Discrete element contact model parameter rapid calibration method for asphalt mixture mixing simulation

Abstract

The invention provides a discrete element contact model parameter rapid calibration method for asphalt mixture mixing simulation. The discrete element contact model parameter rapid calibration method comprises the following steps: S1, constructing a particle discrete element model and calibrating basic contact parameters of aggregate particles; s2, constructing an adhesion parameter prediction model based on asphalt viscosity and asphalt film thickness; and S3, calibrating the adhesion parameter of the aggregate in the virtual mixing test based on the adhesion parameter prediction model. According to the method, the problems that the existing asphalt mixture mixing simulation discrete element model contact parameter calibration process is dispersed and isolated, the calibration difficulty is high, the measurement precision is low, the calibration parameter coupling is high, the calibration process is complex and blind, and the calibration result lacks expansibility are solved, rapid calibration of the contact parameters can be realized, and the calibration efficiency is improved. And meanwhile, rapid prediction of adhesion parameters of different aggregates at different temperatures and asphalt types can be realized, and the calibrated contact parameters can be used for constructing a mixing simulation discrete element model to further research particle movement behaviors in the mixing process.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of numerical simulation of asphalt concrete discrete element mixing process, and particularly relates to a discrete element contact model parameter rapid calibration method for asphalt mixture mixing simulation. BACKGROUND

[0002] The mixing process is a key stage for the formation of asphalt mixture uniformity and is a source guarantee for pavement quality control, directly affecting the distribution state of aggregate and asphalt and further affecting the spatial structure and macroscopic performance of the mixture. Therefore, in-depth understanding of the mixing mechanism is crucial for theoretical breakthrough and practical optimization of road engineering.

[0003] However, the existing research methods face severe challenges in revealing the dynamic mixing process. Traditional test methods can only obtain sample information at the static or termination point, and cannot implement real-time, in-situ and continuous monitoring and quantitative analysis of the real motion trajectory, collision characteristics and structure evolution of particles in the mixing equipment operation. This makes it extremely difficult to establish a dynamic, reliable and quantitative correlation model between mixing parameters, material characteristics and mixture uniformity.

[0004] With the development of computer numerical simulation technology, the discrete element method has become a core numerical simulation tool for studying the mixing process of asphalt mixture, which can dynamically analyze the motion trajectory, velocity field distribution and energy transfer characteristics of particles in the mixing process, and realize the quantitative characterization of particle-particle and particle-boundary interaction at the microscale. However, existing research lacks modeling and contact parameter calibration methods for discrete element models for asphalt mixture mixing simulation. The accuracy of the discrete element model contact parameters significantly affects the accuracy of the simulation, and the core bottleneck is the lack of an efficient, accurate and systematic discrete element model contact parameter calibration method.

[0005] Existing contact parameter calibration methods mainly calibrate basic contact parameters of particles such as the restitution coefficient, static friction coefficient and rolling friction coefficient. Collision tests and friction tests are used for calibration, and the calibration process is often scattered and isolated, and generally relies on a tedious trial-and-error process or limited physical tests. The calibration process is time-consuming, and the calibration accuracy is greatly affected by single particle characteristics. In the calibration system involving adhesion parameters, existing methods use a mixed calibration method for basic contact parameters and adhesion parameters, but lack clear physical basis support in the calibration process, resulting in strong parameter coupling, complex and blind calibration process, and calibration results only applicable to a single simulation system. Each simulation system needs to be calibrated repeatedly, and the results lack scalability. Therefore, it is of great significance to break through the existing calibration method and invent a discrete element contact model parameter rapid calibration method for asphalt mixture mixing simulation. SUMMARY

[0006] In view of the above problems, the present application provides a discrete element contact model parameter rapid calibration method for asphalt mixture mixing simulation.

[0007] The object of the present application can be achieved by the following technical solutions: A discrete element contact model parameter rapid calibration method for asphalt mixture mixing simulation, the discrete element contact model is a Hertz-Mindlin with JKR contact model, the corresponding contact parameters include basic contact parameters and adhesion parameters, wherein the basic contact parameters include the restitution coefficient, the static friction coefficient and the rolling friction coefficient; the calibration of the contact parameters includes the following steps: S1: particle discrete element model construction and aggregate particle basic contact parameter calibration, including the following steps: S1.1: selecting aggregate with a particle size range of 2.36-4.75 mm, carrying out an indoor real aggregate natural slope angle test in a cubic container with a length:width:height ratio of 1:1:1 and an open top, and determining the real aggregate natural slope angle; S1.2: obtaining a particle discrete element model of the real aggregate topography with a particle size range of 2.36-4.75 mm, and determining the particle generation ratio; S1.3: standardizing the size of the particle discrete element model; S1.4: generating particles in the discrete element natural slope angle test device model and preliminarily processing to obtain a preliminarily processed virtual natural slope angle test model; the discrete element natural slope angle test device model is a cubic model with a length:width:height ratio of 1:1:1 and an open top surface; S1.5: determining a test scheme of the response surface design method; S1.6: based on the preliminarily processed virtual natural slope angle test model, carrying out a virtual natural slope angle test according to the test scheme of step S1.5, and determining a response surface model, which is expressed as: In the formula: Angle is the natural slope angle; restitution is the restitution coefficient; static is the static friction coefficient; rolling is the rolling friction coefficient; k 、 k 1、 k 2、 k 3 is a model coefficient; S1.7: determining the basic contact parameters of the aggregate particles according to the real aggregate natural slope angle determined in step S1.1 and the response surface model determined in step S1.6; S2: Using the basic contact parameters calibrated in step S1, a bitumen viscosity and bitumen film thickness-based adhesion parameter prediction model is constructed, which is expressed as: wherein: DA is the adhesion parameter; Angle is the bitumen film thickness; min b is the relative density of the bitumen; V is the rotational viscosity of the bitumen at the mixing temperature, u , u 1、 u 2、 u 3、 u 4、 u 5、 v 、 v 1 、v 2is the model coefficient; S3: Based on the adhesion parameter prediction model constructed in step S2, the adhesion parameters of the aggregate in the virtual mixing test are calibrated.

