Construction method and device of porous asphalt concrete finite element model and application of construction method and device

Abstract

The invention provides a construction method and device of a porous asphalt concrete finite element model and application of the construction method and device. The construction method comprises the following steps: constructing a three-dimensional aggregate database; generating an aggregate accumulation three-dimensional discrete element model based on the three-dimensional aggregate database, and exporting the aggregate accumulation three-dimensional discrete element model in the form of a plurality of STL (Standard Template Library) files; importing the plurality of STL files into MATLAB software, and converting the aggregate accumulation three-dimensional discrete element model into a first mesoscopic model by utilizing the MATLAB software; based on the first mesoscopic model and the preset thickness of the asphalt mortar layer, constructing a second mesoscopic model comprising asphalt mortar and coarse aggregate; discrete element information of the second mesoscopic model is written into a. Py file through MATLAB programming and imported into ABAQUS software for three-dimensional modeling and grid division, grids corresponding to gaps are deleted, and the porous asphalt concrete finite element model is constructed. According to the method, the internal heterogeneity and complex porous structure characteristics of the material are fully considered, and the constructed finite element model can reveal the internal thermal stress distribution rule.

Description

Technical Field

[0001] The present invention belongs to the technical field of civil engineering, and in particular relates to a method and device for constructing a porous asphalt concrete finite element model and applications thereof. Background Art

[0002] In the field of road engineering, porous asphalt concrete has been widely used in urban roads, highways, airport runways and other fields in recent years due to its excellent water permeability, thermal insulation and good noise absorption. This material can not only effectively reduce road surface water and improve driving safety, but also achieve energy conservation and environmental protection by reducing road surface temperature and reducing the heat island effect. However, the thermal performance evaluation of porous asphalt concrete faces many technical challenges, especially due to its complex internal porous structure and heterogeneity, making it difficult for traditional experimental methods to fully and accurately reveal its heat conduction mechanism and thermal behavior characteristics.

[0003] At present, the indoor test methods for the study of thermal conductivity of materials mainly include the flat plate heat source method, steady-state method and transient plane heat source method, etc. These methods can directly measure the thermal conductivity of materials and provide certain data support for the evaluation of thermal performance of materials. However, these test methods have significant limitations when applied to porous asphalt concrete. On the one hand, due to the complexity of the internal structure of porous asphalt concrete, the test measurement results often have large fluctuations, and the test time is long and the repeatability is poor; on the other hand, traditional experimental methods can only provide the overall macroscopic thermal performance parameters of the material, but cannot reveal the influence of its internal porous structure on the thermal conduction behavior.

[0004] In addition, with the rapid development of artificial intelligence and neural network technology, the prediction of material thermal conductivity based on existing experimental data has become an emerging research direction. This method can quickly predict the thermal performance parameters of materials by learning and training a large amount of experimental data. However, this type of method can usually only provide macroscopic parameter prediction results, and it is also impossible to deeply explore the thermal conduction mechanism of porous asphalt concrete from the microscopic structural level. Summary of the invention

[0005] The purpose of the present invention is to provide a method for constructing a finite element model of porous asphalt concrete that fully considers the internal heterogeneity and complex porous structure characteristics of the material. The finite element model constructed by this construction method is applied to thermal performance research, which can more accurately evaluate its thermal performance and suggest the distribution law of its internal thermal stress.

[0006] In a first aspect, the present invention provides a method for constructing a finite element model of a porous asphalt concrete, wherein the porous asphalt concrete is composed of asphalt mortar and coarse aggregate, and the construction method comprises the following steps: Step S10, constructing a three-dimensional aggregate database, wherein the three-dimensional aggregate database comprises a plurality of coarse aggregate geometric models with different particle sizes; Step S20, generating a virtual space, putting the coarse aggregate geometric model in the three-dimensional aggregate database into the virtual space, generating a three-dimensional discrete element model of aggregate stacking, and exporting the three-dimensional discrete element model of aggregate stacking as a plurality of STL files, wherein each STL file also stores the spatial coordinate information of the coarse aggregate geometric model; Step S30, exporting the plurality of STL The file is imported into MATLAB software, and the MATLAB software is used to convert the aggregate accumulation three-dimensional discrete element model into a first mesoscopic model, wherein the first mesoscopic model is composed of N voxels, and the voxel value of each voxel is 0 or 1, and the sum of the volumes of the N voxels is the same as the volume of the virtual space, wherein the voxel with a voxel value of 0 corresponds to the coarse aggregate, and the voxel with a voxel value of 1 corresponds to the void, and N is a positive integer; step S40, based on the first mesoscopic model and the preset thickness of the asphalt mortar layer, a second mesoscopic model including asphalt mortar and coarse aggregate is constructed, and the voxel value in the first mesoscopic model is 1 And the target voxel that meets the preset conditions is marked as the voxel value 2 corresponding to the asphalt mortar layer in the second mesoscopic model, wherein the preset thickness of the asphalt mortar layer is m voxels, and meeting the preset conditions means that the distance between the target voxel and the adjacent voxel with a voxel value of 0 is less than or equal to m voxels, where m is a positive integer; step S50, write each voxel information of the second mesoscopic model into a .py file through MATLAB programming, and import it into ABAQUS finite element software for three-dimensional modeling and meshing, delete the mesh corresponding to the gap, and construct the porous asphalt concrete finite element model.

[0007] In a specific embodiment, the step S10 includes: step 1, obtaining a plurality of coarse aggregates with different particle sizes, and obtaining a two-dimensional projection image of each coarse aggregate based on the AIMS system; step 2, generating a plurality of three-dimensional point cloud data corresponding one-to-one to the two-dimensional projection image of each coarse aggregate based on the trained CGAN model; step 3, constructing a three-dimensional aggregate database based on the plurality of three-dimensional point cloud data generated in step 2, the three-dimensional aggregate database including a plurality of coarse aggregate geometric models with different particle sizes corresponding one-to-one to the plurality of three-dimensional point cloud data.

[0008] In a specific embodiment, the training method of the CGAN model includes: obtaining multiple aggregates for training; using an AIMS system to obtain a two-dimensional projection image of each aggregate; using a laser scanner to obtain three-dimensional aggregate contour point cloud data of each aggregate; obtaining a training set, wherein the training set includes multiple data pairs, each data pair includes a two-dimensional projection image and three-dimensional aggregate contour point cloud data of the same aggregate; using the training set to train the CGAN model, and when the evaluation index is less than a preset value, the training is completed to obtain a trained CGAN model, wherein the evaluation index includes an FID index and a CD index.

