Evaluation method for low-temperature cracking resistance toughness of surface layer of asphalt pavement

Abstract

The invention discloses a method for evaluating low-temperature cracking resistance toughness of a surface layer of an asphalt pavement, and belongs to the technical field of road engineering and material performance evaluation. The evaluation method comprises the following steps: acquiring meteorological data and pavement structure material parameters of a pavement location; taking a temperature fatigue damage factor as a toughness evaluation index; the method comprises the following steps: constructing a seven-level eight-factor orthogonal design table, carrying out thermosetting coupling finite element simulation by utilizing COMSOL Multiphysics, and establishing an orthogonal database containing various environment and structure parameters; and training the database by adopting a neural network optimized by a genetic algorithm to obtain an intelligent evaluation model capable of quickly and accurately predicting the low-temperature cracking resistance toughness of the newly-built asphalt pavement. According to the method, physical simulation and data driving are combined, quantitative evaluation and optimal design of the toughness of the asphalt pavement under the extremely low temperature condition are achieved, and a scientific basis and a practical tool are provided for design and checking calculation of a high-toughness pavement structure in the Tibet region.

Description

Technical Field

[0001] This invention relates to the field of road engineering and materials evaluation, and in particular to an evaluation method for the low-temperature cracking toughness of asphalt pavement surface layers based on finite element simulation and neural networks. Background Technology

[0002] Located in the southwestern part of the Qinghai-Tibet Plateau, Tibet has an average altitude of over 4,000 meters and is known as the "Roof of the World." Its complex geographical environment, undulating terrain, and extreme climate conditions are characterized by strong ultraviolet radiation, significant diurnal temperature variations, extreme low temperatures in winter, and complex regional climate features. These factors combine to cause severe early-stage damage to asphalt pavements, particularly cracking damage (such as low-temperature shrinkage cracks and fatigue cracks) and loosening damage. For example, pavement surveys along the Qinghai-Tibet Highway and in southeastern Tibet indicate that radiation aging, temperature cycling, and low-temperature embrittlement are the main causes of pavement performance degradation, severely restricting the durability and service capacity of local highways.

[0003] Against this backdrop, traditional pavement design and assessment methods have revealed their limitations. On the one hand, existing highway disaster assessment systems tend to focus on macro-level risk control, lacking detailed consideration of regional specificities and the characteristics of pavement materials and structures, making it difficult to accurately reflect the actual resilience of pavements under extreme environments. On the other hand, although some studies have attempted to fill assessment gaps by simulating material performance degradation under extreme environments, these studies have not yet formed a complete assessment system integrating environmental factors, structural response, and performance degradation mechanisms. Furthermore, traditional methods relying on manual inspection and historical data suffer from data lag, inconsistent data types, and high incompleteness, further affecting the accuracy and reliability of the assessment. Summary of the Invention

[0004] To address the problem of low-temperature cracking of asphalt pavements in the extreme environment of Tibet, this invention aims to provide an evaluation method for the low-temperature cracking toughness of asphalt pavement surface layers based on finite element simulation and neural networks. This method aims to improve the tolerance and resilience of highways in Tibetan areas under extreme conditions, thereby supporting the sustainable development of transportation and the healthy operation of the social economy.

[0005] The objective of this invention can be achieved through the following technical solutions: A method for evaluating the low-temperature cracking toughness of asphalt pavement surface layers includes the following steps: S1: Obtain meteorological data for the area where the new asphalt pavement is located, including the lowest monthly average temperature in winter, the lowest annual temperature, the maximum daily temperature difference in a year, the longest duration of low temperature, wind speed, daily average solar radiation, ultraviolet intensity, and other environmental data. S2: Determine the internal parameters such as surface layer modulus, surface layer thermal expansion coefficient, and surface layer thickness based on the proposed asphalt pavement structure material selection; S3: The ratio of the temperature stress at the top of the asphalt pavement surface layer to the splitting strength of the asphalt surface layer is taken as the temperature fatigue damage factor. The temperature fatigue damage factor and the temperature fatigue damage caused by the preset extreme low temperature conditions are used as toughness evaluation indicators. S4: Based on the meteorological data and internal parameters obtained above, set up a seven-level, eight-factor orthogonal design table for finite element simulation modeling to construct an orthogonal database; S5: Regression fitting is performed on the external environment data, pavement structure internal parameters and toughness evaluation index in the database using a neural network; S6: Use the trained neural network to perform toughness calculations on the structure and materials of the newly built asphalt pavement. If the toughness index does not meet the set threshold, then redesign.

