A method for constructing a crushable model of ballast for discontinuous deformation analysis of spherical units

Abstract

The present invention relates to a method for constructing a crushable model of ballast for discontinuous deformation analysis of spherical units, including: a step of scanning the physical object of ballast particles to obtain a triangular mesh model of ballast; a step of voxelizing the obtained triangular mesh model to obtain a voxelized model; and a step of performing scaling processing according to the number of contacts on the obtained voxelized model to obtain a crushable model of ballast. The method of the present invention is simple. During voxelization, a model with an outer contour similar to that of real ballast particles can be obtained through three steps: conservative rasterization, flood filling method, and spherical unit replacement. On the basis of the voxelized model, by means of scaling according to the number of contacts, the internal sub-particles in the voxelized model that have the same arrangement rule and size and cannot reflect real ballast particles are randomly rescaled and arranged again, so as to obtain a crushable model for discontinuous analysis of spherical unit ballast with a real shape.

Description

Technical Field

[0001] The invention relates to the technical field of railway numerical simulation, and in particular to a method for constructing a crushable ballast model for discontinuous deformation analysis of a ball unit. Background Art

[0002] As an important part of railway track, the shape and arrangement of ballast have an important impact on the stability of the track and the safety of train operation. The shape of real ballast is complex and diverse, including irregular polyhedrons and various surface angles, which have an important impact on the mechanical behavior of ballast and track performance.

[0003] For irregularly shaped particles, scholars have developed some modeling methods that can take into account the actual shape of the particles. For example, the clump unit in the discrete element method software PFC can be used with some modeling methods that generate overlapping particles to simulate irregularly shaped particles that cannot be broken. However, this type of modeling method has shortcomings such as overlapping sub-particles or insufficient contact between sub-particles, which makes the established ballast model difficult to use for ballast crushing simulation. For another example, Li Xu [1] The developed discontinuous analysis modeling method based on block units is used to simulate the unbreakable polygonal ballast particles. This type of modeling method uses blocks as calculation units, but the calculation of the contact between different blocks is very complicated, and it is difficult to form a stable and efficient contact algorithm. It is also not suitable for ballast crushing simulation.

[0004] Therefore, this method develops a new discontinuous analysis modeling method for the ball unit ballast model from the perspective of voxelization and particle scaling, ensuring that the ballast sub-particles are in full contact with each other without overlapping, and the model has a crushable property.

[0005] [1] Li Xu, Wang Pengcheng, Ao Guodong, et al. DDA numerical simulation of polygonal ballast particle accumulation and compression characteristics[J]. Journal of the China Railway Society, 2014(4):71-75. DOI:10.3969. Summary of the invention

[0006] In order to solve the above problems, the present invention provides a method for constructing a crushable ballast model for discontinuous deformation analysis of ball units, comprising:

[0007] A step (1) for scanning the physical object of the ballast particles to obtain a triangulated network model of the ballast;

[0008] Step (2) of voxelizing the triangulated mesh model obtained in step (1) to obtain a voxelized model;

[0009] and

[0010] Step (3) for scaling the voxelized model obtained in step (2) according to the number of contacts to obtain a crushable ballast model.

[0011] Based on the above solution, step (2) includes:

[0012] The conservative rasterization step (2-1) for marking the particle contour of the triangular mesh model obtained in step (1) in the voxel space to obtain the outer shell of the ballast particle model;

[0013] The flood filling method step (2-2) for filling the inside of the outer shell of the ballast particle model with voxel units;

[0014] and

[0015] The sphere unit replacement step (2-3) for replacing the voxel units filled inside the outer shell of the ballast particle model with sphere units.

[0016] Based on the above solution, when performing scaling according to the number of contacts in step (3), the number of contacts is used as the condition for determining whether to continue particle enlargement; when the number of contacts of a particle is less than 4, the particle is enlarged; when the number of contacts is equal to 4, the particle size remains unchanged; when the number of contacts is greater than 4, the particle is reduced.

[0017] Based on the above solution, the specific method of step (3) is as follows:

[0018] Based on the voxelized ballast particle model generated, multiply the radius of each sphere unit sub-particle of the voxelized model by a random number uniformly distributed between 0.50000 and 0.99995 generated by the rand function; use the rand function to set the sub-velocities of the sphere unit in the x, y, and z directions to random numbers between -1 and 1, so that the direction of the resultant velocity of the sphere unit is any three-dimensional direction, and its initial resultant velocity magnitude is at most 1;

[0019] The position of each sphere unit is different at each time step, and the number of sphere units in contact with it is also different. According to the number of contacts of the sphere unit at each time step, it is scaled, so that the size and arrangement of the sphere unit sub-particles have randomness.