[0009] In a specific embodiment, the step S20 includes: step (1), generating a virtual space based on preset dimension data; step (2), obtaining the volume fraction of multiple levels of coarse aggregate based on the volume of the virtual space and preset mix ratio information of the porous concrete, wherein different levels of coarse aggregate have different particle size ranges; step (3), based on a preset placement rule, placing the coarse aggregate geometric model in the three-dimensional aggregate database into the virtual space according to the volume fraction of the multiple levels of coarse aggregate determined in step (2), to generate an aggregate stacking three-dimensional discrete element model; step (4), exporting each coarse aggregate geometric model of the aggregate stacking three-dimensional discrete element model as an STL file, and each STL file also stores the spatial coordinate information of the coarse aggregate geometric model.

[0010] In a specific implementation, step S30 includes: step a, importing multiple STL files into MATLAB software; step b, based on a preset grid size, dividing the virtual space of the three-dimensional discrete element model of aggregate accumulation into N voxels through MATLAB programming, and setting the voxel value of each voxel to an initialization value of 1, wherein a voxel value of 1 represents a gap; step c, reading the triangular face information and spatial coordinate information of each coarse aggregate based on each of the STL files, and determining the voxel value of each voxel respectively, when the number of intersections between a ray emitted from the center point of the voxel along the positive direction of the z-axis and the triangular facets on the surface of the coarse aggregate is an odd number, the voxel value of the voxel is changed from the initial value 1 to 0; when the number of intersections between a ray emitted from the center point of the voxel along the positive direction of the z-axis and the triangular facets on the surface of the coarse aggregate is an even number, the voxel value of the voxel is kept unchanged at 1, wherein a voxel with a voxel value of 0 corresponds to the coarse aggregate, and a voxel with a voxel value of 1 corresponds to the gap.

[0011] In a specific embodiment, step S40 includes: step (1), setting the thickness of the asphalt mortar layer to m voxels; step (2), expanding the vertex coordinates of each coarse aggregate in the first mesoscopic model outward by a distance of m voxels to obtain an intermediate mesoscopic model; step (3), re-marking the voxel values ​​of all candidate voxels with a voxel value of 1 in the intermediate mesoscopic model, and when the number of intersections between a ray emitted from the center point of the candidate voxel along the positive direction of the z-axis and the triangular facets on the surface of the coarse aggregate is an odd number, the candidate voxel is used as a target voxel that meets the preset conditions, and the voxel value of the target voxel is changed from 1 to 2; when the number of intersections between a ray emitted from the center point of the candidate voxel along the positive direction of the z-axis and the triangular facets on the surface of the coarse aggregate is an even number, the voxel value of the candidate voxel is kept unchanged at 1, wherein the voxel with a voxel value of 2 corresponds to the asphalt mortar layer.

[0012] In a specific embodiment, step S40 includes: step (1), setting the thickness of the asphalt mortar layer to m voxels; step (2), traversing each first voxel whose voxel value is 0, and obtaining the voxel value of the second voxel whose distance to each first voxel is less than or equal to m voxels, when the voxel value of the second voxel is 1, taking the second voxel as the target voxel, and changing the voxel value of the target voxel from 1 to 2; when the voxel value of the second voxel is 0, keeping the voxel value of the second voxel unchanged at 0, wherein the voxel with a voxel value of 2 corresponds to the asphalt mortar layer.

[0013] In a second aspect, the present invention provides a device for constructing a finite element model of porous asphalt concrete, the device comprising: A first construction module is used to construct a three-dimensional aggregate database, wherein the three-dimensional aggregate database includes a plurality of coarse aggregate geometric models with different particle sizes; a generation module is used to generate a virtual space, and the coarse aggregate geometric model in the three-dimensional aggregate database is put into the virtual space to generate a three-dimensional discrete element model of aggregate accumulation, and the three-dimensional discrete element model of aggregate accumulation is exported as a plurality of STL files, and each STL file also stores the spatial coordinate information of the coarse aggregate geometric model; a model conversion module is used to import the plurality of STL files into MATLAB software, and use the MATLAB software to convert the three-dimensional discrete element model of aggregate accumulation into a first mesoscopic model, wherein the first mesoscopic model is composed of N voxels, and the voxel value of each voxel is 0 or 1, and the sum of the volumes of the N voxels is the same as the volume of the virtual space, wherein the voxel with a voxel value of 0 corresponds to the coarse aggregate, and the voxel with a voxel value of 1 corresponds to the void, and N is a positive integer; a second construction module is used to construct a second mesoscopic model including asphalt mortar and coarse aggregate based on the first mesoscopic model and the preset thickness of the asphalt mortar layer, wherein the voxel value in the first mesoscopic model is 1. And the target voxel that meets the preset conditions is marked as the voxel value 2 corresponding to the asphalt mortar layer in the second mesoscopic model, wherein the preset thickness of the asphalt mortar layer is m voxels, and meeting the preset conditions means that the distance between the target voxel and the adjacent voxel with a voxel value of 0 is less than or equal to m voxels, where m is a positive integer; the third construction module is used to write each voxel information of the second mesoscopic model into a .py file through MATLAB programming, and import it into the ABAQUS finite element software for three-dimensional modeling and meshing, delete the mesh corresponding to the gap, and construct the porous asphalt concrete finite element model.

[0014] In a third aspect, the present invention provides a method for calculating the effective thermal conductivity of a porous asphalt concrete finite element model constructed by the construction method described above, the method comprising the following steps: step (1), importing the porous asphalt concrete finite element model into ABAQUS analysis software, selecting a unit type and setting material parameters and boundary conditions, and then using the software for simulation to obtain node temperature and heat flux, wherein the unit type selects a heat conduction type, the material parameters include density, thermal conductivity and specific heat capacity, and the setting of the boundary conditions includes: using steady-state analysis, setting different temperatures at the upper and lower boundaries of the model to form a temperature gradient, and other boundaries default to insulation conditions; step (2), calculating the equivalent thermal conductivity based on the calculation formula of heat flux and equivalent thermal conductivity.

