A diffusion behavior analysis method for an asphalt multiple regeneration process

Abstract

The application discloses a diffusion behavior analysis method for an asphalt multiple regeneration process and relates to the technical field of asphalt regeneration. First, the aging degree of regenerated asphalt is obtained, weight analysis of diffusion behavior is carried out, and a heavy aging stage and a light aging stage are distinguished. Then, the distribution of diffusion depth and average diffusion rate is generated according to the two stages, and corresponding shallow time series data is collected. The relationship between the shallow diffusion depth and the average diffusion rate is established, and deep diffusion depth time series data is generated through data reconstruction. Finally, the diffusion depth-average diffusion rate deep relationship is constructed by using deep and shallow data, and the diffusion behavior of regenerated asphalt is analyzed. The method quantifies the diffusion phenomenon between new and old asphalts, improves the accuracy of the analysis of the asphalt multiple regeneration process, and has important significance for asphalt regeneration technology.

Description

Technical Field

[0001] The present invention relates to the technical field of asphalt regeneration, and more particularly to a method for analyzing diffusion behavior of asphalt during multiple regeneration processes. Background Art

[0002] Multiple asphalt pavement recycling refers to a paving technique that recycles an already recycled pavement a second or even more times after it reaches the end of its service life. Whether recycled asphalt pavements constructed earlier have reached, or are nearing, the end of their service life can be recycled multiple times has become a key scientific and technical issue facing the development of recycling technology. Compared to single-stage recycling, multiple asphalt recycling differs in that the asphalt coated on the old aggregate is of varying degrees of aging. This means that diffusion occurs between the new asphalt and the varying degrees of aging during the recycling process. Under certain conditions, the diffusion behavior at the interface between the new and old asphalt directly impacts the performance and service life of the recycled asphalt pavement. Furthermore, as the asphalt ages, the diffusion problem caused by partial mixing becomes more pronounced. Therefore, it is necessary to study the more complex diffusion behavior during multiple asphalt recycling. This will provide a theoretical and experimental foundation for the design and performance optimization of multiple-recycled asphalt mixtures.

[0003] Multiple asphalt regeneration is an important direction for the development of regeneration technology. It requires new asphalt, regeneration agent (when necessary) and old asphalt to complete multiple fusion and diffusion, and finally reach a stable whole. Compared with the single regeneration of asphalt pavement, the difference between multiple regeneration is that the asphalt wrapped on the old aggregate is aged asphalt of varying degrees. That is, during the regeneration process, fusion between new asphalt and asphalt of varying degrees of age will occur. The diffusion behavior between the interfaces is more complex, and it is difficult to quantify the diffusion depth and explore the evolution law. This is also one of the difficulties that the current multiple regeneration technology needs to overcome. At present, the research on multiple regeneration of asphalt is mostly focused on the overall macroscopic performance analysis of asphalt and mixture. Liu et al. aged and regenerated asphalt five times. The results showed that the asphalt after each aging and regeneration was able to recover to the performance level of the original asphalt, proving the feasibility of multiple regeneration. Wang Jie et al. aged and regenerated asphalt mixture three times and analyzed the change law of its performance. The results showed that with the increase in the number of regenerations, the low-temperature performance decreased rapidly, while the high-temperature performance showed a logarithmic growth. Zou Guilian et al. used dynamic mechanical analysis (DMA) to study the viscoelastic properties of SBS-modified asphalt before and after multiple regeneration. They found that the viscoelastic properties of SBS-modified asphalt gradually disappear with aging and the number of regenerations. The regeneration agent did not fully regenerate the repeatedly aged SBS-modified asphalt. A macroscopic performance evaluation system can demonstrate the feasibility of multiple regeneration of asphalt pavements and evaluate the pros and cons of multiple regeneration technologies. However, due to the structural complexity of asphalt itself, relying solely on macroscopic performance cannot accurately evaluate the regeneration effect of asphalt and mixtures, nor can it explain the mechanisms that produce the corresponding properties. Consequently, targeted optimization and improvement of material properties are difficult.