[0008] Specifically, step S1.1 is: filling the transparent glass box with a length-width-height ratio of 1:1:1 and an open top with clean and dry aggregate with a particle size range of 2.36-4.75 mm, the length and width of the transparent glass box are greater than 50 mm; carrying out a real aggregate natural slope angle test, the specific steps are: rotating the transparent glass box with the bottom surface edge as the pivot clockwise at a uniform speed, the rotation angle is 90°, the particles form a natural slope angle under the action of gravity, the image is shot and the angle of the real aggregate natural slope angle is obtained, denoted as Angle 0.

[0009] Specifically, step S1.2 is: obtaining the three-dimensional topographic features of the aggregate with a particle size range of 2.36-4.75 mm, drawing a shape parameter distribution curve, determining the shape parameter distribution range according to the 5% and 95% quantile points of the shape parameter cumulative distribution frequency curve, uniformly dividing into N intervals, N>3, selecting a representative particle in each interval, the shape parameter of the representative particle is close to the interval median value, a total of N particles are obtained. Real topography, import discrete element software and use the built-in template filling algorithm to obtain the particle discrete element model, calculate the cumulative distribution frequency of the particles in each interval range as the generation proportion of the particles. Specifically, step S1.3 is: according to the width of the particle discrete element model, the particle discrete element model scaling ratio is converted, the width of all particle discrete element models is unified to a mm, 2.36<a<4.75, the calculation formula of the particle discrete element model scaling ratio is: c = a / b(1) wherein: c is the scaling of the particle discrete element model; a is the width of the particle discrete element model; and b is the actual width of the particle discrete element model.

[0010] Specifically, step S1.4 is that the discrete element natural slope angle test device model is a cubic model with a size ratio of length: width: height of 1:1:1 and an unsealed top surface, and the specific dimensions of the length, width and height are the same as those of the transparent glass box used in the real aggregate natural slope angle test in step S1.1, and are used to perform a virtual natural slope angle test. Specifically, a closed cube with a size ratio of length: width: height in the range of 1:1:1.5-1:1:2 is constructed in the discrete element software, and the closed cube is filled with particles at random positions and directions, the length and width of the closed cube are the same as those of the transparent glass box used in the real natural slope angle test, and the generation ratio of different particles is determined according to the cumulative distribution frequency of step S1.2, the particles are naturally stacked under the action of virtual gravity, the size ratio of the length: width: height of the cube is adjusted to 1:1:1, and the top surface of the cube is set to be unsealed, and the particles exceeding the top surface are scraped flat to flatten the top surface; the contact model of the particles is set to Hertz-Mindlin with JKR contact model, and a preliminary processed virtual natural slope angle test model is obtained, wherein the contact parameters of the Hertz-Mindlin with JKR contact model include basic contact parameters and adhesion parameters, the basic contact parameters include a restitution coefficient, a static friction coefficient and a rolling friction coefficient, and the adhesion parameter of the contact model is a JKR surface energy γ.

[0011] Specifically, step S1.5 includes: S1.5.1: setting the adhesion parameter of the Hertz-Mindlin with JKR contact model, i.e., the JKR surface energy γ, to 0, and setting the value range of the basic contact parameters to be calibrated; S1.5.2: based on the response surface design-central composite (CCD) design method, listing the high, medium and low level values of the parameters to be calibrated, and setting a three-factor five-level test scheme, wherein the three factors refer to the restitution coefficient, the static friction coefficient and the rolling friction coefficient, and the response value in the response surface design method is the virtual natural slope angle.

