Modeling method of railway ballast model and method for quantifying accuracy of railway ballast model

Abstract

The invention discloses a modeling method of a railway ballast model and a method for quantifying the accuracy of the railway ballast model, and belongs to the field of discrete element numerical simulation of a ballast track system. The modeling method of the railway ballast model can be used for realizing modeling of the railway ballast model, and is further used for constructing a fine railway ballast model with required precision; according to the method for quantifying the accuracy of the railway ballast model provided by the invention, the quantification of the accuracy can be realized; thirdly, the modeling parameters of the fine railway ballast model are obtained by analyzing the influence of the modeling parameters on the accuracy of the railway ballast model and considering the contradiction between the accuracy of the model and the calculation efficiency; uniaxial compression DEM simulation is carried out in discrete element software and compared with actual experiments, and the reasonability that the established fine railway ballast model is used for simulating railway ballast degradation is verified.

Description

Technical Field

[0001] The invention relates to a modeling method of a ballast model and a method for quantifying the accuracy of the ballast model, and belongs to the field of discrete element numerical simulation of ballasted track systems. Background Art

[0002] Ballasted track is one of the most widely used forms in railway systems. The roadbed is an important component of the trackbed, which is composed of ballast with a certain gradation. Numerical simulation is often used to analyze the micromechanical properties of the roadbed under train cyclic loads. The commonly used method is to use spherical cluster particles in discrete element software to establish a ballast model and then generate the roadbed. The spherical cluster particle ballast model is composed of a number of overlapping small balls with different particle sizes. Limited by the radius of the small balls, the spherical cluster particle ballast model cannot accurately describe the edges and corners and concave areas of the real ballast. The angular characteristics of the ballast are important properties of its morphological characteristics. The angular characteristics of the ballast will affect the distribution and transmission of the contact force chain of the ballast aggregate. Therefore, the mechanical properties of the ball bed established using the spherical cluster particle ballast model are different from the actual situation. Therefore, it is very meaningful to establish a ballast model that can more comprehensively reflect the geometric shape of the real ballast.

[0003] At present, the accuracy quantification of ballast models and the relationship between ballast model accuracy and computational cost are rarely mentioned; at the same time, no scholar has derived the modeling parameters of the fine ballast model through research. In addition, the spherical cluster particle ballast model has some limitations in simulating the mechanical properties of real ballast.

[0004] Based on the above problems, it is necessary to propose a modeling method for the ballast model and a method to quantify the accuracy of the ballast model. Summary of the invention

[0005] The present invention provides a modeling method of a ballast model, which is used to realize the modeling of the ballast model and further used to construct a fine ballast model with required accuracy; the present invention also provides a method for quantifying the accuracy of the ballast model, through which the quantification of accuracy is realized; furthermore, the modeling parameters of the fine ballast model are obtained by analyzing the influence of the modeling parameters on the accuracy of the ballast model and considering the contradiction between the model accuracy and the calculation efficiency; uniaxial compression DEM simulation is performed in discrete element software and compared with actual experiments to verify the rationality of the established fine ballast model for simulating ballast degradation.

[0006] The technical solution of the present invention is:

[0007] According to a first aspect of the present invention, a modeling method for a ballast model is provided, in which real ballast samples are randomly selected to obtain point cloud data of the real ballast surface contour; based on the point cloud data, a closed first STL contour file is obtained, and then the first STL contour file is imported into discrete element software to generate a rigid sub-block filled three-dimensional contour, thereby obtaining a ballast model formed by bonding rigid sub-blocks.

[0008] Furthermore, the first STL outline file is imported into the discrete element software to generate a three-dimensional outline of rigid sub-block filling, and a ballast model formed by bonding rigid sub-blocks is obtained. Specifically, the first STL outline file is imported into the discrete element software, and the geometric body to be filled is specified according to the discrete element software; then the sub-block type is specified as a tetrahedron; the minimum and maximum side lengths of the filled sub-blocks are specified to control the number N of filled sub-blocks; and the angle of adjacent triangular faces between two tetrahedral blocks is set.

[0009] Furthermore, the ratios of the minimum and maximum side lengths of the sub-block to the equivalent ballast particle size D based on the ballast model are defined as Rs and Rb respectively. Rs and Rb are used as modeling parameters to analyze the changes of the ballast model accuracy with Rs and Rb. The ballast model with an accuracy greater than the preset value is defined as a fine ballast model.

[0010] According to a second aspect of the present invention, a method for quantifying the accuracy of a ballast model is provided, comprising the following steps:

[0011] S1. Obtaining a two-dimensional profile of ballast; wherein the two-dimensional profile of ballast includes a two-dimensional profile based on a ballast model and a two-dimensional profile based on a real ballast;

[0012] S2. determining a first macroscopic morphological parameter based on a ballast model and based on real ballast;

[0013] S3, introduce the area overlap of the 2D ballast profile; take the average value S of the area overlap of the 2D ballast profiles obtained under each 2D profile based on the ballast model a As the area coincidence of the ballast model;

[0014] S4. According to the first macroscopic morphological parameters based on the ballast model and the first macroscopic morphological parameters based on the real ballast, the similarity of each first macroscopic morphological parameter of the ballast model is obtained; and the average value of the similarity of each first macroscopic morphological parameter is used as the macroscopic morphological accuracy A of the ballast model. g ;

[0015] S5. The similarity between the ballast edge angle based on the ballast model and the ballast edge angle based on the real ballast is taken as the edge angle accuracy of the ballast two-dimensional profile; the average of the edge angle accuracies of the ballast two-dimensional profile obtained under each two-dimensional profile of the ballast model is taken as the ballast model edge angle accuracy Ar;

[0016] S6, based on the area overlap Sa of the ballast model and the macroscopic morphological accuracy A of the ballast model g , the ballast model angle accuracy Ar, and assign weights to construct the ballast model accuracy Am expression.