[0004] Therefore, how to propose a diffusion behavior analysis method for the multiple regeneration process of asphalt, take the more complex diffusion behavior of the new and old asphalt fusion interface in the multiple regeneration process as the theme, and quantitatively characterize the diffusion behavior of new asphalt and asphalt of different degrees of aging in the multiple regeneration process is an urgent problem that technical personnel in this field need to solve. Summary of the Invention

[0005] In view of this, the present invention provides a method for analyzing the diffusion behavior of asphalt during multiple regeneration processes. This method focuses on studying the more complex diffusion behavior of the interface between new and old asphalt during multiple regeneration processes, and quantitatively characterizes the diffusion behavior of new asphalt and asphalt of varying degrees of aging during multiple regeneration processes. To achieve the above objectives, the present invention employs the following technical solutions:

[0006] A diffusion behavior analysis method for asphalt multiple regeneration processes, comprising:

[0007] Obtain the asphalt aging degree in the recycled asphalt to conduct a diffusion behavior weight analysis of the asphalt multiple regeneration process, generating a severe aging stage and a mild aging stage;

[0008] Generate diffusion depth distribution and average diffusion rate distribution according to the severe aging stage and the mild aging stage;

[0009] According to the distribution of diffusion depth, shallow diffusion depth time series data are collected; according to the distribution of average diffusion rate, shallow average diffusion rate time series data are collected;

[0010] The average diffusion rate is predicted based on the shallow diffusion depth time series data, and the shallow diffusion depth-average diffusion rate relationship is generated;

[0011] The shallow diffusion depth time series data is reconstructed through the shallow average diffusion rate time series data and the shallow relationship between diffusion depth and average diffusion rate to generate deep diffusion depth time series data;

[0012] Using the deep diffusion depth time series data and the shallow average diffusion rate time series data, the diffusion depth-average diffusion rate deep relationship is constructed;

[0013] The diffusion behavior of asphalt multiple regeneration process was analyzed based on the deep relationship between diffusion depth and average diffusion rate.

[0014] Optionally, the generation of the diffusion depth distribution and the average diffusion rate distribution includes: analyzing the correlation between the diffusion depth distribution and the viscosity, complex shear modulus, storage modulus, loss modulus, and phase angle through a decision tree model, and constructing a first fingerprint library based on strong correlation factors to jointly characterize the diffusion depth distribution; analyzing the correlation between the average diffusion rate distribution and the viscosity, complex shear modulus, storage modulus, loss modulus, and phase angle through a decision tree model, and constructing a second fingerprint library based on strong correlation factors to jointly characterize the average diffusion rate distribution.

[0015] Optionally, obtaining the asphalt aging degree in the regenerated asphalt and performing a diffusion behavior weight analysis of the asphalt multiple regeneration processes to generate a severe aging stage and a mild aging stage includes:

[0016] According to actual conditions, the asphalt multiple regeneration process of the regenerated asphalt is obtained; the asphalt multiple regeneration process is traversed to perform aging evaluation on the regenerated asphalt to obtain decay vector information, wherein the decay vector information includes infection force information; it is determined whether the infection force information is greater than or equal to the infection force threshold; if the infection force information is greater than or equal to the infection force threshold, the asphalt multiple regeneration process is added to the severe aging stage; if the infection force information is less than the infection force threshold, the asphalt multiple regeneration process is added to the mild aging stage.

[0017] Optionally, the predicting of the average diffusion rate based on the shallow diffusion depth time series data to generate a shallow relationship between diffusion depth and average diffusion rate includes:

[0018] Acquiring asphalt material information in the regenerated asphalt, wherein the asphalt material information in the regenerated asphalt corresponds to the shallow diffusion depth time series data;

[0019] Performing average diffusion rate prediction based on the asphalt material information in the regenerated asphalt and the shallow diffusion depth time series data to generate average diffusion rate prediction time series information;

[0020] The shallow layer relationship between diffusion depth and average diffusion rate is generated according to the shallow layer diffusion depth time series data and the average diffusion rate prediction time series information.

[0021] Optionally, the performing of average diffusion rate prediction based on the asphalt material information in the regenerated asphalt and the shallow diffusion depth time series data to generate average diffusion rate prediction time series information includes:

[0022] Acquiring detection record data of multiple asphalt regeneration processes, wherein the detection record data of multiple asphalt regeneration processes includes record data of asphalt material in the regenerated asphalt, record data of diffusion depth of the multiple asphalt regeneration processes, and record data of average diffusion rate of the multiple asphalt regeneration processes;

[0023] Based on the BP neural network, the average diffusion rate prediction model is trained using the asphalt material record data in the regenerated asphalt, the diffusion depth record data of the asphalt multiple regeneration processes, and the average diffusion rate record data of the asphalt multiple regeneration processes; based on the asphalt material information in the regenerated asphalt, the shallow diffusion depth time series data is traversed and input into the average diffusion rate prediction model to generate the average diffusion rate prediction time series information.

