A new type of bonded particle breakage rapid modeling method based on discrete elements and application thereof
By establishing a geometric model of a single particle shape and a template of an unbreakable cluster of shaped particles, a model of a set of bonded particles is generated, which solves the problems of complex modeling of the breakage of bonded particles and inconsistent particle size, and realizes rapid and accurate particle size distribution identification and breakage analysis.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CENT SOUTH UNIV
- Filing Date
- 2023-09-18
- Publication Date
- 2026-06-26
AI Technical Summary
Existing methods for crushing bonded particles are complex to model, requiring the pre-fabrication of bonded particle templates. Furthermore, the particle sizes of sub-particles of the same particle size range are inconsistent, increasing the material's discreteness and affecting the calibration of crushing strength.
Establish a geometric model of a single particle shape, generate an indestructible cluster shape particle template, generate a model of a bonded particle set through the particle surface geometric model, write an algorithm to identify fragments after particle breakage, and perform quantitative analysis.
It can quickly establish particle aggregates that conform to the real shape and gradation, with consistent particle size that does not increase with particle size, high recognition accuracy, avoid particle loss and misjudgment, and accurately identify the evolution of particle gradation during loading.
Smart Images
Figure CN117236151B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of geotechnical engineering technology, and in particular relates to a novel rapid modeling method for the breakage of bonded particles based on discrete element method. Background Technology
[0002] Granular materials are widely used in geotechnical engineering fields such as roadbeds and ballasted track beds due to their low cost and ease of maintenance. Under load, particles break down, altering their particle size distribution and affecting the macroscopic mechanical properties of granular materials. Discrete element method (DEM) is widely used in simulating the breakage of granular materials due to its unique advantages. Existing particle breakage methods mainly include the substitution method and the bonded particle method. The substitution method is heavily influenced by subjective factors and cannot maintain mass conservation, thus having certain drawbacks. In contrast, the bonded particle method is less affected by subjective factors and can reflect the mechanical response under complex loading paths. However, existing bonded particle breakage models are relatively complex, requiring the pre-fabrication of bonded particle templates, and the inconsistent particle sizes of sub-particles of the same particle size range increase the discretization of the material and limit the calibration of the breakage strength of individual particles. Therefore, it is necessary to establish a new rapid modeling method for bonded particle breakage. Summary of the Invention
[0003] To address the shortcomings of existing technologies, this invention discloses a novel rapid modeling method for the breakage of bonded particles based on the discrete element method, the technical solution of which is as follows:
[0004] A novel rapid modeling method for the breakage of cohesive particles based on discrete element method is characterized by:
[0005] S1, Establish a single-particle shape geometric model: Construct a three-dimensional surface morphology model of a single particle based on the single particle used in the indoor test, and construct a single-particle shape geometric model based on the three-dimensional surface morphology model of the single particle.
[0006] S2, Establish a discrete element numerical model for the breakage of particle aggregates: Based on the constructed single particle shape geometric model, generate an unbreakable cluster shape particle template, then generate an unbreakable cluster particle aggregate, transform each unbreakable cluster particle in the unbreakable cluster particle aggregate into a particle surface geometric model, and finally generate a bonded particle set model from the particle surface geometric model to perform numerical simulation of the breakage of bonded particles.
[0007] S3, Obtain the fragment distribution after particle breakage: Write an algorithm to identify the fragments produced after particle breakage and perform quantitative analysis on the size of the fragments.
[0008] This invention also discloses a novel rapid modeling device for the breakage of cohesive particles based on discrete element method, characterized in that:
[0009] Establish a single-particle shape geometric model module: Construct a three-dimensional surface morphology model of a single particle based on the single particle used in the indoor test, and construct a single-particle shape geometric model based on the three-dimensional surface morphology model of the single particle;
[0010] Establish a discrete element numerical model module for particle aggregate breakage: Based on the constructed single particle shape geometric model, generate an indestructible cluster shape particle template, then generate an indestructible cluster particle aggregate, transform each indestructible cluster particle in the indestructible cluster particle aggregate into a particle surface geometric model, and finally generate a bonded particle set model from the particle surface geometric model to perform numerical simulation of bonded particle breakage.