[0020] Based on the above solution, the process of scaling according to the number of contacts is as follows:

[0021] The center of any sphere unit is (x, y, z), and the radius is r. There are several other sphere units around this sphere unit, and these sphere units are: Sphere unit 1 (x 1 , y 1 , z 1 ), with a radius of r 1 ; Sphere unit 2 (x 2 , y 2 , z2 ), with a radius of r 2 ; spherical unit 3(x 3 , y 3 , z 3 ), with a radius of r 3 ; spherical unit 4(x 4 , y 4 , z 4 ), with a radius of r 4 ……; the number of contacts c is initially 0 times;

[0022] If (x - x 1 ) 2 +(y - y 1 ) 2 +(z - z 1 ) 2 =(r + r 1 ) 2 , then the number of contacts c increases by 1 time, otherwise it does not increase;

[0023] If (x - x 2 ) 2 +(y - y 2 ) 2 +(z - z 2 ) 2 =(r + r 2 ) 2 , then the number of contacts c increases by 1 time, otherwise it does not increase;

[0024] If (x - x 3 ) 2 +(y - y 3 ) 2 +(z - z 3 ) 2 =(r + r 3 ) 2 , then the number of contacts c increases by 1 time, otherwise it does not increase;

[0025] If (x - x 4 ) 2 +(y - y 4 ) 2 +(z - z 4 ) 2 =(r + r 4 ) 2 , then the number of contacts c increases by 1 time, otherwise it does not increase;

[0026] And so on...

[0027] Since the spherical unit has 4 parameters (x, y, z, r), the magnification and reduction conditions of the particle are bounded by the number of contacts c = 4;

[0028]

[0029] where: r 0 is the initial radius of the spherical unit; Δr is taken as 5e-7;

[0030] When the contact number of the spherical unit particles is less than 4, the particles are enlarged; when the contact number is equal to 4, the particle size remains unchanged; when the contact number is greater than 4, the particles are reduced.

[0031] On the basis of the above solution, the specific method for replacing the spherical unit in step (2-3) is as follows:

[0032] The coordinates of the voxel unit to be replaced in the voxel space are (R n , C n , L n ), indicating that the voxel is in the R n th row, the C n th column, and the L n th layer. The data of the replaced spherical unit is obtained by the following formula:

[0033]

[0034] In the formula: d vox is the size of the voxel unit; x s、 y s、 z s、 r s are the three-dimensional coordinates and radius of the spherical unit respectively;

[0035] When all the voxel units to be replaced are converted into spherical units, a ballast ball unit DDA model composed of equal-diameter balls with 6 coordination numbers is obtained; then, the model with 6 coordination numbers is converted into a model with 12 coordination numbers.

[0036] On the basis of the above solution, during the three-dimensional scanning in step (1), a blue light scanner is used; during the scanning, the ballast particles are rotated on the rotating table for one full circle, and the data on the particle surface is recorded except for the side in contact with the rotating table. Then, the particles are rotated by an angle and scanned again. When scanning again, all the parts that were blocked during the first scan are recorded, and the final measurement result is obtained by combining the two scan measurement results.

[0037] The modeling method of the true shape model for the discontinuous deformation analysis of crushable ballast of the present invention is simple. During voxelization, a model with an outer contour similar to the true ballast particles can be obtained through three steps: conservative rasterization, flood filling method, and sphere unit replacement. Based on the voxelized model, by means of the method of scaling according to the number of contacts, the internal sub-particles with the same arrangement rules and sizes inside the voxelized model, which cannot reflect the true ballast particles, are randomly rescaled and arranged, ensuring that the ballast sub-particles are in full contact with each other without overlapping, thereby obtaining a discontinuous analysis crushable model of spherical unit ballast with a true shape. Brief Description of the Drawings

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0039] Figure 1 It is a flowchart of the modeling method of this application;

[0040] Figure 2 It is a schematic diagram of three-dimensional scanning in the method of this application;

[0041] Figure 3 It is a schematic diagram of the processing of the invisible part at the bottom of the ballast during three-dimensional scanning in the method of this application;