[0017] Furthermore, the S1 includes:

[0018] The acquisition of the two-dimensional contour based on the real ballast is specifically as follows: randomly selecting a real ballast sample, obtaining the point cloud data of the surface contour of the real ballast, and obtaining a closed first STL contour file based on the point cloud data; then converting the first STL contour file into a surface entity; then converting the surface entity into an engineering drawing, and obtaining the two-dimensional contour of the ballast under three orthogonal views; deleting redundant lines from the two-dimensional contours under the three orthogonal views, and obtaining the two-dimensional contour based on the real ballast;

[0019] The acquisition of the two-dimensional contour based on the ballast model is specifically as follows: according to the established ballast model, the ballast model is exported as a second STL contour file; then the second STL contour file is converted into a surface entity; then the surface entity is converted into an engineering drawing to obtain the two-dimensional contour of the ballast under three orthogonal views; redundant lines inside the two-dimensional contour under the three orthogonal views are deleted to obtain the two-dimensional contour based on the ballast model.

[0020] Furthermore, the first macroscopic morphological parameters include the major axis L, the middle axis I, the minor axis S, the needle index NI, and the flake index SI of the ballast.

[0021] Furthermore, the 2D ballast profile area overlap Sc is expressed as:

[0022]

[0023] Among them, S f is the area of ​​the overlapping part of the two-dimensional profile based on the ballast model and the two-dimensional profile based on the real ballast, and Sr is the area of ​​the two-dimensional profile based on the real ballast.

[0024] Furthermore, the similarity of each first macroscopic morphological parameter of the ballast model is obtained based on the first macroscopic morphological parameter of the ballast model and the real ballast, and the expression is:

[0025]

[0026] Among them, ε i is the similarity of the macroscopic morphological parameter i of the ballast model, Δxi is the difference between the macroscopic morphological parameter i of the ballast model and the macroscopic morphological parameter i of the real ballast, x i is the value of the macroscopic morphological parameter i of the real ballast; i represents the major axis L, the median axis I, the minor axis S, the needle index NI, and the flake index SI.

[0027] Furthermore, the ballast edge angle based on the ballast model and the ballast edge angle based on the real ballast are calculated in the same manner, specifically:

[0028]

[0029] Where A is the ballast edge angle; θ is the polar angle; Δθ is the polar angle interval; R(θ) and R(θ+Δθ) are the polar diameters corresponding to the polar angles θ and (θ+Δθ).

[0030] Furthermore, the area overlap Sa of the ballast model and the macroscopic morphological accuracy A of the ballast model are g , ballast model angle accuracy A r , and assign weights to construct the ballast model accuracy Am expression: Am=0.15*Ag+0.7*Sa+0.15*Ar.

[0031] The beneficial effects of the present invention are:

[0032] First, the present invention provides a modeling method for a bonded block ballast model formed by bonding rigid tetrahedrons. The bonded block ballast model can better simulate the sharp corners and concave features of real ballast, and can more accurately describe the complex shape of real ballast. Compared with the traditional ball cluster model, the bonded block ballast model is not only closer to the real ballast in shape, but also can more accurately simulate the actual contact conditions and ballast crushing forms of the ballast in terms of crushing mechanical behavior.

[0033] Secondly, the present invention proposes a method for quantifying the accuracy of ballast models, which can be widely used in the accuracy evaluation of ballast models; it balances the contradiction between the accuracy of ballast models and the computational cost, and the modeling parameters obtained by analyzing the influence of modeling parameters on model accuracy can meet the modeling accuracy requirements of most discrete ballast particles.

[0034] Third, in the field of railway engineering, the present invention provides a new idea for the existing ballast refinement modeling. The present invention can be used to establish a ballasted track numerical model, and to perform a more accurate and detailed analysis of the macro and micro mechanical properties of the ballasted track. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flow chart of the present invention;

[0036] Figure 2 Obtaining a three-dimensional contour schematic diagram of the real ballast for the present invention;

[0037] Figure 3 A schematic diagram of a basic ballast model is established for the present invention;

[0038] Figure 4 Obtaining a high-precision two-dimensional contour schematic diagram of the ballast model for the present invention;

[0039] Figure 5 A schematic diagram of the macroscopic morphological parameters of ballast defined in the present invention;

[0040] Figure 6 This is a schematic diagram of the influence of the modeling parameters of the present invention on the accuracy of the ballast model;

[0041] Figure 7 This is a schematic diagram of verifying the accuracy of the ballast model based on the acquired modeling parameters in the present invention;

[0042] Figure 8 This is a schematic diagram of uniaxial compression verification of the fine ballast model constructed in the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. It should be noted that the embodiments in this application and the features in the embodiments can be combined with each other arbitrarily without conflict.

