic lattice units.

[0131] The acquired high-mechanical performance filler volume data includes volume ratio and porosity. After extracting mechanical performance parameters from the filler volume data, if the compressive strength reaches 40 MPa, meeting a preset threshold, the final lattice structure dataset is output. For example, the final dataset describes a lattice cell with a wall thickness of 1.1 mm and an inclination angle of 28 degrees, suitable for the design of high-strength negative Poisson's ratio materials.

[0132] S106. Calculate the precise flow path of the mixture in the cracks by using a flow switch guidance algorithm combined with three-dimensional morphological features and lattice structure data.

[0133] like Figure 3 As shown, a fluid guidance algorithm is used to extract the geometric constraints of the cracks from the three-dimensional morphological features and lattice structure data to obtain the initial distribution data of the crack path. Based on the initial distribution data, a discretization model of the crack area is generated by a meshing tool to determine the discretization model data. If the lattice distribution density of the discretization model data deviates from the geometric constraints, the meshing parameters are adjusted to obtain the adjusted discretization model data. Through the data fusion method, the adjusted discretization model data is combined with the fluid dynamic parameters to calculate the flow path of the mixture and obtain the flow path data. Based on the flow path data, a path optimization algorithm is used to process the discontinuous points in the flow path to obtain the optimized flow path data. If the calculation accuracy requirement of the optimized flow path data does not meet the preset threshold, the fluid dynamic parameters are adjusted, the flow path is recalculated, and the final flow path data is determined.

[0134] The core of the fluid guidance algorithm lies in extracting the geometric constraints of the fractures by analyzing 3D morphological features and lattice structure data. For example, in one possible implementation, consider a 3D lattice structure of a porous material and determine the fracture path. First, the width and curvature of the fractures are determined by scanning the surface morphology of the lattice structure.

[0135] For example, the crack width may vary between 0.1 mm and 0.5 mm, and the curvature radius is not less than 2 mm. These geometric constraints provide a basis for subsequent path calculation.

[0136] It should be noted that the fluid guidance algorithm gives priority to the connectivity of the cracks to ensure that the flow path of the fluid in the lattice structure is not blocked.

[0137] When generating the initial distribution data of the crack paths, the algorithm simulates the natural extension of the cracks in the lattice based on geometric constraints.

[0138] Specifically, assuming the lattice unit is a cube with a side length of 1 mm, the cracks may be distributed along the diagonal direction of the lattice. Through statistical analysis, the initial path may show that the cracks are more concentrated in certain areas, such as near the lattice boundaries.

[0139] Preferably, this initial distribution will provide a reference for subsequent meshing to avoid paths being too dispersed. The meshing tool is used to generate a discretized model of the fracture region.

[0140] In one embodiment, meshing may use tetrahedral elements with a cell size of 0.2 mm to capture subtle variations in the fracture. For example, for a fracture path with large curvature, meshing may prioritize increasing mesh density in areas with varying curvature to improve model accuracy.

[0141] It should be noted that if the initial meshing reveals that the point distribution density deviates from the geometric constraints, for example, if the crack width is exaggerated to 0.7 mm in some areas, the mesh parameters need to be adjusted, for example, reducing the cell size to 0.15 mm. This adjustment can better fit the actual crack shape. The data fusion method combines the discretized model data with the fluid dynamic parameters to calculate the flow path of the mixture.

[0142] Illustratively, the fluid dynamic parameters may include a fluid viscosity of 0.01 Pascal seconds and a flow velocity of 0.5 meters per second.

[0143] In one possible implementation, the fusion process determines the flow path by simulating the pressure distribution of the fluid within the fractures. For example, fractures near the center of the lattice may become the primary flow path due to higher pressure. This approach has the advantage of combining geometric and dynamic factors to ensure realistic paths.

[0144] Path optimization algorithms handle discontinuities in flow paths. Specifically, discontinuities can occur at the intersection of fractured paths or at nodes in a lattice structure. For example, if a path breaks at a lattice node, the optimization algorithm smooths and reconnects the broken points, making the path continuous.

[0145] The algorithm also preferably evaluates changes in path curvature to ensure fluid flow stability. This optimization effectively improves the practicality of the path. If the accuracy of the optimized flow path data does not meet the preset threshold, for example, if the path deviation exceeds 0.05 mm, the fluid dynamic parameters must be adjusted. For example, the flow rate can be reduced to 0.3 meters per second and the path recalculated.

[0146] In one embodiment, such adjustment may make the path closer to the actual distribution of cracks, thereby meeting the accuracy requirement.