[0012] Further, the value range of the basic contact parameters to be calibrated is obtained by trial and error method, and the specific steps are as follows: First, each basic contact parameter of the Hertz-Mindlin with JKR contact model in step S1.4 is set to a smaller random initial value maxBased on the preliminarily processed virtual natural slope angle test model, a virtual natural slope angle test is carried out. The specific steps of the virtual natural slope angle test are as follows: the cube is rotated clockwise at a constant speed with the bottom surface edge as the rotation axis, and the rotation angle is 90°. The particles form a natural slope angle under the action of gravity, and the natural slope angle is obtained by linear fitting, which is recorded as Angle 1; Then, set each basic contact parameter of the contact model in step S1.4 to a large random initial value Angle , carry out virtual natural slope angle test and obtain the corresponding natural slope angle, which is recorded as Angle 2; If the angle of the actual aggregate natural slope angle obtained in step S1.1 Angle 0 is located min 1 and max 2, the range of the basic contact parameters is ( min - max ), otherwise for random initial values Angle and restitution Make adjustments and repeat the above steps.

[0013] Specifically, step S1.6 is as follows: referring to the three-factor, five-level experimental scheme set in step S1.4, the contact parameters in the Hertz-Mindlin with JKR contact model are set. The cube is rotated clockwise at a constant speed of 90° with the edge of the bottom surface as the axis of rotation. The particles form a natural slope angle under the action of gravity. The natural slope angle is obtained by linear fitting, and the response surface model of the natural slope angle is determined, which is expressed as: (2) Where: static is the natural slope angle; rolling is the coefficient of restitution; DA is the static friction coefficient; Angle is the rolling friction coefficient; k 、 k 1. k 2. k 3 is the model coefficient.

[0014] Furthermore, step S2 includes the following steps: S2.1: Determine the viscosity-temperature curve of the asphalt being calibrated.

[0015] S2.2: Determine asphalt film thickness.

[0016] The calculation method is: (3) Where: my is the asphalt film thickness; Pbe is the oil-stone ratio; Angle b is the relative density of asphalt, 0.082 is the asphalt film thickness coefficient corresponding to the aggregate particle with a particle size range of 2.36-4.75 mm.

[0017] S2.3: Perform a virtual natural slope angle test based on the preliminary processed virtual natural slope angle test model, and establish a fitting equation between the adhesion parameter and the virtual natural slope angle.

[0018] Specifically, the basic contact parameters in the preliminary processed virtual natural slope angle test model are set as the basic contact parameters calibrated in step S1; The adhesion parameter is set to be a plurality of different values, so as to determine a plurality of test schemes; Based on the above test scheme, the cube is rotated at a uniform speed clockwise around the edge of the bottom surface as the rotation shaft, the rotation angle is 90°, the particles form a natural slope angle under the action of gravity, the virtual natural slope angle is obtained by linear fitting, and a fitting equation between the adhesion parameter and the virtual natural slope angle is established, which is expressed as: DA = Angle + n (4) In the formula: DA is the natural slope angle; Figure 1 is the adhesion parameter; m , n is a model coefficient.

[0019] S2.4: Perform a real aggregate natural slope angle test, obtain aggregate natural slope angles at different temperatures and oil-stone ratios, and construct a multivariate fitting model of temperature, oil-stone ratio and natural slope angle, which is expressed as: (5) In the formula: Figure 2 is the natural slope angle; P be is the oil-stone ratio; T is the asphalt mixing temperature; j , j 1, j 2, j 3, j 4, j 5 is a model coefficient.

[0020] S2.5: Establish an adhesion parameter prediction model based on asphalt viscosity and asphalt film thickness; Specifically, the fitting equation (formula (4)) between the adhesion parameter established based on step S2.3 and the virtual natural slope angle and the multivariate fitting model (formula (5)) of temperature, oil-rock ratio and natural slope angle constructed in step S2.4 are combined with the relationship (formula (3)) between the viscosity-temperature curve measured in step S2.1 and the relationship between the asphalt film thickness and the oil-rock ratio determined in step S2.2 to establish an adhesion parameter prediction model based on the asphalt viscosity and the asphalt film thickness, which is represented as: (6) In the formula: Figure 4 is the adhesion parameter; Figure 5 is the asphalt film thickness; Figure 1 b is the relative density of the asphalt; V is the rotational viscosity of the asphalt at the mixing temperature, u 、 u 1、 u 2、 u 3、 u 4、 u 5、 v 、 v 1 、v 2 are model coefficients.

[0021] Further, step S3 includes the following steps: S3.1: Calculate the average asphalt film thickness of the mixture at the optimal asphalt dosage.

[0022] S3.2: Determine the asphalt viscosity at the mixing temperature corresponding to the virtual mixing test to be carried out based on the asphalt viscosity-temperature curve.

[0023] S3.3: Substitute the average asphalt film thickness obtained in step S3.1 and the asphalt viscosity obtained in step S3.2 into the adhesion parameter prediction model (formula (6)) based on the asphalt viscosity and the asphalt film thickness established in step S2 to obtain the adhesion parameter of the particles at the mixing temperature corresponding to the virtual mixing test to be carried out.

[0024] Preferably, the generation proportion of the particle discrete element model used in step S1.1 is determined according to the actual particle proportion in different shape ranges, and representative particles with significant topographic differences are selected between different particles. The accuracy of the parameter calibration result is positively correlated with the accuracy of the initial particle type number and proportion.

[0025] Preferably, the leveling of the top surface in step S1.3 is recommended to use the leveling method of transverse scraping by an additional virtual scraper model, and the leveling method of longitudinal flattening by a virtual pressing block model cannot be used.