[0044] Embodiment 1:

[0045] According to a first aspect of an embodiment of the present invention, a modeling method for a ballast model is provided, in which real ballast samples are randomly selected to obtain point cloud data of the surface contour of the real ballast; based on the point cloud data, a closed first STL contour file is obtained, and then the first STL contour file is imported into discrete element software to generate a rigid sub-block filled three-dimensional contour, thereby obtaining a ballast model formed by bonding rigid sub-blocks.

[0046] Furthermore, the first STL outline file is imported into the discrete element software to generate a three-dimensional outline of rigid sub-block filling, and a ballast model formed by bonding rigid sub-blocks is obtained. Specifically, the first STL outline file is imported into the discrete element software, and the geometric body to be filled is specified according to the discrete element software; then the sub-block type is specified as a tetrahedron; the minimum and maximum side lengths of the filled sub-blocks are specified to control the number N of filled sub-blocks; and the angle of adjacent triangular faces between two tetrahedral blocks is set.

[0047] Furthermore, the ratios of the minimum and maximum side lengths of the sub-block to the equivalent ballast particle size D based on the ballast model are defined as Rs and Rb respectively. Rs and Rb are used as modeling parameters to analyze the changes of the ballast model accuracy with Rs and Rb. The ballast model with an accuracy greater than the preset value is defined as a fine ballast model.

[0048] According to a second aspect of an embodiment of the present invention, a method for quantifying the accuracy of a ballast model is provided, comprising the following steps:

[0049] S1. Obtaining a two-dimensional profile of ballast; wherein the two-dimensional profile of ballast includes a two-dimensional profile based on a ballast model and a two-dimensional profile based on a real ballast;

[0050] S2. determining a first macroscopic morphological parameter based on a ballast model and based on real ballast;

[0051] S3, the area overlap of the ballast two-dimensional profile is introduced as the second macroscopic morphological parameter; the mean value S of the area overlap of the ballast two-dimensional profile obtained under each two-dimensional profile based on the ballast model is taken a As the area coincidence of the ballast model;

[0052] S4. According to the first macroscopic morphological parameters based on the ballast model and the first macroscopic morphological parameters based on the real ballast, the similarity of each first macroscopic morphological parameter of the ballast model is obtained; and the average value of the similarity of each first macroscopic morphological parameter is used as the macroscopic morphological accuracy A of the ballast model. g ;

[0053] S5. The similarity between the ballast edge angle based on the ballast model and the ballast edge angle based on the real ballast is taken as the edge angle accuracy of the ballast two-dimensional profile; the average of the edge angle accuracies of the ballast two-dimensional profile obtained under each two-dimensional profile of the ballast model is taken as the ballast model edge angle accuracy Ar;

[0054] S6, based on the area overlap Sa of the ballast model and the macroscopic morphological accuracy A of the ballast model g , the ballast model angle accuracy Ar, and assign weights to construct the ballast model accuracy Am expression.

[0055] Furthermore, the S1 includes:

[0056] The acquisition of the two-dimensional contour based on the real ballast is specifically as follows: randomly selecting a real ballast sample, obtaining the point cloud data of the surface contour of the real ballast, and obtaining a closed first STL contour file based on the point cloud data; then converting the first STL contour file into a surface entity; then converting the surface entity into an engineering drawing, and obtaining the two-dimensional contour of the ballast under three orthogonal views; deleting redundant lines from the two-dimensional contours under the three orthogonal views, and obtaining the two-dimensional contour based on the real ballast;

[0057] The acquisition of the two-dimensional contour based on the ballast model is specifically as follows: according to the established ballast model, the ballast model is exported as a second STL contour file; then the second STL contour file is converted into a surface entity; then the surface entity is converted into an engineering drawing to obtain the two-dimensional contour of the ballast under three orthogonal views; redundant lines inside the two-dimensional contour under the three orthogonal views are deleted to obtain the two-dimensional contour based on the ballast model.

[0058] Furthermore, the first macroscopic morphological parameters include the major axis L, the middle axis I, the minor axis S, the needle index NI, and the flake index SI of the ballast.

[0059] Furthermore, the 2D ballast profile area overlap Sc is expressed as:

[0060]

[0061] Among them, S f is the area of ​​the overlapping part of the two-dimensional profile based on the ballast model and the two-dimensional profile based on the real ballast, and Sr is the area of ​​the two-dimensional profile based on the real ballast.

[0062] Furthermore, the similarity of each first macroscopic morphological parameter of the ballast model is obtained based on the first macroscopic morphological parameter of the ballast model and the real ballast, and the expression is:

[0063]

[0064] Among them, ε i is the similarity of the macroscopic morphological parameters i of the ballast model, Δx i is the difference between the macroscopic morphological parameter i of the ballast model and the macroscopic morphological parameter i of the real ballast, x i is the value of the macroscopic morphological parameter i of the real ballast; i is the major axis L, the median axis I, the minor axis S, the needle index NI, and the flake index SI.

[0065] Furthermore, the ballast edge angle based on the ballast model and the ballast edge angle based on the real ballast are calculated in the same manner, specifically:

[0066]

[0067] Where A is the ballast edge angle; θ is the polar angle; Δθ is the polar angle interval.

[0068] Furthermore, the area overlap Sa of the ballast model and the macroscopic morphological accuracy A of the ballast model are g , ballast model angle accuracy A r , and assign weights to construct the ballast model accuracy Am expression: Am=0.15*Ag+0.7*Sa+0.15*Ar.