[0147] It should be noted that multiple iterative adjustments can gradually approach the optimal solution. The advantage of this method is that it significantly improves the reliability of the calculation results through fine-tuning of parameters.

[0148] S107. Obtain flow path data and monitor the displacement changes of the goaf in real time. If the displacement changes exceed the preset range, update the lattice structure parameters and mixture ratio based on the real-time monitoring analysis results to generate an optimized accurate filling plan.

[0149] Real-time monitoring data of goaf subsidence displacement is obtained, and displacement change information is collected through a sensor network to determine the displacement change trend. If the displacement change trend exceeds the preset threshold range, a lattice structure analysis tool is used to extract the geometric features of the goaf and obtain updated lattice structure parameters. Using a data fusion method, the updated lattice structure parameters are combined with the mixture ratio data to generate initial filling plan data. A random forest algorithm is used to classify the initial filling plan data, determine the applicability of the filling plan, and obtain classified filling plan data. If the matching degree of the classified filling plan data does not meet the preset standard, the mixture ratio parameters are adjusted through a grid search tool to obtain adjusted filling plan data. Based on the adjusted filling plan data, an optimized filling plan is generated.

[0150] Real-time monitoring of subsidence and displacement in mined-out areas is a key component of ensuring safety in underground engineering projects. For example, a sensor network could consist of multiple laser displacement sensors deployed on the top and sidewalls of the goaf to collect displacement data in real time. Assuming displacement is recorded once per second, the data indicates that an area has sunk 3 mm in 24 hours.

[0151] It's important to note that this high-frequency acquisition captures subtle changes, providing a reliable foundation for subsequent analysis. Displacement trends are determined through time series analysis. For example, if the subsidence rate exceeds 0.5 mm / hour for three consecutive hours, it is considered to have exceeded a preset threshold, triggering further processing. If displacement exceeds the threshold, a lattice structure analysis tool is used to extract geometric features of the goaf.

[0152] In one possible implementation, the tool generates a lattice model of the goaf based on 3D laser scanning data. The lattice elements might be cubes with sides measuring 2 cm, recording the width, height, and curvature of the cavities. For example, analysis revealed a region with a cavity width of 1.5 meters and a curvature radius of 5 meters. These parameters capture the complex morphology of the goaf and provide a basis for designing the filling plan.

[0153] It is understandable that the update of the lattice structure can accurately depict the dynamic changes of the goaf. The data fusion method combines the lattice structure parameters with the mixture ratio data to generate the initial filling plan.

[0154] Specifically, the mix might include cement, fly ash, and water in a ratio of 1:2:0.5. The fusion process takes into account cavity geometry, such as increasing the cement ratio to improve strength in wider areas. An initial solution might recommend 0.8 tons of mix per cubic meter of cavity. This approach ensures that the filling solution matches the actual conditions of the goaf. A random forest algorithm is used to classify the suitability of the initial filling solutions.

[0155] In one embodiment, the algorithm trains a model based on historical data, with input parameters including cavity volume, mix strength, and construction cost. The classification results may show that a certain solution is 80% suitable, but does not meet the preset standard of 90%.

[0156] The algorithm also ideally identifies reasons for non-suitability, such as insufficient mix flowability. This categorization improves the efficiency of solution screening. If a match is insufficient, the grid search tool adjusts the mix parameters. For example, to address a flowability issue, the grid search might test a solution with a cement ratio reduced from 1 to 0.8, or a water ratio increased to 0.6. After these adjustments, the solution might demonstrate improved flowability and a reduced fill time of up to 12 hours.

[0157] It should be noted that the grid search gradually approaches the optimal ratio through multiple rounds of iteration to ensure the feasibility of the solution. Based on the adjusted data, an optimized filling solution is generated.

[0158] In one possible implementation, the optimization plan might specify a layered filling strategy that prioritizes filling the bottom cavity, with each layer controlled to a thickness of 0.5 meters. This strategy evenly distributes the mix and avoids localized subsidence.

[0159] It is understandable that the optimization plan may also be combined with construction equipment parameters, such as adjusting the pumping pressure to 2 MPa to meet the filling requirements of different areas. This approach improves construction efficiency and filling uniformity.

[0160] Example 2

[0161] This embodiment further provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method.

[0162] Example 3

[0163] This embodiment also provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the method when executed by a processor.

