A granular flow based polycrystalline rock modeling method, system and apparatus

Abstract

The application discloses a kind of polycrystal rock modeling method, system and equipment based on particle flow, it is related to rock mechanics and numerical simulation technical field, including the numerical model of polycrystal rock is built;In numerical model, it is filled by coarse grain, based on tessellation algorithm, rigid block rblock is generated;Rigid block rblock is traversed, and it is converted into geometric body crystal cluster;The coarse grain and rigid block rblock filled are deleted, and numerical model is filled again using fine grain, according to the position of the geometric body crystal cluster where fine grain is located, fine grain is grouped, and the grouping result is determined;According to the grouping result, the required simulation parameter is set, stress loading and crack propagation simulation operation is carried out, and simulation result is obtained.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of rock mechanics and numerical simulation, and in particular, relates to a multi-crystal rock modeling method, system and device based on particle flow. BACKGROUND

[0002] Granite, as a typical multi-crystal rock, is mainly composed of quartz, feldspar and mica. Due to its unique mineral composition and microstructure characteristics, multi-crystal rock often shows significant micro-heterogeneity. Taking granite as an example, its micro-heterogeneity is mainly related to the types and proportions of mineral components, the size and arrangement of crystal grains, and the distribution and development degree of micro-cracks, and these factors jointly affect the mechanical properties and crack propagation evolution characteristics of granite.

[0003] The particle flow software (Particle Flow Code, PFC) with the discrete element method as the core can effectively simulate the non-continuous deformation and progressive failure process of multi-crystal rock (such as granite). However, the current use of the Voronoi polygon and rigid block rblock module in the PFC software to simulate multi-crystal rock has the following shortcomings: First, when constructing Voronoi polygons based on the initial particle center and filling particles, especially in three-dimensional models and when the particle size difference is large, it is easy to cause particle filling misplacement or mixing into non-corresponding mineral areas, which leads to inaccurate establishment of particle contact relationships and further affects the modeling efficiency. In addition, the particle filling process is computationally inefficient, especially for large models or small particle filling, and the time difference in reaching equilibrium for different crystal clusters is huge. Moreover, the block name formed by filling based on the block discrete element idea cannot be directly associated with the mineral grouping, and the subsequent grouping operation programming is complex and has poor universality, which seriously affects the modeling efficiency. SUMMARY

[0004] The present application provides a multi-crystal rock modeling method, system, medium and device based on particle flow, to solve the above-mentioned problems existing in the prior art, i.e. how to effectively improve the modeling efficiency of multi-crystal rock in the prior art. The present application provides a multi-crystal rock modeling method based on particle flow, which comprises:

[0005] Constructing a numerical model of multi-crystal rock;

[0006] In the numerical model, filling it with coarse particles, generating rigid blocks rblock based on the Voronoi polygon space subdivision algorithm;

[0007] Traversing the rigid blocks rblock and converting them into geometric crystal clusters;

[0008] Delete the filled coarse particles and rigid blocks rblock, and refill the numerical model with fine particles, group the fine particles according to the position of the geometric crystal cluster where the fine particles are located, and determine the grouping result;

[0009] According to the grouping result, set the required simulation parameters, perform stress loading and crack propagation simulation operation, and obtain the simulation result.

[0010] Optionally, the Voronoi tessellation algorithm is used to generate the rigid block rblock, and specifically includes:

[0011] ;

[0012] Wherein, is the Euclidean distance, is the Euclidean space, represents the Voronoi crystal cluster, is the center point of the coarse particle, and represent the different coarse particle center point corresponding number.

[0013] Optionally, the numerical model of the polycrystalline rock specifically includes:

[0014] The numerical model of the semi-disc, disc, cylinder and cube of the granite with different grain sizes.