[0015] In a fourth aspect, the present invention provides a method for analyzing the pavement temperature field using a porous asphalt concrete finite element model constructed using the construction method described above, the method comprising: importing the porous asphalt concrete finite element model into ABAQUS analysis software, setting material parameters and boundary conditions, and then using the software for simulation to obtain the pavement temperature field distribution and heat flux, wherein the material parameters include density, thermal conductivity and specific heat capacity, and the setting of the boundary conditions includes: using transient analysis; defining the initial temperature field; setting the temperature, solar radiation and wind speed on the surface of the model.

[0016] The beneficial effects of the present invention include at least: 1. The construction method provided by the present invention first constructs a three-dimensional discrete element model of aggregate accumulation based on the coarse aggregate geometric model of the aggregate database, and then converts the three-dimensional discrete element model of aggregate accumulation into multiple STL The file is exported, and then multiple STL files are imported into MATLAB software. The MATLAB software is used to convert the aggregate stacking three-dimensional discrete element model into a first mesoscopic model including coarse aggregate, and then a second mesoscopic model including coarse aggregate and asphalt mortar is constructed based on the first mesoscopic model. Each voxel in the second mesoscopic model is marked with a voxel value. When the voxel value is marked as 1, it represents that the voxel corresponds to a void. When the voxel value is marked as 0, it represents that the voxel corresponds to coarse aggregate. When the voxel value is marked as 2, it represents that the voxel corresponds to asphalt mortar. Finally, each voxel information of the second mesoscopic model is written into a .py file through MATLAB programming, and imported into ABAQUS finite element software for three-dimensional modeling and meshing, and the mesh corresponding to the void is deleted to construct the porous asphalt concrete finite element model. The porous asphalt concrete finite element model can truly reproduce the three-dimensional mesoscopic characteristics of porous asphalt concrete, and can more accurately reflect the complex pore distribution and coarse aggregate arrangement characteristics inside the material. When applied to thermal performance research, it can more accurately evaluate its thermal performance and indicate its internal thermal stress distribution law.

[0017] 2. The present invention adopts grid discretization and grid mapping technology, which can efficiently establish a finite element model of porous asphalt concrete and perform fine grid division on complex three-dimensional structures. Compared with traditional finite element modeling methods, the present invention significantly improves the efficiency and accuracy of model construction and can better adapt to the complex geometric characteristics of porous materials.

[0018] In addition to the above-described purposes, features and advantages, the present invention has other purposes, features and advantages. The present invention will be further described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic diagram of the steps of a method for constructing a porous asphalt concrete finite element model provided in one embodiment of the present invention; Figure 2 This is a finite element model diagram of porous asphalt concrete constructed by the present invention, wherein: Figure 2 (a) is a three-dimensional diagram of the finite element model of porous asphalt concrete. Figure 2 (b) Figure 2 (a) Enlarged view of part A shown; Figure 3 Aggregate distribution diagram inside the porous asphalt concrete finite element model constructed by the present invention; Figure 4 A module diagram of a device for constructing a finite element model of porous asphalt concrete provided in another embodiment of the present invention. DETAILED DESCRIPTION

[0020] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0021] The method for constructing a finite element model of porous asphalt concrete provided by the present invention fully considers the internal heterogeneity and complex porous structure characteristics of the material, and applies the finite element model to thermal performance research, which can more accurately evaluate its thermal performance and suggest the distribution law of its internal thermal stress, thereby providing a scientific basis for optimizing pavement material design and improving road service life.

[0022] According to a first aspect of the present invention, the present invention provides a method for constructing a finite element model of porous asphalt concrete, wherein the porous asphalt concrete consists of asphalt mortar and coarse aggregate.

[0023] In an optional embodiment, the asphalt mortar is a continuous uniform mixture of asphalt and fine aggregate with a particle size less than 2.36 mm, and the coarse aggregate is aggregate with a particle size greater than 2.36 mm.

[0024] See also Figure 1 The construction method provided by the present invention comprises the following steps: Step S10: constructing a three-dimensional aggregate database, wherein the three-dimensional aggregate database includes a plurality of coarse aggregate geometric models with different particle sizes.

[0025] In the present invention, the coarse aggregates in the three-dimensional aggregate database are all convex polyhedral aggregates.

[0026] In an optional implementation, step S10 includes: Step 1: Obtain a plurality of coarse aggregates, and obtain a two-dimensional projection image of each of the coarse aggregates based on the AIMS system.

[0027] In the present invention, the two-dimensional projection image of the coarse aggregate is a two-dimensional projection binarization image.

[0028] In this embodiment, the coarse aggregate refers to aggregate with a particle size of ≥2.36 mm, and the particle size refers to an effective particle size.

[0029] In this embodiment, aggregates with a particle size of ≥2.36 mm are defined as coarse aggregates. In other embodiments, aggregates with a particle size greater than 2.56 mm, 3 mm, or other particle size values ​​may be defined as coarse aggregates.

[0030] In an optional embodiment, the plurality of coarse aggregates are divided into multiple grades of coarse aggregates, and coarse aggregates of different grades have different particle size ranges.

[0031] In this embodiment, multiple coarse aggregates are divided into primary coarse aggregate, secondary coarse aggregate, tertiary coarse aggregate and quaternary coarse aggregate, wherein the particle size range of the primary coarse aggregate is 13.2~16mm, the particle size range of the secondary coarse aggregate is 9.5~13.2mm, the particle size range of the tertiary coarse aggregate is 4.75~9.5mm, and the particle size range of the quaternary coarse aggregate is 2.36~4.75mm.

[0032] In other embodiments, multiple coarse aggregates can also be divided into two levels of coarse aggregate, three levels of coarse aggregate, five levels of coarse aggregate, eight levels of coarse aggregate, etc. The particle size range of each level of coarse aggregate can also be adjusted. The particle size range of the same level of coarse aggregate can be customized and the number of coarse aggregate levels can also be customized.

[0033] Specifically: First, the crushed stones are screened and grouped according to the preset multi-level coarse aggregate size range, and a preset number of crushed stones are selected from each group as aggregate templates. The image measurement system AIMS2 is used to collect data on the aggregate templates, and two-dimensional projection images of different aggregate templates are obtained in batches, and the shape parameters of each aggregate template are analyzed and recorded, wherein the shape parameters include edge angles and aspect ratios.