[0164] The above are merely preferred embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for filling a mined-out area with a negative Poisson's ratio mixture, characterized in that: The following steps are involved: Collecting three-dimensional morphological data of the goaf and constructing a three-dimensional digital model of the goaf, wherein the three-dimensional digital model of the goaf includes crack distribution and spatial characteristics; According to the three-dimensional digital model of the goaf, the volume and distribution density of the cracks in the goaf are obtained, and the filling amount of the negative Poisson's ratio mixture and the flow switch layout plan are determined; According to the rheological properties of negative Poisson's ratio mixture, the relationship between viscosity and shear rate is analyzed to obtain the fluidity parameters of crack filling; If the fluidity parameter meets the preset threshold, the negative Poisson's ratio mixture ratio is optimized through numerical simulation to obtain the optimal value of the lateral expansion characteristic; Based on the optimal value of the lateral expansion characteristics, the parameters of the lattice structure are designed to generate the lattice structure data of the filling body; By using the flow switch guidance algorithm, combined with the three-dimensional digital model of the goaf and the lattice structure data, the flow path of the negative Poisson's ratio mixture in the cracks is obtained, and the flow switch layout plan is updated. Based on the updated flow turnout layout plan, the subsidence displacement changes of the goaf are monitored in real time. According to the real-time monitoring results, the lattice structure data and negative Poisson's ratio mixture ratio are updated to generate an optimized filling plan.

2. The method according to claim 1, characterized in that The method of collecting three-dimensional morphological data of the goaf and constructing a three-dimensional digital model of the goaf includes: Acquire 3D morphological data from the goaf and use stereo microscope scanning technology to obtain original point cloud data; For the original point cloud data, if the point cloud density is lower than the preset threshold, the interpolation algorithm is used to supplement the data to obtain uniform point cloud data; Through uniform point cloud data, the volume segmentation algorithm is used to extract the crack distribution characteristics and determine the crack spatial location information; Based on the spatial location information of the cracks, the spatial characteristic information including the crack distribution is constructed to obtain a preliminary digital model; From the preliminary digital model, a surface fitting algorithm is used to generate a smooth spatial surface and obtain a three-dimensional digital model of the goaf.

3. The method according to claim 1, characterized in that The method of obtaining the volume and distribution density of cracks in the goaf based on the three-dimensional digital model of the goaf, and determining the filling amount of the negative Poisson's ratio mixture and the flow switch layout plan includes: The crack distribution data is obtained from the three-dimensional digital model of the goaf, and the crack positioning points are extracted using the volume analysis method to obtain the crack space coordinate set; The total volume of the fracture is calculated using the fracture spatial coordinate set. If the total volume exceeds a preset threshold, a segmentation algorithm is used to divide the fracture area and determine the fracture sub-area set. According to the set of fracture sub-regions, the distribution density value of each sub-region is calculated to obtain the density distribution matrix; Extract high-density areas from the density distribution matrix and generate a mixture filling distribution map using a mapping algorithm; Based on the mixture filling distribution map, candidate position points of the mobile turnout are obtained. If the distance between the candidate position points is less than a preset threshold, the position points are merged to obtain the optimized turnout location set; Generate turnout connection lines through turnout location sets, adjust connection line distribution using path optimization methods, and determine the mobile turnout layout plan; The spatial feature information is extracted from the turnout layout scheme, and the corresponding relationship between the filling amount and the turnout distribution is generated, so as to obtain the final negative Poisson's ratio mixture filling amount and the flow turnout layout scheme.

4. The method according to claim 1, wherein The method of analyzing the relationship between viscosity and shear rate based on the rheological properties of the negative Poisson's ratio mixture to obtain the fluidity parameters for crack filling includes: Based on rheological experimental data, a viscosity-shear rate relationship model of negative Poisson's ratio mixtures was constructed; According to the relationship model between viscosity and shear rate, the experimental data were fitted to obtain the curve of viscosity changing with shear rate; According to the change curve, key turning points are extracted and viscosity parameter sets are generated; If the viscosity value in the viscosity parameter set exceeds the preset fluidity threshold, the fitting parameters are adjusted and the curve is regenerated to obtain an optimized viscosity parameter set; Extracting the fluidity parameters that match the crack filling requirements from the optimized viscosity parameter set; Using a classification algorithm, the liquidity parameters are divided into high liquidity group and low liquidity group, and the parameter range of the high liquidity group is determined; Generate a formula adjustment plan for the mixture through the parameter range of the high fluidity group; If the deviation between the adjusted liquidity parameter of the formula and the target liquidity is greater than a preset threshold, the formula is iteratively optimized to obtain the adjusted liquidity parameter; According to the adjusted fluidity parameters, the flow behavior of the mixture in the cracks is simulated; Use numerical simulation tools to generate flow path distribution maps and determine the flow coverage corresponding to the mobility parameters; Extract the mobility parameter scheme that matches the fracture filling requirements from the mobility coverage; If the coverage range meets the preset filling ratio, the final fluidity parameter solution is output to obtain the final fluidity parameter of the crack filling.