[0011] The module for obtaining the fragment distribution after particle breakage identifies the fragments produced after particle breakage and performs quantitative analysis on the size of the fragments.
[0012] The present invention also discloses a non-volatile storage medium, characterized in that the non-volatile storage medium includes a stored program, wherein the program, when running, controls the device where the non-volatile storage medium is located to execute the above-mentioned novel rapid modeling method for the breakage of cohesive particles based on discrete element method.
[0013] Beneficial effects
[0014] (1) By importing particle templates of different shapes, an unbreakable cluster model is first generated, and then a particle bond is generated at the original position of the unbreakable cluster model. The particle bond inherits the size and shape of the original unbreakable cluster model. This method can greatly speed up the creation of particle bonds that conform to the real shape and gradation, and the size of sub-particles of the same particle size can be consistent and no longer increases with the increase of particle size.
[0015] (2) The particle size distribution identification algorithm developed in this invention has higher identification accuracy than the algorithm built into the particle flow discrete element software. It will not cause particle loss or misjudgment and can accurately identify the particle size distribution evolution process during loading. The algorithm idea can be extended to other discrete element software. Attached Figure Description
[0016] Figure 1 Schematic diagram of particle shape template;
[0017] Figure 2 Schematic diagram of an indestructible cluster assembly (three particle shape templates);
[0018] Figure 3 A schematic diagram of the geometric model of the particle surface (including templates for three particle shapes);
[0019] Figure 4 Schematic diagram of granular binder;
[0020] Figure 5 Schematic diagram of particle-bonded body contact;
[0021] Figure 6 Particle crushing and gradation changes diagram. Detailed Implementation
[0022] A novel rapid modeling method for the breakage of cohesive particles based on discrete element method (DEM) is characterized by the following steps:
[0023] S1. Establish a geometric model of a single particle shape. Based on the single particle used in the indoor experiment, construct a three-dimensional surface morphology model of the single particle, and then construct a geometric model of the single particle shape based on the three-dimensional surface morphology model. The three-dimensional surface morphology of the single particle can be obtained through three-dimensional scanning. Constructing the geometric model of the single particle shape includes the following steps:
[0024] S11. Based on the single-particle three-dimensional surface morphology model (STL file format) obtained from indoor experiments, the single-particle three-dimensional surface morphology model is imported into the discrete element method software to form a single-particle shape geometric model, and this model is assigned a group name. Different particle shapes are processed sequentially until all particle shape geometric models are established, and different particle shape geometric models have different group names.
[0025] S2. Establish a discrete element numerical model for the breakage of particle assemblies. Based on the constructed single-particle shape geometric model, generate an indestructible cluster shape particle template, then generate an indestructible cluster particle assembly, transform each indestructible cluster particle in the indestructible cluster particle assembly into a particle surface geometric model, and finally generate a bonded particle set model from the particle surface geometric model to perform numerical simulation of bonded particle breakage;
[0026] The bonded particle set model is obtained by generating sub-particles inside the particle surface geometry model, and the volume of a single sub-particle should be smaller than the volume of the particle surface geometry model.
[0027] Establishing the model of the bonded particle set includes the following steps:
[0028] S21, Establish the sample model. The sample model is established, and the mass ratio m of each grade can be determined from the experimental gradation curve. i Based on the total mass m and density ρ of the material, calculate the number of particles n corresponding to each particle size using the following formula. i :
[0029]
[0030] In the formula: D i D represents the lower limit particle diameter for this gradation. i This represents the upper limit of particle diameter for this gradation.
[0031] S22, Generate Indestructible Cluster Particles. Based on the calculated number of particles corresponding to each particle size, generate indestructible cluster particles of the corresponding size inside the sample, and assign the same group name to cluster particles of the same size. Iterate the calculations on the model until the overall unbalanced force ratio of the model is less than 1e-3, at which point the model is considered to have reached equilibrium.