[0034] It is understood that the same group of aggregate templates have the same particle size range.

[0035] Step 2: Generate multiple three-dimensional point cloud data corresponding to the two-dimensional projection image of each coarse aggregate based on the trained CGAN model.

[0036] In an optional embodiment, the training method of the CGAN model includes: obtaining multiple aggregates for training; using the AIMS system to obtain a two-dimensional projection image of each aggregate; using a laser scanner to obtain three-dimensional aggregate contour point cloud data of each aggregate; obtaining a training set, wherein the training set includes multiple data pairs, each data pair includes a two-dimensional projection image and three-dimensional aggregate contour point cloud data of the same aggregate; using the training set to train the CGAN model, and when the evaluation index is less than a preset value, the training is completed to obtain a trained CGAN model.

[0037] In the present invention, the evaluation indicators include the FID (Frechet Inception Distance) indicator and the CD (Chamfer Distance) indicator. The preset value of the FID indicator is 10, and the preset value of the CD indicator is 0.05.

[0038] In the present invention, the generator of the CGAN model receives a two-dimensional projection image as input, uses a convolutional neural network (CNN) to extract image features, and generates corresponding three-dimensional point cloud data; the discriminator compares the generated three-dimensional point cloud data (output value) with the input three-dimensional aggregate contour point cloud data (true value), and guides the generator to optimize based on the comparison results. During the training process, the FID index and CD index are used as evaluation indicators to evaluate the generation quality and continuously adjust the learning rate, network architecture and training strategy according to the evaluation results to improve the performance of the model and ensure that the generated three-dimensional point cloud data is highly consistent with the actual aggregate shape.

[0039] Step 3: construct a three-dimensional aggregate database based on the multiple three-dimensional point cloud data generated in step 2, wherein the three-dimensional aggregate database includes multiple coarse aggregate geometric models with different particle sizes corresponding to the multiple three-dimensional point cloud data.

[0040] It can be understood that each 3D point cloud data corresponds to a coarse aggregate geometric model. If 1000 aggregates are needed in the 3D aggregate database, the 2D projection images of 1000 coarse aggregates are obtained in step 1, and 1000 3D point cloud data corresponding to the 2D projection images of the 1000 coarse aggregates are generated in step 2.

[0041] The specific steps are as follows: connecting the points in each three-dimensional point cloud data to form a three-dimensional aggregate entity composed of spatial tetrahedrons according to a certain rule, so as to obtain the contour surface of the three-dimensional aggregate that is closed in space; based on all the surface triangular mesh information composed of the vertices of the contour surface of the three-dimensional aggregate, the conversion between text documents or other files in other specific formats is performed through MATLAB code programming operation, and then imported into the discrete element software PFC to construct multiple aggregate geometric models, and all the aggregate geometric models are collected to construct a three-dimensional aggregate database.

[0042] In another optional implementation, other existing technologies may be used to construct a three-dimensional aggregate database.

[0043] Step S20, generating a virtual space, according to the preset mix ratio information of the porous concrete, putting the coarse aggregate geometric model in the three-dimensional aggregate database into the virtual space, generating an aggregate stacking three-dimensional discrete element model, and exporting the aggregate stacking three-dimensional discrete element model as multiple STL files, and each STL file also stores the spatial coordinate information of the coarse aggregate geometric model.

[0044] In the present invention, PFC3D software is used to generate a three-dimensional discrete element model of aggregate accumulation.

[0045] This step includes: Step (1), generating a virtual space based on preset dimension data.

[0046] In the present invention, the virtual space is a cuboid or a cube, and the volume of the virtual space is V 总 .

[0047] Step (2): based on the volume of the virtual space and the preset mix ratio information of the porous concrete, the volume fractions of the multiple grades of coarse aggregate are obtained, and the coarse aggregates of different grades have different particle size ranges.

[0048] In the present invention, the preset mix ratio information of the porous concrete includes aggregate gradation, asphalt-stone ratio and target porosity.

[0049] For ease of understanding, this step uses an example to illustrate the calculation process of the volume of multi-grade coarse aggregate. Assume that the density of asphalt is 1000kg / m 3 , aggregate density is 2600kg / m 3 The porosity is 20%, the oil-stone ratio is 4.5%, and the aggregate grading is shown in Table 1.

[0050] Table 1 Aggregate gradation data of porous concrete The volume calculation process for each grade of coarse aggregate is as follows: first, the sum of the volumes of aggregate and asphalt is calculated based on the porosity, then the volume of asphalt is calculated based on the oil-stone ratio, then the volume of aggregate is calculated based on the volume of asphalt and the total volume, then the total volume of coarse aggregate is calculated based on the volume of aggregate and the grading data of aggregate, and finally the volume of each grade of coarse aggregate is calculated based on the grading data of aggregate and the total volume of coarse aggregate.

[0051] The specific calculation is as follows: The sum of the volumes of aggregate and asphalt is calculated as: V 骨料 +V 沥青= (1-20%) V 总 =80%V 总 .

[0052] The volume of asphalt is calculated as: (1000×V 沥青 ) / (2600×V 骨料 )=4.5%;V 沥青 = (2600 × 4.5%) V 骨料 / 1000.

[0053] The volume of aggregate is calculated as: V 骨料 =80%V 总 / [1+(2600×4.5%) / 1000].

[0054] The total volume of coarse aggregate is calculated as: V 粗骨料 = (1-10.4%) V 骨料 =89.6%×80% V 总 / [1+(2600×4.5%) / 1000]=64.17%V 总 .

[0055] The volume of coarse aggregate in each particle size range is as follows: V 13.2~16mm =[(100-92.7) / 89.6]V 粗骨料 =8.2%V 粗骨料 =8.2%×64.17%V 总 .

[0056] V 9.5~13.2 mm =[(92.7-58.2) / 89.6]V 粗骨料 =38.5%V 粗骨料 =38.5%×64.17%V 总 .

[0057] V 4.75~9.5mm =[(58.2-16.7) / 89.6]V 粗骨料 =46.3%V 粗骨料 =46.3%×64.17% V 总 .

[0058] V 2.36~4.75mm =[(16.7-10.4) / 89.6]V 粗骨料 =7.0%V 粗骨料 =7.0%×64.17%V 总 .

