An epoxy asphalt compatibility quantitative analysis and formula optimization method based on interface trajectory identification

Abstract

The application relates to an epoxy asphalt compatibility quantitative analysis and formula optimization method based on interface trajectory identification. A composite system molecular model containing an epoxy system crosslinking network and asphalt four-component molecules is constructed, molecular dynamics trajectory characteristics of molecules in an interface region are obtained, and single-molecule trajectories are automatically classified based on a machine learning algorithm, so that identification of an epoxy phase, an asphalt phase and an interface mixed state is realized. On this basis, the proportion of mixed state molecules, the cross-phase diffusion depth and the trajectory distribution characteristics are extracted, and a compatibility index I is constructed, which is used for quantitatively evaluating the compatibility level under different formulas. The method introduces a "default preferred threshold + reference system adaptive calibration" strategy, so as to ensure the robustness and migration of trajectory classification under different temperatures, formulas and simulation conditions.

Description

Technical Field

[0001] This application relates to the interdisciplinary fields of road material modification, molecular simulation, and intelligent analysis, specifically to a method for quantitative analysis and formulation optimization of epoxy asphalt compatibility based on interface trajectory recognition. The method is used to identify the interfacial state between the epoxy system and the asphalt system at the molecular scale and quantitatively assess their compatibility level. It can be widely applied to the formulation optimization of epoxy resins, curing agents, asphalt, and compatibilizers, and falls within the scope of compatibility evaluation and structural characterization methods for polymer composite materials. Background Technology

[0002] Epoxy asphalt is a composite material formed by the reaction of epoxy resin and a curing agent in an asphalt matrix to create a three-dimensional cross-linked structure. Due to its excellent mechanical strength, heat resistance, and fatigue resistance, it has been widely used in bridge paving, heavy-duty roads, airport pavements, and high-temperature environment road surfaces. However, epoxy systems and asphalt differ fundamentally in polarity, molecular size, chain structure, and interaction energy, making them highly susceptible to phase separation, interfacial debonding, and even structural defects during curing or service, thus severely impacting the service life and reliability of the composite material.

[0003] Currently, compatibility evaluation mainly relies on macroscopic characterization methods such as softening point difference, storage stability tests, and fluorescence microscopy. These methods are significantly influenced by sample preparation and subjective judgment, and only provide "outcome-based" assessments, failing to reveal the dynamic formation process of compatibility. Some studies have attempted to use molecular simulations to calculate thermodynamic parameters such as solubility parameters, interaction energies, or enthalpy of mixing to predict compatibility trends. However, these parameters are mostly static averages of the system and cannot reflect cross-phase diffusion, microscopic miscibility, and dynamic interfacial structural behavior within local regions.

[0004] In other words, existing technologies still have two shortcomings: First, there is a lack of molecular-level criteria for dynamically identifying the "epoxy phase-asphalt phase-interfacial mixture"; second, there is a lack of a quantitative compatibility evaluation index that can be compared across material systems. This means that material screening still relies on empirical experiments and multiple rounds of formulation adjustments, resulting in long R&D cycles and a lack of clear mechanistic support. Therefore, there is an urgent need for an analytical method that addresses the needs of materials engineering, possesses dynamic molecular-scale discrimination capabilities, and can quantify compatibility levels to guide formulation design.

[0005] Based on this need, this invention introduces molecular dynamics trajectory analysis and machine learning classification methods, using the dynamic characteristics of molecules in the interface region as the core identification basis, to achieve automatic identification, quantitative evaluation, and formulation optimization of the compatibility of epoxy asphalt systems. Summary of the Invention

[0006] This invention aims to overcome two key shortcomings in existing epoxy asphalt compatibility evaluation methods: first, existing technologies lack the means to dynamically identify the epoxy phase, asphalt phase, and interfacial mixture at the molecular scale; second, there is currently no unified quantitative index system for compatibility that can be compared across material systems and used for formulation screening and parameter optimization. To address these shortcomings, this invention proposes a method based on molecular dynamics trajectory analysis and machine learning classification. This method identifies phase distributions from molecular motion behavior, constructs a compatibility index, and forms a closed-loop path for material design, upgrading compatibility from a qualitative judgment to a calculable and optimizable engineering quantity.

[0007] Technical issues

[0008] This invention mainly solves the following three technical problems:

[0009] (1) How to distinguish the three states between epoxy system and asphalt system at the molecular scale: fully cured cross-linked phase, free-moving asphalt phase and transitional interface mixed state;

[0010] (2) How to extract characteristic parameters with material physics significance from molecular dynamics trajectories and achieve automatic phase classification;

[0011] (3) How to transform the classification results into quantitative indicators that can be compared across formulations and systems, so that they can be used for material screening and parameter optimization.

