Modeling method of breakable particle model with real particle shape based on controllable grid division number and rule degree

Abstract

The invention provides a modeling method of a breakable particle model with a real particle shape based on a controllable grid division number and a rule degree, and belongs to the technical field of breakable particle models.The method comprises the steps that firstly, the real particle shape is photographed, extracted and scanned, particle contour information is obtained, then the particle contour information is read through an editor, and the breakable particle model is obtained. The method comprises the following steps: generating random points and circumscribed rectangles in a contour by using a grid rule degree control method, exporting circumscribed rectangle information of the random points, then importing coordinate information of the random points into neper software to carry out Voronio division on the circumscribed rectangles, exporting divided unit aggregate information, importing the information into a python editor again, cutting and retaining the rectangles by using a particle contour, and finally obtaining a final product. And finally, assembling the processed Voronoi units into a crushable particle model. On the basis of real particle data obtained through photographing or scanning, modeling of the real particle shape is achieved through a cell element bonding modeling method, and it is guaranteed that the model can simulate parameters such as the shape, the surface roughness and the size of the real particle more accurately.

Description

Technical Field

[0001] The invention relates to the technical field of breakable particle models, and in particular to a modeling method of a breakable particle model based on a real particle shape with controllable grid division quantity and regularity. Background Art

[0002] Granular materials are widely used in various civil engineering projects, such as water conservancy projects, mining, tunnel construction, and building foundations. Due to their excellent compressive resistance and good permeability, they have become an important part of engineering construction. However, during engineering construction and use, granular materials often encounter overload or impact, resulting in particle breakage. Particle breakage not only changes the macroscopic mechanical properties of the material, but also causes changes in the internal mechanical properties of the material, thereby affecting the stability and safety of the engineering structure. Traditional physical experiments are difficult to observe and accurately analyze the crushing process in real time at the particle scale due to the limitations of equipment accuracy and experimental conditions. Therefore, how to accurately simulate the mechanism of particle crushing at the microscale has become an important topic in granular material research.

[0003] With the continuous development of numerical simulation technology, the discrete element method (DEM) has gradually become one of the main tools for studying particle crushing. The discrete element method can effectively reveal the mechanical behavior of particles during crushing by treating them as independent particles and simulating the interactions between them. In order to better simulate the particle crushing phenomenon and improve the reliability of the simulation results, it is urgent to develop a discrete element model based on the real particle shape. This model can take into account the shape, surface roughness and micromechanical characteristics of the particles during the crushing process, thereby providing a more realistic and effective tool for the numerical study of particle crushing. This is not only of great significance for the performance analysis of granular materials, but also provides more scientific theoretical support for related engineering applications.

[0004] In the study of discrete element particle crushing, the construction of particle model is the primary task. There are two traditional modeling methods: one is the replacement particle method, which is to construct a crushable particle model by using a real contour and filling the interior of the contour with many small spherical particles bonded together. The modeling of the replacement particle method is convenient and simple, the particle model has a small calculation load and a fast running speed, but its crushing standard and crushing mode have not been clearly determined, and the most critical defect of the model is that the volume enclosed by the contour and the total volume of the sub-particles are not equal, and the shape characteristics of the particles cannot be reflected after crushing.

[0005] Compared with the modeling method of particle replacement, polyhedral cells can satisfy the volume conservation during modeling, and the external load can be decomposed into tensile and shear forces acting on the polyhedron. The crushing path of tetrahedral cells is always smooth, while polyhedral cells provide more crushing methods through micro-convex bodies and rough failure paths. These paths cause some cells to undergo tensile and shear sliding, which more accurately reflects the actual process of particle crushing from a mechanical perspective and reproduces the crushing path more realistically. Although existing open source software (such as Neper, Voro++) and commercial tools (such as PFC3D, 3DEC, Matlab) provide polyhedral segmentation functions, they can usually only handle simple convex geometries. Real particles have complex shapes and uneven surfaces. Existing tools either cannot directly split them into polyhedral cells suitable for discrete element analysis, or the divided polyhedral cells cannot control the regularity of the mesh.