5. The method according to claim 1, wherein If the fluidity parameter satisfies the preset threshold, the negative Poisson's ratio mixture ratio is optimized by numerical simulation to obtain the optimal value of the lateral expansion characteristic, including: If the fluidity parameter meets the preset threshold, the initial proportioning scheme of the negative Poisson's ratio mixture is generated through numerical simulation; According to the initial mix ratio, the finite element analysis tool is used to simulate the lateral expansion behavior of the mixture and obtain the expansion behavior data; extracting a lateral expansion characteristic parameter from the expansion behavior data, and adjusting the mix ratio scheme if the lateral expansion characteristic parameter deviates from a preset range to obtain an adjusted mix ratio scheme; The Monte Carlo algorithm is used to simulate the expansion behavior of the mixture under different working conditions through the adjusted proportion scheme to determine the expansion behavior distribution; Extract key distribution features from the expansion behavior distribution. If the key distribution features meet the preset expansion requirements, output the optimized ratio scheme. According to the optimized mix ratio scheme, the mechanical response curve of the mixture is generated to obtain the optimized value of lateral expansion.

6. The method according to claim 1, characterized in that The method of designing the parameters of the lattice structure based on the optimal value of the lateral expansion characteristic and generating the lattice structure data of the filling body includes: The geometric parameters of the lattice structure are extracted from the optimal values of the lateral expansion characteristics, and the initial lattice model is generated using a meshing tool; Based on the initial lattice model, a finite element analysis tool is used to simulate the mechanical response of the lattice structure under loading conditions to obtain mechanical response data; Extracting stress distribution parameters from the mechanical response data; if the stress distribution parameters deviate from a preset range, adjusting the geometric parameters to obtain an adjusted lattice model; Based on the adjusted lattice model, an optimized layout of the lattice structure is generated by a topology optimization algorithm, and optimized layout data is determined; Extract key geometric features from the optimized layout data, use the mesh smoothing tool to process the optimized layout, and obtain smoothed lattice structure data; Generate a three-dimensional model data set of the filling body based on the smoothed lattice structure data to obtain filling body data with high mechanical properties; The mechanical performance parameters are extracted from the filling volume data. If the performance parameters meet the preset threshold, the final lattice structure data is output.

7. The method according to claim 1, characterized in that The flow switch guidance algorithm is combined with the three-dimensional digital model of the goaf and the lattice structure data to obtain the flow path of the negative Poisson's ratio mixture in the cracks and update the flow switch layout plan, including: Using the flow switch guidance algorithm, the geometric constraints of the cracks are extracted from the three-dimensional digital model of the goaf and the lattice structure data to obtain the initial distribution data of the crack paths. Based on the initial distribution data, a discretization model of the crack area is generated by a meshing tool, and the discretization model data is determined; If the point distribution density of the discretized model data deviates from the geometric constraint conditions, the grid division parameters are adjusted to obtain the adjusted discretized model data; The adjusted discretized model data is combined with the fluid dynamic parameters through data fusion method to obtain the flow path data of the mixture; According to the flow path data of the mixture, the path optimization algorithm is used to process the discontinuous points in the flow path to obtain the final flow path data, and then the flow turnout layout plan is updated.

8. The method according to claim 1, characterized in that The updated flow turnout layout plan is used to monitor the displacement changes of the goaf in real time. Based on the real-time monitoring results, the lattice structure data and the negative Poisson's ratio mixture ratio are updated to generate an optimized filling plan, including: Obtain real-time monitoring data of goaf subsidence displacement, collect displacement change information through the sensor network, and determine the displacement change trend; If the displacement change trend exceeds the preset threshold range, the lattice structure analysis tool is used to extract the geometric characteristics of the goaf and obtain the updated lattice structure parameters; By using data fusion method, the updated lattice structure parameters are combined with the mixture proportion data to generate the initial filling scheme data; The random forest algorithm is used to classify the initial filling scheme data to obtain the classified filling scheme data; If the matching degree of the classified filling scheme data does not meet the preset standard, the mixture ratio parameters are adjusted through the grid search tool to generate an optimized filling scheme.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory, wherein: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images