[0015] The present application provides a kind of polycrystalline rock modeling system based on particle flow, including:

[0016] The construction module is used to construct the numerical model of polycrystalline rock;

[0017] The division module is used to fill it with coarse particles in the numerical model, and generate rigid block rblock based on the Voronoi tessellation algorithm;

[0018] The traversal module is used to traverse the rigid block rblock and convert it into geometric crystal cluster;

[0019] The grouping module is used to delete the filled coarse particles and rigid blocks rblock, and refill the numerical model with fine particles, group the fine particles according to the position of the geometric crystal cluster where the fine particles are located, and determine the grouping result;

[0020] The simulation module is used to set the required simulation parameters according to the grouping result, perform stress loading and crack propagation simulation operation, and obtain the simulation result.

[0021] The application provides a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned granular flow-based polycrystalline rock modeling method when executing the program.

[0022] The beneficial effects of the present application relative to the prior art are as follows: the present application provides a granular flow-based polycrystalline rock modeling method, which has the following advantages by adopting the modeling logic of "coarse first, then fine, and then classification": first, the rigid block framework is formed by using coarse particles to preliminarily fill the numerical model, which greatly reduces the calculation amount and optimizes the calculation resources; second, after the rigid block is converted into a geometric crystal cluster, fine particles are used for re-filling, which simplifies the internal geometric structure of the model and ensures the simulation accuracy; finally, the fine particles are grouped based on the position of the geometric crystal cluster, which effectively solves the problems of inaccurate particle grouping and complex modeling process and cumbersome command flow in the traditional method. The present method is suitable for PFC2D and PFC3D software, and is simple to operate, efficient in calculation and strong in adaptability, which can accurately simulate the mechanical behavior and crack propagation process of polycrystalline rocks such as granite, and has important significance for understanding the microstructure characteristics and failure mode of polycrystalline rocks under complex mechanical environment. BRIEF DESCRIPTION OF DRAWINGS

[0023] The accompanying drawings, which are incorporated into and form part of the specification, illustrate embodiments consistent with the present application and, together with the specification, serve to explain the principles of the application.

[0024] Figure 1 A flowchart of a granular flow-based polycrystalline rock modeling method provided by an embodiment of the present application is shown in the figure.

[0025] Figure 2 A technical roadmap of a granular flow-based polycrystalline rock modeling method provided by an embodiment of the present application is shown in the figure.

[0026] Figure 3 is a schematic diagram of a two-dimensional numerical model coarse particle filling to preliminarily construct a semicircular disc, a circular disc and a rectangular numerical model provided by an embodiment of the present application;

[0027] Figure 4 is a schematic diagram of a two-dimensional numerical model region division and rigid block rblock generation by using the Voronoi polygon subdivision method provided by an embodiment of the present application;

[0028] Figure 5 is a schematic diagram of converting the rigid block output of a two-dimensional numerical model into a geometric crystal cluster provided by an embodiment of the present application;

[0029] Figure 6 is a schematic diagram of a two-dimensional numerical model fine particle filling to construct a semicircular disc, a circular disc and a rectangular numerical model provided by an embodiment of the present application;

[0030] Figure 7 is a schematic diagram of grouping particles of a two-dimensional numerical model using a loop nesting function provided by an embodiment of the present application;

[0031] Figure 8 is a schematic diagram of adding cement to internal mineral components of a two-dimensional numerical model provided by an embodiment of the present application;

[0032] Figure 9 is a schematic diagram of constructing a semi-disc, disc, and cubic numerical model using coarse particle filling of a three-dimensional numerical model provided by an embodiment of the present application;

[0033] Figure 10 is a schematic diagram of dividing a three-dimensional numerical model into regions and generating rigid blocks rblock using a Voronoi tessellation method provided by an embodiment of the present application;

[0034] Figure 11 is a schematic diagram of converting the output of rigid blocks of a three-dimensional numerical model into geometric crystal clusters provided by an embodiment of the present application;

[0035] Figure 12 is a schematic diagram of constructing a semi-disc, disc, and cubic numerical model using fine particle filling of a three-dimensional numerical model provided by an embodiment of the present application;

[0036] Figure 13 is a schematic diagram of grouping particles of a three-dimensional numerical model using a loop nesting function provided by an embodiment of the present application;