[0006] The existing bonded cell modeling method is to use the discrete element software PFC to import the outline of the particle model and convert it into a wall with a constraint effect; then generate a number of spherical small-diameter sub-particles that do not touch each other inside the wall, and then use the radius expansion method to gradually enlarge the radius of the spherical sub-particles for filling. After filling, the sub-particle core is used as a seed to divide the circumscribed cube of the particle outline into a Voronoi polyhedron aggregate, and then use the PFC command to cut the cube through the particle outline. Another method is to import the particle outline into the finite element software and generate a particle model through the Vorono meshing of the finite element software. Although these methods can achieve the modeling of real particles, there is no way to control the regularity of the Voronoi meshing and the number of units in the Voronoi meshing.

[0007] The prior art has the following technical defects:

[0008] (1) The model generated by the particle replacement method cannot simulate the shape, surface roughness, volume and other parameters of real particles. (2) Neither discrete element software nor finite element software has a unified method to control the number of grids and the degree of grid regularity during mesh division. (3) The existing modeling methods cannot achieve the unification of two-dimensional modeling and three-dimensional modeling. Therefore, in response to this problem, this paper proposes a division method that controls the degree of grid regularity, aiming to accurately describe the real particle shape and provide effective support for the numerical simulation of the particle crushing process. Summary of the invention

[0009] The purpose of the present invention is to provide a modeling method of a crushable particle model based on a real particle shape with controllable number and regularity of grid divisions, so as to solve the technical problems mentioned in the background technology.

[0010] The method provided includes first taking photos to extract and scan the real particle shape to obtain the particle contour information, then reading the particle contour information through a python editor, using the method of controlling the degree of grid rules within the contour to generate random points and a circumscribed rectangle (a cube in three dimensions), exporting the circumscribed rectangle (a cube in three dimensions) information of the random points, and then importing the coordinate information of the random points into the neper software to perform Voronio division on the circumscribed rectangle (a cube in three dimensions), then exporting the divided unit assembly information, and once again importing it into the python editor to use libraries such as trimesh, numpy, shapely, etc. to complete the cutting and retention of the rectangle (a cube in three dimensions) with the particle contour, and finally assembling the processed Voronoi units into a breakable particle model.

[0011] The bonded cell modeling method is used to achieve modeling based on the real particle shape, ensuring that the model can more accurately simulate the shape, surface roughness, volume and other parameters of the real particles. A method that can control the degree of grid regularity is proposed to ensure the uniformity of the number of sub-particles and the size of sub-particles during particle modeling. It is beneficial to provide a method that can control the number of sub-particles and the uniformity of the size of sub-particles in the study of particle crushing in discrete element software. This modeling method is not only applicable to two-dimensional but also to three-dimensional, achieving the unity of two-dimensional and three-dimensional modeling.

[0012] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0013] A modeling method for a real particle shape breakable particle model based on controllable meshing quantity and regularity, the method comprising the following steps:

[0014] Step 1: Select real particles, take photos with a camera and use deep learning image recognition technology to obtain the two-dimensional contour boundary of the real particles, scan with a scanner and use 3D scanning software to obtain the real three-dimensional particle contour that needs to be modeled;

[0015] Step 2: Read the 2D and 3D contour information of the particles as the model boundary, generate random points inside the particle model by controlling the degree of grid regularity, then generate contour information, and finally extract the random point coordinate information and contour information;

[0016] Step 3: Import the random point coordinate information and contour information into the software to generate a unit aggregate, and extract all the information of the unit aggregate;

[0017] Step 4: Extract the vertex coordinates of each unit in the unit set, and number each unit body, read the information of several face units and particle contour information of one of the numbers, if the unit intersects with the contour, use the convex hull segmentation algorithm to complete the cutting of several face bodies through the data point replacement method, judge the information located inside and outside the contour after cutting according to the particle contour, extract the vertex information inside the contour, if the unit is inside the contour, extract all the vertex information of the unit;

[0018] Step 5: Repeat step 4, cut all polygons or three-dimensional polyhedral units according to their numbers, and save the vertex information in the same file in the same way;

[0019] Step 6: Import the file data information from the previous step into the discrete element software, generate all the polygons or three-dimensional polyhedrons to break the real particle model, use the parallel bonding method to give the bonding force between each unit, and then you can simulate the particle crushing.