[0006] Optionally, in step S1, meteorological environmental data is collected. The data can be meteorological data for the region in the next 10 years as predicted during the design. If there is no predicted data, historical data from the past 10 years is used. The data accuracy is at the county / district level, and the data interval is in hours.

[0007] Optionally, in step S2, the commonly used road surface structure in Tibetan areas is selected as follows: ① For Class III highways and below: 18~20cm semi-rigid base course + 4~5cm asphalt mixture layer; ② For Class II highways: 20~30cm semi-rigid base course + 6cm asphalt mixture lower layer + 4cm asphalt mixture upper layer; ③ For Class I highways and above: 32~40cm semi-rigid base course + 7~12cm asphalt mixture or asphalt macadam lower layer + 6cm asphalt mixture intermediate layer + 4cm asphalt mixture upper layer.

[0008] Optionally, in step S4, the process of designing the seven-level, eight-factor orthogonal design table includes: The eight factors are minimum temperature (°C), daily temperature range (°C), and daily total solar radiation (MJ / m²). 2 ), wind speed (m / s), pavement thickness (cm), surface modulus (MPa), surface thermal expansion coefficient (10) -5 ) and UV intensity (based on region classification); The lowest temperature is obtained by taking the lowest monthly average temperature and the annual maximum daily temperature difference in the county where the road is located during winter. The average temperature is used as the fourth level, the average temperature plus half of the annual maximum daily temperature difference is the first level, the average temperature minus half of the annual maximum daily temperature difference is the seventh level, and the intermediate levels are divided evenly. The daily temperature range is set as the seventh level with the maximum annual daily temperature range, the first level with the minimum annual daily temperature range, and the intermediate levels are evenly divided. Wind speed levels are divided into 0, 1, 2, 3, 4, 5, and 6, corresponding to levels one through seven. The total daily solar radiation is 2.5, 5, 7.5, 10, 12.5, 15, and 17.5, corresponding to levels one through seven. The surface layer thickness is determined automatically by the preset road structure or by using recommended values: 3, 4, 5, 6, 7, 3+5, 4+6, corresponding to the first to the seventh level. The surface layer modulus is determined automatically by the preset pavement structure or by using recommended values: 3000, 3500, 4500, 5500, 6500, 7500, and 8500, corresponding to the first to seventh levels. The surface layer has thermal expansion coefficients of 1.25, 1.5, 1.75, 2, 2.25, 2.5, and 2.75, corresponding to levels one through seven. Ultraviolet intensity was analyzed based on the ultraviolet spectrum.

[0009] Optionally, in step S4, the finite element simulation process includes: Step S4-1: Construct a three-dimensional geometric model using the COMSOL Multiphysics finite element simulation model; Step S4-2: Construct a thermo-mechanical coupling model and calculate the maximum temperature fatigue damage factor for each set of working conditions; Step S4-3: Simulate the temperature fatigue damage to the road surface caused by different low-temperature durations by setting the calculation time; Step S4-4: Export the temperature fatigue damage database corresponding to all operating conditions.

[0010] Optionally, in step S5, the neural network model used is the GA-BP model, i.e., the genetic algorithm, with the following parameter settings: number of generations: 5; population size: 10; real number encoding: [1e... -6 [1], the selection function parameter is 0.09, the crossover function parameter is 2, and the mutation function parameter is [2 gen 3].

[0011] Optionally, in step S5, the setup of the training set and the test set should take into account that the data parameters are orthogonally designed, and the selection of the training set and the test set should be random.