[0037] Figure 14 is a schematic diagram of adding cement to internal mineral components of a three-dimensional numerical model provided by an embodiment of the present application;

[0038] Figure 15 is a schematic diagram of constructing a semi-disc, disc, and cubic numerical model of different grain sizes using a three-dimensional numerical model provided by an embodiment of the present application;

[0039] Figure 16 is a schematic diagram of dividing a three-dimensional numerical model into regions and generating rigid blocks rblock using a Voronoi tessellation method provided by an embodiment of the present application;

[0040] Figure 17 is a schematic diagram of converting the output of a three-dimensional numerical model of different grain sizes into geometric crystal clusters provided by an embodiment of the present application;

[0041] Figure 18 is a schematic diagram of a semi-disc, disc, and cubic three-dimensional numerical model of different grain sizes of granite provided by an embodiment of the present application;

[0042] Figure 19Is the preliminary construction of different grain size two-dimensional semicircular disc, disc, rectangular numerical model schematic diagram provided by the embodiment of the application;

[0043] Figure 20 Is the two-dimensional numerical model is divided into regions and rigid block rblock is generated by the tessellation method provided by the embodiment of the application;

[0044] Figure 21 Is the two-dimensional numerical model output conversion geometry crystal cluster schematic diagram provided by the embodiment of the application;

[0045] Figure 22 Is the two-dimensional numerical model schematic diagram of different grain size granite semicircular disc, disc, rectangle provided by the embodiment of the application;

[0046] Figure 23 The computer device schematic diagram of the polycrystalline rock efficient modeling method based on particle flow provided by the embodiment of the application. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical scheme and advantages of the present application clearer, the technical scheme in the present application will be described clearly and completely below in combination with the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0048] Firstly, the terms involved in the present application are explained:

[0049] Particle flow: PFC (Particle Flow Code) software is an application example of discrete element method, which can effectively simulate the complex dynamic process of particle medium, including the movement, collision and other contact effects of particles. PFC software can handle the complex large displacement and rotation phenomenon inside the rock mass by means of discrete element method, and even simulate the block separation process, which provides a practical and feasible simulation means for studying the nonlinear large deformation characteristics in jointed rock mass.

[0050] Traversal: in particle flow, "traversal" is a key calculation process, which involves systematic inspection and data processing of key components of the model, including particles (balls), contact points (contacts), walls (walls) and rigid clusters (clumps). By executing the "loop foreach" command, the software can access each element within the defined range along the preset search path and extract its key attributes such as position and velocity as data pointers for further analysis.

[0051] Pointers: Pointers play a crucial role in the core mechanisms of PFC particle flow simulation. Pointers are special variables used to track and reference particles or collections of particles, providing direct access to specific data points within the particle data structure. This mechanism enables the simulator to quickly retrieve and adjust key particle properties, such as their position, velocity, and force state. Pointers are not limited to static data access; they can also dynamically reference newly generated particles or mark those to be removed from the simulation. When performing massively parallel computations, another key role of pointers is coordinating task allocation across different processors or compute nodes. Intelligent use of pointers can optimize computing resource utilization and accelerate simulations, especially when working with complex models containing millions of particles. Pointers also support data export, allowing information about particle states and interactions to be exported to text files (e.g., txt) or data files (e.g., dat).

[0052] rblock rigid block: In PFC software, rblock rigid block is an important model component used to simulate geological bodies or structural units in geotechnical mechanics. It has the advantages of significantly reducing computing time and resource consumption, improving computing efficiency, adapting to complex geometries, and having good compatibility with other modules. It can accurately represent complex geometric shapes and mechanical properties, and can be created in a variety of ways, such as importing from geometries, combining command generation, etc. For large-scale engineering simulations, it can simulate geological bodies with complex geometries, such as slopes and rock masses, allowing PFC to more accurately simulate complex structures in actual engineering and provide a more reliable basis for engineering design and analysis; from the perspective of small-scale indoor tests, the computational efficiency of standard cylinders, cubes and other specimens constructed by rblock and generated with the help of rblock as a channel is high, and it plays an important role in simulating the mechanical behavior of geotechnical materials and describing crack initiation.