[0042] Figure 4 It is a schematic diagram of surface reconstruction and model simplification after three-dimensional scanning in the method of this application;

[0043] Figure 5 It is a schematic diagram of the flood filling method in the method of this application;

[0044] Figure 6 It is a schematic diagram of sphere unit replacement in the method of this application;

[0045] Figure 7 It is a schematic diagram of the scaling method according to the number of contacts for three-dimensional irregular ballast in the method of this application;

[0046] Figure 8 It is a schematic diagram of the structure of the bonding unit described in this application;

[0047] Figure 9 It is a result diagram of the ballast crushing simulation analysis in this application. Detailed Embodiments

[0048] Next, in combination with the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0049] Embodiment 1

[0050] As Figure 1 shown, the present application provides a method for modeling a true shape model for discontinuous deformation analysis of crushable ballast. Specifically, as Figure 1 shown, the present invention realizes the modeling of crushable ballast mainly through three steps: 3D scanning, voxelization, and scaling according to the number of contacts. The ballast is also gradually converted from real ballast particles into a triangular mesh model, a voxelized model, and finally a crushed model.

[0051] Specifically, the method of the present invention includes the following steps:

[0052] Step (1) 3D scanning

[0053] As Figure 2 shown, the present invention uses a blue light scanner to scan the physical object of the ballast particles to obtain the triangular mesh pattern of the ballast.

[0054] During the blue light scanning, the part where the particle contacts the rotating table is invisible. By rotating and scanning the particle on the rotating table for one circle, except for the side that contacts the rotating table, the data on the particle surface is recorded. Then, the particle is rotated by an angle (the purpose is to capture the part that was blocked during the first shooting during the second shooting), and measured again. The two measurement results are combined to obtain the final measurement result, as Figure 3 shown.

[0055] The number of triangular meshes of the ballast can be adjusted as needed. In this article, it is expected that the ballast particles are composed of 150 - 300 sub-particles. Therefore, the triangular mesh model is simplified to be composed of 3000 triangular faces. The result is as Figure 4 shown.

[0056] In the above process, surface reconstruction is a function built into the blue light scanner, and an stl file can be obtained; the number of triangular faces of the model can be modified through Geomagic software.

[0057] Step (2) Voxelization

[0058] The entire voxelization process includes three parts: conservative rasterization, flood filling method, and sphere element replacement.

[0059] (2 - 1) Conservative rasterization processing

[0060] Conservative rasterization is a method of converting triangular patches into voxels. A triangular mesh model is composed of individual triangular patches. By calculating the voxel units corresponding to each triangular patch, a voxelized shell can be constructed and combined. Conservative rasterization processing belongs to the conventional technical means of those skilled in the art and will not be elaborated here.

[0061] Through the algorithm of conservative rasterization, the particle contour can be marked in the voxel space to obtain the shell of the ballast particle model.

[0062] (2-2) Flood filling method processing

[0063] After voxelization to obtain the shell of the ballast, the next step is to fill the inside. Taking the marked particle contour as the boundary, the voxel space is divided into the particle interior and the particle exterior.

[0064] Through the flood filling algorithm, it is possible to fill the inside of the voxelization result with particles. Here, the two-dimensional case is taken as an example for illustration: As shown in -a, the red is the particle contour boundary obtained in the previous step, and the blue voxels are the starting points of the flood filling method. Generally, voxels that must be outside the particle contour boundary are selected as the starting points. Figure 5 As shown in -a, the red is the particle contour boundary obtained in the previous step, and the blue voxels are the starting points of the flood filling method. Generally, voxels that must be outside the particle contour boundary are selected as the starting points.

[0065] Figure 5 -b is a schematic diagram of the loop body of the flood filling method. Before the filling is completed, the program always repeats this process. In this algorithm, the voxel space is stored as a three-dimensional matrix. The white, red, and blue in the schematic diagram correspond to the numbers 0, 1, and 2 respectively. Starting from the selected starting point, the colors of the adjacent voxels above, below, left, and right (six adjacent voxels in the three-dimensional case) of the blue voxel (the voxel marked with a light blue dot in the center) are judged in turn. If it is white, the adjacent voxel is modified to blue and pushed onto the stack. The stack stores the voxels to be looped; if it is red or blue, nothing is done. After the above operations are completed on the initial voxels, the same operations are performed on the voxels in the stack in turn until there are no remaining voxels in the stack. After the flood filling method is completed, the blue voxels are the particle exterior, the red voxels are the particle boundary, and the white voxels are the particle interior. The ballast voxel model is the part of the white and red voxels in the figure. After being processed by the flood filling method, the inside of the obtained voxelized model has been filled with particles, and the shape of the outer contour is close to that of the particles of the example object.