[0026] Further, the rapid calibration method proposed by the present application is also applicable to the discrete element contact model parameter calibration of the following mixing systems: recycled asphalt mixture mixing system, warm mix asphalt mixture mixing system.

[0027] The present application proposes a rapid calibration method for the discrete element contact model for asphalt mixture mixing simulation, which considers the influence of asphalt contact adhesion behavior and aggregate particle topography on contact parameters. By simulating the geometric features of aggregate, based on the rapid natural slope angle simulation test and indoor test means, the basic contact parameters of aggregate particles and the adhesion parameters of coated asphalt aggregate are calibrated. The present application can be used for rapid calibration of contact parameters for conventional hot mix asphalt mixture mixing simulation. In addition, it is also applicable to recycled asphalt mixture mixing system and warm mix asphalt mixture mixing system, and is used for simulation and analysis of particle motion and distribution behavior in the mixing process. Compared with the prior art, the present application has the following characteristics and advantages: 1) In the process of conventional parameter calibration, the particles used in the virtual test are spherical particles, ignoring the influence of aggregate topography on the parameter calibration results. The present application method selects 2.36-4.75 mm aggregate for parameter calibration, which fully reflects the influence of aggregate topography while ensuring the reliability of the measurement results, and realizes the rapid calibration of the basic contact parameters of aggregate based on the real geometric topography information of aggregate.

[0028] 2) In the conventional adhesion parameter calibration method, the basic contact parameters and adhesion parameters are calibrated together, and there is a lack of clear physical basis in the calibration process, resulting in strong parameter coupling, complex and blind calibration process. The present application method is based on the key theoretical assumption that the basic contact parameters are mainly controlled by the inherent topography of aggregate particles, and the film thickness and viscosity of asphalt as the main body of adhesion dominate the adhesion parameter size. The complex parameter calibration process is divided into two logically clear and physically clear steps: first, calibrate the basic contact parameters of aggregate, and then calibrate the adhesion parameters of asphalt.

[0029] 3) The step-by-step calibration of the present application reduces the dimension and complexity of each stage of calibration, simplifies the calibration process, avoids the calibration difficulty and inaccuracy caused by the interaction of multiple parameters in the traditional method, and improves the calibration efficiency and accuracy.

[0030] 4) The present application method establishes a reliable adhesion parameter prediction model based on physical mechanism, which can realize the rapid calibration of adhesion parameters of the discrete element contact model for mixing simulation of different asphalt types and mixing temperatures without repeated experimental calibration, and solves the problem of non-reproducible and difficult to migrate application of the calibration results of the traditional method.

[0031] 5) The rapid calibration method proposed by the present application is also applicable to recycled asphalt mixture mixing system, warm mix asphalt mixture mixing system, and has strong application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Angle The present invention is a flow chart of a method for rapid calibration of discrete element contact model parameters for asphalt mixture mixing simulation.

[0033] Figure 2 is the particle shape parameter distribution curve in the embodiment of the present invention.

[0034] FIG3( a ) is a front view of a virtual natural slope angle test model in a natural accumulation state according to an embodiment of the present invention.

[0035] FIG3( b ) is an oblique view of a virtual natural slope angle test model in a natural accumulation state according to an embodiment of the present invention.

[0036] Figure 4 Schematic diagram of a method for obtaining a natural slope angle in an embodiment of the present invention.

[0037] Angle : is a fitting image between the adhesion parameter and the virtual natural slope angle in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The technical solution provided by this application will be further described below in conjunction with specific embodiments and accompanying drawings. The advantages and features of this application will become more apparent with reference to the following description.

[0039] Example like DA As shown in FIG, a method for rapid calibration of discrete element contact model parameters for asphalt mixture mixing simulation includes the following steps: S1: Construction of particle discrete element model and calibration of basic contact parameters of aggregate particles; S2: Using the basic contact parameters calibrated in step S1, an adhesion parameter prediction model based on asphalt viscosity and asphalt film thickness is constructed; S3: Calibrate the adhesion parameters of the coated asphalt aggregate in the virtual mixing test based on the adhesion parameter prediction model constructed in step S2.

[0040] Specifically, in step S1, the specific process of constructing the particle discrete element model and calibrating the basic contact parameters of aggregate particles is as follows: S1.1: Conduct indoor tests on the natural slope angle of real aggregate to determine the natural slope angle of the aggregate.

[0041] Specifically, in a transparent glass box with a size of 100 mm x 100 mm x 100 mm and an open top, the natural slope angle test of real aggregate is carried out by filling clean and dry aggregate with a particle size range of 2.36-4.75 mm. The specific steps are as follows: the transparent glass box is rotated clockwise at a uniform speed with the bottom surface edge as the rotation axis, the rotation angle is 90°, and the particles form a natural slope angle under the action of gravity. The image is taken and the angle of the natural slope angle of the real aggregate is obtained, denoted as Figure 5 0.

[0042] S1.2: Obtain a particle discrete element model based on the topography of real aggregate, and determine the particle generation ratio.