[0020] Furthermore, the specific process of step 1 is as follows: the real particles come from the blasting gravel, crushed stone, or pebbles in the riverbed in the actual project. The two-dimensional real particle contour boundary is obtained by taking pictures with a camera using deep learning image recognition technology. The data file is saved as a JSON file. The three-dimensional contour information of the real particles is obtained by a scanner. The particle model consistent with the real particle shape and volume is generated by point cloud modeling and triangular mesh splicing, and saved as an STL format file.

[0021] Furthermore, in step 2, the two-dimensional and three-dimensional contour information of the particles is read through the Python compiler, and random points are generated inside the particle model. After the random points are generated, the circumscribed rectangle of the contour or the three-dimensional cube is generated, and the coordinate information of the random points and the length, width and height information of the circumscribed rectangle or the three-dimensional cube are extracted.

[0022] Furthermore, the specific process of the method for controlling the regularity of the grid in step 2 is as follows:

[0023] a. First calculate the area S enclosed by the particle outline or the volume V in three-dimensional cases;

[0024] a. Set the number of random points or seed points Np based on the Voronio algorithm;

[0025] c. Calculate the average equivalent radius r of the divided grid unit body, Np = S / (π*r 2 ), in three-dimensional case: Np=V / (4 / 3*π*r 3 );

[0026] d. Define the grid rule parameter α, where 0.0≤α≤0.8, to ensure that the minimum distance Lmin between random points or seed points is greater than α*2*r, and that random points are generated within the contour;

[0027] e. Control the size of parameter α to control the regularity of the grid.

[0028] Furthermore, in step 3, the random point information is imported into the neper software through the command of the neper software to divide the circumscribed rectangle of the contour or the three-dimensional cube into a number of Voronoi edge grids or a three-dimensional polyhedral unit aggregate, and the information of the unit aggregate is extracted. The file is a tess file, which contains the point, face, and edge information of each unit.

[0029] Furthermore, in step 4, the tess file of the unit collection is read using the Python compiler, the vertex coordinates of each unit in the unit collection of several polygons or three-dimensional polyhedrons are extracted, and each unit body is numbered, and the vertex coordinates of each unit are saved under the unit body number. When it is a two-dimensional case, the unit information of several polygons numbered 1 and the particle contour information are read. If the unit intersects with the contour, all the vertex information of the unit is extracted, and the data point replacement method is directly adopted. The trimesh, numpy and shapely libraries are used to replace the vertex coordinates outside the contour of several polygons intersecting with the contour with the contour points contained in the polygon and the intersection formed by the contour point connection, and the vertices inside the contour point are retained to complete the reconstruction of the polygon. If the unit is inside the contour, all the vertex information of the unit is extracted and saved directly.

[0030] Furthermore, in step 5, all polygons or three-dimensional polyhedral units are looped according to the method of step 4, and all information is extracted and converted into a command language format readable by discrete element software, and saved in txt format.

[0031] Furthermore, in step 6, the txt document of step 5 is imported into the discrete element software for running, and a plurality of polygons or three-dimensional polyhedrons are given bonding force for particle crushing.

[0032] Furthermore, in step 3, the number and volume differences of the neutron particles in the particle crushing can be adjusted by controlling the Voronio grid division, thereby reducing the simulation error.

[0033] Furthermore, in step 2, the method for controlling the degree of grid regularity can ensure the uniformity of the number of sub-particles and the volume size of sub-particles during particle modeling.

[0034] The present invention has the following beneficial effects due to the adoption of the above technical solution:

[0035] (1) The present invention uses the bonded cell modeling method based on the real particle data obtained by photographing or scanning to achieve modeling based on the real particle shape, ensuring that the model can more accurately simulate the parameters such as the shape, surface roughness, and volume of the real particles.