[0012] Technical solution

[0013] To achieve the above objectives, this application provides a method for identifying and optimizing the compatibility of epoxy asphalt based on molecular dynamics and machine learning. The preparation method includes the following steps:

[0014] Step S1 Model Construction: Construct a three-dimensional molecular model containing an epoxy system crosslinking network and an asphalt model. The epoxy system uses epoxy resin monomers and curing agents as the reaction system, and forms a three-dimensional crosslinking network through a crosslinking reaction. The asphalt model is constructed using asphalt four-component molecules, which include saturated components, aromatic components, resins, and asphaltenes. The asphalt four-component molecules are configured according to the mass ratio.

[0015] Step S2 Molecular Dynamics Simulation: The three-dimensional molecular model constructed in Step S1 is used to perform evolution calculations using molecular dynamics simulation methods to obtain molecular trajectory data after the system reaches thermodynamic equilibrium. Under the set temperature and pressure conditions, the system is subjected to energy minimization and equilibrium simulation to bring the system to thermodynamic equilibrium. During the equilibrium stage, the motion trajectory data of each molecule is collected as input data for subsequent interface position identification and trajectory feature extraction.

[0016] Step S3 determines the interface position between epoxy and asphalt: Based on the molecular trajectory data obtained in step S2, the molecular composition ratio of each molecule in its neighborhood and its spatial distribution characteristics are statistically calculated to identify the epoxy phase, asphalt phase and interface region; the number density or mass density of different types of molecules are statistically analyzed along the Z direction of the model to form a molecular density distribution curve, and the interface position between the epoxy system and the asphalt system is determined according to the density gradient change.

[0017] Step S4 Trajectory Feature Extraction: Based on the molecular trajectory data obtained in step S2, the motion trajectories of each molecule in the interface region determined in step S3 are statistically analyzed, and trajectory features that can characterize molecular dynamics are calculated and extracted; wherein, the trajectory features include self-diffusion coefficient D, time dependence index α of mean square displacement (MSD), molecular trajectory envelope volume, and the proportion of cross-phase residence time of molecules in different phase regions.

[0018] Step S5 Phase Classification: Based on the trajectory features obtained in Step S4, a machine learning algorithm is used to classify the system components within the interface region determined in Step S3 to determine the phase state of each component. The system components include molecules derived from the epoxy crosslinking network and the asphalt model, and all molecules are treated as objects to be classified in this step. An adaptive threshold method is used to determine the phase state of each molecule, using the pure epoxy crosslinking system and the pure asphalt system as control systems. Statistical boundaries of molecular diffusion behavior are extracted to dynamically determine the classification boundaries of the epoxy phase, asphalt phase, and interface mixed state.

[0019] Step S6 Compatibility Analysis: Based on the phase classification results obtained in Step S5, calculate the compatibility index I, which is used to quantitatively evaluate the compatibility level of epoxy asphalt. The evaluation is based on one or more of the following: the proportion of mixed molecules at the interface, the depth of cross-phase diffusion, and the trajectory distribution characteristics.

[0020] Step S7 Formulation optimization: Adjust the formulation of epoxy resin, curing agent, asphalt or compatibilizer based on the compatibility index I, and repeat steps S1 to S6 after adjustment to form a closed-loop iterative optimization process based on compatibility index I.

[0021] This application also provides an epoxy asphalt compatibility analysis system for implementing the epoxy asphalt compatibility identification and optimization method based on molecular dynamics and machine learning. The system includes: a model building module, a molecular dynamics simulation module, a trajectory feature extraction module, a machine learning classification module, a compatibility index calculation module, and a formulation optimization module. Each module communicates with the other in sequence and outputs the compatibility index and formulation optimization suggestions.

[0022] This application also provides a computer-readable storage medium having program instructions stored thereon, which, when executed by a processor, cause the processor to perform the method for identifying and optimizing the compatibility of epoxy asphalt based on molecular dynamics and machine learning.

[0023] Beneficial effects

[0024] Compared with the prior art, the present invention has the following advantages:

[0025] This invention is the first to establish a unified methodology integrating "interface trajectory recognition + adaptive threshold determination + compatibility index construction." Compared to traditional methods that only observe macroscopic structures or calculate energy differences, this invention can reveal the molecular motion behavior and mutual solubility mechanism of the epoxy-asphalt interface, and can transform microscopic-scale judgments into quantitative indicators applicable to engineering design. It can be used for:

[0026] (1) Replace empirical formulation screening and realize model-based material design;

[0027] (2) Supports high-throughput simulation and algorithm optimization, significantly shortening the R&D cycle of epoxy asphalt materials;

[0028] (3) Establish an evaluation system that links experiment, simulation and optimization, and establish a micro-correspondence for experimental indicators such as softening point difference, segregation height and FM similarity;

[0029] (4) Enable cross-system comparison capabilities so that different epoxy resin types, curing agent systems, asphalt sources or compatibilizer ratios can all be uniformly included in the evaluation framework;

[0030] (5) Construct a scalable technology platform that is not only applicable to epoxy asphalt, but also extends to sulfurized asphalt, polyurethane asphalt and other polymer-asphalt composite systems.