[0059] Taking the total volume of the virtual space as 100%, the volume fraction of each grade of coarse aggregate is: the volume fraction of the first-level coarse aggregate with a particle size range of 13.2~16mm is 5.26%, the volume fraction of the second-level coarse aggregate with a particle size range of 9.5~13.2mm is 24.7%, the volume fraction of the third-level coarse aggregate with a particle size range of 4.75~9.5mm is 29.71%, and the volume fraction of the fourth-level coarse aggregate with a particle size range of 2.36~4.75mm is 4.49%.

[0060] Step (3): Based on a preset placement rule, the coarse aggregate geometric model in the three-dimensional aggregate database is placed into the virtual space according to the volume fraction of the multi-level coarse aggregate determined in step (2), so as to generate a three-dimensional discrete element model of aggregate accumulation.

[0061] In the present invention, the preset placement rule is specifically: placing in grades according to the level of coarse aggregate, and placing in each grade based on the particle size value of the coarse aggregate in a regular order from large to small.

[0062] Specifically: Start with the coarse aggregate with the largest particle size, add a certain number of particles each time, and perform layered processing according to the particle size range. Through the discrete element software PFC 3D command stream, after each coarse aggregate particle is added, the volume fraction of the coarse aggregate in the current particle size range is automatically calculated. If the current volume fraction of coarse aggregate is less than the target volume fraction, continue to add coarse aggregate in this particle size range until the target volume fraction is reached. After completing the placement of coarse aggregate in one particle size range, move on to the coarse aggregate in the next particle size range.

[0063] It can be understood that the target volume fraction is the volume fraction of the multi-grade coarse aggregate determined in step (2).

[0064] For the convenience of understanding, an example is given. First, the first-level coarse aggregate with a particle size range of 13.2~16mm is added. When the volume fraction of the first-level coarse aggregate reaches 5.26%, the second-level coarse aggregate with a particle size range of 9.5~13.2mm is added. When the volume fraction of the second-level coarse aggregate reaches 24.7%, the third-level coarse aggregate with a particle size range of 4.75~9.5mm is added. When the volume fraction of the third-level coarse aggregate reaches 29.71%, the fourth-level coarse aggregate with a particle size range of 2.36~4.75mm is added. When the volume fraction of the fourth-level coarse aggregate reaches 4.49%, the addition of all coarse aggregates is completed. When adding each level of coarse aggregate, the aggregate with a larger particle size range is added first, such as coarse aggregate G1 with a particle size of 14mm and coarse aggregate G2 with a particle size of 15mm. When adding, coarse aggregate G2 is added first.

[0065] Step (4), exporting each coarse aggregate geometric model of the aggregate stacking three-dimensional discrete element model as an STL file, and each STL file also stores the spatial coordinate information of the coarse aggregate geometric model.

[0066] In the present invention, the STL file is a 3D model including the shape and geometric information of the coarse aggregate, and also stores the spatial coordinate information of the coarse aggregate.

[0067] It can be understood that if the aggregate accumulation three-dimensional discrete element model includes 1200 coarse aggregate geometric models, the number of exported STL files is 1200, that is, each coarse aggregate geometric model corresponds to one STL file.

[0068] Each STL file also stores the spatial coordinate information of the coarse aggregate geometric model, so that the position of each coarse aggregate can be kept unchanged during subsequent format conversion.

[0069] Step S30, importing the multiple STL files into MATLAB software, and using the MATLAB software to convert the aggregate stacking three-dimensional discrete element model into a first mesoscopic model, wherein the first mesoscopic model is composed of N voxels, and the voxel value of each voxel is 0 or 1, and the sum of the volumes of the N voxels is the same as the volume of the virtual space, wherein the voxels corresponding to the coarse aggregate are marked as 0, and the voxels corresponding to the voids are marked as 1, and N is a positive integer.

[0070] This step includes: Step a: Import multiple STL files into MATLAB software.

[0071] Step b: Based on a preset grid size, the virtual space of the three-dimensional discrete element model of aggregate accumulation is divided into N voxels through MATLAB programming, and the voxel value of each voxel is set to an initialization value of 1, wherein a voxel value of 1 represents a gap.

[0072] In the present invention, each voxel is a regular cubic grid.

[0073] In the present invention, the preset grid size must be able to divide the length, width, and height of the model to ensure that the size of each voxel is consistent and avoid irregular or uneven divisions during the grid division process. For example, the length, width, and height of the model need to be able to be divided by the specified voxel size.

[0074] Step c, based on each of the STL files, read the triangular face information and spatial coordinate information of each coarse aggregate, and determine the voxel value of each voxel respectively. When the number of intersections between a ray emitted by the center point of the voxel along the positive direction of the z-axis and the triangular facets on the surface of the coarse aggregate is an odd number, the voxel value of the voxel is changed from an initial value of 1 to 0; when the number of intersections between a ray emitted by the center point of the voxel along the positive direction of the z-axis and the triangular facets on the surface of the coarse aggregate is an even number, the voxel value of the voxel is kept unchanged at 1, wherein the voxels corresponding to the coarse aggregate are marked as 0, and the voxels corresponding to the gaps are marked as 1.

[0075] In the present invention, the spatial coordinate information of the coarse aggregate includes vertex coordinate information of each vertex of the coarse aggregate.

[0076] In the present invention, the triangular face information and spatial coordinate information of each coarse aggregate are read in turn, and each voxel is traversed. When the voxel is located inside the aggregate, a ray emitted from the center point of the voxel along the positive direction of the z-axis has an odd number of intersections with the triangular facets on the surface of the coarse aggregate; when the voxel is located outside the aggregate, a ray emitted from the center point of the voxel along the positive direction of the z-axis has an even number of intersections with the triangular facets on the surface of the coarse aggregate.

[0077] Step S40: Based on the first meso model and the preset thickness of the asphalt mortar layer, a second meso model including asphalt mortar and coarse aggregate is constructed, and the target voxel whose voxel value is 1 in the first meso model and meets the preset conditions is marked as a voxel value 2 corresponding to the asphalt mortar layer in the second meso model, wherein the preset thickness of the asphalt mortar layer is m voxels, and meeting the preset conditions means that the distance between the target voxel and the adjacent voxel with a voxel value of 0 is less than or equal to m voxels, where m is a positive integer.