[0043] Specifically, the three-dimensional topographic features of aggregate with a particle size range of 2.36-4.75 mm are obtained. The natural slope angle results obtained during the natural slope angle test of particles in this particle size range can fully reflect the influence of contact parameters and topographic changes, and can also obtain stable and easy-to-determine (clear edge) natural slope angles. The shape parameter distribution curve is drawn as shown in Angle According to the 5% and 95% quantile points of the shape parameter cumulative distribution frequency curve, the shape parameter distribution range is determined, which is uniformly divided into 5 intervals. One representative particle (the shape parameter of the particle is close to the middle value of the interval) is selected in each interval, a total of 5 real topographies of particles are obtained, and the particle discrete element model is obtained by importing the discrete element software and using the built-in template filling algorithm. The cumulative distribution frequency of the particles in each interval range is calculated as the generation ratio of the corresponding particles in the subsequent discrete element natural slope angle test device model, as shown in Table 1.

[0044] Table 1: Particles with different characteristic shapes and corresponding ratios S1.3: Standardize the size of the particle discrete element model.

[0045] Specifically, according to the width of the particle discrete element model, the scaling ratio of the particle discrete element model is calculated. Preferably, the width of all particle discrete element models is uniformly 2.36 mm. Substituting formula (1), the scaling ratio of the particle discrete element model is obtained as: c = 2.36 / b (1′) In the formula, c is the scaling ratio of the particle discrete element model, and b is the actual width of the particle discrete element model.

[0046] S1.4: Generate particles in the discrete element natural slope angle test device model and perform preliminary processing.

[0047] Specifically, in a closed cube with a size of 100 mm x 100 mm x 150 mm, particles are filled in random positions and directions, and the generation proportion of different particles is determined according to the cumulative distribution frequency of step S1.2, the particles are naturally stacked under the action of gravity, the cube is adjusted to 100 mm x 100 mm x 100 mm (at this time the particles are all stacked in the lower part of the initial cube, and the height direction is reduced to 100 mm), and the top surface of the cube is set to be not closed, and the particles exceeding the top surface are scraped flat, as shown in FIG. 3(a) and FIG. 3(b). The contact model of the particles is set to the Hertz-Mindlin with JKR contact model to obtain a preliminary processed virtual natural slope angle test model. The Hertz-Mindlin with JKR contact model is a model used to describe the contact behavior of particles in DEM (discrete element method), which combines the Hertz contact theory and the Mindlin-Deresiewicz tangential force model, and introduces the JKR surface energy parameter to simulate the cohesive force between particles. The contact parameters of the Hertz-Mindlin with JKR contact model include basic contact parameters and adhesion parameters, the basic contact parameters include the restitution coefficient, the static friction coefficient and the rolling friction coefficient, and the adhesion parameter of the contact model is the JKR surface energy γ.

[0048] The asphalt mixture simulation is different from the traditional mechanical simulation, and therefore is not applicable to the parallel bonding model and the Bergs model commonly used in the discrete element simulation of mechanical properties, and the calibration method is completely different. In the present application, the Hertz-Mindlin with JKR contact model is applied to the asphalt mixture simulation by referring to the method of granular mechanics (usually the Hertz-Mindlin with JKR contact model is used to represent the contact characteristics of surface water-containing particles, i.e. surface energy).

[0049] S1.5: Determine the test scheme of the response surface design method.

[0050] The adhesion parameter of the Hertz-Mindlin with JKR contact model, i.e. the JKR surface energy γ, is set to 0, the value range of the basic contact parameters to be calibrated is set by trial and error, the high, medium and low level values of the parameters to be calibrated are listed based on the response surface design-central composite (CCD) design method, and a three-factor five-level test scheme is set, as shown in Table 2, and the response value in the response surface design method is the virtual natural slope angle.

[0051] Table 2 Parameter setting of response surface design three-factor five-level S1.6: Perform virtual natural slope angle test and determine response surface model; Specifically, referring to the three-factor five-level test scheme set in step S1.4, the contact parameters in the Hertz-Mindlin with JKR contact model are set. The cube is rotated clockwise at a constant speed with the bottom surface edge as the rotation axis. The rotation speed is controlled to 10° / s and the rotation angle is 90°. The particles form a natural slope angle under the action of gravity, and the natural slope angle is obtained by linear fitting. DA As shown in Figure 1, the natural slope angle is usually not a straight line. In order to obtain the data of the natural slope angle more easily and accurately, and to ensure the stability of the natural slope angle results under different test schemes, the points on the lower diagonal of the cube are fitted during the entire calibration process, and the response surface model of the natural slope angle is determined as follows: (2′) Where: Angle is the natural slope angle; restitution is the coefficient of restitution; static is the coefficient of static friction; rolling is the coefficient of rolling friction.

[0052] S1.7: The actual natural slope angle of the aggregate determined in step S1.1 ( Angle 0) and the response surface model determined in step S1.6 (Equation (2′)), the basic contact parameters of aggregate particles, including the restitution coefficient, static friction coefficient, and rolling friction coefficient, are determined, as shown in Table 3.