[0031] Therefore, this invention realizes a continuous path from "microscopic identification - quantitative evaluation - formula optimization", which fills the technical gap of existing technologies that only have static images or macroscopic conclusions but lack dynamic mechanism descriptions and calculable index systems, and has significant theoretical significance and engineering application value. Attached Figure Description

[0032] To more clearly illustrate the technical solution and implementation process of the present invention, the accompanying drawings are used to assist in understanding the method flow, interface molecular trajectory classification method and compatibility index construction mechanism of the present invention.

[0033] Figure 1 This diagram illustrates the overall method flowchart of the present invention.

[0034] Figure 2 A schematic diagram of the overall method according to an embodiment of the present invention

[0035] Figure 3A schematic diagram showing the classification of molecular trajectories in the interface region is shown.

[0036] Figure 4 This diagram illustrates how the compatibility index I varies with different formulation variables.

[0037] Figure 5 This shows the correlation between the compatibility index I and the softening point difference.

[0038] The accompanying drawings are for illustrative purposes only and do not constitute a limitation on the scope of protection of this invention. Detailed Implementation

[0039] The preferred embodiments of the present invention will now be described in detail with reference to specific examples. It should be understood that the following examples are given for illustrative purposes only and are not intended to limit the scope of the invention. Those skilled in the art can make various modifications and substitutions to the present invention without departing from its spirit and essence.

[0040] Unless otherwise specified, the experimental methods used in the following examples are conventional methods.

[0041] Unless otherwise specified, all materials and reagents used in the following examples are commercially available.

[0042] One embodiment of this application provides a method for quantitative analysis of the compatibility of epoxy asphalt and formulation optimization based on interface trajectory recognition. The preparation method includes the following steps:

[0043] Step S1 Model Construction: Construct a three-dimensional molecular model containing an epoxy system crosslinking network and an asphalt model. The epoxy system uses epoxy resin monomers and curing agents as the reaction system, and forms a three-dimensional crosslinking network through a crosslinking reaction. The asphalt model is constructed using asphalt four-component molecules, which include saturated components, aromatic components, resins, and asphaltenes. The asphalt four-component molecules are configured according to the mass ratio.

[0044] Step S2 Molecular Dynamics Simulation: The three-dimensional molecular model constructed in Step S1 is used to perform evolution calculations using molecular dynamics simulation methods to obtain molecular trajectory data after the system reaches thermodynamic equilibrium. Under the set temperature and pressure conditions, the system is subjected to energy minimization and equilibrium simulation to bring the system to thermodynamic equilibrium. During the equilibrium stage, the motion trajectory data of each molecule is collected as input data for subsequent interface position identification and trajectory feature extraction.

[0045] Step S3 determines the interface position between epoxy and asphalt: Based on the molecular trajectory data obtained in step S2, the molecular composition ratio of each molecule in its neighborhood and its spatial distribution characteristics are statistically calculated to identify the epoxy phase, asphalt phase and interface region; the number density or mass density of different types of molecules are statistically analyzed along the Z direction of the model to form a molecular density distribution curve, and the interface position between the epoxy system and the asphalt system is determined according to the density gradient change.

[0046] Step S4 Trajectory Feature Extraction: Based on the molecular trajectory data obtained in step S2, the motion trajectories of each molecule in the interface region determined in step S3 are statistically analyzed, and trajectory features that can characterize molecular dynamics are calculated and extracted; wherein, the trajectory features include self-diffusion coefficient D, time dependence index α of mean square displacement (MSD), molecular trajectory envelope volume, and the proportion of cross-phase residence time of molecules in different phase regions.

[0047] Step S5 Phase Classification: Based on the trajectory features obtained in Step S4, a machine learning algorithm is used to classify the system components within the interface region determined in Step S3 to determine the phase state of each component. The system components include molecules derived from the epoxy crosslinking network and the asphalt model, and all molecules are treated as objects to be classified in this step. An adaptive threshold method is used to determine the phase state of each molecule, using the pure epoxy crosslinking system and the pure asphalt system as control systems. Statistical boundaries of molecular diffusion behavior are extracted to dynamically determine the classification boundaries of the epoxy phase, asphalt phase, and interface mixed state.