[0078] In an optional implementation, the step S40 includes: Step (1), setting the thickness of the asphalt mortar layer to m voxels.

[0079] Step (2), expanding the vertex coordinates of each coarse aggregate in the first mesoscopic model outward by a distance of m voxels to obtain an intermediate mesoscopic model.

[0080] Step (3), re-marking the voxel values ​​of all candidate voxels whose voxel values ​​of the intermediate mesoscopic model are 1, when the number of intersections between a ray emitted from the center point of the candidate voxel along the positive direction of the z-axis and the triangular facets on the surface of the coarse aggregate is an odd number, the candidate voxel is used as a target voxel that meets the preset conditions, and the voxel value of the target voxel is changed from 1 to 2; when the number of intersections between a ray emitted from the center point of the candidate voxel along the positive direction of the z-axis and the triangular facets on the surface of the coarse aggregate is an even number, the voxel value of the candidate voxel is kept unchanged at 1, wherein the voxel with a voxel value of 2 corresponds to the asphalt mortar layer.

[0081] In another optional implementation, the step S40 includes: Step (1), setting the thickness of the asphalt mortar layer to m voxels.

[0082] Step (2), traverse each first voxel whose voxel value is 0, and obtain the voxel value of the second voxel whose distance to each first voxel is less than or equal to m voxels, when the voxel value of the second voxel is 1, take the second voxel as the target voxel, and change the voxel value of the target voxel from 1 to 2; when the voxel value of the second voxel is 0, keep the voxel value of the second voxel unchanged to 0, wherein the voxel with a voxel value of 2 corresponds to the asphalt mortar layer.

[0083] It should be noted that the distance between the first voxel and the second voxel refers to the distance between the center point of the first voxel and the center point of the second voxel.

[0084] For ease of understanding, an example is given, assuming m=2, then the voxel value of the second voxel that is 1 voxel or 2 voxels away from the first voxel is obtained. If the voxel value of the second voxel is 1, it means that in the first mesoscopic model, the voxel represents a void, then the voxel value is changed to 2, and at this time, the voxel represents the mortar layer; if the voxel value of the second voxel is 0, it means that in the first mesoscopic model, the voxel represents coarse aggregate, then the voxel value remains unchanged, that is, the voxel still represents the coarse aggregate.

[0085] Step S50, write each voxel information of the second mesoscopic model into a .py file through MATLAB programming, and import it into ABAQUS finite element software for three-dimensional modeling and meshing, delete the meshes corresponding to the gaps, and construct the porous asphalt concrete finite element model.

[0086] In the present invention, each voxel information includes the spatial coordinates and voxel value of the voxel.

[0087] This step includes: Step (1), open ABAQUS finite element software, run the imported .py file, and obtain a finite element model with mesh division completed, wherein the finite element model includes coarse aggregate, asphalt mortar and voids.

[0088] Specifically, start the ABAQUS / CAE interface, import the .py file through the Python script interface, run the .py file, visually check the imported model geometry in the Viewport window, confirm the correct identification of the coarse aggregate and asphalt mortar parts in the model, and check the quality of the unit division. Use the Assembly module to create a model instance and ensure the correct model position.

[0089] Step (2), converting the finite element model into a mesh display, and converting all meshes into isolated meshes for easy editing.

[0090] Step (3), retaining the meshes corresponding to the coarse aggregate and the asphalt mortar, and deleting all other meshes corresponding to the voids, to obtain the porous asphalt concrete finite element model.

[0091] See also Figure 2 and Figure 3 ,in, Figure 2 This is a finite element model diagram of porous asphalt concrete constructed by the present invention. Figure 2 (a) is a three-dimensional diagram of the finite element model of porous asphalt concrete. Figure 2 (b) Figure 2 (a) Enlarged view of part A shown; Figure 3 This is the internal aggregate distribution diagram of the porous asphalt concrete finite element model. Figure 3 201 in the formula represents coarse aggregate. Figure 3 The 202 indicates asphalt mortar.

[0092] See also Figure 4 According to a second aspect of the present invention, the present invention provides a device for constructing a porous asphalt concrete finite element model, the construction device 100 comprising: The first construction module 101 is used to construct a three-dimensional aggregate database, wherein the three-dimensional aggregate database includes a plurality of coarse aggregate geometric models with different particle sizes.

[0093] The generation module 102 is used to generate a virtual space, put the coarse aggregate geometric model in the three-dimensional aggregate database into the virtual space, generate an aggregate stacking three-dimensional discrete element model, and export the aggregate stacking three-dimensional discrete element model as multiple STL files, and each STL file also stores the spatial coordinate information of the coarse aggregate geometric model.

[0094] The model conversion module 103 is used to import the multiple STL files into MATLAB software, and use the MATLAB software to convert the aggregate stacking three-dimensional discrete element model into a first mesoscopic model, wherein the first mesoscopic model is composed of N voxels, and the voxel value of each voxel is 0 or 1, and the sum of the volumes of the N voxels is the same as the volume of the virtual space, wherein the voxels corresponding to the coarse aggregate are marked as 0, and the voxels corresponding to the gaps are marked as 1, and N is a positive integer.

[0095] The second construction module 104 is used to construct a second meso model including asphalt mortar and coarse aggregate based on the first meso model and the preset thickness of the asphalt mortar layer. The target voxel with a voxel value of 1 in the first meso model and meeting the preset conditions is marked as a voxel value 2 corresponding to the asphalt mortar layer in the second meso model, wherein the preset thickness of the asphalt mortar layer is m voxels, and meeting the preset conditions means that the distance between the target voxel and the adjacent voxel with a voxel value of 0 is less than or equal to m voxels, where m is a positive integer.

[0096] The third construction module 105 is used to write each voxel information of the second microscopic model into a .py file through MATLAB programming, and import it into ABAQUS finite element software for three-dimensional modeling and meshing, delete the meshes corresponding to the gaps, and construct the porous asphalt concrete finite element model.

[0097] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the working process of constructing the device described above can refer to the contents in the aforementioned method embodiment and will not be repeated here.

[0098] According to a third aspect of the present invention, the present invention provides a method for calculating the effective thermal conductivity of a porous asphalt concrete finite element model constructed by the construction method described above, the method comprising the following steps: Step (1), importing the porous asphalt concrete finite element model into ABAQUS analysis software, selecting the unit type and setting the material parameters and boundary conditions, and then using the software to perform simulation to obtain the node temperature and heat flux.