[0069] Embodiment 2:

[0070] like Figure 1-8As shown, according to the modeling method of the ballast model and the method for quantifying the accuracy of the ballast model provided by the present invention, the modeling of the railway ballast particle refinement model is realized, and the modeling parameters of the refined ballast model are obtained by analyzing the influence of the modeling parameters on the accuracy of the ballast model and considering the contradiction between the model accuracy and the calculation efficiency; the uniaxial compression DEM simulation is carried out in the discrete element software and compared with the actual experiment to verify the rationality of the established refined ballast model. The specific description is as follows:

[0071] 1. Select a real ballast sample and establish a ballast model consisting of rigid sub-blocks, including:

[0072] like Figure 2-Figure 3 As shown in the figure, a real ballast sample is randomly selected, and a 3D laser scanner is used to obtain the point cloud data of the real ballast surface contour. The point cloud data is imported into the reverse 3D software to obtain a closed first STL contour file, and then the first STL contour file is imported into the discrete element software. The discrete element software has a built-in R-block module. The rblockconstruct command is used to generate a rigid sub-block to establish a ballast model. Specifically, the from-geometry keyword is used to specify the geometry to be filled (the contour in the stl format imported from the discrete element software); the tetrahedral keyword is used to specify The stator block type is tetrahedron; the minimum-edge and maximum-edge keywords are used to specify the minimum and maximum edge lengths of the filling sub-blocks respectively, thereby controlling the number of filling sub-blocks N. The smaller the minimum-edge and maximum-edge values ​​are, the larger N is, and the closer the ballast model is to the real ballast; the patch-angle-tolerance keyword is used to control the angle of adjacent triangular faces between two tetrahedral blocks to improve the surface quality of the ballast model (the angle of adjacent triangular faces is set to 20°~45°, and 30° is taken in the embodiment of the present invention).

[0073] For example, Figure 3 As shown, by specifying the minimum and maximum side lengths of the filling sub-blocks and thus controlling the number of filling sub-blocks, two ballast models with 727 and 19529 sub-block numbers are generated respectively.

[0074] Experimental verification shows that the ballast model formed by bonding rigid tetrahedrons has higher shape accuracy than other sub-block types.

[0075] 2. Principles of Quantifying Ballast Model Accuracy

[0076] The ballast morphological characteristics are described, focusing on the macroscopic morphological characteristics and angular characteristics of the ballast, and a method for quantifying the accuracy of the ballast model based on the ballast macroscopic morphological accuracy and angular accuracy is proposed; the contradiction between the ballast model accuracy and the computational efficiency during numerical simulation is balanced, and the modeling parameters are determined by analyzing the influence of the modeling parameters on the ballast model accuracy, including:

[0077] refer to Figure 4 The present invention quantifies the accuracy of the ballast model from a two-dimensional perspective and proposes a method for obtaining a high-precision two-dimensional profile of the ballast, comprising:

[0078] The acquisition of the two-dimensional contour based on the real ballast is specifically as follows: a real ballast sample is randomly selected, and point cloud data of the surface contour of the real ballast is obtained by using a three-dimensional laser scanner, and the point cloud data is imported into the reverse three-dimensional software to obtain a closed first STL contour file; then the first STL contour file is imported into the three-dimensional software, and the imported first STL contour file is converted into a surface entity; then the surface entity is converted into an engineering drawing to obtain the two-dimensional contour of the ballast under three orthogonal views; the two-dimensional contour under the three orthogonal views is imported into Auto CAD, and redundant lines (the redundant lines are the lines located within the contour) are deleted to obtain the two-dimensional contour based on the real ballast.

[0079] The acquisition of the two-dimensional contour based on the ballast model is as follows: establish the ballast model in the discrete element software, and use the geometry export command to export it as a second STL contour file; then import the second STL contour file into the three-dimensional software. At this time, the imported second STL contour file is a graphic rather than an entity, and the three-dimensional software cannot recognize the features, so the graphic needs to be converted into a surface entity; then the surface entity is converted into an engineering drawing to obtain the two-dimensional contour of the ballast under three orthogonal views; import the two-dimensional contour under the three orthogonal views into Auto CAD, delete the redundant lines (the redundant lines are the lines within the contour), and obtain the two-dimensional contour based on the ballast model. Figure 4 An example is shown.

[0080] Description of macroscopic morphological characteristics:

[0081] The macroscopic morphological characteristics of ballast refer to its size and shape. Thanks to the development of software technology, the smallest outer envelope cuboid of ballast can be accurately obtained based on the surface entity, and then the major axis L, middle axis I, and minor axis S of the ballast are obtained to describe the size of the ballast, and the ballast needle index is derived. Flaky index Used to describe the shape of ballast. If only L, I, S, NI, and SI are used to describe the macroscopic shape of ballast, the constructed ballast model will be highly similar to the actual ballast in parameter values, but the actual shape is quite different. Therefore, the 2D ballast contour area overlap is introduced. The Sc mainly quantifies the overall filling of the ballast model to the real ballast, where S f is the area of ​​the overlap between the two-dimensional profile based on the ballast model and the two-dimensional profile based on the real ballast, Sr is the area of ​​the two-dimensional profile based on the real ballast; for the convenience of calculation, the average S of the area overlap of the two-dimensional profiles of the ballast obtained under the three two-dimensional profiles based on the ballast model is taken a As the area overlap of the ballast model. Figure 5 An example is shown.