[0048] Step S6 Compatibility Analysis: Based on the phase classification results obtained in Step S5, calculate the compatibility index I, which is used to quantitatively evaluate the compatibility level of epoxy asphalt. The evaluation is based on one or more of the following: the proportion of mixed molecules at the interface, the depth of cross-phase diffusion, and the trajectory distribution characteristics.

[0049] Step S7: Formulation optimization: Adjust the formulations of epoxy resin, curing agent, asphalt, or compatibilizer based on the compatibility index I, and repeat steps S1 to S6 after adjustment to form a closed-loop iterative optimization process based on the compatibility index I. See the flowchart for the relevant method. Figure 1 .

[0050] In one embodiment, during the specific modeling process in step S1, the mass fraction can be equivalently converted into a volume fraction or molecular number ratio based on the molecular density of each component for the construction of the molecular model.

[0051] In one embodiment, the asphalt model in step S1 is constructed according to the mass or volume fraction ratio of the actual asphalt sample. The asphalt model is constructed by selecting molecules that can represent the structural characteristics of saturated components, aromatic components, resins and asphaltenes as representative molecules of the four components of asphalt. The relative number of each representative molecule is determined according to the mass or volume fraction ratio of each component in the actual asphalt sample to form a molecular model corresponding to the composition of the actual asphalt.

[0052] In one embodiment, the curing agent includes amine curing agents, acid anhydride curing agents, and polymer curing agents.

[0053] In one embodiment, the four components of asphalt in step S1 are configured according to a mass ratio, wherein the mass fraction ratio of the four components is: saturated component 10-30%, aromatic component 20-40%, resin 20-40%, and asphaltene 10-30%.

[0054] In one embodiment, in step S1, the relative contents of the epoxy system and the asphalt system in the three-dimensional molecular model are set according to the mass ratio, and the mass fraction of the epoxy system in the composite system is 20-50%.

[0055] In one embodiment, the molecular dynamics simulation described in step S2 is performed using a temperature- and pressure-isothermal ensemble (NPT ensemble) for equilibration.

[0056] In one embodiment, the temperature range of the molecular dynamics simulation in step S2 is 25–180°C, the pressure is 1 atm, the time step is 0.5–2.0 fs, and the equilibrium time is not less than 10 ns.

[0057] In one embodiment, step S5, determining the phase state of a molecule, employs an adaptive threshold determination method based on molecular dynamics behavior characteristics, specifically including:

[0058] The self-diffusion coefficient D and the time-dependent index α of mean square displacement (MSD) are calculated for the trajectories of molecules in the interface region. Phase classification is then performed according to the following rules: Statistical analysis is conducted on the motion trajectories of the system's constituent molecules in the interface region to calculate the self-diffusion coefficient D and the time-dependent index α of the mean square displacement (MSD) for each molecule. Phase classification is then performed based on the range of values ​​for the self-diffusion coefficient D and the time-dependent index α of the MSD.

[0059] (1) When D is below the first diffusion threshold and / or α is below the first exponential threshold, it is determined to be an epoxy phase molecule;

[0060] (2) When D is higher than the second diffusion threshold and / or α is higher than the second exponential threshold, it is determined to be a bitumen phase molecule;

[0061] (3) When a molecule simultaneously satisfies the conditions between the first threshold and the second threshold, it is determined to be an interface mixed state molecule;

[0062] The first diffusion threshold and the second diffusion threshold are adaptively determined based on the statistical valley or geometric mean of the logarithmic distribution of the diffusion coefficient obtained under the same simulation conditions as the pure epoxy crosslinked body and the pure asphalt melt reference system.

[0063] In one embodiment, the first diffusion threshold is 10⁻¹³ cm² / s, and the second diffusion threshold is 10⁻¹³ cm² / s. 0 cm² / s.

[0064] In one embodiment, the first threshold value of α is 0.2, and the second threshold value is 0.8.

[0065] In one embodiment, in step S5, the machine learning algorithm used to classify the trajectories of the system components within the interface region includes a trajectory classification algorithm, which is an unsupervised clustering, semi-supervised learning, or a deep learning model based on trajectory feature vectors.

[0066] In one embodiment, in step S6, the compatibility index I is calculated based on the trajectory classification result. The compatibility index I is calculated by weighting the normalized P', R', and D' according to the following formula:

[0067] I = w1·P' + w2·R' + w3·D',

[0068] Where P' is the proportion of molecules in the interface mixed state, R' is the cross-phase diffusion depth, D' is the trajectory dispersion, and w1, w2, and w3 are non-negative weighting coefficients in the range of 0 to 1 and satisfy w1+w2+w3=1.