[0066] (2-3) Sphere unit replacement

[0067] The last step of voxelization modeling is to replace voxel units with spherical units. Voxel units are generally considered to be square, and inside the computer, the storage format of the voxel space is a three-dimensional matrix. When an element in the voxel matrix is 0, this voxel unit should be converted into the corresponding spherical unit. Let the coordinates of this voxel in the voxel space be (R n , C n , L n ), indicating that the voxel is in the R n -th row, the C n -th column, and the L n -th layer. The data of the spherical unit is obtained by the following formula:

[0068]

[0069] where: d vox is the size of the voxel unit; x s , y s , z s , r s are the three-dimensional coordinates and radius of the spherical unit respectively.

[0070] After all white voxels are converted into spherical units, a ballast ball unit DDA model composed of equal-diameter balls with 6 coordination numbers is obtained. In this model composed of equal-diameter balls with 6 coordination numbers, there are large voids. In fact, this 6-coordination number arrangement structure is considered to be the loosest arrangement. And the densest arrangement of equal-diameter balls is 12 coordination numbers. In crystal theory, these two packing methods are called simple cubic packing and hexagonal close packing respectively. By deleting some balls in the 6-coordination number model and enlarging some other balls, the 6-coordination number model can be converted into a 12-coordination number model, as shown in Figure 6 . The 12-coordination number model has very small voids and is a densest arrangement structure.

[0071] The particles generated by voxelization are similar to real particles in the outer contour, but cannot reflect the internal situation of the particles. Since the sub-particles inside the particles are arranged regularly and have the same size, when the particles are broken under the action of external forces, the particles will inevitably crack only along the regular planes, which is quite different from the actual particle breakage.

[0072] Step (3) Scaling according to the contact number

[0073] The particle magnification method is a commonly used method for generating specimens in discrete element simulations. In previous applications, the magnification ratio was usually set to enable the particles to finally reach the specified gradation. The advantage of this approach is that the gradation of the final specimen can be determined, but there will be relatively large gaps between the particles. Therefore, it is suitable for simulating soil samples piled together but not for simulating the whole crushed ballast particles to be broken. The number of contacting particles among the particles in the specimen can reflect the degree of compaction of the particles in the specimen to a certain extent. In the method of the present invention, the contact number is used as a condition for determining whether to continue particle magnification, with the aim of having the particles closely arranged after filling is completed.

[0074] The ideal arrangement of sub-particles should be able to reflect the outer contour of the particles and have sub-particles with random sizes and random arrangements, with the sub-particles contacting each other pairwise or being at a very small distance from each other. Therefore, based on the ballast particle model generated by voxelization, the radius of each spherical unit sub-particle in the voxelization model is multiplied by a random number uniformly distributed between 0.50000 and 0.99995 generated using the rand function; the rand function is used to set the sub-velocities of the spherical unit in the x, y, and z directions to random numbers from -1 to 1. To improve its accuracy, it can be taken to 5 decimal places, so that the direction of the resultant velocity of the spherical unit is in any three-dimensional direction, and the maximum magnitude of its initial resultant velocity is 1.

[0075]

[0076] The position of each spherical unit is different at each time step, and the number of spherical units in contact with it is also different. According to the contact number of the spherical unit at each time step, it is scaled, so that the size and arrangement of the spherical unit sub-particles have randomness. The specific contact number scaling method is as follows:

[0077] The center of any spherical unit is (x, y, z), and the radius is r. There are several other spherical units around this spherical unit, and these spherical units are: spherical unit 1 (x 1 , y 1 , z 1 ), with a radius of r 1 ; spherical unit 2 (x 2 , y 2 , z 2 ), with a radius of r 2 ; spherical unit 3 (x 3 , y 3 , z 3 ), with a radius of r 3 ; spherical unit 4 (x 4 , y 4 , z 4 ), with a radius of r 4 ……. The initial contact number c is 0 times.

