A method for constructing a discrete element model of geocell reinforced asphalt mixture
By constructing a three-dimensional discrete element model of geocell-reinforced asphalt mixture, the problem that existing technologies cannot explain the reinforcement mechanism is solved, enabling microscopic research and efficient calculation, and characterizing the true morphological features of coarse aggregate.
Patent Information
- Application Number
- CN202310465555.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-04-27
- Publication Date
- 2025-12-16
- Estimated Expiration
- 2043-04-27
AI Technical Summary
Existing macroscopic mechanical property tests cannot explain the reinforcement mechanism of geocells on asphalt mixtures at the microscopic level, and it is difficult to study the interaction relationship between phase components through indoor mechanical property tests and overall homogenized numerical models.
CT scanning technology was used to obtain the morphological characteristics of coarse aggregate. A three-dimensional discrete element model was constructed using PFC3D software to generate rigid cluster geocells and coarse aggregate. A particle spherical element contact model was set up, and a virtual uniaxial compaction test was conducted to calibrate the microscopic parameters and establish a three-dimensional discrete element model of geocell reinforced asphalt mixture.
This study enables the exploration of the reinforcement mechanism of geocells in asphalt mixtures from a microscopic perspective, improves computational efficiency, avoids the problems of fixed coarse aggregate positions and large computational load, effectively characterizes the true morphological features of coarse aggregates, and simplifies the calculation process.
Smart Images
Figure CN116434892B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of road engineering technology, specifically relating to a method for constructing a three-dimensional discrete element model of asphalt mixture with geocell reinforcement. Background Technology
[0002] Asphalt mixtures can be viewed as a three-phase composite material consisting of coarse aggregate, asphalt mortar (a mixture of fine aggregate and asphalt), and voids. The interactions between these phases become even more complex after geocell reinforcement. The physical properties of this multiphase system are largely influenced by its individual phases, such as the shape, texture, and skeletal structure of the coarse aggregate, the relaxation characteristics, cohesion, and bond strength of the asphalt mortar, and the content, size, and distribution of voids. These internal structural characteristics are closely related to the macroscopic mechanical properties of geocell-reinforced asphalt mixtures, but are difficult to study through laboratory mechanical property tests and overall homogenized numerical models.
[0003] The Discrete Element Method (DEM), as an emerging numerical simulation method, was first applied to soil and rock materials. After the advent of the bond contact model, it began to be applied to other types of granular or cemented solid materials, especially asphalt mixtures. By constructing the internal phase components of reinforced asphalt mixtures through a granular system and based on the interparticle contact model and Newton's second law, it is possible to effectively simulate the interaction process between the internal phase components of reinforced asphalt mixtures under stress. This allows for a better understanding of the reinforcement mechanism of geocells applied to asphalt mixtures at the microstructural level. Summary of the Invention
[0004] To address the limitation of existing macroscopic mechanical property tests in explaining the reinforcement mechanism of geocells in asphalt mixtures at the microscopic level, this invention aims to provide a method for constructing a three-dimensional discrete element model of asphalt mixtures containing geocell reinforcement. This allows for the exploration of the reinforcement mechanism of geocells in asphalt mixtures at the microscopic level.
[0005] A method for constructing a three-dimensional discrete element model of asphalt mixture containing geocell reinforcement, comprising the following steps:
[0006] Step S1: Use CT scanning technology to obtain the morphological characteristics of 10 types of coarse aggregates and draw them into 3D images, then import them into PFC. 3D The software establishes a rigid cluster template library for 10 different types of coarse aggregate.
[0007] Step S2, using PFC 3DThe "generate" command in the software generates rigid cluster geocells in the reinforced area of the model space. Then, it randomly selects non-overlapping templates from the coarse aggregate rigid cluster template library constructed in step S1 to generate rigid cluster coarse aggregate with the true contour under the target gradation.
[0008] Step S3: Fill the entire model space with granular spheres of the same size in a non-overlapping manner. Identify the granular spheres that overlap with the rigid cluster geocells and rigid cluster coarse aggregates generated in Step S2 as geocells and coarse aggregates, respectively, and identify the other granular spheres as asphalt mortar.
[0009] Step S4 involves deleting the target number of asphalt mortar ball elements to simulate voids, while simultaneously deleting all rigid cluster geocells and rigid cluster coarse aggregates from step S2. The contact models between the interior of the asphalt mortar, between coarse aggregate and asphalt mortar, and between geocells and asphalt mortar in step S3 are set to parallel bond models, while the contact models between coarse aggregates and between geocells and coarse aggregates are set to linear models.