[0069] In one embodiment, I ≥ 0.70 indicates high compatibility, 0.40 ≤ I < 0.70 indicates medium compatibility, and I < 0.40 indicates low compatibility.

[0070] One embodiment of this application provides an epoxy asphalt compatibility analysis system for implementing the epoxy asphalt compatibility quantitative analysis and formulation optimization method based on interface trajectory recognition. The system comprises: a model building module, a molecular dynamics simulation module, a trajectory feature extraction module, a machine learning classification module, a compatibility index calculation module, and a formulation optimization module. Each module communicates sequentially and outputs a compatibility index and formulation optimization suggestions.

[0071] One embodiment of this application provides a computer-readable storage medium storing program instructions thereon, which, when executed by a processor, cause the processor to perform the described method for quantitative analysis of epoxy asphalt compatibility and formulation optimization based on interface trajectory recognition.

[0072] Figure 2 The diagram illustrates a method according to an embodiment of the present invention, comprising seven consecutive steps. First, a molecular model of a composite system containing an epoxy resin crosslinking network and asphalt four-component molecules is constructed, and molecular trajectory data under equilibrium conditions is obtained through molecular dynamics simulation. Then, interface regions are identified, and molecular trajectory features (such as self-diffusion coefficient D, MSD time exponent α, trajectory envelope volume, and residence ratio) are extracted. The trajectories are then classified using a machine learning algorithm to automatically determine the phase state of the molecules. Further, a compatibility index I is constructed based on the classification results, serving as a core indicator for evaluating the compatibility level under different formulations. Finally, the I value is used to guide formulation adjustments, forming a closed-loop path of "modeling—analysis—feedback—optimization," achieving quantifiable, interpretable, and comparable compatibility identification and design of the epoxy asphalt material system.

[0073] The key innovation of this invention lies in the fact that it transforms molecular trajectory data into structural phase information for the first time, and further condenses it into an engineering-usable compatibility index I. This transforms the compatibility evaluation process, which originally relied on subjective judgment or experimental experience, into a quantitative analysis method based on simulation data and algorithm identification, which is suitable for high-throughput material screening and quantitative formulation optimization tasks.

[0074] Example 1: Compatibility Identification of Baseline Epoxy Asphalt Systems

[0075] This embodiment constructs a typical epoxy asphalt composite system model to verify the ability of the method of the present invention to identify and quantitatively analyze the compatibility level of the system under the condition of no external compatibilizer.

[0076] (1) Model building

[0077] Bisphenol A type epoxy resin monomer (EP) and aliphatic amine curing agent (EDA) were selected as the reaction system. The asphalt component included 60 representative molecules of the four SARA components, configured by mass ratio as follows: 20% saturated component, 35% aromatic component, 25% resin, and 20% asphaltenes. The system contained 15 epoxy resin molecules and 30 curing agent molecules, with the total number of asphalt molecules matched by volume fraction, resulting in a cross-linked network of approximately 35% in the epoxy system.

[0078] (2) Crosslinking reaction setup

[0079] The simulation was performed using Materials Studio software, and the system was set to automatically form covalent bonds when the distance between the epoxy and amine groups was less than 3.5 Å. The crosslinking temperature was set to 160 °C, the heating phase lasted for 500 ps, ​​and the final degree of curing was approximately 75%.

[0080] (3) Molecular dynamics simulation

[0081] The COMPASS force field was used for simulation under NPT conditions (temperature 60℃, pressure 1 atm) with a time step of 1.0 fs. After minimizing the system energy, equilibrium simulation was performed for 12 ns, and the molecular trajectory in the last 8 ns was used for subsequent feature extraction and analysis.

[0082] (4) Interface area recognition

[0083] The molecular density distribution along the Z-direction is calculated to determine the interface location between epoxy and asphalt. The interface region is defined as a band-shaped region within ±1.5 nm of the density gradient region.

[0084] (5) Feature extraction and classification

[0085] A total of 86 molecules were identified within the interface band, and six features were extracted from them: neighborhood composition ratio, self-diffusion coefficient D, MSD time exponent α, trajectory envelope volume, and cross-phase residence time ratio. A K-means clustering algorithm with K=3 was used for phase classification, supplemented by a soft constraint of "neighborhood ratio ≥ 70%" to improve classification stability.

[0086] (6) Results and Analysis

[0087] The classification results show:

[0088] Epoxy phase molecules account for 41%;

[0089] The asphalt phase molecules account for 36%;

[0090] Interfacial mixed-state molecules account for 23%.

[0091] The compatibility index formula of this invention is used to calculate:

[0092] P=0.23, R=0.78 (nm), D=0.51 (normalized value)

[0093] A compatibility index of I = 0.58 indicates a "moderate compatibility" level.