[0078] If (x - x1 ) 2 +(y - y 1 ) 2 +(z - z 1 ) 2 =(r + r 1 ) 2 , the contact number c increases by 1, otherwise it does not increase;

[0079] If (x - x 2 ) 2 +(y - y 2 ) 2 +(z - z 2 ) 2 =(r + r 2 ) 2 , the contact number c increases by 1, otherwise it does not increase;

[0080] If (x - x 3 ) 2 +(y - y 3 ) 2 +(z - z 3 ) 2 =(r + r 3 ) 2 , the contact number c increases by 1, otherwise it does not increase;

[0081] If (x - x 4 ) 2 +(y - y 4 ) 2 +(z - z 4 ) 2 =(r + r 4 ) 2 , the contact number c increases by 1, otherwise it does not increase;

[0082] And so on...

[0083] Since the spherical unit has 4 parameters (x, y, z, r), the magnification and reduction conditions of the spherical unit particles are bounded by the contact number c = 4.

[0084]

[0085] Where: r 0 is the initial radius of the spherical unit; when the radius of the ballast particle is 0.03 - 0.05 m, Δr = 5e - 7. Δr should be selected according to the length of the calculation time step, the particle radius and other conditions. Δr can be modified in the scaling program according to the contact number. The radius of the ballast particles is 0.03 - 0.05 m, the time step is 1e - 6, and Δr is taken as 5e - 7.

[0086] When the number of contacts of the ball unit particles is less than 4, enlarge the particles; when the number of contacts is equal to 4, keep the particle size unchanged; when the number of contacts is greater than 4, shrink the particles. In the three-dimensional case, the calculation frame is composed of triangular faces, and inside the boundary are randomly moving spheres, as Figure 7 shown. As the scaling continues, the radius of the ball unit particles in the model tends to be stable. At this time, the sub-particles of the ballast particle model established have random arrangements, sizes, and are in close contact with each other, and can well simulate the real shape of the particles.

[0087] Through the above steps (1)-(3), the method of the present invention can finally construct a crushable model of ballast for discontinuous deformation analysis of ball units.

[0088] The model established by the method of the present invention simulates the crushable properties of particles by setting bonding units. A bonding unit is added between two contacting spheres to bond the spheres together, making a group of discrete spheres into a continuous medium. The bonding unit is a solid circular frustum, and each bonding unit has a fracture surface that is the same as the middle cross-section of the bonding unit. During the calculation process, a fracture algorithm is used to monitor the stress of each bonding unit. When the force on the bonding unit reaches a certain value, or the distance between two ball units exceeds the threshold, the bonding unit fractures, and the fractured bonding unit no longer restricts the relative movement between the bonded particles. The schematic diagram of the bonding unit is as Figure 8 shown.

[0089] Due to the limitation of simulating ballast crushing by bonding units, there should be no overlap between the sub-particles of the ballast model, and they should just be in contact with each other, so that on average each sub-particle has a relatively large number of contacts. Such close contact can better simulate the crushing characteristics of ballast.

[0090] The method of the present invention can make the ball units non-overlapping and just in contact with each other, which is a necessary condition for adding bonding units. After adding bonding units, the model has the crushable properties.

[0091] Use a large-scale consolidometer to load a ballast specimen until it breaks, and the force-displacement curve of the specimen can be obtained. At the same time, use the method of the present invention to generate a model of the ballast specimen and conduct a crushing simulation test identical to the indoor test, and obtain the force-displacement curve through simulation under the selected calculation parameters.

[0092] The specific parameters are as follows:

[0093]

[0094] The comparison between the indoor test and the simulation test is as Figure 9As shown, it can be seen that the curves of the simulation test fit well. The simulated peak force and the deformation of the particles during fragmentation are very close to the peak fragmentation force measured in the test. There is only a slight deviation in the slope of the force-displacement curve from the test value. The form of the fragments formed after fragmentation is as Figure 9 shown. The position where fragmentation occurs in the simulation is the same as that in the actual test, and the shapes are similar.

[0095] Therefore, it can be seen that the method of the present invention can preferably establish a discontinuous analysis breakable model of ballast with a ball unit having a real shape.