[0036] (2) Based on the bonded cell modeling method, a method for controlling the degree of grid regularity is proposed to ensure the uniformity of the number and volume of sub-particles during particle modeling, which provides new ideas and methods for subsequent research on the influence of the degree of differentiation of the number and volume of sub-particles on particle crushing. Finally, this modeling method is not only applicable to two-dimensional but also to three-dimensional, achieving the unification of two-dimensional and three-dimensional modeling, and providing an effective technical means for discrete element research. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a schematic diagram of the process steps of a modeling method of a real particle shape breakable particle model capable of controlling the degree of regularity of mesh division according to an embodiment of the present invention;

[0038] Figure 2 This is a real particle image obtained by taking photos with a camera according to an embodiment of the present invention;

[0039] Figure 3 is a two-dimensional contour image extracted by the deep learning technology of an embodiment of the present invention;

[0040] Figure 4 The embodiment of the present invention imports the two-dimensional contour image into Matlab to perform binarization calculation of the contour data information graph;

[0041] Figure 5 The embodiment of the present invention generates a random point graph;

[0042] Figure 6 The Voronoi unit aggregate diagram generated by importing random point and rectangle information into neper software in the embodiment of the present invention;

[0043] Figure 7 This is a diagram generated by importing polygonal data of the Voronoi unit aggregate of an embodiment of the present invention into a python editor;

[0044] Figure 8 It is a graph of a Voronoi unit aggregate of a particle contour cyclic cutting according to an embodiment of the present invention;

[0045] Fig. 9 It is a discrete element real particle breakable model diagram generated by importing polygonal data after cutting into PFC2D in an embodiment of the present invention;

[0046] Fig.10 The random point comparison graph generated by the embodiment of the present invention;

[0047] Fig.11 is a comparison diagram of the discrete element real particle breakable model generated by an embodiment of the present invention;

[0048] Fig.12 is a three-dimensional contour image of real particles obtained by scanning with a scanner according to an embodiment of the present invention;

[0049] Fig.13 The embodiment of the present invention imports particle contour information, generates a circumscribed cube, and generates a random point map according to a grid rule parameter α=0.0;

[0050] Fig.14 It is a Voronoi unit aggregate diagram generated by importing neper software into an embodiment of the present invention;

[0051] Fig.15 It is a discrete element real particle crushable model diagram generated by importing polygon data after cutting into PFC3D in an embodiment of the present invention;

[0052] Fig.16 The random point comparison diagram is generated by the grid number Np=100, grid rule parameters α=0.0 and α=0.8 in the embodiment of the present invention;

[0053] Fig.17 : is a comparison diagram of discrete element real particle breakable models generated when the number of grids Np=100, grid rule parameters α=0.0 and α=0.8 in the embodiment of the present invention;

[0054] Fig.18 This is a diagram comparing the single particle breakage conditions of the test and simulation of the embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and preferred embodiments. However, it should be noted that many details listed in the specification are only for the purpose of enabling the reader to have a thorough understanding of one or more aspects of the present invention, and these aspects of the present invention can be implemented even without these specific details.

[0056] A modeling method for a real particle shape breakable particle model based on controllable meshing quantity and regularity, the method comprising the following steps:

[0057] Step 1: Select real particles, take photos with a camera, and use deep learning image recognition technology to obtain the two-dimensional real particle contour boundaries. Scan with a scanner and use 3D scanning software to obtain the real three-dimensional particle contours that need to be modeled. Real particles come from blasting materials such as gravel and crushed stone in actual projects, or pebbles in riverbeds. Take photos with a camera and use deep learning image recognition technology to obtain the two-dimensional real particle contour boundaries. The data file is generally saved as a "JSON" file. Use a scanner (such as the RevoPointPOP2 scanner) to obtain the three-dimensional contour information of the real particles. The particle model that is consistent with the real particle shape, volume, etc. is generated through point cloud modeling and triangular mesh stitching, and is usually saved as an "STL" format file. For example Figure 2 It is the real particle image obtained by taking a photo with a camera, such as Figure 3 is a two-dimensional contour image extracted by deep learning technology, such as Figure 4 Import the two-dimensional contour image into Matlab to binarize and calculate the contour data information graph. Fig.12 This is the three-dimensional contour image of the real particles obtained by scanning.