[0094] (7) Verification of technical effects and thresholds

[0095] See the schematic diagram of molecular trajectory classification in the experimental interface region. Figure 3 In the figure, different colors or trajectory shapes represent molecular trajectories identified as epoxy phase, asphalt phase, and interface mixture, respectively, to illustrate the principle and effect of this invention in achieving phase differentiation by extracting molecular trajectory features and combining them with machine learning algorithms. The logical relationships and processing flow between each step are schematic representations and do not limit the parameter values ​​or calculation results in specific embodiments.

[0096] The results show that the method of the present invention can effectively identify interfacial mixed-state molecules and output a quantitative compatibility index, possessing the capability for benchmark evaluation of material formulations. To verify the universality of the adopted D and α thresholds, pure epoxy crosslinked and pure asphalt systems were simulated under the same MD parameter conditions, and logarithmic values ​​were calculated. 10 The results show a D-distribution. The results exhibit a bimodal structure, with the middle valley corresponding to D≈3.1×10⁻¹²cm² / s, falling within the interface classification interval (10⁻¹³~10⁻¹) used in this invention. 0 Within cm² / s). The mean α value of the pure epoxy system is about 0.05, and the mean α value of the pure asphalt system is about 0.95, further verifying that the sub-diffusion behavior of α∈(0,1) is reasonable and generalizable as a criterion for interface state.

[0097] Example 2: Analysis of the effect of different compatibilizer contents on the compatibility index

[0098] This embodiment demonstrates the application capability of the method of the present invention in the screening of compatibilizer formulations. It is a schematic diagram showing the change of compatibility index I with different formulation variables, such as adjustments to epoxy resin content, compatibilizer dosage, or curing agent type, reflecting the trend of the compatibility index change corresponding to these adjustments. This illustrates that the present invention can be used for quantitative comparison and screening between different formulation systems. Experimental results are shown in [link to experimental results]. Figure 4 .

[0099] (1) System setting

[0100] Continuing with the system of Example 1, amphiphilic compatibilizers (designed to have both epoxy affinity and aromatic affinity ends) were added at mass fractions of 0%, 0.5%, 1.0%, and 2.0%, respectively.

[0101] (2) Simulation and Analysis Process

[0102] All systems used the same modeling and molecular dynamics simulation conditions, and employed a consistent trajectory feature extraction and phase classification process.

[0103] (3) Results of the I index

[0104] Compatibilizer content Compatibility Index 0 0.58 0.5% 0.66 1.0% 0.74 2.0% 0.68

[0105] (4) Trend description and mechanism analysis

[0106] The compatibility index first increases and then decreases with the addition of compatibilizer, indicating the existence of an optimal dosage range. Appropriate amounts of compatibilizer can enhance interfacial miscibility, improve the proportion of mixed states and diffusion coordination, while excessive compatibilizer may lead to molecular aggregation, disrupt the ordered structure of the interface, and cause a decrease in the R and D indices, resulting in a drop in the I value.

[0107] (5) Technical effects

[0108] The method of this invention can directly output the optimal dosage range, significantly reducing the reliance on traditional formulation test paths.

[0109] Example 3: Comparison of the effects of different types of curing agents on compatibility

[0110] (1) Setting up a comparison system

[0111] Crosslinking systems were constructed using aliphatic amine curing agent EDA and acid anhydride curing agent MHHPA, respectively, with all other parameters kept consistent, and the degree of curing controlled at 75% ± 3%. In this embodiment, to compare the effects of different curing agent types on interfacial compatibility, the proportion of mixed-state molecules at the interface P' and the cross-phase diffusion depth R' were selected as the main indicators to simplify the calculation of the compatibility index I. Although trajectory dispersion D' can also be used as a supplementary parameter, this example focuses on the direct impact of structural compactness on interfacial behavior.

[0112] (2) Trajectory analysis results

[0113] Curing agent type The proportion of molecules in the mixed state at the interface, P' Transphase diffusion depth R' / nm Compatibility Index I Amine-based curing agents (EDA) 0.26 0.85 0.67 Anhydride curing agents (MHHPA) 0.15 0.49 0.53

[0114] (3) Conclusion

[0115] The denser the cured structure (e.g., in anhydride systems), the thinner the interfacial transition zone, resulting in a decrease in both P' and R', and consequently a significant drop in the compatibility index I. The method of this invention can quantitatively analyze the impact of different curing systems on compatibility behavior and provide a basis for structural design optimization (e.g., introducing flexible segments or reducing crosslinking density).