[0082] Angular feature description:

[0083] The angular characteristics of ballast mainly refer to its irregular changes such as sharp edges, sharp corners and concave, which have a great influence on the force chain transmission of ballast aggregate. The present invention uses the Fourier morphological analysis method to describe the angular characteristics of ballast, specifically: first, the distance from the boundary of the two-dimensional contour of the ballast to the geometric center point of the contour is expanded using the Fourier series as follows:

[0084]

[0085] Among them, the 2D profile of ballast includes the 2D profile based on the ballast model and the 2D profile based on the real ballast; R(θ) is the polar diameter corresponding to the polar angle θ, a 0 is the average radius of the two-dimensional profile of the ballast (i.e., the average value of the distances from all points on the two-dimensional profile of the ballast to the geometric center point of the profile. In the embodiment of the present invention, the points on the two-dimensional profile of the ballast are divided into 1° intervals, i.e., a total of 360 points), a(n) and b(n) are Fourier coefficients, and n is the order.

[0086] The contour area integral is solved as:

[0087]

[0088] Define the contour equivalent radius:

[0089]

[0090] Define ballast characteristic parameters:

[0091]

[0092] In the formula, α s , α r , α t Characterize the characteristic factors of shape, edge angle and texture scale respectively.

[0093] The present invention modifies the method proposed by Al-Rousan et al. to obtain Fourier coefficients at discrete angles with known polar diameters, and derives a 0, a(n), b(n) expressions:

[0094]

[0095] Wherein, θ is the polar angle, 0≤θ≤2π; Δθ is the polar angle interval.

[0096] The final formula for calculating the ballast edge angle is:

[0097]

[0098] Quantifying the accuracy of the ballast model refers to quantifying the macroscopic morphological accuracy and angular accuracy of the ballast model respectively, and then assigning different weights to the macroscopic morphological accuracy and angular accuracy to obtain the accuracy of the ballast model. The quantification method of the macroscopic morphological parameters of the ballast model, the major axis L, the middle axis I, the minor axis S, the needle index NI, and the flake index SI is as follows:

[0099]

[0100] Among them, ε i is the similarity of the macroscopic morphological parameters i of the ballast model, Δx i is the difference between the macroscopic morphological parameter i of the ballast model and the macroscopic morphological parameter i of the real ballast, x i is the value of the macroscopic morphological parameter i of the real ballast; i is L, I, S, NI, SI; the mean of the similarity of the five parameters is defined as the macroscopic morphological accuracy A of the ballast model g ,Right now:

[0101]

[0102] The angular accuracy of the 2D ballast profile is the similarity between the angular angle of the ballast based on the ballast model and the angular angle of the ballast based on the real ballast. The angular accuracy of the ballast model is defined as A. r is the average of the angular accuracy of the 2D profile of the ballast obtained under the three 2D profiles of the ballast model, that is:

[0103]

[0104] Among them, A aj is the edge angle accuracy of the 2D ballast profile obtained under the j-th 2D profile based on the ballast model, j = 1, 2, 3; ΔA(j) is the difference between the ballast edge angle based on the ballast model and the ballast edge angle based on the real ballast under the j-th 2D profile, A(j) is the ballast edge angle based on the real ballast under the j-th 2D profile; the ballast edge angle based on the ballast model and the ballast edge angle based on the real ballast are calculated by Perform calculations.

[0105] According to the area overlap Sa of the ballast model and the macroscopic morphological accuracy A of the ballast model g , ballast model angle accuracy A r , and assign weights to construct the ballast model accuracy Am expression, which is:

[0106] Am=0.15*Ag+0.7*Sa+0.15*Ar

[0107] From the above, we can know that the weight assigned to Sa is 0.7, the weight assigned to Ag is 0.15, and the weight assigned to Ar is 0.15. Based on this design, the accuracy of the ballast model can be better described.

[0108] In order to balance the contradiction between the accuracy of the ballast model and the computational efficiency in numerical simulation, the modeling parameters are determined by analyzing their influence on the accuracy of the ballast model, including:

[0109] In order to reduce the number of inefficient contacts between points and points and between edges in the ballast model, which have negligible contribution to resisting the deformation of the ballast model, rblockerode is used for rounded corner corrosion when the bonding bond is activated. The corrosion factor η is set to 0.02, which can reduce a large number of inefficient contacts without affecting the overall morphology of the ballast model. The computational efficiency of the discrete element software is subject to the number of contacts. Analysis shows that the number of contacts in the ballast model is strongly linearly related to the number of sub-blocks N. Therefore, reducing some inefficient contacts will help improve the computational efficiency.

[0110] Considering that the actual ballast size is different, in order to establish the relationship between the sub-block size and the ballast size and thus simplify the selection of the sub-block size during modeling, the minimum and maximum side lengths of the sub-block and the equivalent particle size of the ballast based on the ballast model are defined. The ratios of are Rs and Rb respectively. Rs and Rb are used as modeling parameters to analyze the change of ballast model accuracy with Rs and Rb. The ballast model with an accuracy greater than 95% is defined as a fine ballast model, and Rs and Rb are taken as 0.2 and 0.5 respectively.