[0010] Step S5 involves conducting virtual uniaxial compaction tests and indoor uniaxial compression tests. The stress and strain curves from the virtual and indoor tests are fitted using a "back-calculation simulation" method to calibrate the micro-parameters of the parallel bond model and the linear model, thereby establishing a three-dimensional discrete element model of the geocell reinforced asphalt mixture.
[0011] Further, in step S1, the established coarse aggregate rigid cluster template library specifically involves selecting 10 different morphological characteristics of coarse aggregate with a particle size of 4.75mm to 16mm, using a CT scanner to collect their three-dimensional contour information, and drawing them into a 3D image DAT file.
[0012] Further, in step S2, the rigid cluster geocells use regularly arranged spherical particles with a particle size of 1.25 mm to simulate the actual size. Based on the mass M of the asphalt mixture specimen, the gradation of the coarse aggregate (16 mm, 13.2 mm, 9.5 mm, 4.75 mm corresponding to a1, a2, a3, a4) and density ρ, the total volume of coarse aggregate in each particle size range is calculated. The calculation formula is taken as an example for the particle size of 4.75 mm to 9.5 mm, as follows:
[0013]
[0014] The total volume V with a particle size of 4.75mm to 9.5mm 4.75mm～9.5mm To control the conditions, PFC is used. 3DThe "generate" command in the software randomly selects a template from the library built in step S1 to generate non-overlapping coarse aggregate rigid clusters with particle sizes ranging from 4.75 mm to 9.5 mm. Simultaneously, the volumes of all coarse aggregate rigid clusters within this particle size range are superimposed until a volume V is reached. 4.75mm～9.5mm Then stop generating; similarly, coarse aggregates of target mass within other particle size ranges can be generated.
[0015] Further, in step S3, the size of the particle sphere unit is determined to be 1.25 mm based on the computing power. The particle sphere units are arranged regularly to avoid overlapping and filling the entire model space; PFC is used. 3D In Fish language, a custom program was developed to determine the positional relationship between rigid cluster geocells and rigid cluster coarse aggregates and granular spheres, and to group the granular spheres into coarse aggregates, asphalt mortar, and geocells.
[0016] Furthermore, in step S4, the target number of asphalt mortar spherical elements is equal to the product of the total number of spherical elements filling the entire model space and the porosity. The main mesoscopic parameters of the linear model are the elastic modulus Emod, Poisson's ratio Kratio, and friction coefficient μ. In addition, the parallel bond model also has the parallel elastic modulus Pb_Emod, the parallel Poisson's ratio Pb_Kratio, and cohesion. and bond strength The internal contact between the geocell and the coarse aggregate is set with a parallel bond model with extremely high bond strength to ensure the integrity of its own structure during the stress process.
[0017] Further, in step S5, the indoor uniaxial compression test specifically involves preparing geocell-reinforced asphalt mixture cubic specimens (100mm × 100mm × 100mm) using the static pressure method. Unconfined uniaxial compression is then applied to the specimens using an MTS testing machine at a compression rate of 50mm / min and a test temperature of 20℃. Load and displacement data during the compression process are collected and converted into stress-strain curves. The virtual uniaxial compression test utilizes a PFC (Polymerized Fiber Optic Compressor). 3DThe discrete element method (DEM) simulation software is used to import the virtual geocell reinforced asphalt mixture specimen obtained in step S4. The parameters are calibrated using the "back-calculation exercise" method in step S5. Specifically, the micro-parameters of the linear model and parallel bond model in step S5 are first assumed based on experience. Then, two walls are generated at the top and bottom of the virtual specimen. The bottom wall is fixed, and the top wall is moved downwards at a rate of 50 mm / min to apply a vertical load to the virtual specimen for uniaxial compression. At the same time, the virtual stress-strain curve is monitored. If the virtual stress-strain curve couples with the stress-strain curve in the indoor test, the assumed parameters are used as the final micro-parameters of the linear model and parallel bond model. If the two curves differ significantly, the assumed parameters are adjusted, and the compression process is repeated until all parameters are calibrated.