[0012] Optionally, in step S6, the toughness standard is set according to the designed asphalt pavement grade. If it is a highway, the toughness index threshold should not exceed 0.7. If the road grade is reduced, the threshold can be appropriately increased.

[0013] Compared with the prior art, the present invention has the following beneficial effects: 1. Achieved accurate quantitative assessment of multi-factor coupling in extremely complex environments: This invention is the first to systematically couple and model multiple key environmental factors in Tibet, such as extreme low temperatures, huge diurnal temperature ranges, strong solar radiation, and high ultraviolet intensity, with pavement structure and material parameters. Through finite element simulation, the pavement temperature stress field under the combined effect of these complex factors was accurately quantified, overcoming the shortcomings of traditional methods that consider only one factor and are difficult to reflect the true coupling effect.

[0014] 2. A smart evaluation model deeply integrating mechanism and data-driven approaches was constructed: This invention innovatively adopts a hybrid modeling strategy of "physical mechanism simulation (finite element method) + data-driven fitting (neural network)". The finite element model reveals the intrinsic physical laws of temperature damage, while the neural network efficiently learns the complex nonlinear relationships between multiple parameters and toughness indicators, combining the accuracy of mechanism analysis with the efficiency of intelligent algorithms, thus achieving a leap from "theoretical simulation" to "engineering application".

[0015] 3. An innovative toughness index suitable for high-altitude environments was proposed: abandoning general structural performance indicators, a "temperature fatigue damage factor" (the ratio of surface layer temperature stress to splitting strength) was creatively proposed as the core toughness evaluation index. This index has a clear physical meaning and can directly reflect the fatigue damage accumulation and cracking risk of asphalt pavement under temperature cycling and extreme low temperatures, making it more targeted and sensitive than traditional indicators.

[0016] 4. Orthogonal experimental design and database construction significantly improve optimization efficiency: A high-quality orthogonal database covering a broad parameter space is constructed using a seven-level, eight-factor orthogonal design table with minimal and scientific finite element simulations. This provides a solid data foundation for solving multi-parameter optimization problems, avoids the "curse of dimensionality" caused by full factorial design, and significantly reduces computational and time costs.

[0017] 5. Order-of-magnitude improvement in evaluation speed achieved through optimized neural networks: A BP neural network optimized using a genetic algorithm (GA) (GA-BP) is trained on the database, overcoming the drawback of traditional BP networks being prone to getting trapped in local optima. The trained neural network model can instantly predict the resilience of new pavement schemes, reducing the evaluation time from several hours of simulation to seconds, enabling rapid iteration and optimization design.

[0018] 6. A dynamic toughness threshold standard linked to road grade was established: This invention directly links toughness assessment results with engineering practice, proposing a criterion for dynamically setting toughness index thresholds based on road grade (such as expressways and ordinary roads). This makes the assessment standard more flexible and provides engineering guidance, ensuring that the design results meet both safety and durability requirements and economic principles.

[0019] 7. A closed-loop optimization system from assessment to design has been formed: This invention is not only an assessment method, but also a complete "design-verification-feedback-optimization" closed-loop system. When the neural network verification result does not meet the threshold, the system will give a redesign instruction, thereby guiding designers to adjust material or structural parameters until the toughness requirements are met, truly realizing the prediction and avoidance of low-temperature cracking risks in the design stage. Attached Figure Description

[0020] Figure 1 This is a flowchart of the method steps of the present invention.

[0021] Figure 2 This is a commonly used road surface design drawing for the Tibet region in step S2 of the method of the present invention.