[0058] Step 2: Read the two-dimensional and three-dimensional contour information of the particles as the model boundary through the Python compiler, generate random points inside the particle model through the method of controlling the degree of grid regularity proposed in the present invention, then generate the circumscribed rectangle of the contour (the three-dimensional one is a circumscribed cube), and finally extract the random point coordinate information and the circumscribed rectangle information.

[0059] The information of the two-dimensional and three-dimensional contours of the particles is read by the Python compiler, and random points are generated inside the particle model. The method for controlling the degree of grid regularity proposed by the present invention is as follows: Figure 1 As shown:

[0060] a. First calculate the area S (volume V in three-dimensional case) enclosed by the particle outline.

[0061] b. Set the number of random points (seed points) Np based on the Voronio algorithm.

[0062] c. Calculate the average equivalent radius r of the divided grid unit body: Np = S / (π*r 2 ). In three-dimensional case: Np=V / (4 / 3*π*r 3 )

[0063] d. Define the grid rule parameter α (0.0≤α≤0.8) to ensure that the minimum distance between random points (seed points) Lmin>α*2*r, and the random points are generated within the contour.

[0064] e. Control the size of parameter α to control the regularity of the grid.

[0065] After generating random points, continue to generate the circumscribed rectangle of the outline (a cube in three dimensions). Extract the coordinate information of the random points and the length, width and height information of the circumscribed rectangle (a cube in three dimensions). Fig.13 The figure shows the import of particle contour information, the generation of a circumscribed cube, and the generation of a random point map according to the grid rule parameter α = 0.0. Fig.16 Shown is a random point comparison diagram generated when the number of grids Np=100 and the grid rule parameters α=0.0 and α=0.8.

[0066] like Figure 5 The figure shows the import of particle contour information, the generation of the circumscribed rectangle, the input of the number of grids Np = 100, and the grid rule parameter α = 0.0 to generate a random point map. As can be seen from the figure, according to the Voronoi algorithm, inputting Np random points can generate Np grids, so the number of grids can be controlled. Inputting the grid rule parameter α = 0.0 can control the regularity of the grid.

[0067] Step 3: Import the random point coordinate information and the circumscribed rectangle (circumscribed cube in three dimensions) information into the neper software to generate a Voronoi polygonal grid (polyhedral grid in three dimensions) unit aggregate, and extract all the information of the unit aggregate, such as vertices, faces, and edges. Through the command of the neper software, import the random point information into the neper software to divide the circumscribed rectangle (cube in three dimensions) of the contour into a Voronoi polygonal grid (polyhedral grid in three dimensions), and extract the information of the unit aggregate. The file is a tess file, which contains the point, face, edge and other information of each unit. Figure 6 The figure shows the Voronoi unit aggregate diagram generated by importing random points and rectangle information into the neper software, which can quickly realize Voronoi division. Fig.14 A diagram of the Voronoi cell cluster generated for importing random point and cube information into the neper software.

[0068] Step 4: Use the Python compiler to extract the vertex coordinates of each unit in the polygonal (polyhedron in three dimensions) unit set, and number each unit. Read the polyhedral unit information and particle contour information numbered 1. If the unit intersects with the contour: use the convex hull segmentation algorithm to complete the cutting of the polyhedron by replacing the data point method. According to the particle contour, determine which part is inside the contour and which part is outside the contour after cutting; extract the vertex information of the polygonal (polyhedron in three dimensions) inside the contour. If the unit is inside the contour: extract all vertex information of the unit.

[0069] Use the Python compiler to read the tess file of the unit collection, extract the vertex coordinates of each unit in the polygon (polyhedron in three dimensions) unit collection, number each unit, and save the vertex coordinates of each unit under the unit number. In the two-dimensional case: read the polygon unit information numbered 1 and the particle contour information. If the unit intersects with the contour: extract all the vertex information of the unit, directly use the data point replacement method, and use trimesh, numpy, shapely and other libraries to replace the vertex coordinates outside the contour of the polygon that intersects with the contour with the contour points contained in the polygon and the intersection formed by the contour point connection, retain the vertices inside the contour point, and complete the reconstruction of the polygon. If the unit is inside the contour: extract all the vertex information of the unit and save it directly. The three-dimensional method is the same. If Figure 7 Import polygonal data for a Voronoi cell collection into a graph generated by the Python editor.