[0024] Optionally, reconstructing shallow diffusion depth time series data using shallow average diffusion rate time series data and the shallow relationship between diffusion depth and average diffusion rate to generate deep diffusion depth time series data includes:

[0025] Obtaining an average diffusion rate deviation vector according to the shallow average diffusion rate time series data and the diffusion depth-average diffusion rate shallow relationship, wherein the average diffusion rate deviation vector includes average diffusion rate magnitude information and average diffusion rate direction information;

[0026] Performing correlation analysis based on the average diffusion rate magnitude information and the average diffusion rate direction information to obtain missing diffusion depth information;

[0027] The shallow diffusion depth time series data is reconstructed according to the missing diffusion depth information to generate the deep diffusion depth time series data.

[0028] Optionally, performing a correlation analysis based on the average diffusion rate magnitude information and the average diffusion rate direction information to obtain the missing diffusion depth information includes:

[0029] According to the average diffusion rate size information and the average diffusion rate direction information, the average diffusion rate detection record data of the asphalt multiple regeneration process is collected, wherein the average diffusion rate detection record data of the asphalt multiple regeneration process includes diffusion depth detection record data; the diffusion depth detection record data is traversed to obtain the derived time length parameter and the diffusion depth detection vector parameter; according to the derived time length parameter and the diffusion depth detection vector parameter, the missing diffusion depth information is filtered.

[0030] Optionally, the analysis of the diffusion behavior of the asphalt multiple regeneration process based on the deep relationship between the diffusion depth and the average diffusion rate includes:

[0031] Constructing a teacher analysis model and a student analysis model, predicting the deep relationship between diffusion depth and average diffusion rate using the teacher analysis model to obtain a first prediction result, obtaining the credibility of the teacher analysis model for the first prediction result, updating the first prediction result based on the credibility, and using the updated first prediction result as the second prediction result;

[0032] Obtain the difference between the second prediction result and the result of the student analysis model predicting the deep relationship between diffusion depth and average diffusion rate. Based on the difference, update the parameters of the student analysis model to train the student analysis model, and perform diffusion behavior analysis and prediction using the trained student analysis model.

[0033] Optionally, obtaining the credibility of the teacher analysis model for the first prediction result, updating the first prediction result based on the credibility, and using the updated first prediction result as the second prediction result includes: comparing each of the different credibility levels one by one; and determining the first prediction result corresponding to the minimum credibility as the second prediction result.

[0034] Optionally, obtaining the credibility of the teacher analysis model for the first prediction result, updating the first prediction result according to the credibility, and using the updated first prediction result as the second prediction result also includes: calculating the weighted value corresponding to each first prediction result according to the credibility of each first prediction result; performing weighted summation on each first prediction result combined with the weighted value, and performing weighted averaging on the weighted sum and the second prediction result to obtain the updated second prediction result as the final prediction result.

[0035] It can be seen from the above technical solution that, compared with the prior art, the present invention discloses a method for analyzing the diffusion behavior of asphalt during multiple regeneration processes, which has the following beneficial effects:

[0036] The present invention proposes a method for analyzing the diffusion behavior of an asphalt multiple regeneration process, comprising: obtaining the degree of asphalt aging in the regenerated asphalt to perform a diffusion behavior weight analysis of the asphalt multiple regeneration process, and generating a severe aging stage and a mild aging stage; generating a diffusion depth distribution and an average diffusion rate distribution according to the severe aging stage and the mild aging stage; collecting shallow diffusion depth time series data according to the diffusion depth distribution; collecting shallow average diffusion rate time series data according to the average diffusion rate distribution; predicting the average diffusion rate based on the shallow diffusion depth time series data, and generating a shallow relationship between diffusion depth and average diffusion rate; reconstructing the shallow diffusion depth time series data through the shallow average diffusion rate time series data and the shallow relationship between diffusion depth and average diffusion rate, and generating deep diffusion depth time series data; constructing a deep relationship between diffusion depth and average diffusion rate using the deep diffusion depth time series data and the shallow average diffusion rate time series data; and analyzing the diffusion behavior of the asphalt multiple regeneration process based on the diffusion depth and average diffusion rate deep relationship. The present invention takes the more complex diffusion behavior of the new and old asphalt fusion interface during multiple regeneration processes as its theme, quantitatively characterizes the diffusion behavior of new asphalt and asphalt of different degrees of aging during multiple regeneration processes, and performs a weighted analysis of the diffusion behavior of asphalt during multiple regeneration processes by obtaining the degree of asphalt aging in the regenerated asphalt to generate a severe aging stage and a mild aging stage; based on the severe aging stage and the mild aging stage, the diffusion depth distribution and the average diffusion rate distribution are generated, and the diffusion depth and diffusion rate between new asphalt of different aging degrees and asphalt of different degrees of aging during multiple regeneration processes are quantified, thereby improving the accuracy of the diffusion behavior analysis of asphalt during multiple regeneration processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0038] Figure 1 A schematic flow chart of a diffusion behavior analysis method for the multiple asphalt regeneration process provided by the present invention. DETAILED DESCRIPTION

[0039] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0040] The embodiment of the present invention discloses a method for analyzing the diffusion behavior of asphalt during multiple regeneration processes. Figure 1 Shown, including:

[0041] Obtain the asphalt aging degree in the recycled asphalt to conduct a diffusion behavior weight analysis of the asphalt multiple regeneration process, generating a severe aging stage and a mild aging stage;

[0042] Generate diffusion depth distribution and average diffusion rate distribution according to the severe aging stage and the mild aging stage;

[0043] According to the distribution of diffusion depth, shallow diffusion depth time series data are collected; according to the distribution of average diffusion rate, shallow average diffusion rate time series data are collected;

[0044] The average diffusion rate is predicted based on the shallow diffusion depth time series data, and the shallow diffusion depth-average diffusion rate relationship is generated;

[0045] The shallow diffusion depth time series data is reconstructed through the shallow average diffusion rate time series data and the shallow relationship between diffusion depth and average diffusion rate to generate deep diffusion depth time series data;

[0046] Using the deep diffusion depth time series data and the shallow average diffusion rate time series data, the diffusion depth-average diffusion rate deep relationship is constructed;

[0047] The diffusion behavior of asphalt multiple regeneration process was analyzed based on the deep relationship between diffusion depth and average diffusion rate.

[0048] Furthermore, the generation of the diffusion depth distribution and the average diffusion rate distribution includes: analyzing the correlation between the diffusion depth distribution and the viscosity, complex shear modulus, storage modulus, loss modulus, and phase angle through a decision tree model, and constructing a first fingerprint library based on strong correlation factors to jointly characterize the diffusion depth distribution; analyzing the correlation between the average diffusion rate distribution and the viscosity, complex shear modulus, storage modulus, loss modulus, and phase angle through a decision tree model, and constructing a second fingerprint library based on strong correlation factors to jointly characterize the average diffusion rate distribution.

[0049] In a specific embodiment, constructing a first fingerprint library based on strongly correlated factors to jointly characterize the diffusion depth distribution includes:

[0050] The diffusion depth distribution is represented by the combination of viscosity, complex shear modulus and loss modulus, the size of the quantitative characteristics of the diffusion behavior at different positions is collected, it is assumed that the diffusion behavior occurs at the measured point P i , and the diffusion depth distribution of the measured point P i is represented by the combination of the measured diffusion depth distribution of the measured point P

[0051]

[0052] Among them, represents the joint representation strength of the t-th measurement, k represents the number of measurements, i represents the i-th measured point, j represents the j-th quantitative characteristic, the diffusion depth distribution of the i-th measured point is obtained by j quantitative characteristics, and the first fingerprint library is constructed

[0053]

[0054] Among them, m is the number of quantitative characteristics, n is the number of diffusion behavior groups, The i-th row vector of the measured point P i is That is, the fingerprint characteristic value of the measured point P i measured by m quantitative characteristics, and a multi-dimensional characteristic dimension is formed by different types of quantitative characteristics, and each quantitative characteristic is the intensity of the joint representation of the diffusion depth distribution when the diffusion behavior occurs at the measured point P i . Similarly, the second fingerprint library is determined in the same way as the first fingerprint library.