[0053] The following describes in detail the technical solution of the present invention and how the technical solution of the present invention solves the above-mentioned technical problems using specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The following embodiments of the present invention are described in conjunction with the accompanying drawings.

[0054] Example 1

[0055] like Figure 1 and Figure 2 As shown, this embodiment shows a polycrystalline rock modeling method based on granular flow, including:

[0056] S1: Construct numerical models of polycrystalline rocks.

[0057] Exemplarily, a numerical model can be constructed in the particle flow according to the test requirements, such as a semi-disk numerical model, a disk numerical model, a cylindrical numerical model, and the like.

[0058] S2: In the numerical model, the coarse particles are filled in the numerical model, and the rigid block rblock is generated based on a Voronoi space subdivision algorithm.

[0059] Exemplarily, the coarse particles with a relatively large size (such as 10%-15% of the size of the sample side length or radius) are filled in the constructed numerical model, which is used for subsequent construction of rblock and geometric crystal clusters; and the suspended particles (i.e., particles not in contact with other particles) and particles outside the model range are identified and deleted, and after the corresponding mechanical parameters are preliminarily assigned, a cyclic calculation is performed to release internal stress and reduce unbalanced force to a relatively small value (such as 1e-5).

[0060] Exemplarily, the center position of each particle of the constructed sample is defined as the Voronoi seed point position, and the following Voronoi space subdivision method is used to construct the rigid block rblock model:

[0061] ;

[0062] wherein, is the Euclidean distance, is the Euclidean space, represents the Voronoi crystal cluster, i.e., the rigid block rblock, is the seed, and represent different seed numbers.

[0063] S3: The rigid block rblock is traversed and converted into a geometric crystal cluster.

[0064] Exemplarily, by traversing all rblock pointers, each rigid block is given a standardized number name (such as "geo_1", "geo_2", and the like), and the identification and statistics of all rblocks are completed; then the rblocks with these structured numbers are converted into geometric crystal clusters; by executing the above steps in a loop, all rigid blocks are converted into geometric crystal clusters of corresponding mineral types, ensuring that the converted geometric bodies maintain the spatial characteristics of the original rigid blocks and have specific mineral property identifiers, to establish a dynamic association mechanism with the particle system, and lay a foundation for subsequent key processes such as particle accurate filling, mineral grouping, and parameter setting.

[0065] Exemplarily, before converting the rigid block rblock into the geometric crystal cluster, the mineral types of the rigid block rblock can be grouped according to the actual mineral components and contents, and the specific steps can include:

[0066] Traverse the rigid block rblock list, and perform the following operations on each rigid block: obtain a random number in the range of 0-1; according to the preset mineral content interval, map the random number to the corresponding mineral type; the specific mapping rule is: when the random number is less than the content of plagioclase, the rigid block is grouped as plagioclase; when the random number is greater than the content of plagioclase and less than the sum of the contents of plagioclase and potassium feldspar, the rigid block is grouped as potassium feldspar; when the random number is greater than the sum of the contents of plagioclase and potassium feldspar and less than the sum of the contents of plagioclase, potassium feldspar and quartz, the rigid block is grouped as quartz; when the random number is greater than 1 minus the content of mica, the rigid block is grouped as mica.

[0067] Through the above condition judgment, the process of gradually accumulating judgment according to each mineral component is realized, and it is ensured that the generated mineral distribution meets the preset mineral component content ratio.

[0068] S4: Delete the filled coarse particles and rigid blocks rblock, and refill the numerical model with fine particles. According to the position of the fine particles in the geometric crystal cluster, the fine particles are grouped, and the grouping result is determined.