[0070] Step 5: Loop through the previous step, cut all polygonal (3D is polyhedron) units according to the number, and save the vertex information in the same way in the same file. Loop through all polygonal (3D is polyhedron) units according to the method in step 4, extract all the information and convert it into a command language format readable by the discrete element software, and save it in txt format. Figure 8 The figure shows the Voronoi unit assembly diagram of the particle contour circular cutting, which realizes the Voronoi division of the particle contour.

[0071] Step 6: Import the file data information of the previous step into the discrete element software, generate all polygons (three-dimensional is a polyhedron) to break the real particle model. Use the parallel bonding method to give the bonding force between each unit, and then you can simulate the particle crushing. Import the txt file of step 5 into the discrete element software and run it. Give the bonding force between the polygons (three-dimensional is a polyhedron) to crush the particles. Fig. 9 The figure shows the discrete element real particle breakable model diagram generated by importing the polygonal data after cutting into PFC2D. Fig.10 Comparison diagram of discrete element real particle breakable models generated for the number of grids Np = 100, grid rule parameters α = 0.0 and α = 0.8, where the number of sub-particles in both models is 100; the difference in sub-particle volume in the model with α = 0.0 is large; the difference in sub-particle volume in the model with α = 0.8 is small. Through the comparison diagram, we can see the specific effect of the number of grids Np and the grid rule parameter α on discrete element modeling, which provides new ideas and methods for subsequent research on the influence of the degree of differentiation of the number of sub-particles and the sub-particle volume on particle breakage.

[0072] like Fig.15The discrete element real particle crushable model diagram generated by importing the cut polygon data into PFC3D. Fig.16 Shown is a comparison diagram of the discrete element real particle breakable model generated when the grid number Np=100 and the grid rule parameters α=0.0 and α=0.8. Fig.18 A single particle crushing diagram is provided to compare the test and simulation to achieve the simulation effect.

[0073] Based on the Voronio algorithm, a creative method is proposed to control the number of random points (seed points) to control the number of grid cells after Voronio grid division, control the distribution of random points (seed points) and thus control the regularity of the Voronio grid. By controlling the Voronio grid division, the number and volume differences of sub-particles in particle crushing can be adjusted to reduce simulation errors, providing an effective technical means for in-depth research on the effects of sub-particle shape and number on particle crushing. Step 4 is a key step in implementing step 2. It uses computer two-dimensional and three-dimensional geometry knowledge to achieve the use of particle contour cutting of Voronio unit aggregates, build a real particle Voronio grid division model, and realize real particle modeling in particle crushing research.

[0074] Matters not covered by the present invention are known technologies.

[0075] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A modeling method for a real particle shape breakable particle model based on a controllable number and regularity of mesh divisions, characterized by: The method comprises the following steps: Step 1: Select real particles, take photos with a camera and use deep learning image recognition technology to obtain the two-dimensional contour boundary of the real particles, scan with a scanner and use 3D scanning software to obtain the real three-dimensional particle contour that needs to be modeled; Step 2: Read the 2D and 3D contour information of the particles as the model boundary, generate random points inside the particle model by controlling the degree of grid regularity, then generate contour information, and finally extract the random point coordinate information and contour information; Step 3: Import the random point coordinate information and contour information into the software to generate a unit aggregate, and extract all the information of the unit aggregate; Step 4: Extract the vertex coordinates of each unit in the unit set, and number each unit body, read the information of several face units and particle contour information of one of the numbers, if the unit intersects with the contour, use the convex hull segmentation algorithm to complete the cutting of several face bodies through the data point replacement method, judge the information located inside and outside the contour after cutting according to the particle contour, extract the vertex information inside the contour, if the unit is inside the contour, extract all the vertex information of the unit; Step 5: Repeat step 4, cut all polygons or three-dimensional polyhedral units according to their numbers, and save the vertex information in the same file in the same way; Step 6: Import the file data information from the previous step into the discrete element software, generate all the polygons or three-dimensional polyhedrons to break the real particle model, use the parallel bonding method to give the bonding force between each unit, and then you can simulate the particle crushing.