[0053] Table 3 Calibration results of basic contact parameters of particles Specifically, in step S2, the specific construction process of the adhesion parameter prediction model based on asphalt viscosity and asphalt film thickness is as follows: S2.1: Determine the viscosity-temperature curve of the asphalt being calibrated.

[0054] Specifically, the viscosity-temperature curve of the calibrated asphalt material is measured according to the specification "Test Procedures for Asphalt and Asphalt Mixtures in Highway Engineering" (JTG E20-2011) T0625.

[0055] S2.2: Determine asphalt film thickness.

[0056] Specifically, the asphalt film thickness is determined based on the specific surface area of ​​2.36 mm particles in the Technical Specifications for Highway Asphalt Pavement Construction (JTG F40-2004). The calculation method is: (3′) Where: DA is the asphalt film thickness; P be is the oil-stone ratio; Angle bRelative density of asphalt.

[0057] S2.3: Establish a fitting equation between the adhesion parameter and the virtual natural slope angle based on the virtual natural slope angle test.

[0058] Specifically, the basic contact parameters in the preliminarily processed virtual natural slope angle test model are set as the basic contact parameters calibrated in step S1; the adhesion parameters are set as 0, 0.5, 1, 1.5, and 2 respectively, and five test schemes are determined; the cube is rotated clockwise at a uniform speed with the bottom surface edge as the rotation shaft, the rotation speed is controlled to be 10 ° / s, the rotation angle is 90°, the particles form a natural slope angle under the action of gravity, the virtual natural slope angle is obtained by linear fitting, and a fitting equation between the adhesion parameter and the virtual natural slope angle is established, as shown in DA The fitting equation is as follows: ​ = 6.428 ​ + 48.318(4′) In the formula: ​ is the natural slope angle; ​ is the adhesion parameter.

[0059] S2.4: Perform a real aggregate natural slope angle test to obtain aggregate natural slope angles at different temperatures and oil aggregate ratios, and construct a multivariate fitting model of temperature, oil aggregate ratio, and natural slope angle.

[0060] Specifically, asphalt is added to clean and dry aggregate with a particle size range of 2.36-4.75 mm, the mass of the asphalt is determined according to 1%, 2%, 3%, and 4% of the mass of the aggregate, the asphalt mixture with different oil aggregate ratios is uniformly mixed at 140°C, 160°C, and 180°C, and the real aggregate natural slope angle test is performed within 1 min after the mixing is completed to obtain aggregate natural slope angles at different oil aggregate ratios, and a multivariate fitting model of temperature, oil aggregate ratio, and natural slope angle is constructed, as follows: (5′) In the formula: ​ is the natural slope angle; P be is the oil aggregate ratio; T is the asphalt mixing temperature.

[0061] S2.5: Establish an adhesion parameter prediction model based on the viscosity of asphalt and the thickness of asphalt film.

[0062] Specifically, the fitting equation (formula (4')) between the adhesion parameter established based on step S2.3 and the virtual natural slope angle and the multiple fitting model (formula (5')) of temperature, oil-stone ratio and natural slope angle constructed in step S2.4, combined with the relationship (formula (3')) between the viscosity-temperature curve measured in step S2.1 and the asphalt film thickness and the oil-stone ratio determined in step S2.2, establish an adhesion parameter prediction model based on the asphalt viscosity and the asphalt film thickness, as follows: (6') In the formula: ​ is the adhesion parameter; ​ is the asphalt film thickness; ​ b is the relative density of asphalt; V is the rotational viscosity of asphalt at the mixing temperature.

[0063] Specifically, in step S3, the specific calibration process of the adhesion parameter of the aggregate in the virtual mixing test based on the adhesion parameter prediction model of step S2 is as follows: S3.1: Calculate the average asphalt film thickness of the mixture under the optimum asphalt content.

[0064] Specifically, based on the synthetic mix proportion of the asphalt mixture, the average asphalt film thickness of the mixture under the optimum asphalt content is calculated according to the specification "Technical Specification for Construction of Highway Asphalt Pavement" (JTG F40-2004). Taking the standard gradation AC16 as an example, assuming that the optimum oil-stone ratio is 4.7%, the average asphalt film thickness is: S3.2: Determine the asphalt viscosity at the mixing temperature corresponding to the virtual mixing test based on the asphalt viscosity-temperature curve. Assuming that the viscosities of asphalt at 150℃, 160℃ and 170℃ are 0.19 Pa·s, 0.13 Pa·s and 0.096 Pa·s respectively.

[0065] S3.3: Put the average asphalt film thickness obtained in step S3.1 and the asphalt viscosity obtained in step S3.2 into the adhesion parameter prediction model based on the asphalt viscosity and the asphalt film thickness established in step S2 to obtain the adhesion parameter of the particles at the mixing temperature corresponding to the virtual mixing test, as shown in Table 4.

[0066] Table 4 Adhesion parameters corresponding to different mixing temperatures The above description is only a description of the preferred embodiments of the present application, and is not any limitation on the scope of the present application. Any modification or modification made by any person skilled in the art according to the above disclosed technical content shall be regarded as an equivalent effective embodiment, and shall fall within the scope of the technical scheme protected by the present application.