[0032] S23, Export the particle surface geometry model of the cluster particles: Export the particle surface geometry model of each unbreakable cluster particle using the built-in command of the discrete element software, and assign the group name of the cluster particle to the particle surface geometry model. After all unbreakable cluster particles have been exported, delete all cluster particles; Convert the particle surface geometry model into a wall that can restrict particle displacement using the built-in command of the software. At this point, there will be a particle surface geometry model and a wall of the same size and shape at the original position of each unbreakable cluster particle.
[0033] S24, Generate breakable sub-particles: Traverse all particle surface geometry models, and generate sub-particles within each particle surface geometry model according to a target porosity of 0.3 to 0.36. The target porosity is the ratio of the total volume of the sub-particles to the volume of the particle surface geometry model. Multiple sub-particles can be of equal or unequal size. Sub-particles within particle surface geometry models of the same size have the same size. The group name of the sub-particles is the same as that of the particle surface geometry model, and a unique sequence ID number is assigned to each sub-particle within the particle surface geometry model. The sequence ID numbers of sub-particles within different particle surface geometry models are all different. The model is calculated and iterated again until the overall unbalanced force ratio of the model is less than 1e-3, at which point the model can be considered to have reached a preliminary equilibrium state.
[0034] S25, Identify contact features: First, delete all walls derived from unbreakable cluster particles. Then, traverse the contacts and distinguish all contacts. When the sequence ID numbers of the particles at both ends of the contact are the same, it indicates that the contact is between small particles within the same large particle. At the same time, determine the group name of the particles and group the contact according to the group name of the particles at both ends. When the sequence ID numbers of the particles at both ends of the contact are different, it indicates that the contact is between large particles, i.e., the interface contact of large particles. At this point, all contact grouping is completed. Contacts within the same particle surface geometry model are intraparticle contacts, and contacts between different particle surface geometry models are interface contacts.
[0035] S26, assign microscopic contact parameter properties to generate a breakable aggregate. Different microscopic contact parameters are assigned to contacts with different contact group names based on the contact group name. Intragranular contacts are bonded contacts and are assigned bonding parameters, while interfacial contacts are non-bonded contacts and are assigned non-bonding parameters.
[0036] S3. Obtain the fragment distribution after particle breakage. Develop an algorithm to identify the fragments produced after particle breakage and perform quantitative analysis on the fragment size.
[0037] S31, Store particle pointers. Traverse all particles, find particles with the same sequence ID number, and store the pointers of this batch of particles in one memory location; find the next batch of particles with the same sequence ID number, and store the pointers of this batch of particles in the next memory location. Continue until all particles have been traversed and all particle pointers have been allocated in their respective memory locations.
[0038] S32 identifies single bonded particles. First, it identifies the first particle a within the first memory segment. i The pointer is used for identification, and the memory is traversed to find the first particle 'a'. i Adhesive particles, then particle a i and a i The attribute 2 of the bonded particle is assigned the value "sequence ID + 1", and the internal memory particle a is removed. i and a i The pointers of the bonded particles; iterate through the memory again for particles bonded to "sequence ID + 1" with attribute 2, and continue to assign attribute 2 of the bonded particles to "sequence ID + 1", while removing the pointers of the bonded particles in the memory, until the number of particles with attribute 2 "sequence ID + 1" no longer changes; continue for the first particle b remaining in the first memory. i To identify the memory, iterate through the remaining memory chips and find the one that matches the first chip b. i The bonded particles, then particle b i and in that memory with b i The bonded particle's attribute 2 is assigned the value "sequence ID + 2". The memory is then iterated through again for particles bonded to "sequence ID + 2" with attribute 2, and the attribute 2 of these bonded particles is reassigned as "sequence ID + 2". Simultaneously, the pointers to these bonded particles in the memory are removed until the number of particles with attribute 2 "sequence ID + 2" no longer changes. This process is repeated for the remaining particles in the memory until no more particle pointers are present. The value of "sequence ID + n" at this point is recorded. id ".