[0116] Example 4: Correlation verification between the method of the present invention and macroscopic performance

[0117] This diagram illustrates the correlation between the compatibility index I constructed in this invention and conventional engineering characterization indicators (such as softening point difference ΔT). It shows that there is a good quantitative correspondence between the I value and the experimentally measured macroscopic performance indicators, which is used to establish a linkage analysis framework of "molecular simulation - experimental verification - performance evaluation".

[0118] (1) System and indicator setting

[0119] To verify the correspondence between the micro-compatibility index and the macro-performance index, three formulations were selected for storage stability tests and softening point difference ΔT measurements, and compared with the I index.

[0120] Formula number Compatibility Index I Softening point difference ΔT / ℃ A 0.54 18.2 B 0.63 11.6 C 0.74 6.8

[0121] (2) Results Analysis and Technical Effects

[0122] The results show that the I value is significantly negatively correlated with ΔT; the higher the I value, the smaller the softening point difference, and the more compatible the system is. The method of this invention can be used as an experimental performance prediction tool for pre-screening potential formulations, reducing experimental rounds, and improving the efficiency of material development.

[0123] This invention can be used for:

[0124] (1) Development and screening of epoxy asphalt pavement material formulations;

[0125] (2) Performance evaluation and content optimization of compatibilizer;

[0126] (3) Assessment of the substitutability of asphalt from different sources;

[0127] (4) Provide "virtual testing" capabilities to reduce physical testing by 40% to 60%;

[0128] (5) Establish an enterprise-level "digital R&D platform for road materials".

[0129] See results Figure 5 The diagram illustrates the relationship between the compatibility index I and conventional engineering characterization indicators such as the softening point difference ΔT. It is used to explain that the compatibility index constructed in this invention can be used to characterize the correlation between molecular-scale compatibility evaluation results and macroscopic performance indicators. The relationship shown in the diagram is illustrative and does not limit the specific distribution of experimental data points.

[0130] The above are merely preferred embodiments of the present invention. It should be noted that, for those skilled in the art, numerous improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for quantitative analysis of epoxy asphalt compatibility and formulation optimization based on interface trajectory recognition, characterized in that, The preparation method includes the following steps: Step S1 Model Construction: Construct a three-dimensional molecular model containing an epoxy system crosslinking network and an asphalt model. The epoxy system uses epoxy resin monomers and curing agents as the reaction system, and forms a three-dimensional crosslinking network through a crosslinking reaction. The asphalt model is constructed using asphalt four-component molecules, which include saturated components, aromatic components, resins, and asphaltenes. The asphalt four-component molecules are configured according to the mass ratio. Step S2 Molecular Dynamics Simulation: The three-dimensional molecular model constructed in Step S1 is used to perform evolution calculations using molecular dynamics simulation methods to obtain molecular trajectory data after the system reaches thermodynamic equilibrium. Under the set temperature and pressure conditions, the system is subjected to energy minimization and equilibrium simulation to bring the system to thermodynamic equilibrium. During the equilibrium stage, the motion trajectory data of each molecule is collected as input data for subsequent interface position identification and trajectory feature extraction. Step S3 determines the interface position between epoxy and asphalt: Based on the molecular trajectory data obtained in step S2, the molecular composition ratio of each molecule in its neighborhood and its spatial distribution characteristics are statistically calculated to identify the epoxy phase, asphalt phase and interface region; the number density or mass density of different types of molecules are statistically analyzed along the Z direction of the model to form a molecular density distribution curve, and the interface position between the epoxy system and the asphalt system is determined according to the density gradient change. Step S4 Trajectory Feature Extraction: Based on the molecular trajectory data obtained in step S2, the motion trajectories of each molecule in the interface region determined in step S3 are statistically analyzed, and trajectory features that can characterize molecular dynamics are calculated and extracted; wherein, the trajectory features include self-diffusion coefficient D, time dependence index α of mean square displacement (MSD), molecular trajectory envelope volume, and the proportion of cross-phase residence time of molecules in different phase regions. Step S5 Phase Classification: Based on the trajectory features obtained in Step S4, a machine learning algorithm is used to classify the system components within the interface region determined in Step S3 to determine the phase state of each component. The system components include molecules derived from the epoxy crosslinking network and the asphalt model, and all molecules are treated as objects to be classified in this step. An adaptive threshold method is used to determine the phase state of each molecule, using the pure epoxy crosslinking system and the pure asphalt system as control systems. Statistical boundaries of molecular diffusion behavior are extracted to dynamically determine the classification boundaries of the epoxy phase, asphalt phase, and interface mixed state. Step S6 Compatibility Analysis: Based on the phase classification results obtained in Step S5, calculate the compatibility index I, which is used to quantitatively evaluate the compatibility level of epoxy asphalt. The evaluation is based on one or more of the following: the proportion of mixed molecules at the interface, the depth of cross-phase diffusion, and the trajectory distribution characteristics. Step S7 Formulation optimization: Adjust the formulation of epoxy resin, curing agent, asphalt or compatibilizer based on the compatibility index I, and repeat steps S1 to S6 after adjustment to form a closed-loop iterative optimization process based on compatibility index I.