[0055] Specifically, a double-layer test piece of new and old asphalt overlap (old asphalt at the bottom) is used. The dynamic change of viscosity of new and old asphalt during fusion at different test temperatures (same mixing temperature) is measured by a Brookfield viscometer, and the viscosity result is recorded every 5s until the result is stable. Frequency scanning, time scanning and temperature scanning tests of dynamic shear rheometer (DSR) are carried out, and the rheological parameters such as complex shear modulus, storage modulus, loss modulus and phase angle of the sample are recorded in stress (variable) control mode. The difference between the test results in each asphalt fusion process is analyzed.

[0056] Further, the diffusion behavior weight analysis of the asphalt multiple regeneration process of the aging degree of the reclaimed asphalt is carried out, and the heavy aging stage and the light aging stage are generated, including:

[0057] According to actual conditions, the asphalt multiple regeneration process of the regenerated asphalt is obtained; the asphalt multiple regeneration process is traversed to perform aging evaluation on the regenerated asphalt to obtain decay vector information, wherein the decay vector information includes infection force information; it is determined whether the infection force information is greater than or equal to the infection force threshold; if the infection force information is greater than or equal to the infection force threshold, the asphalt multiple regeneration process is added to the severe aging stage; if the infection force information is less than the infection force threshold, the asphalt multiple regeneration process is added to the mild aging stage.

[0058] Specifically, the degree of aging is determined by a decay vector (or decay vector), including: defining a decay vector: a decay vector is a mathematical model that describes the change in the state of a system under time or usage conditions. It can be regarded as a collection of different aging indicators, each of which represents the decay of different aspects of the material. The steps of defining a decay vector include: selecting indicators, determining key indicators related to aging, such as fatigue strength, hardness, bonding strength, material density, etc. Standardization, standardizing the indicators so that they are on the same order of magnitude; data collection: collecting data related to the defined decay vector. Through measured data after long-term use and termination test data obtained in accelerated aging experiments; establishing a decay model: constructing the relationship between feature data and the degree of aging through decision trees, random forests or neural networks; calculating the degree of aging: calculating the degree of aging through the established decay model, and combining the different indicators in the decay vector to calculate a comprehensive "aging index"; state estimation: using the Kalman filter state estimation method, updating the estimate of the degree of aging based on time series data to obtain decay vector information. Among them, the infectivity threshold is determined based on the maximum regeneration possibility of the recycled asphalt.

[0059] Furthermore, the average diffusion rate is predicted based on the shallow diffusion depth time series data to generate a shallow diffusion depth-average diffusion rate relationship, which includes:

[0060] Acquiring asphalt material information in the regenerated asphalt, wherein the asphalt material information in the regenerated asphalt corresponds to the shallow diffusion depth time series data;

[0061] Performing average diffusion rate prediction based on the asphalt material information in the regenerated asphalt and the shallow diffusion depth time series data to generate average diffusion rate prediction time series information;

[0062] The shallow layer relationship between diffusion depth and average diffusion rate is generated according to the shallow layer diffusion depth time series data and the average diffusion rate prediction time series information.

[0063] Furthermore, the average diffusion rate prediction is performed based on the asphalt material information in the regenerated asphalt and the shallow diffusion depth time series data to generate the average diffusion rate prediction time series information, including:

[0064] Acquiring detection record data of multiple asphalt regeneration processes, wherein the detection record data of multiple asphalt regeneration processes includes record data of asphalt material in the regenerated asphalt, record data of diffusion depth of the multiple asphalt regeneration processes, and record data of average diffusion rate of the multiple asphalt regeneration processes;

[0065] Based on the BP neural network, the average diffusion rate prediction model is trained using the asphalt material record data in the regenerated asphalt, the diffusion depth record data of the asphalt multiple regeneration processes, and the average diffusion rate record data of the asphalt multiple regeneration processes; based on the asphalt material information in the regenerated asphalt, the shallow diffusion depth time series data is traversed and input into the average diffusion rate prediction model to generate the average diffusion rate prediction time series information.