[0039] S33, Identify all bonded particles. Identify the particles inside the second memory module, repeating step S32; continue identifying the particles inside the next memory module, repeating step S32, until all particles inside the memory module have been identified.
[0040] S34, calculate the sample gradation. Particles with the same particle attribute 2 are considered bonded particles. Iterate through the particles starting from "first sequence ID number + m", where m is from 1 to n. idCalculate the total volume and equivalent diameter (the diameter of a sphere with the same total volume) of the particles with the first sequence ID number + m, and add this total accumulation to the corresponding gradation range according to the particle size range. Then repeat operation S34 to traverse the particles with the second sequence ID number + m, and continue to add the total accumulation to the corresponding gradation range, until all sequence ID numbers have been traversed and the gradation curve is recorded.
[0041] Example
[0042] Combination Figure 1-6 The following example and parameters illustrate a novel rapid modeling method for the breakage of bonded particles according to the present invention. It includes the following steps:
[0043] 1. See Figure 1 Taking three single-particle three-dimensional surface morphology models as an example, the single-particle three-dimensional surface morphology model (STL file) is imported into PFC using the command "geometryimport" to construct the corresponding three single-particle shape geometry models. Then, the shape template is imported into the indestructible cluster template using the software's built-in command "clump template create".
[0044] 2. See Figure 2 Based on the total mass and gradation curve, the number of particles in each grade is calculated by S21, and then the corresponding number of indestructible cluster models are generated and iterated continuously until the unbalanced stress ratio is less than 1e-3.
[0045] 3. See also Figure 3 The software uses the built-in command "clump export geometry" to export the particle surface geometry model of each unbreakable cluster and assigns it a different group name. At the same time, the software uses the built-in command "wall import from-geometry" to convert the particle surface geometry model into a wall that restricts particle displacement. After the particle surface geometry models of all unbreakable clusters are exported, all unbreakable clusters are deleted. At the original location of each unbreakable cluster, there will be a particle surface geometry model and a wall of the same shape and size.
[0046] 4. See Figure 4Sub-particles are generated on each particle surface geometry model. The particle size of sub-particles in each surface geometry part can be set by the user. Unlike the existing methods for breaking up bonded particles (where the particle size of sub-particles increases with the particle size of the parent particle), this paper takes the case where the particle size of sub-particles in different particle surface geometry models is the same. Sub-particles are generated in each particle surface geometry model according to a specified porosity of 0.36. The group name of the sub-particles is the same as the group name of the particle surface geometry model. Then, the process is repeated until the equilibrium state is reached. Sub-particles that have escaped from the particle surface geometry model during the process are deleted, and the process is repeated until the equilibrium state is reached again.
[0047] 5. Delete all walls derived from the particle surface geometry model. Refer to S25 and S26 to bond particles within the same group name; the bonding parameters can be set manually. Contact between particles from different groups is linear and not adhesive. Figure 5 This concludes the method for rapid modeling of the breakage of bonded particles.
[0048] 6. Load the model and monitor the gradation. Apply a downward velocity to the upper wall and an upward velocity to the lower wall to load the model. Then, identify the gradation using algorithms S31-S34. The final particle breakage and gradation changes are as follows: Figure 6 As shown.
[0049] This invention addresses the problems of complex, time-consuming, and highly discrete modeling of bonded particle breakage. Compared to traditional methods, this method can more quickly establish particle aggregates with different particle shapes and sizes, eliminates the need for pre-fabricated bonded particle templates, and allows for the same particle size range for sub-particles of the same size, without increasing with particle size. It also facilitates the calibration of microscopic mechanical parameters, providing an effective technical means for further understanding the microscopic mechanical mechanisms of particle breakage and the macroscopic mechanical behavior of breakable particulate materials.
[0050] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.