2. The modeling method of the real particle shape breakable particle model based on controllable meshing quantity and regularity according to claim 1, characterized in that: The specific process of step 1 is as follows: the real particles come from the blasting gravel, crushed stone, or pebbles in the riverbed in actual engineering projects. The two-dimensional contour boundaries of the real particles are obtained by taking pictures with a camera using deep learning image recognition technology. The data file is saved as a JSON file. The three-dimensional contour information of the real particles is obtained through a scanner. The particle model consistent with the shape and volume of the real particles is generated through point cloud modeling and triangular mesh splicing and saved as an STL format file.

3. The modeling method of a real particle shape breakable particle model based on controllable meshing quantity and regularity according to claim 1, characterized in that: In step 2, the two-dimensional and three-dimensional contour information of the particle is read through the Python compiler, and random points are generated inside the particle model. After the random points are generated, the circumscribed rectangle of the contour or the three-dimensional cube is generated, and the coordinate information of the random points and the length, width and height information of the circumscribed rectangle or the three-dimensional cube are extracted.

4. The modeling method of a real particle shape breakable particle model based on controllable meshing quantity and regularity according to claim 1, characterized in that: The specific process of the method for controlling the degree of grid regularity in step 2 is as follows: a. First calculate the area S enclosed by the particle outline or the volume V in three-dimensional cases; a. Set the number of random points or seed points Np based on the Voronio algorithm; c. Calculate the average equivalent radius r of the divided grid unit body, Np = S / (π*r 2 ), in three-dimensional case: Np=V / (4 / 3*π*r 3 ); d. Define the grid rule parameter α, where 0.0≤α≤0.8, to ensure that the minimum distance between random points or seed points Lmin>α*2*r, and that the random points are generated within the contour; e. Control the size of parameter α to control the regularity of the grid.

5. The modeling method of a real particle shape breakable particle model based on controllable meshing quantity and regularity according to claim 1, characterized in that: In step 3, the random point information is imported into the neper software through the command of the neper software to divide the circumscribed rectangle of the contour or the three-dimensional cube into a number of Voronoi edge grids or a three-dimensional polyhedral unit aggregate, and the information of the unit aggregate is extracted. The file is a tess file, which contains the point, face, and edge information of each unit.

6. The modeling method of a real particle shape breakable particle model based on controllable meshing quantity and regularity according to claim 1, characterized in that: In step 4, the Python compiler is used to read the tess file of the unit collection, extract the vertex coordinates of each unit in the unit collection of several polygons or three-dimensional polyhedrons, and number each unit body. The vertex coordinates of each unit are saved under the unit body number. When it is a two-dimensional case, read the information of several polygon units numbered 1 and the particle contour information. If the unit intersects with the contour, extract all the vertex information of the unit, and directly adopt the data point replacement method. Use trimesh, numpy and shapely libraries to replace the vertex coordinates outside the contour of several polygons intersecting with the contour with the contour points contained in the polygon and the intersection formed by the contour point connection, retain the vertices inside the contour point, and complete the reconstruction of the polygon. If the unit is inside the contour, extract all the vertex information of the unit and save it directly.

7. The modeling method of a real particle shape breakable particle model based on controllable meshing quantity and regularity according to claim 1, characterized in that: In step 5, all polygons or three-dimensional polyhedral units are looped according to the method in step 4, and all information is extracted and converted into a command language format readable by discrete element software and saved in txt format.

8. The modeling method of a real particle shape breakable particle model based on controllable meshing quantity and regularity according to claim 1, characterized in that: In step 6, the txt file of step 5 is imported into the discrete element software for running, and a plurality of polygons or three-dimensional polyhedrons are given bonding force for particle crushing.

9. The modeling method of a real particle shape breakable particle model based on controllable meshing quantity and regularity according to claim 1, characterized in that: In step 3, the number and volume differences of the neutron particles in particle crushing can be adjusted by controlling the Voronio grid division, thereby reducing the simulation error.

10. The modeling method of a real particle shape breakable particle model based on controllable meshing quantity and regularity according to claim 1, characterized in that: In step 2, the method of controlling the degree of grid regularity can ensure the uniformity of the number of sub-particles and the size of the sub-particles during particle modeling.

Images