Claims

1. A method for rapid calibration of discrete element contact model parameters for asphalt mixture mixing simulation, characterized in that: The discrete element contact model is a Hertz-Mindlin with JKR contact model. The corresponding contact parameters include basic contact parameters and adhesion parameters. The basic contact parameters include the restitution coefficient, static friction coefficient, and rolling friction coefficient. The calibration of the contact parameters includes the following steps: S1: Construction of particle discrete element model and calibration of basic contact parameters of aggregate particles, including the following steps: S1.1: Select aggregates with a particle size range of 2.36–4.75 mm and conduct an indoor real aggregate natural slope angle test in a cubic container with a length × width × height ratio of 1:1:1 and an open top to determine the real aggregate natural slope angle; S1.2: Obtain a particle discrete element model of realistic aggregate morphology in the size range of 2.36–4.75 mm and determine the particle generation ratio; S1.3: Standardize the particle discrete element model size; S1.4: Generate and preliminarily process particles in a discrete element natural slope angle test device model to obtain a preliminarily processed virtual natural slope angle test model; the discrete element natural slope angle test device model is a cubic model with a length×width×height ratio of 1:1:1 and an open top surface; S1.5: Determine the experimental plan for the response surface design method; S1.6: Based on the preliminarily processed virtual natural slope angle test model, carry out a virtual natural slope angle test according to the test plan of step S1.5, and determine a response surface model. The response surface model is expressed as: （2） Where: Angle is the natural slope angle; restitution is the coefficient of restitution; static is the static friction coefficient; rolling is the rolling friction coefficient; k 、 k 1. k 2. k 3 is the model coefficient; S1.7: Determine the basic contact parameters of the aggregate particles based on the actual natural slope angle of the aggregate determined in step S1.1 and the response surface model determined in step S1.6; S2: Using the basic contact parameters calibrated in step S1, an adhesion parameter prediction model based on asphalt viscosity and asphalt film thickness is constructed. The adhesion parameter prediction model based on asphalt viscosity and asphalt film thickness is expressed as: （6） Where: γ is the adhesion parameter; DA is the asphalt film thickness; γ b is the relative density of asphalt; V is the rotational viscosity of asphalt at mixing temperature, u 、 u 1. u 2. u 3. u 4. u 5. v 、 v 1 、v 2 is the model coefficient; S3: Calibrate the adhesion parameters of the aggregate in the virtual mixing test based on the adhesion parameter prediction model constructed in step S2.

2. A method for rapid calibration of discrete element contact model parameters for asphalt mixture mixing simulation according to claim 1, characterized in that: Step S1.1 is as follows: A transparent glass box with a length × width × height ratio of 1:1:1 and an open top was filled with clean, dry aggregate with a particle size range of 2.36–4.75 mm. The specific dimensions of the length and width of the transparent glass box were greater than 50 mm. A real aggregate natural slope angle test was conducted. The specific steps were as follows: the transparent glass box was rotated clockwise at a constant speed with the bottom surface edge as the rotation axis. The rotation angle was 90°. The particles formed a natural slope angle under the action of gravity. An image was taken to obtain the angle of the real aggregate natural slope angle, which was recorded as Angle 0.

3. The method for rapid calibration of discrete element contact model parameters for asphalt mixture mixing simulation according to claim 1, characterized in that: Step S1.2 is specifically as follows: The three-dimensional morphological characteristics of aggregates with a particle size range of 2.36–4.75 mm were obtained, and the shape parameter distribution curve was drawn. The shape parameter distribution range was determined based on the 5% and 95% quantiles of the cumulative distribution frequency curve of the shape parameter, and the range was evenly divided into N intervals, where N>3. A representative particle was selected from each interval, and the shape parameter of the representative particle was close to the middle value of the interval. The true morphology of N particles was obtained in total, and the discrete element software was imported and the built-in template filling algorithm of the software was used to obtain the particle discrete element model. The cumulative distribution frequency of the particles in each interval was calculated as the particle generation ratio.

4. A method for rapid calibration of discrete element contact model parameters for asphalt mixture mixing simulation according to claim 1, characterized in that: Step S1.3 is as follows: According to the width of the particle discrete element model, the scaling ratio of the particle discrete element model is converted, and the width of all particle discrete element models is unified to a mm, 2.36 < a < 4.

75. The calculation formula of the scaling ratio of the particle discrete element model is: c = a / b (1) Where c is the scaling ratio of the particle discrete element model; a is the width of the quasi-unified particle discrete element model; and b is the actual width of the particle discrete element model.

5. The method for rapid calibration of discrete element contact model parameters for asphalt mixture mixing simulation according to claim 1, characterized in that: Step S1.4 is as follows: The discrete element natural slope angle test device model is a cubic model with a length × width × height ratio of 1:1:1 and an open top surface. The specific dimensions of length, width, and height are the same as those of the transparent glass box used in the real aggregate natural slope angle test in step S1.1, and is used to perform a virtual natural slope angle test. In discrete element software, a closed cube with a length × width × height ratio of 1:1:1.5 to 1:1:2 was constructed and filled with particles at random positions and directions. The length and width of the closed cube were the same as those of the transparent glass box used in the actual natural slope angle test. The generation ratio of different particles was determined according to the cumulative distribution frequency in step S1.