Claims
1. A novel rapid modeling method for the breakage of cohesive particles based on discrete element method, characterized in that: S1, Establish a single-particle shape geometric model: Construct a three-dimensional surface morphology model of a single particle based on the single particle used in the indoor test, and construct a single-particle shape geometric model based on the three-dimensional surface morphology model of the single particle. S2, Establish a discrete element numerical model for the breakage of particle aggregates: Based on the constructed single particle shape geometric model, generate an unbreakable cluster shape particle template, then generate an unbreakable cluster particle aggregate, transform each unbreakable cluster particle in the unbreakable cluster particle aggregate into a particle surface geometric model, and finally generate a bonded particle set model from the particle surface geometric model to perform numerical simulation of the breakage of bonded particles. S21, Establish the sample model: Based on the experimental gradation curve, the mass ratio m of each grade can be determined. i Based on the total mass m and density of the material The number of particles n corresponding to each particle size is calculated using the following formula. i : ； In the formula: This is the lower limit particle diameter for this gradation. This represents the upper limit of particle diameter for this grade of gradation; S22, Generate indestructible cluster particles: Based on the calculated number of particles corresponding to each particle size, generate indestructible cluster particles of the corresponding size inside the sample, and assign the same group name to the cluster particles of the same particle size; perform calculation iterations on the model until the overall unbalanced force ratio of the model is less than 1e-3, then the model can be considered to have reached the equilibrium state. S23, Export the particle surface geometry model of the cluster particles: Export the particle surface geometry model of each unbreakable cluster particle using the built-in command of the discrete element software, and assign the group name of the cluster particle to the particle surface geometry model. After all unbreakable cluster particles have been exported, delete all cluster particles; Convert the particle surface geometry model into a wall that can restrict particle displacement using the built-in command of the software. At this point, there will be a particle surface geometry model and a wall of the same size and shape at the original position of each unbreakable cluster particle. S24, Generate breakable sub-particles: Traverse all particle surface geometry models, and generate sub-particles within each particle surface geometry model according to a target porosity of 0.3 to 0.
36. The target porosity is the ratio of the total volume of the sub-particles to the volume of the particle surface geometry model. Multiple sub-particles can be of equal or unequal size. Sub-particles within particle surface geometry models of the same size have the same size. The group name of the sub-particles is the same as that of the particle surface geometry model, and a unique sequence ID number is assigned to each sub-particle within the particle surface geometry model. The sequence ID numbers of sub-particles within different particle surface geometry models are all different. The model is calculated and iterated again until the overall unbalanced force ratio of the model is less than 1e-3, at which point the model can be considered to have reached a preliminary equilibrium state. S25, Identify contact features: First, delete all walls derived from unbreakable cluster particles. Then, traverse the contacts and distinguish all contacts. When the sequence ID numbers of the particles at both ends of the contact are the same, it indicates that the contact is between small particles within the same large particle. At the same time, determine the group name of the particles and group the contact according to the group name of the particles at both ends. When the sequence ID numbers of the particles at both ends of the contact are different, it indicates that the contact is between large particles, i.e., the interface contact of large particles. At this point, all contact grouping is completed. Contacts within the same particle surface geometry model are intraparticle contacts, and contacts between different particle surface geometry models are interface contacts. S26, Assign microscopic contact parameter properties to generate a breakable aggregate. Based on the contact group name, assign different microscopic contact parameters to contacts with different contact group names. Intragranular contacts are bonded contacts and are assigned bonding parameters, while interfacial contacts are non-bonded contacts and are assigned non-bonded parameters. S3, Obtain the fragment distribution after particle breakage: Identify the fragments produced after particle breakage and perform quantitative analysis on the size of the fragments; S31, Store particle pointers: Traverse all particles, find particles with the same sequence ID number, and store the pointers of this batch of particles in one memory; find the next batch of particles with the same sequence ID number, and store the pointers of this batch of particles in the next memory, until all particles have been traversed and the particle pointers are allocated in the corresponding memory. S32, Identify a single bonded particle: First, identify the first particle a inside the first memory. i The pointer is used for identification, and the memory is traversed to find the first particle 'a'. i