[0018] The beneficial effects of this invention are as follows: On the one hand, the method of constructing flexible cluster coarse aggregate can not only be randomly generated at any position inside the model, avoiding the disadvantages of fixed position and single shape of coarse aggregate in virtual specimens caused by direct CT scanning, but also fully characterize the real morphological features of coarse aggregate, simplify the huge amount of calculation caused by the superposition of a large number of unit particles in rigid cluster coarse aggregate, and improve the calculation efficiency; on the other hand, this invention regards geocell as a special coarse aggregate to construct flexible cluster geocell, which can effectively provide reinforcement to asphalt mixture, avoid the large unbalanced force that rigid cluster geocell is prone to form inside asphalt mixture, and facilitate the exploration of the reinforcement mechanism of geocell to asphalt mixture from the microscopic level. Attached Figure Description
[0019] Figure 1 Schematic diagrams of coarse aggregates in different real-world forms;
[0020] Figure 2 Image showing the effect of collecting 3D contour information of coarse aggregate;
[0021] Figure 3 A three-dimensional discrete element rigid cluster simulation of a real-form coarse aggregate;
[0022] Figure 4 Rendering of a three-dimensional discrete element rigid cluster simulation of a geocell;
[0023] Figure 5 Rendering of the three-dimensional discrete element rigid cluster simulation of coarse aggregate in a virtual specimen;
[0024] Figure 6 Renderings of three-dimensional discrete element flexible cluster simulation of coarse aggregate, geocells, and asphalt mortar in a virtual specimen;
[0025] Figure 7 Schematic diagram of internal micro-contact of a three-dimensional discrete element model of geocell-reinforced asphalt mixture;
[0026] Figure 8Schematic diagram of components composed of linear contact model and parallel bonding model;
[0027] Figure 9 Virtual uniaxial compression test of three-dimensional discrete element model of geocell-reinforced asphalt mixture;
[0028] Figure 10 Comparison of stress-strain curves between virtual uniaxial compression test and indoor uniaxial compression test of geocell reinforced asphalt mixture. Detailed Implementation
[0029] The specific embodiments of the present invention will now be described with reference to the accompanying drawings.
[0030] This embodiment describes a method for constructing a three-dimensional discrete element model of asphalt mixture containing geocell reinforcement, including the following steps:
[0031] Step S1: Select 10 coarse aggregates with different morphological characteristics within the particle size range of 4.75mm to 16mm, such as... Figure 1 As shown, its three-dimensional contour information was then acquired using a CT scanner and drawn into a 3D image DAT file, as shown. Figure 2 As shown, finally import it into PFC. 3D The software includes a library of rigid cluster templates for 10 different types of coarse aggregates, such as... Figure 3 As shown.
[0032] Step S2, using PFC 3D The "generate" command in the software generates rigid clusters of geocells in the reinforced region of the model space. Regularly arranged spherical particles with a particle size of 1.25 mm are used to simulate the actual size. Figure 4 As shown, based on the mass M of the asphalt mixture specimen, the gradation of the coarse aggregate (16mm, 13.2mm, 9.5mm, 4.75mm corresponding to a1, a2, a3, a4) and density ρ, the total volume of coarse aggregate in each particle size range is calculated; taking the particle size of 4.75mm to 9.5mm as an example, the formula is as follows:
[0033]
[0034] The total volume V with a particle size of 4.75mm to 9.5mm 4.75mm～9.5mm To control the conditions, PFC is used. 3D The "generate" command in the software randomly selects a template from the library built in step S1 to generate non-overlapping coarse aggregate rigid clusters with particle sizes ranging from 4.75mm to 9.5mm. A custom program developed using the Fish language simultaneously generates these coarse aggregate rigid clusters and superimposes the volumes of all coarse aggregate rigid clusters within that particle size range until a volume V is reached. 4.75mm～9.5mmGeneration was then stopped; with a gradation type of AC-13 (16mm, 13.2mm, 9.5mm, 4.75mm corresponding to 100%, 96.2%, 77.2%, 54.1% respectively), a mass of 2491g, and a basalt coarse aggregate density of 3.1g / cm³, the reaction proceeded. 3 Taking a cubic asphalt mixture sample with dimensions of 100mm×100mm×100mm as an example, two other types of coarse aggregate rigid clusters with particle sizes of 9.5mm~13.2mm and 13.2mm~16mm are generated in the model space, such as... Figure 5 As shown.
[0035] Step S3, the asphalt mixture sample size is V 试件 If the radius of the sphere element is r, then the number N of sphere elements required to fill the entire model space is as follows:
[0036]
[0037] If we assume the radius of the spherical element is 1.18 mm, then according to the above calculation, the number of spherical elements needed to fill the entire model space is close to 125,000. This exceeds the computing power of a typical computer after applying contact modeling, resulting in low computational efficiency. Therefore, assuming the radius of the spherical element is 1.25 mm, the required number of spherical elements is only 64,000, which basically meets the computational requirements. After filling the entire model space with spherical elements of the same size in a non-overlapping manner...