2. The particles were naturally accumulated under the action of virtual gravity. The length × width × height ratio of the cube was adjusted to 1:1:1, and the top surface of the cube was set to be open. The particles that exceeded the top surface were scraped to level the top surface. The contact model of the particles was set to the Hertz-Mindlin with JKR contact model to obtain a preliminary processed virtual natural slope angle test model. The contact parameters of the Hertz-Mindlin with JKR contact model include basic contact parameters and adhesion parameters. The basic contact parameters include the restitution coefficient, static friction coefficient, and rolling friction coefficient. The adhesion parameter of the contact model is the JKR surface energy γ.

6. A method for rapid calibration of discrete element contact model parameters for asphalt mixture mixing simulation according to claim 1, characterized in that: Step S1.5 specifically includes: S1.5.1: Set the adhesion parameter of the Hertz-Mindlin with JKR contact model, i.e., the JKR surface energy γ, to 0, and set the value range of the basic contact parameters to be calibrated; S1.5.2: Based on the response surface design-central composite design method, list the high, medium and low level values ​​of the parameters to be calibrated, and set a three-factor five-level test plan. The three factors are the restitution coefficient, static friction coefficient and rolling friction coefficient. The response value in the response surface design method is the virtual natural slope angle.

7. The method for rapid calibration of discrete element contact model parameters for asphalt mixture mixing simulation according to claim 1, characterized in that: Step S1.6 is as follows: Referring to the three-factor, five-level experimental scheme set in step S1.4, the contact parameters in the Hertz-Mindlin with JKR contact model were set. The cube was rotated clockwise at a constant speed of 90° about the edge of the bottom surface as the axis of rotation. The particles formed a natural slope angle under the action of gravity. The natural slope angle was obtained by linear fitting, and the response surface model of the natural slope angle was determined.

8. The method for rapid calibration of discrete element contact model parameters for asphalt mixture mixing simulation according to claim 1, characterized in that: Step S2 includes: Step S2 includes the following steps: S2.1: Determine the viscosity-temperature curve of the asphalt being calibrated; S2.2: Determine the asphalt film thickness using the following calculation: （3） Where: DA is the asphalt film thickness; P be is the oil-stone ratio; γ b is the relative density of asphalt, 0.082 is the asphalt film thickness coefficient corresponding to aggregate particles with a particle size range of 2.36–4.75 mm; S2.3: Conduct a virtual natural slope angle test based on the preliminarily processed virtual natural slope angle test model and establish a fitting equation between the adhesion parameters and the virtual natural slope angle; specifically: The basic contact parameters in the preliminarily processed virtual natural slope angle test model are set to the basic contact parameters calibrated in step S1; The adhesion parameters are set to multiple different values, thereby determining multiple test plans; Based on the above experimental scheme, the cube is rotated clockwise at a constant speed with the bottom surface edge as the rotation axis. The rotation angle is 90°. The particles form a natural slope angle under the action of gravity. The virtual natural slope angle is obtained by linear fitting, and the fitting equation between the adhesion parameter and the virtual natural slope angle is established, which is expressed as: Angle = mγ + n （4） Where: Angle is the natural slope angle; γ is the adhesion parameter; m 、 n is the model coefficient; S2.4: Conduct real aggregate natural slope angle tests to obtain the natural slope angles of aggregates at different temperatures and asphalt-aggregate ratios. Construct a multivariate fitting model of temperature, asphalt-aggregate ratio, and natural slope angle, expressed as: （5） Where: Angle is the natural slope angle; P be is the oil-stone ratio; T is the asphalt mixing temperature; j 、 j 1. j 2. j 3. j 4. j 5 is the model coefficient; S2.5: Based on the fitting equation between the adhesion parameters and the virtual natural slope angle established in step S2.3 and the multivariate fitting model of temperature, asphalt-stone ratio and natural slope angle constructed in step S2.4, combined with the viscosity-temperature curve measured in step S2.1 and the relationship between asphalt film thickness and asphalt-stone ratio determined in step S2.2, an adhesion parameter prediction model based on asphalt viscosity and asphalt film thickness is established.

9. The method for rapid calibration of discrete element contact model parameters for asphalt mixture mixing simulation according to claim 1, characterized in that: Step S3 includes: S3.1: Calculate the average asphalt film thickness of the mixture at the optimum asphalt dosage; S3.2: Determine the asphalt viscosity at the mixing temperature corresponding to the virtual mixing test to be conducted based on the asphalt viscosity-temperature curve; S3.3: Substitute the average asphalt film thickness obtained in step S3.1 and the asphalt viscosity obtained in step S3.2 into the adhesion parameter prediction model based on asphalt viscosity and asphalt film thickness established in step S2 to obtain the adhesion parameters of the particles at the mixing temperature corresponding to the virtual mixing test to be carried out.

Images