[0069] Exemplarily, after determining the geometric range of the numerical model, the key parameters such as particle radius and porosity can be set, and the fine particles with relatively small size (about 0.5%-1% of the sample length or radius) can be used to accurately fill the sample; then, the random number generation function is used to map the content of different mineral components (such as quartz, mica, feldspar, etc.), which provides the basis for subsequent geometric crystal cluster grouping; further, through the optimized loop nesting function, all geometric crystal cluster information is accurately obtained and systematically classified, and grouping is realized according to the spatial position of the fine particles in the crystal cluster, ensuring that all fine particles in a specific geometric crystal cluster are correctly identified, such as all fine particles (quartz particles) in the quartz crystal cluster are correctly identified, or by traversing all geometric crystal clusters, the particle grouping operation is performed on each independent crystal cluster, and different colored particles (such as potassium feldspar, quartz, plagioclase, mica, etc.) in the crystal cluster are classified.

[0070] Exemplarily, all particle contacts generated are traversed, and corresponding contact force parameters are given to the same mineral inside and between different minerals, including contact models, mechanical parameters, etc., to perfect the polycrystal rock numerical model.

[0071] S5: According to the grouping result, set the required simulation parameters, perform stress loading and crack propagation simulation operations, and obtain the simulation result.

[0072] Example 2

[0073] Step one: three numerical models were established in particle flow software PFC2D respectively, the geometric properties and size information are as follows: semi-circular disc (SCB) numerical model, radius 25 mm, slot width 1 mm, length 12 mm, the particle size distribution of coarse particles in the filling sample is 0.8 mm-1.2 mm, the porosity is 0.08, a total of 267 particles; disc numerical model, radius 25 mm, the particle size distribution of coarse particles in the filling sample is 0.8 mm-1.2 mm, the porosity is 0.08, finally 633 particles are generated; rectangular numerical model, width 50 mm, height 100 mm, the particle size distribution of coarse particles in the filling sample is 1 mm-1.5 mm, the porosity is 0.08, finally 934 particles are generated, the coarse particles filling the preliminary constructed sample is as shown in Figure 3 ;

[0074] Step two: write a function to identify suspended particles, and delete them together with the particles outside the geometric area range of the numerical model;

[0075] Step three: give the particles generated in step two effective stiffness, normal-tangent stiffness ratio and other mechanical parameters for subsequent modeling;

[0076] Step four: divide the seed points (coarse particle center points) generated in step one according to the division rule of the tessellation polygon, and generate rblock rigid blocks, the effect diagram after division is as shown in Figure 4 ;

[0077] Step five: write a parameterized function to perform global traversal of rigid block rblock, systematically number and name each rblock entity generated in step four, and build a complete identification system from "geo_1" to "geo_n"; after completing the naming, batch convert these structured identified rblock rigid blocks into geometric crystal clusters, establish an efficient geometric-particle linkage mechanism, lay a foundation for subsequent accurate particle filling and intelligent grouping, the spatial distribution of the converted crystal clusters is as shown in Figure 5 ;

[0078] Step six: delete the coarse particles filled in step one and all rigid blocks rblock in step four, and fill the three numerical models with smaller size fine particles. The fine particle size distribution of the SCB numerical model is 0.2mm-0.334, the porosity is 0.08, and finally 3857 particles are generated; the fine particle size distribution of the disc numerical model is 0.2mm-0.334, the porosity is 0.08, and finally 7938 particles are generated; the fine particle size distribution of the rectangular numerical model is 0.2mm-0.334, the porosity is 0.08, the minimum radius of the fine particles used to fill the sample is 0.2mm, the maximum to minimum particle size ratio is 1.67, the porosity is 0.08, and finally 20114 particles are generated. The fine particle filled numerical model is shown in Figure 6 ;

[0079] Step seven: use the 0-1 range random number function provided by PFC software, combined with the different content intervals of the four minerals, the random number interval corresponds to the mineral species. For example, the content of mica mineral is 0.017, so the group specified by the random number in the range of 0-0.017 is mica; the content of quartz mineral is 0.214, so the group specified by the random number in the range of 0.017-0.231 is quartz, and so on until all mineral species are specified;