[0096] Each embodiment in this specification is described in a related manner. For the same and similar parts between the embodiments, reference can be made to each other. The key point of each embodiment is to illustrate the differences from other embodiments. The above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims

1. A method for constructing a crushable ballast model for discontinuous deformation analysis of ball units, characterized in that: include: A step (1) for scanning the physical object of ballast particles to obtain a triangulated network model of the ballast; Step (2) of voxelizing the triangulated mesh model obtained in step (1) to obtain a voxelized model; and Step (3) of scaling the voxelized model obtained in step (2) according to the number of contacts to obtain a crushable ballast model; Step (2) comprises: A conservative rasterization step (2-1) for marking the particle contours of the triangulated mesh model obtained in step (1) in voxel space, thereby obtaining the outer shell of the ballast particle model; A flood filling method step (2-2) for filling the interior of the shell of the ballast particle model with voxel units; and A spherical unit replacement step (2-3) for replacing voxel units filled inside the shell of the ballast particle model with spherical units; When scaling by contact number is performed in step (3), the contact number is used as a condition for determining whether to continue to scale up the particle; when the contact number of the particle is less than 4, the particle is scaled up; when the contact number is equal to 4, the particle size is kept unchanged; when the contact number is greater than 4, the particle is scaled down; The specific method of step (3) is as follows: Based on the ballast particle model generated by voxelization, the radius of each ball unit sub-particle of the voxelized model is multiplied by a random number uniformly distributed between 0.50000 and 0.99995 generated by the rand function; the rand function is used to set the component velocities of the ball unit in the x, y, and z directions to random numbers between -1 and 1, so that the direction of the combined velocity of the ball unit is any three-dimensional direction, and the maximum magnitude of its initial combined velocity is 1; The position of each ball unit is different in each time step, and the number of ball units in contact with it is also different. The ball unit is scaled according to the number of contacts at each time step, so that the size and arrangement of the ball unit sub-particles are random; The process of contact number scaling is as follows: The center of any ball unit is (x, y, z) and the radius is r. There are several other ball units around the ball unit, which are: ball unit 1 (x1, y1, z1), with a radius of r1; ball unit 2 (x2, y2, z2), with a radius of r2; ball unit 3 (x3, y3, z3), with a radius of r3; ball unit 4 (x4, y4, z4), with a radius of r4...; the number of contacts c is initially 0; If (x-x1) 2 +(y-y1) 2 +(z-z1) 2 =(r+r1) 2 , then the contact number c increases by 1, otherwise it does not increase; If (x-x2) 2 +(y-y2) 2 +(z-z2) 2 =(r+r2) 2 , then the contact number c increases by 1, otherwise it does not increase; If (x-x3) 2 +(y-y3) 2 +(z-z3) 2 =(r+r3) 2 , then the contact number c increases by 1, otherwise it does not increase; If (x-x4) 2 +(y-y4) 2 +(z-z4) 2 =(r+r4) 2 , then the contact number c increases by 1, otherwise it does not increase; And so on… Since the spherical unit has four parameters (x, y, z, r), the enlargement and reduction conditions of the particles are limited to the contact number c = 4; Where: r0 is the initial radius of the spherical unit; Δr is 5e-7; When the contact number of the ball unit particles is less than 4, the particles are enlarged; when the contact number is equal to 4, the particle size remains unchanged; when the contact number is greater than 4, the particles are reduced.

2. The method for constructing a crushable ballast model for discontinuous deformation analysis of ball units according to claim 1, characterized in that: The specific method of the ball unit replacement step (2-3) is: The coordinates of the voxel unit to be replaced in the voxel space are (R n , C n , L n ), indicating that the voxel is in the Rth n Row, C n Column, L n The data of the replaced ball unit is obtained by the following formula: Where: d vox is the size of the voxel unit; x s、 y s、 z s、 r s are the three-dimensional coordinates and radius of the spherical unit respectively; When all the voxel units to be replaced are converted into ball units, a ballast ball unit DDA model consisting of 6-coordinated equal-diameter balls is obtained; then, the 6-coordinated model is converted into a 12-coordinated model.

3. The method for constructing a crushable ballast model for discontinuous deformation analysis of ball units according to claim 1, characterized in that: In step (1), a blue light scanner is used for three-dimensional scanning. During scanning, the ballast particles are rotated and scanned on a rotating table for one circle. The data of the particle surface except the side in contact with the rotating table is recorded. The particle is then rotated by an angle and re-scanned. During the re-scanning, all the parts that were blocked for the first time are recorded. The two scanning measurement results are combined to obtain the final measurement result.

4. A crushable ballast model constructed by the method according to any one of claims 1 to 3.

Images