[0038] The model space is filled with non-overlapping granular spherical elements of the same size as cubes (100mm × 100mm × 100mm). Granular spherical elements overlapping with the rigid cluster geocells and rigid cluster coarse aggregates generated in step S2 are identified as geocells and coarse aggregates, respectively; other granular spherical elements are identified as asphalt mortar. A program edited in Fish language determines whether the positions of the rigid cluster geocells and rigid cluster coarse aggregates overlap with the granular spherical elements in the model space. Granular spherical elements overlapping with the rigid cluster geocells and rigid cluster coarse aggregates are identified as flexible cluster geocells and flexible cluster coarse aggregates, respectively; non-overlapping granular spherical elements are identified as asphalt mortar. Figure 6 As shown.
[0039] Step S4: Based on the porosity of the laboratory-molded asphalt mixture specimen being 5.5%, the number of asphalt mortar particle sphere units to be deleted when simulating porosity is 64000 × 5.5%. After deleting the asphalt mortar particle sphere units, all rigid cluster geocells and rigid cluster coarse aggregates from step S2 are also deleted. Then, the contact models between coarse aggregates and between geocells and coarse aggregates in step S3 are set as linear models, and the contact models between coarse aggregates and asphalt mortar and between geocells and asphalt mortar are set as parallel bond models, such as... Figure 7 The main mesoscopic parameters of the linear model are the elastic modulus Emod, Poisson's ratio Kratio, and the coefficient of friction μ. The main mesoscopic parameters of the parallel bond model, in addition to those of the linear model, include the parallel elastic modulus Pb_Emod, the parallel Poisson's ratio Pb_Kratio, and the cohesion. and bond strength Among them, Emod, Pb_Emod, Kratio, and Pb_Kratio can be related to the normal stiffness k. n , and tangential stiffness k s , The conversion between them is shown in the following formulas, where A and L are the contact area and length, respectively. Linear model and parallel bonding model, as follows: Figure 8 As shown. The internal contact setting of the geocell and coarse aggregate establishes the bond strength. A large parallel bond model is used to ensure the integrity of its own structure during the stress process.
[0040]
[0041] Step S5: First, conduct an indoor uniaxial compression test. Based on the gradation of AC-13 asphalt mixture in Table 1, prepare 100mm×100mm×100mm geocell reinforced asphalt mixture cube specimens using the static compression method. Perform unconfined uniaxial compression on the specimens using an MTS testing machine at a compression rate of 50mm / min and a test temperature of 20℃. Collect the load and displacement data during the compression process and convert them into stress-strain curves, as shown below. Figure 9 .
[0042] Table 1. Passing Rate of AC-13 Asphalt Mixtures at Various Sieve Sizes
[0043] Sieve aperture diameter (mm) 16 13.2 9.5 4.75 2.36 1.18 0.6 0.3 0.15 0.075 Pass rate (%) 100 96.2 77.2 54.1 37.8 27.5 19.1 14.6 10.3 6.1
[0044] The virtual uniaxial compression test utilizes PFC3D discrete element simulation software. The virtual geocell-reinforced asphalt mixture specimen obtained in step S4 is imported. A "back-calculation exercise" approach is adopted. First, based on empirical assumptions, the mesoscopic parameters of the linear model and parallel bond model from step S5 are used as estimated values. Then, two walls are generated at the top and bottom of the virtual specimen. The bottom wall is fixed, and the top wall is moved downwards at a rate of 50 mm / min to apply a vertical load for uniaxial compression. Figure 9As shown, the virtual stress-strain curve is simultaneously monitored to match the stress-strain curve from the laboratory test. If the two curves differ significantly, the estimated adjustment parameters are returned, and the compression process is repeated. Finally, the final values of the micro-parameters that best match the stress-strain curve from the laboratory test are determined. The adjustment parameters are shown in Table 2. The uniaxial compression stress-strain curve of the discrete element model of geocell-reinforced asphalt mixture under the final values of the micro-parameters are shown in Table 2. Figure 10 As shown.
[0045] Table 2 Results of Microscopic Parameter Adjustment for Contact Model
[0046]
[0047] This specification has described the invention in detail with general descriptions and specific embodiments. However, some modifications or refinements can be made to the invention, and these modifications and refinements should also be considered within the scope of protection of the invention.