[0080] Step eight: implement systematic traversal of all geometric crystal clusters through iterative algorithm (from the second cluster to the final cluster), and perform accurate particle grouping logic for each independent crystal cluster to realize intelligent identification and accurate classification of different color characteristic particles (corresponding to potassium feldspar, quartz, plagioclase, mica and other mineral components), and form a visual model of mineral spatial distribution, as shown in Figure 7 ;

[0081] Step nine: write a function to traverse the contact list between local spherical particles and particles, obtain the two particle groups contacted by the contact pointer, and the particle group will play a role in identifying particles in the next step. Under the premise that the contact groups of the two minerals are the same, if both particle groups are plagioclase, then assign "plagioclase" to the group of the contact pointer. Similarly, the groups of the remaining three minerals can be determined; if the contact groups of the two minerals are not the same, then the group of the contact pointer is recorded as "joint". After completing all the above operations, the establishment of non-homogeneous rock is realized;

[0082] Step ten: define three arrays to store the effective modulus, stiffness ratio and other parameters of the linear part and parallel bond part, which are used for subsequent assignment.

[0083] For the contact group in step nine, if it is identified as "plagioclase", the contact pointer type is given as "linear parallel bonding model"; then according to the attributes, modes, etc. of the linear parallel bonding model, the parameters such as effective modulus, stiffness ratio, tensile strength, cohesion, internal friction angle, friction coefficient are defined by using array assignment, that is, the mechanical parameter definition of "plagioclase" mineral is completed, and the other three minerals and joint groups are assigned by analogy with the above method, that is, the complete setting of the internal contact model of the sample is completed, and the effect diagram is shown in Figure 8 , wherein the dark gray contact is a parallel bonding model, the black is a smooth joint bonding model, the contact assignment is normal, and there are few gaps.

[0084] Example 3

[0085] Step one: three numerical models are respectively established in the particle flow software PFC3D, and the geometric properties and size information are as follows: the SCB numerical model has a radius of 25 mm, a thickness of 25 mm, a notch width of 1 mm, and a length of 12 mm; the coarse particle size distribution of the filled sample is 1.8 mm-2.52 mm, and the porosity is 0.1, and finally 542 particles are generated; the radius of the disc numerical model is 25 mm, the thickness is 25 mm, the coarse particle size distribution of the filled sample is 2 mm-3 mm, and the porosity is 0.05, and finally 895 particles are generated; the bottom surface length of the cubic numerical model is 25 mm, the height is 50 mm, the coarse particle size distribution of the filled sample is 2 mm-3 mm, and the porosity is 0.1, and finally 412 particles are generated, and the coarse particle filling completes the preliminary construction of the sample as shown in Figure 9 ;

[0086] Step two: write a function to identify suspended particles, and delete them together with the particles outside the geometric area range of the numerical model;

[0087] Step three: give the particles generated in step two effective stiffness, normal and tangential stiffness ratio and other mechanical parameters for subsequent modeling;

[0088] Step four: use the particle flow self-contained command to divide the seed points (coarse particle center points) generated in step one according to the division rules of the Voronoi polygon, and generate rblock rigid blocks, and the effect diagram after division is as shown in Figure 10 ;

[0089] Step five: write a function to traverse the rblock rigid block pointer, and number and name each rblock rigid block in step four until all rblock rigid blocks are named, and then output all rblock rigid blocks to crystal clusters, which can be linked with particles for subsequent particle filling grouping and other operations, and the effect diagram after exporting the crystal clusters is as shown in Figure 11 ;

[0090] Step six: delete the coarse particles filled in step one and all rigid blocks rblock in step four, and fill the three numerical models with smaller size fine particles. The fine particle size distribution of the SCB numerical model is 0.2mm-0.334, the porosity is 0.08, and finally 3857 particles are generated; the fine particle size distribution of the disc numerical model is 0.2mm-0.334, the porosity is 0.08, and finally 7938 particles are generated; the fine particle size distribution of the rectangular numerical model is 0.2mm-0.334, the porosity is 0.08, the minimum radius of the fine particles used to fill the sample is 0.2mm, the maximum to minimum particle size ratio is 1.67, the porosity is 0.08, and finally 20114 particles are generated. The fine particle filled sample is shown in Figure 12