[0022] Figure 3 This is the ultraviolet spectrum of the Tibet region in step S3 of the method of the present invention. Detailed Implementation

[0023] The present invention will now be described in detail with reference to specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0024] A method for evaluating the low-temperature cracking toughness of asphalt pavement surface layers includes one or more of the following: meteorological data collection in the Tibet Autonomous Region, structural and material design of asphalt pavement in Tibet, orthogonal design tables, finite element simulation models, selection of toughness indices, and neural network fitting. Figure 1 As shown, the specific steps are as follows: S1: Obtain meteorological data for the area where the new asphalt pavement is located, including the lowest monthly average temperature in winter, the lowest annual temperature, the maximum daily temperature difference in a year, the longest duration of low temperature, wind speed, daily average solar radiation, ultraviolet intensity, and other environmental data. In step S1, meteorological environmental data is collected. The data can be the meteorological data for the region predicted during the design phase for the next 10 years. If no predicted data is available, historical data from the past 10 years will be used. The data accuracy is at the county / district level, and the data interval is hourly. For example, hourly data for Chayu County, Linzhi City: 2025-01-01 08:00: Temperature -18.52℃, Net solar irradiance (net, J / m²) 2 ): 0.0, UV intensity (J / m 2 ): 0.0; 2025-01-01 09:00: Temperature -19℃, Net solar irradiance (net, J / m²) 2 ): 29440, UV intensity (J / m 2 ): 9192.

[0025] S2: Determine the internal parameters such as surface layer modulus, surface layer thermal expansion coefficient, and surface layer thickness based on the proposed asphalt pavement structure material selection; Currently, the commonly used road surface structures in Tibetan areas are as follows: ① For Class III highways and below: 18~20cm semi-rigid base course + 4~5cm asphalt mixture layer; ② For Class II highways: 20~30cm semi-rigid base course + 6cm asphalt mixture lower layer + 4cm asphalt mixture upper layer; ③ For Class I highways and above: 32~40cm semi-rigid base course + 7~12cm asphalt mixture or asphalt macadam lower layer + 6cm asphalt mixture intermediate layer + 4cm asphalt mixture upper layer, as shown in the attached diagram. Figure 2 As shown.

[0026] S3: The ratio of the temperature stress at the top of the asphalt pavement surface layer to the splitting strength of the asphalt surface layer is taken as the temperature fatigue damage factor. The temperature fatigue damage factor and the temperature fatigue damage caused by the preset extreme low temperature conditions are used as toughness evaluation indicators. The temperature fatigue damage factor is used as the criterion. If the factor is greater than 1, the material will crack. If it is less than 1, the single-cycle damage is calculated based on Miner's linear cumulative damage theory, and the cumulative fatigue damage is obtained by accumulation. This is used to comprehensively evaluate the crack resistance of the pavement under long-term temperature cycling.

[0027] The preset extreme low temperature condition refers to the low temperature condition corresponding to the meteorological data obtained in step S1. Damage refers to the temperature fatigue damage caused under such extreme conditions. The calculation formulas for temperature fatigue damage and temperature fatigue damage factor are as follows: (1) (2) In the formula: —Number of loading cycles at which the specimen fails; —Temperature fatigue damage factor; —The material coefficient can be determined experimentally; D—Temperature fatigue damage under a single action.

[0028] S4: Based on the meteorological data and internal parameters obtained above, set up a seven-level, eight-factor orthogonal design table for finite element simulation modeling to construct an orthogonal database; The process of designing a seven-level, eight-factor orthogonal design table includes: The eight factors are minimum temperature (°C), daily temperature range (°C), and daily total solar radiation (MJ / m²). 2 ), wind speed (m / s), pavement thickness (cm), surface modulus (MPa), surface thermal expansion coefficient (10) -5) and UV intensity (based on region classification); The lowest temperature is obtained by taking the lowest monthly average temperature and the annual maximum daily temperature difference in the county where the road is located during winter. The average temperature is used as the fourth level, the average temperature plus half of the annual maximum daily temperature difference is the first level, the average temperature minus half of the annual maximum daily temperature difference is the seventh level, and the intermediate levels are divided evenly. The daily temperature range is set as the seventh level with the maximum annual daily temperature range, the first level with the minimum annual daily temperature range, and the intermediate levels are evenly divided. Wind speed levels are divided into 0, 1, 2, 3, 4, 5, and 6, corresponding to levels one through seven. The total daily solar radiation is 2.5, 5, 7.5, 10, 12.5, 15, and 17.5, corresponding to levels one through seven. The surface layer thickness is determined automatically by the preset road structure or by using recommended values: 3, 4, 5, 6, 7, 3+5, 4+6, corresponding to the first to the seventh level. Based on the preset road surface structure, the thickness of each layer of the road surface can be obtained (single layer such as 4, 5, 6; double layer such as 4+6; triple layer such as 4+6+7). Then, seven levels are set, with the preset road surface thickness set as the fourth level. If the preset structure is a single layer, the difference between levels is 20% × the measured thickness (rounded to the nearest integer). At the same time, two common double-layer structures are set for comparison (3+5, 4+6), increasing from the first level to the seventh level. If the preset surface layer thickness is 5, orthogonally set as 3, 4, 5, 6, 7, 3+5, 4+6.