2. The method for quantitative analysis and formulation optimization of epoxy asphalt compatibility based on interface trajectory recognition according to claim 1, characterized in that, The four components of asphalt in step S1 are configured according to the following mass ratio: saturated component 10-30%, aromatic component 20-40%, resin 20-40%, and asphaltene 10-30%.

3. The method for quantitative analysis and formulation optimization of epoxy asphalt compatibility based on interface trajectory recognition according to claim 1, characterized in that, In step S1, the relative contents of the epoxy system and the asphalt system in the three-dimensional molecular model are set according to the mass ratio, and the mass fraction of the epoxy system in the composite system is 20-50%.

4. The method for quantitative analysis and formulation optimization of epoxy asphalt compatibility based on interface trajectory recognition according to claim 1, characterized in that, The molecular dynamics simulation described in step S2 uses an isothermal and isobaric ensemble (NPT ensemble) for equilibrium. The temperature range of the molecular dynamics simulation is 25–180 °C, the pressure is 1 atm, the time step is 0.5–2.0 fs, and the equilibrium time is not less than 10 ns.

5. The method for quantitative analysis and formulation optimization of epoxy asphalt compatibility based on interface trajectory recognition according to claim 1, characterized in that, In step S5, the determination of the phase state of the molecule adopts an adaptive threshold determination method based on molecular dynamics behavior characteristics, specifically including: The self-diffusion coefficient D and the time-dependent index α of mean square displacement (MSD) are calculated for the trajectories of molecules in the interface region. Phase classification is then performed according to the following rules: Statistical analysis is conducted on the motion trajectories of the system's constituent molecules in the interface region to calculate the self-diffusion coefficient D and the time-dependent index α of the mean square displacement (MSD) for each molecule. Phase classification is then performed based on the range of values ​​for the self-diffusion coefficient D and the time-dependent index α of the MSD. (1) When D is below the first diffusion threshold and / or α is below the first exponential threshold, it is determined to be an epoxy phase molecule; (2) When D is higher than the second diffusion threshold and / or α is higher than the second exponential threshold, it is determined to be a bitumen phase molecule; (3) When a molecule simultaneously satisfies the conditions between the first threshold and the second threshold, it is determined to be an interface mixed state molecule; The first diffusion threshold and the second diffusion threshold are adaptively determined based on the statistical valley or geometric mean of the logarithmic distribution of the diffusion coefficient obtained under the same simulation conditions as the pure epoxy crosslinked body and the pure asphalt melt reference system.

6. The method for quantitative analysis and formulation optimization of epoxy asphalt compatibility based on interface trajectory recognition according to claim 1, characterized in that, In step S5, the machine learning algorithm used to classify the trajectories of the system components within the interface region includes a trajectory classification algorithm, which is an unsupervised clustering, semi-supervised learning, or a deep learning model based on trajectory feature vectors.

7. The method for quantitative analysis and formulation optimization of epoxy asphalt compatibility based on interface trajectory recognition according to claim 1, characterized in that, In step S6, the compatibility index I is calculated based on the trajectory classification results. The compatibility index I is calculated by weighting the normalized P', R', and D' according to the following formula: I = w1·P' + w2·R' + w3·D', Where P' is the proportion of molecules in the interface mixed state, R' is the cross-phase diffusion depth, D' is the trajectory dispersion, and w1, w2, and w3 are non-negative weighting coefficients in the range of 0 to 1 and satisfy w1+w2+w3=1.

8. The method for quantitative analysis and formulation optimization of epoxy asphalt compatibility based on interface trajectory recognition according to claim 7, characterized in that, A value of I ≥ 0.70 indicates high compatibility, 0.40 ≤ I < 0.70 indicates medium compatibility, and I < 0.40 indicates low compatibility.

9. A system for analyzing the compatibility of epoxy asphalt for implementing the method of claim 1, characterized in that, include: The system consists of a model building module, a molecular dynamics simulation module, a trajectory feature extraction module, a machine learning classification module, a compatibility index calculation module, and a formulation optimization module. Each module communicates with the others in sequence and outputs the compatibility index and formulation optimization suggestions.

10. A computer-readable storage medium having stored thereon program instructions that, when executed by a processor, cause to perform the method as described in any one of claims 1 to 8.

Images