[0111] Combination Figure 6 The acquisition process of the above parameters is further described: for the selected ballast, Rb is set to 0.5 according to the maximum side length and the equivalent particle size of the ballast, and the accuracy analysis of the ballast model of the present invention is performed for different values ​​of the minimum side length. Figure 6 It can be seen that when the accuracy is greater than 95%, Rs and Rb are selected with the minimum number of sub-blocks N, that is, Rs and Rb are taken as 0.2 and 0.5 respectively.

[0112] It should be noted that the larger N is, the closer the ballast model is to the real ballast, but the lower the calculation efficiency will be. In order to balance the contradiction between the accuracy of the ballast model and the calculation efficiency in the numerical simulation, when the accuracy is greater than 95%, Rs and Rb with the minimum number of sub-blocks N can be selected to achieve the above purpose.

[0113] Furthermore, the present invention uses a uniaxial compression and crushing test of ballast to analyze the mechanical properties of the ballast model, performs uniaxial compression DEM simulation in discrete element software and compares it with actual experiments to verify the rationality of the established fine ballast model, specifically: actual experiment: select ballast, perform uniaxial compression and crushing test, and record the force-displacement curve of the press loading plate and the ballast crushing form during the test; DEM simulation: establish a corresponding fine ballast model in discrete element software based on the obtained fine modeling parameters Rs and Rb, and perform uniaxial compression DEM simulation, record the force-displacement curve of the loading wall and the ballast crushing form and compare them with the actual experiment, so as to calibrate the contact parameters of the ballast model, so that the mechanical properties of the ballast model are consistent with the real ballast; it should be noted that the contact parameters of the ballast model corresponding to ballasts of different shapes and sizes are different, which are mainly manifested in the differences in bonding strength and tensile strength. Attention should be paid to the differentiated calibration of these parameters during modeling.

[0114] Furthermore, the optional specific implementation modes of the present invention are described as follows:

[0115] Step 1: Randomly select a ballast and use a 3D laser scanner to obtain the 3D contour point cloud data of the real ballast. In order to improve the calculation efficiency, the point cloud data needs to be simplified. Based on the simplified point cloud data, a closed 3D contour surrounded by several triangular faces is generated in the reverse 3D software and saved as an STL format file, namely the first STL contour file.

[0116] Step 2: Obtain the various parameter values ​​of the real ballast. Import the real ballast STL format file into Solidworks, use the envelope command to obtain the ballast major axis L, middle axis I, and minor axis S, and calculate the ballast equivalent particle size based on the real ballast. And according to The shape parameters of the real ballast are calculated, and the L, I and S of the ballast are measured to be 67.25 mm, 59.47 mm and 49.47 mm, respectively. The corresponding NI index is 1.13 and SI index is 0.83, and the ballast is classified as block ballast according to its shape. The method for obtaining the high-precision two-dimensional profile of ballast proposed in the present invention is used to obtain the two-dimensional profile based on the real ballast, and the area of ​​the first two-dimensional profile based on the real ballast is calculated to be 2528.52 mm 2 , based on the second 2D contour area of ​​the real ballast is 2351.98mm 2, based on the third 2D contour area of ​​the real ballast is 3151.17mm 2 ; Further, using the method for calculating the ballast edge angle proposed in the present invention, it is calculated that the ballast edge angle based on the first two-dimensional contour of the real ballast is 0.0168, the ballast edge angle based on the second two-dimensional contour of the real ballast is 0.0045, and the ballast edge angle based on the third two-dimensional contour of the real ballast is 0.0112.

[0117] Step 3: Determine the modeling parameters of the fine ballast model. Specifically, for the selected sample ballast, the relationship between the number of sub-blocks and the number of contacts between sub-blocks is first analyzed, and the corresponding relationship between the number of contacts between sub-blocks Y and the number of sub-blocks N is obtained as Y = -56.35 + 1.87N, which is approximately linear. The calculation efficiency of the discrete element software is subject to the detection of the number of contacts, so it is necessary to reduce the existence of some contacts; when activating the bonding bond between sub-blocks, rblockerode is used for rounded corner corrosion, and the corrosion factor η is set to 0.02, which can reduce the number of inefficient contacts by about 92.5% without affecting the morphology of the ballast model.

[0118] Furthermore, the values ​​of the minimum-edge and maximum-edge of the sub-block and the equivalent particle size D are defined as Rs and Rb respectively. Rs and Rb are used as modeling parameters to study the influence of changing their values ​​on the accuracy of the ballast model. The influence of changing Rs and Rb on the number N of sub-blocks in the ballast model is analyzed respectively. The analysis shows that N is more sensitive to the change of Rs. Therefore, Rb is set to a larger value (0.5). The purpose of this is to use as few sub-blocks as possible to fill the model. The value of Rs is changed to analyze the influence of different modeling parameters on the number and accuracy of sub-blocks in the ballast model. The results are shown in Figure 2. Figure 6 As shown in FIG. 1 , overall, the macroscopic morphological accuracy, angular accuracy, and ballast model accuracy of the ballast model all decrease with the increase of Rs. In order to balance the contradiction between model accuracy and computational efficiency, a strategy is adopted to minimize the number of sub-blocks when the model accuracy meets certain conditions. The present invention identifies the ballast model with an accuracy greater than 95% as a fine ballast model, so the maximum Rs is 0.2, and the modeling parameters of the fine ballast model are Rs=0.2, Rb=0.5; further, the number of filling sub-blocks N can be determined, for example, as Figure 6 The N shown below is 503.