[0099] Among them, the unit type selects the heat conduction type, the material parameters include density, thermal conductivity and specific heat capacity, and the boundary condition settings include: using steady-state analysis; setting different temperatures at the upper and lower boundaries of the model to form a temperature gradient; other boundaries default to insulation conditions.

[0100] In an optional embodiment, the material parameters are based on experimentally acquired data, which can further improve the accuracy of the model calculation results and effectively reduce the model error.

[0101] Step (2): Based on the calculation formula of heat flux and equivalent thermal conductivity, the equivalent thermal conductivity is calculated.

[0102] In the present invention, the calculation formula of the equivalent thermal conductivity is: in, Q is the heat flux, L is the specimen thickness, A is the heat transfer area, ∆T is the temperature difference between the upper and lower ends of the model.

[0103] According to the fourth aspect of the present invention, the present invention provides a method for analyzing the pavement field temperature using a porous asphalt concrete finite element model constructed using the construction method described above, the method comprising: importing the porous asphalt concrete finite element model into ABAQUS analysis software, setting material parameters and boundary conditions, and then using the software for simulation to obtain the pavement field temperature field distribution and heat flux.

[0104] The material parameters include density, thermal conductivity and specific heat capacity, and the boundary conditions are set up by: using transient analysis; defining an initial temperature field; and setting temperature, solar radiation and wind speed on the surface of the model.

[0105] In an optional embodiment, the material parameters are based on experimentally acquired data, which can further improve the accuracy of the model calculation results and effectively reduce the model error.

[0106] The present invention greatly improves the efficiency of thermal performance data acquisition and reduces test costs and time consumption by constructing a finite element model to calculate the effective thermal conductivity and perform pavement field temperature analysis. At the same time, the present invention can generate thermal performance simulation data of porous concrete under different temperature conditions, providing efficient and accurate technical support for material design and engineering applications.

[0107] It can be understood that the porous asphalt concrete finite element model construction method and thermal performance evaluation method provided by the present invention have strong versatility and adaptability, and can also be extended to the thermal performance research of other porous composite materials.

[0108] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions and substitutions can be made without departing from the concept of the present invention, which should be regarded as belonging to the protection scope of the present invention.

Claims

1. A method for constructing a finite element model of porous asphalt concrete, wherein the porous asphalt concrete is composed of asphalt mortar and coarse aggregate, characterized in that: The construction method comprises the following steps: Step S10, constructing a three-dimensional aggregate database, wherein the three-dimensional aggregate database includes a plurality of coarse aggregate geometric models with different particle sizes; Step S20, generating a virtual space, putting the coarse aggregate geometric model in the three-dimensional aggregate database into the virtual space, generating a three-dimensional discrete element model of aggregate accumulation, and exporting the three-dimensional discrete element model of aggregate accumulation as a plurality of STL files, wherein each STL file also stores the spatial coordinate information of the coarse aggregate geometric model; Step S30, importing a plurality of STL files into MATLAB software, and using the MATLAB software to convert the aggregate stacking three-dimensional discrete element model into a first mesoscopic model, wherein the first mesoscopic model is composed of N voxels, and the voxel value of each voxel is 0 or 1, and the sum of the volumes of the N voxels is the same as the volume of the virtual space, wherein the voxel with a voxel value of 0 corresponds to the coarse aggregate, the voxel with a voxel value of 1 corresponds to the void, and N is a positive integer; Step S40: Based on the first mesoscopic model and the preset thickness of the asphalt mortar layer, a second mesoscopic model including asphalt mortar and coarse aggregate is constructed, and the target voxel with a voxel value of 1 in the first mesoscopic model and meeting the preset conditions is marked as a voxel value 2 corresponding to the asphalt mortar layer in the second mesoscopic model, wherein the preset thickness of the asphalt mortar layer is m voxels, and meeting the preset conditions means that the distance between the target voxel and the adjacent voxel with a voxel value of 0 is less than or equal to m voxels, and m is a positive integer; Step S50, write each voxel information of the second mesoscopic model into a .py file through MATLAB programming, and import it into ABAQUS finite element software for three-dimensional modeling and meshing, delete the meshes corresponding to the gaps, and construct the porous asphalt concrete finite element model.

2. The method for constructing a porous asphalt concrete finite element model according to claim 1, characterized in that: The step S10 comprises: Step 1, obtaining a plurality of coarse aggregates with different particle sizes, and obtaining a two-dimensional projection image of each of the coarse aggregates based on the AIMS system; Step 2: Generate multiple three-dimensional point cloud data corresponding to the two-dimensional projection image of each coarse aggregate based on the trained CGAN model; Step 3: Based on the multiple three-dimensional point cloud data generated in step 2, a three-dimensional aggregate database is constructed, wherein the three-dimensional aggregate database includes multiple coarse aggregate geometric models with different particle sizes corresponding to the multiple three-dimensional point cloud data.

3. The method for constructing a porous asphalt concrete finite element model according to claim 2, characterized in that: The training methods of the CGAN model include: Get multiple training sets; The AIMS system is used to obtain a two-dimensional projection image of each aggregate; Use a laser scanner to obtain three-dimensional aggregate contour point cloud data of each aggregate; Acquire a training set, wherein the training set includes a plurality of data pairs, each data pair including a two-dimensional projection image and three-dimensional aggregate contour point cloud data of the same aggregate; The CGAN model is trained using the training set. When the evaluation index is less than a preset value, the training is completed to obtain a trained CGAN model, wherein the evaluation index includes a FID index and a CD index.

4. The method for constructing a porous asphalt concrete finite element model according to claim 1, characterized in that: The step S20 comprises: Step (1), generating a virtual space based on preset dimension data; Step (2), based on the volume of the virtual space and the preset mix ratio information of the porous concrete, obtaining the volume fractions of multiple grades of coarse aggregate, wherein different grades of coarse aggregate have different particle size ranges; Step (3), based on a preset placement rule, the coarse aggregate geometric model in the three-dimensional aggregate database is placed into the virtual space according to the volume fraction of the multi-level coarse aggregate determined in step (2), so as to generate a three-dimensional discrete element model of aggregate accumulation; Step (4), exporting each coarse aggregate geometric model of the aggregate stacking three-dimensional discrete element model as an STL file, and each STL file also stores the spatial coordinate information of the coarse aggregate geometric model.