[0066] Furthermore, reconstructing the shallow diffusion depth time series data by using the shallow average diffusion rate time series data and the shallow relationship between diffusion depth and average diffusion rate to generate the deep diffusion depth time series data includes:

[0067] Obtaining an average diffusion rate deviation vector according to the shallow average diffusion rate time series data and the diffusion depth-average diffusion rate shallow relationship, wherein the average diffusion rate deviation vector includes average diffusion rate magnitude information and average diffusion rate direction information;

[0068] Performing correlation analysis based on the average diffusion rate magnitude information and the average diffusion rate direction information to obtain missing diffusion depth information;

[0069] The shallow diffusion depth time series data is reconstructed according to the missing diffusion depth information to generate the deep diffusion depth time series data.

[0070] Furthermore, performing correlation analysis based on the average diffusion rate magnitude information and the average diffusion rate direction information to obtain missing diffusion depth information includes:

[0071] According to the average diffusion rate size information and the average diffusion rate direction information, the average diffusion rate detection record data of the asphalt multiple regeneration process is collected, wherein the average diffusion rate detection record data of the asphalt multiple regeneration process includes diffusion depth detection record data; the diffusion depth detection record data is traversed to obtain the derived time length parameter and the diffusion depth detection vector parameter; according to the derived time length parameter and the diffusion depth detection vector parameter, the missing diffusion depth information is filtered.

[0072] Furthermore, the analysis of the diffusion behavior of the asphalt multiple regeneration process based on the deep relationship between the diffusion depth and the average diffusion rate includes:

[0073] Constructing a teacher analysis model and a student analysis model, predicting the deep relationship between diffusion depth and average diffusion rate using the teacher analysis model to obtain a first prediction result, obtaining the credibility of the teacher analysis model for the first prediction result, updating the first prediction result based on the credibility, and using the updated first prediction result as the second prediction result;

[0074] Obtain the difference between the second prediction result and the result of the student analysis model predicting the deep relationship between diffusion depth and average diffusion rate. Based on the difference, update the parameters of the student analysis model to train the student analysis model, and perform diffusion behavior analysis and prediction using the trained student analysis model.

[0075] In a specific embodiment, obtaining the credibility of the teacher analysis model for the first prediction result includes: obtaining the second probability value and prediction times of each teacher model, the second probability value representing the accuracy of each teacher model in predicting the sample data under different target states; summing the second probability values ​​corresponding to the number of prediction times to obtain the sum of the second probability values; dividing the sum of the second probability values ​​obtained by the number of predictions to obtain the first probability value; obtaining the number of categories of each first prediction result, the number of categories representing the number of prediction categories contained in each first prediction result; calculating the product value of the first probability value and its logarithm; summing the product values ​​corresponding to the number of prediction times to obtain the sum of the product values, and taking the negative of the sum of the product values ​​as the entropy of the probability value; determining the entropy as the credibility of the teacher model for the first prediction result; the credibility is used to represent the credibility of the first prediction result; the first probability value represents the accuracy of the teacher model in predicting the sample data under the target state; wherein, the teacher model under the target state includes at least one randomly masked neural node.

[0076] Furthermore, obtaining the credibility of the teacher analysis model for the first prediction result, updating the first prediction result according to the credibility, and using the updated first prediction result as the second prediction result includes: comparing each of the different credibility one by one; and determining the first prediction result corresponding to the minimum credibility as the second prediction result.

[0077] Furthermore, obtaining the credibility of the teacher analysis model for the first prediction result, updating the first prediction result according to the credibility, and using the updated first prediction result as the second prediction result also includes: calculating the weighted value corresponding to each first prediction result according to the credibility of each first prediction result; performing weighted summation on each first prediction result combined with the weighted value, and performing weighted averaging on the weighted sum and the second prediction result to obtain the updated second prediction result as the final prediction result.

[0078] In a specific embodiment, a comprehensive evaluation is performed on the second prediction result and the final prediction result, including: based on the grid attribute model, using the interpolation algorithm to interpolate the second prediction result and the final prediction result to obtain the first model and the second model for studying the diffusion behavior, respectively; if the diffusion behavior belongs to the high-quality asphalt multiple regeneration process in both the first model and the second model, then it is the high-quality asphalt multiple regeneration process; or comparing the second prediction results and the final prediction results at different vertical layers, if the study layer section satisfies both the second prediction result and the final prediction result, then it is the high-quality asphalt multiple regeneration process.