[0091] Step seven: use the 0-1 range random number function provided by PFC software, combined with the different content intervals of the four minerals, and the random number interval corresponds to the mineral category. For example, the content of mica mineral is 0.017, so the group specified by the random number in the range of 0-0.017 is mica; the content of quartz mineral is 0.214, so the group specified by the random number in the range of 0.017-0.231 is quartz, and so on until all mineral categories are specified;

[0092] Step eight: by looping through all the geometric body crystal clusters (from the second cluster to the last cluster), perform particle grouping operation on each crystal cluster, and classify different colored particles (representing potassium feldspar, quartz, plagioclase, mica) respectively, as shown in Figure 13

[0093] Step nine: write a function to traverse the contact list between local spherical particles and particles, obtain the two particle groups contacted by the contact pointer, and the particle group will play a role in identifying particles in the next step. Under the premise that the contact groups of the two minerals are the same, if both particle groups are plagioclase, then assign "plagioclase" to the group of the contact pointer. Similarly, the groups of the remaining three minerals can be determined; if the contact groups of the two minerals are not the same, then the group of the contact pointer is recorded as "joint". After completing all the above operations, the establishment of non-homogeneous rock is realized;

[0094] Step ten: define three arrays to store the effective modulus, stiffness ratio and other parameters of the linear part and parallel bond part, which are used for subsequent assignment.

[0095] ​​For the contact group in step nine, if identified as "plagioclase", it is given a contact pointer type of "linear parallel bonding model"; then according to the attributes, modes, etc. of the linear parallel bonding model, the parameters such as effective modulus, stiffness ratio, tensile strength, cohesion, internal friction angle, friction coefficient are defined by using array assignment, that is, the mechanical parameters of "plagioclase" mineral are defined, and the other three minerals and joint groups are assigned by analogy with the above method, that is, the complete setting of the internal contact model of the sample is completed, and the effect diagram is as shown in Figure 14 , wherein the dark gray contact is a parallel bonding model, the black is a smooth joint bonding model, the contact assignment is normal, and there are few gaps.

[0096] Example 4

[0097] Step one: build a granite numerical model with different grain sizes in the particle flow software PFC3D, and select a cube as the reference numerical model. The bottom surface is 25mm long, 50mm high, and the minimum particle radius distribution of the coarse particles for filling the coarse-grained, medium-grained and fine-grained granite numerical models is set to 3mm, 1.5mm and 0.75mm, respectively. The preliminary model is as shown in Figure 15 .

[0098] Step two: use the command provided by the particle flow to divide the seed points (center points of coarse particles) of the sample generated in step one according to the rules of the Thiessen polygon and generate rblock rigid blocks, and then output them as geometric crystal clusters after completion, as shown in Figure 16 , 17 .

[0099] Step three: fill the original sample with fine particles with a minimum particle radius of 0.6mm, and group the filled fine particles using the grouping function. The numerical models of granite with three different grain sizes are as shown in Figure 18 .

[0100] Example 5

[0101] Step one: build a granite numerical model with different grain sizes in the particle flow software PFC2D, and select a rectangular numerical model as the reference sample. The width is 50mm, the height is 100mm, and the minimum particle radius distribution of the coarse particles for filling the coarse-grained, medium-grained and fine-grained granite numerical models is set to 3mm, 1.5mm and 0.75mm, respectively. The preliminary model is as shown in Figure 19 .

[0102] Step two: use the command provided by the particle flow to divide the seed points (center points of coarse particles) of the sample generated in step one according to the rules of the Thiessen polygon and generate rblock rigid blocks, and then output them as geometric crystal clusters after completion, as shown in Figure 20 , 21 .

[0103] Step three: fill the original sample with fine particles with a minimum particle radius of 0.3mm, group the filled fine particles using the function, and the numerical model of the three different grain sizes of granite is as shown in Figure 22 .