5. The method for constructing a porous asphalt concrete finite element model according to claim 1, characterized in that: The step S30 comprises: Step a, import multiple STL files into MATLAB software; Step b, based on a preset grid size, the virtual space of the three-dimensional discrete element model of aggregate accumulation is divided into N voxels through MATLAB programming, and the voxel value of each voxel is set to an initialization value of 1, wherein a voxel value of 1 represents a gap; Step c, based on each of the STL files, read the triangular face information and spatial coordinate information of each coarse aggregate, and determine the voxel value of each voxel respectively. When the number of intersections between a ray emitted by the center point of the voxel along the positive direction of the z-axis and the triangular facets on the surface of the coarse aggregate is an odd number, the voxel value of the voxel is changed from an initial value of 1 to 0; when the number of intersections between a ray emitted by the center point of the voxel along the positive direction of the z-axis and the triangular facets on the surface of the coarse aggregate is an even number, the voxel value of the voxel is kept unchanged at 1, wherein the voxel with a voxel value of 0 corresponds to the coarse aggregate, and the voxel with a voxel value of 1 corresponds to the gap.

6. The method for constructing a porous asphalt concrete finite element model according to any one of claims 1 to 5, characterized in that: The step S40 comprises: Step (1), setting the thickness of the asphalt mortar layer to m voxels; Step (2), expanding the vertex coordinates of each coarse aggregate in the first mesoscopic model outward by a distance of m voxels to obtain an intermediate mesoscopic model; Step (3), re-marking the voxel values ​​of all candidate voxels whose voxel values ​​of the intermediate mesoscopic model are 1, when the number of intersections between a ray emitted from the center point of the candidate voxel along the positive direction of the z-axis and the triangular facets on the surface of the coarse aggregate is an odd number, the candidate voxel is used as a target voxel that meets the preset conditions, and the voxel value of the target voxel is changed from 1 to 2; when the number of intersections between a ray emitted from the center point of the candidate voxel along the positive direction of the z-axis and the triangular facets on the surface of the coarse aggregate is an even number, the voxel value of the candidate voxel is kept unchanged at 1, wherein the voxel with a voxel value of 2 corresponds to the asphalt mortar layer.

7. The method for constructing a porous asphalt concrete finite element model according to any one of claims 1 to 5, characterized in that: The step S40 comprises: Step (1), setting the thickness of the asphalt mortar layer to m voxels; Step (2), traverse each first voxel whose voxel value is 0, and obtain the voxel value of the second voxel whose distance to each first voxel is less than or equal to m voxels, when the voxel value of the second voxel is 1, take the second voxel as the target voxel, and change the voxel value of the target voxel from 1 to 2; when the voxel value of the second voxel is 0, keep the voxel value of the second voxel unchanged to 0, wherein the voxel with a voxel value of 2 corresponds to the asphalt mortar layer.

8. A device for constructing a finite element model of porous asphalt concrete, characterized in that: The construction device comprises: A first construction module is used to construct a three-dimensional aggregate database, wherein the three-dimensional aggregate database includes a plurality of coarse aggregate geometric models with different particle sizes; A generation module is used to generate a virtual space, put the coarse aggregate geometric model in the three-dimensional aggregate database into the virtual space, generate a three-dimensional discrete element model of aggregate accumulation, and export the three-dimensional discrete element model of aggregate accumulation as multiple STL files, and each STL file also stores the spatial coordinate information of the coarse aggregate geometric model; A model conversion module, used to import the multiple STL files into MATLAB software, and use the MATLAB software to convert the aggregate stacking three-dimensional discrete element model into a first mesoscopic model, wherein the first mesoscopic model is composed of N voxels, and the voxel value of each voxel is 0 or 1, and the sum of the volumes of the N voxels is the same as the volume of the virtual space, wherein the voxel with a voxel value of 0 corresponds to the coarse aggregate, the voxel with a voxel value of 1 corresponds to the void, and N is a positive integer; A second construction module is used to construct a second mesoscopic model including asphalt mortar and coarse aggregate based on the first mesoscopic model and the preset thickness of the asphalt mortar layer, wherein the target voxel with a voxel value of 1 in the first mesoscopic model and meeting the preset condition is marked as a voxel value of 2 corresponding to the asphalt mortar layer in the second mesoscopic model, wherein the preset thickness of the asphalt mortar layer is m voxels, and meeting the preset condition means that the distance between the target voxel and the adjacent voxel with a voxel value of 0 is less than or equal to m voxels, where m is a positive integer; The third construction module is used to write each voxel information of the second microscopic model into a .py file through MATLAB programming, and import it into the ABAQUS finite element software for three-dimensional modeling and meshing, delete the meshes corresponding to the gaps, and construct the porous asphalt concrete finite element model.

9. A method for calculating effective thermal conductivity using a porous asphalt concrete finite element model constructed by the construction method according to any one of claims 1 to 7, characterized in that: The method comprises the following steps: Step (1), importing the porous asphalt concrete finite element model into ABAQUS analysis software, selecting the unit type and setting the material parameters and boundary conditions, and then using the software to perform simulation to obtain the node temperature and heat flux, wherein the unit type selects the heat conduction type, the material parameters include density, thermal conductivity and specific heat capacity, and the boundary conditions are set including: using steady-state analysis; setting different temperatures at the upper and lower boundaries of the model to form a temperature gradient; and other boundaries are set to insulation conditions by default; Step (2): Based on the calculation formula of heat flux and equivalent thermal conductivity, the equivalent thermal conductivity is calculated.

10. A method for analyzing the pavement field temperature using a porous asphalt concrete finite element model constructed by the construction method according to any one of claims 1 to 7, characterized in that: The method includes: importing the porous asphalt concrete finite element model into ABAQUS analysis software, setting material parameters and boundary conditions, and then using the software to perform simulation to obtain the pavement temperature field distribution and heat flux, wherein the material parameters include density, thermal conductivity and specific heat capacity, and the setting of the boundary conditions includes: using transient analysis; defining the initial temperature field; and setting the temperature, solar radiation and wind speed on the surface of the model.

Images