[0079] In a specific embodiment, a diffusion behavior analysis system for asphalt multiple regeneration processes includes:

[0080] Acquisition module: used to obtain the aging degree of asphalt in the recycled asphalt, conduct diffusion behavior weight analysis of the asphalt multiple regeneration process, and generate heavy aging stage and light aging stage;

[0081] Generation module: used to generate diffusion depth distribution and average diffusion rate distribution according to the severe aging stage and the mild aging stage;

[0082] Acquisition module: used to collect shallow diffusion depth time series data based on the distribution of diffusion depth; collect shallow average diffusion rate time series data based on the distribution of average diffusion rate;

[0083] Prediction module: used to predict the average diffusion rate based on shallow diffusion depth time series data and generate the shallow relationship between diffusion depth and average diffusion rate;

[0084] Reconstruction module: used to reconstruct shallow diffusion depth time series data through shallow average diffusion rate time series data and the shallow relationship between diffusion depth and average diffusion rate to generate deep diffusion depth time series data;

[0085] Construction module: used to construct the deep-layer diffusion depth-average diffusion rate deep relationship using deep-layer diffusion depth time series data and shallow-layer average diffusion rate time series data;

[0086] Analysis module: used to analyze the diffusion behavior of asphalt multiple recycling processes based on the deep relationship between diffusion depth and average diffusion rate.

[0087] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0088] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for analyzing diffusion behavior during multiple asphalt regeneration processes, characterized in that: include: Obtain the asphalt aging degree in the recycled asphalt to conduct a diffusion behavior weight analysis of the asphalt multiple regeneration process, generating a severe aging stage and a mild aging stage; Generate diffusion depth distribution and average diffusion rate distribution according to the severe aging stage and the mild aging stage; According to the distribution of diffusion depth, shallow diffusion depth time series data are collected; according to the distribution of average diffusion rate, shallow average diffusion rate time series data are collected; The average diffusion rate is predicted based on the shallow diffusion depth time series data, and the shallow diffusion depth-average diffusion rate relationship is generated; The shallow diffusion depth time series data is reconstructed through the shallow average diffusion rate time series data and the shallow relationship between diffusion depth and average diffusion rate to generate deep diffusion depth time series data; The data reconstruction of the shallow diffusion depth time series data by using the shallow average diffusion rate time series data and the shallow relationship between diffusion depth and average diffusion rate to generate the deep diffusion depth time series data includes: Obtaining an average diffusion rate deviation vector according to the shallow average diffusion rate time series data and the diffusion depth-average diffusion rate shallow relationship, wherein the average diffusion rate deviation vector includes average diffusion rate magnitude information and average diffusion rate direction information; Performing correlation analysis based on the average diffusion rate magnitude information and the average diffusion rate direction information to obtain missing diffusion depth information; Reconstructing the shallow diffusion depth time series data according to the missing diffusion depth information to generate the deep diffusion depth time series data; Performing correlation analysis based on the average diffusion rate magnitude information and the average diffusion rate direction information to obtain missing diffusion depth information includes: According to the average diffusion rate magnitude information and the average diffusion rate direction information, collecting the average diffusion rate detection record data of the asphalt multiple regeneration process, wherein the average diffusion rate detection record data of the asphalt multiple regeneration process includes the diffusion depth detection record data; traversing the diffusion depth detection record data to obtain the derived duration parameter and the diffusion depth detection vector parameter; according to the derived duration parameter and the diffusion depth detection vector parameter, filtering the missing diffusion depth information; Using the deep diffusion depth time series data and the shallow average diffusion rate time series data, the diffusion depth-average diffusion rate deep relationship is constructed; The diffusion behavior of asphalt multiple regeneration process was analyzed based on the deep relationship between diffusion depth and average diffusion rate.

2. The method for analyzing diffusion behavior of asphalt during multiple regeneration processes according to claim 1, characterized in that: The generating of the diffusion depth distribution and the average diffusion rate distribution includes: analyzing the correlation between the diffusion depth distribution and the viscosity, complex shear modulus, storage modulus, loss modulus, and phase angle through a decision tree model, and constructing a first fingerprint library based on strong correlation factors to jointly characterize the diffusion depth distribution; analyzing the correlation between the average diffusion rate distribution and the viscosity, complex shear modulus, storage modulus, loss modulus, and phase angle through a decision tree model, and constructing a second fingerprint library based on strong correlation factors to jointly characterize the average diffusion rate distribution.