[0104] The above is the particle flow-based polycrystalline rock modeling method provided by one or more embodiments of the present specification. Based on the same idea, the present specification also provides a corresponding particle flow-based polycrystalline rock modeling system, which comprises:

[0105] A construction module for constructing a numerical model of a polycrystalline rock;

[0106] A division module for filling the numerical model with coarse particles, generating rigid blocks rblock based on the Voronoi space subdivision algorithm;

[0107] A traversal module for traversing the rigid blocks rblock and converting them into geometric body crystal clusters;

[0108] A grouping module for deleting the filled coarse particles and rigid blocks rblock, and re-filling the numerical model with fine particles, grouping the fine particles according to their positions in the geometric body crystal clusters, and determining the grouping results;

[0109] A simulation module for setting the required simulation parameters according to the grouping results, performing stress loading and crack propagation simulation operations, and obtaining simulation results.

[0110] The specific limitations of the particle flow-based polycrystalline rock modeling system can be referred to the limitations of the particle flow-based polycrystalline rock modeling method described above, which will not be repeated here. Each module in the particle flow-based polycrystalline rock modeling system described above can be realized by software, hardware and their combination. The above modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0111] The present application also provides a computer device as shown in Figure 23 , and the structure diagram of the computer device is as shown in Figure 23 . At the hardware level, the computer device comprises a processor, an internal bus, a network interface, a memory and a non-volatile memory, and of course can also comprise other hardware required by the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs to realize the particle flow-based polycrystalline rock modeling method provided by the above embodiments.

[0112] The technical features of the above embodiments can be combined in any manner. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described, however, as long as the combinations of the technical features do not contradict each other, they should be considered to be within the scope of the present application.

Claims

1. A polycrystalline rock modeling method based on granular flow, characterized in that: include: Construct numerical models of polycrystalline rocks; In the numerical model, the coarse particles are filled in it, and the rigid blocks rblock are generated based on the Thiessen polygon space partitioning algorithm; The method of generating a rigid block rblock based on the Thiessen polygon space partitioning algorithm specifically includes: ； in, is the Euclidean distance, For European space, Represents a Voronoi crystal cluster of Thiessen polygons, is the center point of the coarse particle, and Represents the numbers corresponding to the center points of different coarse particles; Traverse the rigid block rblock and convert it into a geometric crystal cluster; Delete the filled coarse particles and rigid blocks rblock, and refill the numerical model with fine particles. Group the fine particles according to the position of the geometric crystal cluster where the fine particles are located, and determine the grouping results. According to the grouping results, the required simulation parameters are set, the stress loading and crack propagation simulation operations are performed, and the simulation results are obtained.

2. The polycrystalline rock modeling method based on granular flow according to claim 1, characterized in that: The numerical model of the polycrystalline rock specifically includes: Semi-disk numerical models, disk numerical models, cylinder numerical models and cube numerical models of granite with different grain sizes.

3. A polycrystalline rock modeling system based on granular flow, characterized in that: include: Building blocks for constructing numerical models of polycrystalline rocks; The partitioning module is used to fill the numerical model with coarse particles and generate rigid blocks rblock based on the Thiessen polygon space partitioning algorithm; The method of generating a rigid block rblock based on the Thiessen polygon space partitioning algorithm specifically includes: ； in, is the Euclidean distance, For European space, Represents a Voronoi crystal cluster of Thiessen polygons, is the center point of the coarse particle, and Represents the numbers corresponding to the center points of different coarse particles; The traversal module is used to traverse the rigid block rblock and convert it into a geometric crystal cluster; The grouping module is used to delete the filled coarse particles and rigid blocks rblock and refill the numerical model with fine particles. The fine particles are grouped according to the position of the geometric crystal cluster where the fine particles are located and the grouping results are determined; The simulation module is used to set the required simulation parameters according to the grouping results, perform simulation operations of stress loading and crack propagation, and obtain simulation results.

4. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for modeling polycrystalline rock based on particle flow according to any one of claims 1 to 2 is implemented.

Images