Claims
1. A method for constructing a three-dimensional discrete element model of asphalt mixture reinforced with geocell, characterized in that, Specifically comprising the following steps: Step S1, using CT scanning technology to obtain the morphological characteristics of 10 kinds of coarse aggregate and draw 3D image, then import PFC3D software to establish the rigid cluster template library of 10 kinds of different morphological coarse aggregate; Step S2, using the "generate" command in PFC3D software to generate rigid cluster geocell in the reinforced area of the model space, and then randomly select the template in the rigid cluster template library of coarse aggregate established in step S1 to generate the real outline of the rigid cluster coarse aggregate with non-overlapping target gradation; Step S3, fill the entire model space with the same size of particle ball unit in a non-overlapping manner, identify the particle ball unit overlapping with the rigid cluster geocell and the rigid cluster coarse aggregate generated in step S2 as geocell and coarse aggregate respectively, and identify the other particle ball unit as asphalt mortar; Step S4, by deleting the target number of asphalt mortar ball unit to simulate the void, at the same time deleting all the rigid cluster geocell and the rigid cluster coarse aggregate in step S2, and setting the contact model between the asphalt mortar inside, the coarse aggregate and the asphalt mortar, the geocell and the asphalt mortar in step S3 as parallel bond model, and the coarse aggregate and the coarse aggregate, the geocell and the coarse aggregate as linear model; Step S5, carry out virtual uniaxial compaction test and indoor uniaxial compression test, adopt the "back calculation exercise" mode to fit the stress and strain curves in virtual test and indoor test, to calibrate the mesoscopic parameters of parallel bond model and linear model, and to establish the three-dimensional discrete element model of geocell reinforced asphalt mixture; The target number of asphalt mortar ball units in step S4 is equal to the product of the total number of particle ball units in the entire model space and the void ratio; the mesoscopic parameters of the linear model are the elastic modulus , Poisson's ratio , and friction coefficient µ; in addition to the above, the parallel bond model has parallel elastic modulus , parallel Poisson's ratio , cohesion , and bond strength ; the internal contact of the geocell and the coarse aggregate is provided with a parallel bond model with great bond strength, so as to ensure the integrity of its own structure during the stress process; In step S5, the indoor uniaxial compression test is performed. Specifically, a static pressure method is used to prepare a geocell reinforced asphalt mixture cubic test piece with a size of 100 mm x 100 mm x 100 mm. An MTS testing machine is used to perform unconfined uniaxial compression on the test piece at a compression rate of 50 mm / min and a test temperature of 20°C. The load and displacement during compression are collected and converted into a stress-strain curve. The virtual uniaxial compression test uses PFC 3D discrete element simulation software to import the virtual geocell reinforced asphalt mixture test piece obtained in step S4. The parameters are calibrated in the manner of "back calculation exercise" in step S5. Specifically, the mesoscopic parameters of the linear model and the parallel bond model in step S5 are first assumed based on experience. Then, two walls are generated at the top and bottom of the virtual test piece. The bottom wall is fixed, and the top wall is moved downward at a rate of 50 mm / min to apply a vertical load to the virtual test piece to perform uniaxial compression. At the same time, the virtual stress-strain curve is monitored. If the virtual stress-strain curve is coupled with the stress-strain curve during the indoor test, the above-mentioned assumed parameters are used as the final mesoscopic parameters of the linear model and the parallel bond model. If the two curves do not match, the assumed parameters are adjusted and the compression process is repeated until all parameters are calibrated.
2. The method according to claim 1, wherein, In step S1, the established rigid cluster template library of coarse aggregate is specifically that, 10 kinds of coarse aggregate with different morphological characteristics and particle size of 4.75mm~16mm are selected, the three-dimensional contour information of which is collected by using CT scanner, and 3D image DAT file is drawn.
3. The method according to claim 1, wherein, In step S2, the rigid cluster geocell is simulated by particle ball units with a particle size of 1.25 mm and regular arrangement; according to the mass M of the asphalt mixture test piece, the gradation of coarse aggregate: 16 mm, 13.2 mm, 9.5 mm and 4.75 mm respectively correspond to , the density p, the total volume of coarse aggregate in each particle size range is calculated; the calculation formula is taken as an example for the particle size of 4.75 mm~9.5 mm, as follows: ； Taking the total volume V4.75mm~9.5mm of 4.75mm~9.5mm particle size as the control condition, using the "generate" command in PFC3D software, the template in the library established in step S1 is randomly selected, and the rigid cluster of coarse aggregate with particle size of 4.75mm~9.5mm is generated without overlapping, at the same time the volume of all the rigid clusters of coarse aggregate in this particle size range is superimposed, and the generation is stopped until V4.75mm~9.5mm is reached, and the target mass of coarse aggregate in other particle size range can be generated in the same way.
4. The method according to claim 1, wherein, In step S3, the size of the particle ball unit is determined as 1.25 mm according to the calculation capacity, and the particle ball units are arranged regularly to fill the entire model space without overlapping. A self-defined program is developed in Fish language in PFC 3D to determine the positional relationship between the rigid cluster geocell, the rigid cluster coarse aggregate, and the particle ball unit, group the particle ball units, and identify them as coarse aggregate, asphalt mortar, and geocell.