3. The method for analyzing diffusion behavior of asphalt during multiple regeneration processes according to claim 1, characterized in that: The method of obtaining the aging degree of asphalt in the regenerated asphalt and performing a diffusion behavior weight analysis of the asphalt multiple regeneration process to generate a severe aging stage and a mild aging stage includes: According to actual conditions, the asphalt multiple regeneration process of the regenerated asphalt is obtained; the asphalt multiple regeneration process is traversed to perform aging evaluation on the regenerated asphalt to obtain decay vector information, wherein the decay vector information includes infection force information; it is determined whether the infection force information is greater than or equal to the infection force threshold; if the infection force information is greater than or equal to the infection force threshold, the asphalt multiple regeneration process is added to the severe aging stage; if the infection force information is less than the infection force threshold, the asphalt multiple regeneration process is added to the mild aging stage.

4. The method for analyzing diffusion behavior of asphalt during multiple regeneration processes according to claim 1, characterized in that: The method of predicting the average diffusion rate based on the shallow diffusion depth time series data and generating the shallow diffusion depth-average diffusion rate relationship includes: Acquiring asphalt material information in the regenerated asphalt, wherein the asphalt material information in the regenerated asphalt corresponds to the shallow diffusion depth time series data; Performing average diffusion rate prediction based on the asphalt material information in the regenerated asphalt and the shallow diffusion depth time series data to generate average diffusion rate prediction time series information; The shallow layer relationship between diffusion depth and average diffusion rate is generated according to the shallow layer diffusion depth time series data and the average diffusion rate prediction time series information.

5. The method for analyzing diffusion behavior of asphalt during multiple regeneration processes according to claim 4, characterized in that: The predicting of the average diffusion rate based on the asphalt material information in the regenerated asphalt and the shallow diffusion depth time series data to generate the average diffusion rate prediction time series information includes: Acquiring detection record data of multiple asphalt regeneration processes, wherein the detection record data of multiple asphalt regeneration processes includes record data of asphalt material in the regenerated asphalt, record data of diffusion depth of the multiple asphalt regeneration processes, and record data of average diffusion rate of the multiple asphalt regeneration processes; Based on the BP neural network, the average diffusion rate prediction model is trained using the asphalt material record data in the regenerated asphalt, the diffusion depth record data of the asphalt multiple regeneration processes, and the average diffusion rate record data of the asphalt multiple regeneration processes; based on the asphalt material information in the regenerated asphalt, the shallow diffusion depth time series data is traversed and input into the average diffusion rate prediction model to generate the average diffusion rate prediction time series information.

6. The method for analyzing diffusion behavior of asphalt during multiple regeneration processes according to claim 1, characterized in that: The analysis of the diffusion behavior of the asphalt multiple regeneration process based on the deep relationship between the diffusion depth and the average diffusion rate includes: Constructing a teacher analysis model and a student analysis model, predicting the deep relationship between diffusion depth and average diffusion rate using the teacher analysis model to obtain a first prediction result, obtaining the credibility of the teacher analysis model for the first prediction result, updating the first prediction result based on the credibility, and using the updated first prediction result as the second prediction result; Obtaining a difference between the second prediction result and a result of the Student analysis model predicting the diffusion depth-average diffusion rate deep relationship, based on the difference, The parameters of the student analysis model are updated to train the student analysis model, and diffusion behavior analysis and prediction are performed using the trained student analysis model.

7. The method for analyzing diffusion behavior of asphalt during multiple regeneration processes according to claim 6, characterized in that: The obtaining of the credibility of the teacher analysis model for the first prediction result, updating the first prediction result according to the credibility, and using the updated first prediction result as the second prediction result includes: comparing each of the different credibility levels one by one; and determining the first prediction result corresponding to the minimum credibility as the second prediction result.

8. The method for analyzing diffusion behavior of asphalt during multiple regeneration processes according to claim 7, characterized in that: The obtaining of the credibility of the teacher analysis model for the first prediction result, updating the first prediction result according to the credibility, and using the updated first prediction result as the second prediction result also includes: calculating the weighted value corresponding to each first prediction result according to the credibility of each first prediction result; performing weighted summation on each first prediction result combined with the weighted value, and performing weighted averaging on the weighted sum and the second prediction result to obtain the updated second prediction result as the final prediction result.

Images