[0119] Step 4: Establish a fine ballast model based on the modeling parameters of the fine ballast model. Import the first STL contour file into the discrete element software, use the rblockconstruct command to generate a rigid sub-block to establish the ballast model, use the from-geometry keyword to specify the geometry to be filled; use the tetrahedral keyword to specify the sub-block type as a tetrahedron; use the minimum-edge and maximum-edge keywords to specify the minimum and maximum edge lengths of the filled sub-blocks respectively to control the number of filled sub-blocks N; use the patch-angle-tolerance keyword to control the angle of adjacent triangular faces between two tetrahedral blocks to improve the surface quality of the ballast model. The parameters of the obtained fine ballast model are: L of the fine ballast model is 65.26mm, I is 58.44mm, S is 49.95mm, the corresponding NI index is 1.117, and SI index is 0.855; the first two-dimensional contour area based on the fine ballast model is 2438.78mm 2 The second 2D contour area based on the fine ballast model is 2266.07 mm 2 The third 2D contour area based on the fine ballast model is 3046.99 mm 2 ; Further, using the method for calculating the ballast edge angle proposed in the present invention, it is calculated that the ballast edge angle of the first two-dimensional contour based on the fine ballast model is 0.0148, the ballast edge angle of the second two-dimensional contour based on the fine ballast model is 0.0052, and the ballast edge angle of the third two-dimensional contour based on the fine ballast model is 0.0106.

[0120] Step 5: Verify model accuracy. Figure 7 As shown in the figure, the quantification method proposed in the present invention is to quantify the accuracy of the ballast model based on three randomly selected two-dimensional contours. The selection of contours has a great influence on the evaluation of model accuracy. Therefore, the ballast model is rotated around the x-axis and the z-axis at intervals of 30 degrees to obtain two-dimensional contours at different angles. The two-dimensional contours are compared with the real ballast two-dimensional contours at the corresponding angles to verify the feasibility of using random two-dimensional contours to quantify the accuracy of the ballast model. Furthermore, about 15 block, needle and sheet ballasts of different sizes are randomly selected, and the corresponding fine ballast models are established based on the obtained modeling parameters, and their accuracy is quantified. The results show that the accuracy of most ballast models can reach a high level of accuracy, which verifies the universality of this modeling method in establishing fine ballast models.

[0121] Step 6: Verify whether the mechanical properties of the fine ballast model can accurately simulate the crushing response of the real ballast, and obtain the microscopic contact parameters of the fine ballast model. The uniaxial compression crushing test and DEM numerical simulation are designed. The test results are as follows: Figure 8 The specific steps include:

[0122] Uniaxial compression test:

[0123] ①The main equipment used is a microcomputer-controlled multi-purpose pressure testing machine with a maximum test force of 100kN and computer data acquisition equipment;

[0124] ②Select ballast samples and scan them to obtain their high-precision three-dimensional profiles;

[0125] ③ Wrap the ballast with plastic film to prevent the fragments generated during the loading process from splashing, and place the ballast stably in the center of the loading base plate, with its long axis as parallel to the loading base plate as possible;

[0126] ④ Slowly move the loading plate so that it just contacts the top of the ballast, apply a force of about 200N to fix the ballast, and then apply a constant loading speed of 0.6mm / min to the loading plate to apply an axial load to the ballast. The computer records the force-displacement curve of the loading plate in real time, and observes the crushing of the ballast with the naked eye. Stop loading when the force-displacement curve drops sharply or the ballast is obviously damaged. The force-displacement curve and the ballast crushing form can accurately reflect the crushing mechanical behavior of the ballast, and use them as the main basis for calibrating the ballast model parameters.

[0127] DEM numerical simulation:

[0128] ① Establish the corresponding fine ballast model in discrete element software;

[0129] ② Generate corresponding fixed walls and movable walls in the discrete element software, which correspond to the loading base plate and loading plate in the actual test respectively; adjust the relative position of the ballast model and the wall to be consistent with the actual test;

[0130] ③ After the ballast model and the fixed wall are in stable contact, a loading rate of 0.1 m / s is applied to the moving wall, the force-displacement curve of the moving wall is monitored, and the ballast crushing form is observed; through the trial and error method, the internal mesoscopic contact parameters of the ballast model are continuously adjusted so that the force-displacement curve of the moving wall and the crushing results of the fine ballast model are consistent with the actual test, and the mesoscopic contact parameters are calibrated; the details are shown in Table 1:

[0131] Table 1 DEM parameters of fine ballast model

[0132]

[0133] ④ It should be noted that the ballast particles have large differences in shape and size, and their corresponding contact parameters are not fixed and need to be adjusted according to actual conditions. The parameter differences are mainly reflected in the bonding strength and tensile strength between sub-blocks.

[0134] pass Figure 8It can be seen that the fine ballast model established by the method of the present invention has a high degree of consistency with the crushing mechanical characteristics of the real ballast, indicating that the fine model established by the present invention is relatively consistent with the real ballast and can be further used in research such as numerical simulation of railway engineering.

[0135] The specific implementation modes of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above implementation modes, and various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

Claims

1. A ballast modeling method, characterized in that: Real ballast samples are randomly selected to obtain point cloud data of the real ballast surface contour. Based on the point cloud data, a closed first STL contour file is obtained, and then the first STL contour file is imported into the discrete element software to generate a rigid sub-block filled three-dimensional contour, and a ballast model formed by bonding rigid sub-blocks is obtained.