Adhesive particles, then particle a i and a i The 2nd attribute of the bonded particle is assigned the value "sequence ID + 1", and the internal memory particle a is removed. i and a i A pointer to bonded particles; Iterate through the memory again to find the particles in the memory that are bound to "sequence ID + 1" with attribute 2. Continue assigning attribute 2 of the bound particles to "sequence ID + 1" while removing the pointers to the bound particles in the memory, until the number of particles with attribute 2 "sequence ID + 1" no longer changes. Then continue iterating through the remaining first particle b in the first memory location. i To identify the memory, iterate through the remaining memory chips and find the one that matches the first chip b. i The bonded particles, then particle b i and in that memory with b i The bonded particle's attribute 2 is assigned the value "sequence ID + 2". The memory is then iterated through again for particles bonded to "sequence ID + 2" with attribute 2, and the attribute 2 of these bonded particles is reassigned as "sequence ID + 2". Simultaneously, the pointers to these bonded particles in the memory are removed until the number of particles with attribute 2 "sequence ID + 2" no longer changes. This process is repeated for the remaining particles in the memory until no more particle pointers are present. The value of "sequence ID + n" at this point is recorded. id ”; S33, Identify all bonded particles: Identify the particles inside the second memory module, repeating the steps in S32; continue identifying the particles inside the next memory module until all particles inside the memory module have been identified; S34, Calculate the sample gradation: Particles with the same particle attribute 2 are bonded particles. Iterate through the particles from "first sequence ID number + m", where m is from 1 to n. id Calculate the total volume and equivalent diameter of the particles with the first sequence ID number + m, and add the total accumulation to the corresponding gradation range according to the particle size range; then repeat operation S34 to traverse the particles with the second sequence ID number + m, and continue to add the total accumulation to the corresponding gradation range, until all sequence ID numbers have been traversed and the gradation curve is recorded.
2. The novel rapid modeling method for the breakage of cohesive particles based on discrete element method according to claim 1, characterized in that: In step S1, the single-particle three-dimensional surface morphology model is obtained through three-dimensional scanning.
3. The novel rapid modeling method for the breakage of cohesive particles based on discrete element method according to claim 1, characterized in that: In step S1, the establishment of the geometric model of a single particle shape includes the following steps: S11. Based on the single-particle three-dimensional surface morphology model obtained from indoor experiments, the three-dimensional surface morphology model is imported into the discrete element software to form a single-particle shape geometric model, and the model group name is assigned; different shaped particles are operated in sequence until the single-particle shape geometric models of all shapes are established, and different single-particle shape geometric models have different group names.
4. The novel rapid modeling method for the breakage of cohesive particles based on discrete element method according to claim 1, characterized in that: In step S2, the bonded particle set model is obtained by generating sub-particles inside the particle surface geometry model, and the volume of a single sub-particle should be smaller than the volume of the corresponding particle surface geometry model.
5. The novel rapid modeling method for the breakage of cohesive particles based on discrete element method according to claim 1, characterized in that: Multiple sub-particles are of equal size; the total number of the sub-particles is 50 to 2000.
6. A novel rapid modeling device for the crushing of cohesive particles based on discrete element method, the device executing the novel rapid modeling method for the crushing of cohesive particles based on discrete element method as described in claim 1, characterized in that: Establish a single-particle shape geometric model module: Construct a three-dimensional surface morphology model of a single particle based on the single particle used in the indoor test, and construct a single-particle shape geometric model based on the three-dimensional surface morphology model of the single particle; Establish a discrete element numerical model module for particle aggregate breakage: Based on the constructed single particle shape geometric model, generate an indestructible cluster shape particle template, then generate an indestructible cluster particle aggregate, transform each indestructible cluster particle in the indestructible cluster particle aggregate into a particle surface geometric model, and finally generate a bonded particle set model from the particle surface geometric model to perform numerical simulation of bonded particle breakage. The module for obtaining the fragment distribution after particle breakage identifies the fragments produced after particle breakage and performs quantitative analysis on the size of the fragments.
7. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored program, wherein the program, when running, controls the device where the non-volatile storage medium is located to execute the novel rapid modeling method for the breakage of cohesive particles based on discrete element method as described in any one of claims 1 to 5.