2. The ballast modeling method according to claim 1, characterized in that: The first STL outline file is imported into the discrete element software, and a three-dimensional outline of rigid sub-block filling is generated to obtain a ballast model formed by bonding rigid sub-blocks. Specifically, the first STL outline file is imported into the discrete element software, and the geometric body to be filled is specified according to the discrete element software; then the sub-block type is specified as a tetrahedron; the minimum and maximum side lengths of the filled sub-blocks are specified to control the number N of filled sub-blocks; and the angles of adjacent triangular faces between two tetrahedral blocks are set.

3. The method for modeling a ballast model according to claim 1, characterized in that: The ratios of the minimum and maximum side lengths of the sub-block to the equivalent ballast particle size D based on the ballast model are defined as Rs and Rb respectively. Rs and Rb are used as modeling parameters to analyze the changes of the ballast model accuracy with Rs and Rb. The ballast model with an accuracy greater than the preset value is defined as a fine ballast model.

4. A method for quantifying the accuracy of a ballast model, characterized in that: The following steps are involved: S1. Obtaining a two-dimensional profile of ballast; wherein the two-dimensional profile of ballast includes a two-dimensional profile based on a ballast model and a two-dimensional profile based on a real ballast; S2. determining a first macroscopic morphological parameter based on a ballast model and based on real ballast; S3, introduce the area overlap of the 2D ballast profile; take the average value S of the area overlap of the 2D ballast profiles obtained under each 2D profile based on the ballast model a As the area coincidence of the ballast model; S4. According to the first macroscopic morphological parameters based on the ballast model and the first macroscopic morphological parameters based on the real ballast, the similarities of the first macroscopic morphological parameters of the ballast model are obtained; and the average of the similarities of the first macroscopic morphological parameters is used as the macroscopic morphological accuracy Ag of the ballast model; S5. The similarity between the ballast edge angle based on the ballast model and the ballast edge angle based on the real ballast is taken as the edge angle accuracy of the ballast two-dimensional profile; the average of the edge angle accuracies of the ballast two-dimensional profile obtained under each two-dimensional profile of the ballast model is taken as the ballast model edge angle accuracy Ar; S6. Based on the area overlap Sa of the ballast model, the macroscopic morphology accuracy Ag of the ballast model, the angular accuracy Ar of the ballast model, and assigning weights, construct an expression for the ballast model accuracy Am.

5. The method for quantifying the accuracy of ballast model according to claim 4, characterized in that: The S1 includes: The acquisition of the two-dimensional contour based on the real ballast is specifically as follows: randomly selecting a real ballast sample, obtaining the point cloud data of the surface contour of the real ballast, and obtaining a closed first STL contour file based on the point cloud data; then converting the first STL contour file into a surface entity; then converting the surface entity into an engineering drawing, and obtaining the two-dimensional contour of the ballast under three orthogonal views; deleting redundant lines from the two-dimensional contours under the three orthogonal views, and obtaining the two-dimensional contour based on the real ballast; The acquisition of the two-dimensional contour based on the ballast model is specifically as follows: according to the established ballast model, the ballast model is exported as a second STL contour file; then the second STL contour file is converted into a surface entity; then the surface entity is converted into an engineering drawing to obtain the two-dimensional contour of the ballast under three orthogonal views; redundant lines inside the two-dimensional contour under the three orthogonal views are deleted to obtain the two-dimensional contour based on the ballast model.

6. The method for quantifying the accuracy of ballast model according to claim 4, characterized in that: The first macroscopic morphological parameters include the major axis L, the central axis I, the minor axis S, the needle index NI, and the flake index SI of the ballast.

7. The method for quantifying the accuracy of ballast model according to claim 4, characterized in that: The 2D ballast profile area overlap Sc is expressed as: Among them, S f is the area of ​​the overlapping part of the two-dimensional profile based on the ballast model and the two-dimensional profile based on the real ballast, and Sr is the area of ​​the two-dimensional profile based on the real ballast.

8. The method for quantifying the accuracy of ballast model according to claim 4, characterized in that: The similarity of each first macroscopic morphological parameter of the ballast model is obtained based on the ballast model and the first macroscopic morphological parameter of the real ballast, and the expression is: Among them, ε i is the similarity of the macroscopic morphological parameter i of the ballast model, Δxi is the difference between the macroscopic morphological parameter i of the ballast model and the macroscopic morphological parameter i of the real ballast, x i is the value of the macroscopic morphological parameter i of the real ballast; i represents the major axis L, the median axis I, the minor axis S, the needle index NI, and the flake index SI.

9. The method for quantifying the accuracy of ballast model according to claim 4, characterized in that: The ballast edge angle based on the ballast model and the ballast edge angle based on the real ballast are calculated in the same way, specifically: Where A is the ballast edge angle; θ is the polar angle; Δθ is the polar angle interval; R(θ) and R(θ+Δθ) are the polar diameters corresponding to the polar angles θ and (θ+Δθ).

10. The method for quantifying the accuracy of ballast model according to claim 4, characterized in that: According to the area overlap Sa of the ballast model, the macroscopic morphology accuracy Ag of the ballast model, the angular accuracy Ar of the ballast model, and the weights are assigned, the expression of the ballast model accuracy Am is constructed as follows: Am=0.15*Ag+0.7*Sa+0.15*Ar.

Images