[0029] If the preset structure is double-layered, the difference between the horizontal layers is 20% × the actual measured thickness of each layer (rounded to the nearest integer). At the same time, two common single-layer structures are set for comparison (5, 6), increasing from the first level to the seventh level. If the preset surface layer thickness is 4+6, the orthogonal settings are 5, 6, 2+4, 3+5, 4+6, 5+7, 6+8.

[0030] If the preset structure is three layers, the difference between levels is 20% × the measured thickness of each layer (rounded to the nearest integer). Simultaneously, a common single-layer structure and a double-layer structure are set for comparison (5, 4+6), increasing from the first level to the seventh level. If the preset surface layer thickness is 4+6+7, orthogonally set to 5, 4+6, 3+4+5, 4+5+5, 4+6+7, 5+7+8, 6+7+10.

[0031] The surface layer modulus is determined automatically by the preset pavement structure or by using recommended values: 3000, 3500, 4500, 5500, 6500, 7500, and 8500, corresponding to levels one through seven. Alternatively, based on the preset pavement structure, the surface layer modulus of the corresponding surface layer material can be measured experimentally, and then seven levels can be set. The measured modulus value is set as level four, and the difference between levels is 20% × the measured surface layer modulus value, increasing from level one to level seven.

[0032] The surface layer has thermal expansion coefficients of 1.25, 1.5, 1.75, 2, 2.25, 2.5, and 2.75, corresponding to levels one through seven. Ultraviolet intensity was analyzed based on the ultraviolet spectrum, as shown in the attached figure. Figure 3 As shown, the corresponding ultraviolet intensity values ​​are adopted according to the cities and counties where highways are constructed, and five levels are set from the shallowest to the deepest. Among them, the ultraviolet intensity directly affects the toughness index (temperature fatigue damage factor), which is controlled by the ultraviolet aging coefficient. The coefficients for each level are 1.1, 1.15, 1.2, 1.25 and 1.3, respectively. The ultraviolet aging correction is performed by multiplying the maximum temperature fatigue damage factor obtained by simulation.

[0033] The process of finite element simulation includes: S4-1: Constructing a three-dimensional geometric model using COMSOL Multiphysics finite element simulation model; S4-2: Construct a thermo-mechanical coupling model and calculate the maximum temperature fatigue damage factor for each set of working conditions; S4-3: Simulate the temperature fatigue damage to the road surface caused by different low temperature durations by setting the calculation time; S4-4: Export the temperature fatigue damage corresponding to all working conditions to build a database.

[0034] The temperature fatigue damage factor is used as the criterion for judgment. The maximum temperature fatigue damage factor in a day represents the most unfavorable moment of the day and reflects the situation where the road surface is most prone to cracking.

[0035] S5: Regression fitting is performed on the external environment data, pavement structure internal parameters and toughness evaluation index in the database using a neural network; In step S5, the training and test sets are set considering that the data parameters are orthogonally designed, and the selection of the training and test sets should be random. For example, the training set should account for 70%-80% of the total samples, and the test set should account for 20%-30% of the total samples. The samples are shuffled when dividing the training and test sets.

[0036] In step S5, the neural network model used is the GA-BP model, i.e., the genetic algorithm, with the following parameter settings: number of generations: 5; population size: 10; real number encoding: [1e... -6[1], the selection function parameter is 0.09, the crossover function parameter is 2, and the mutation function parameter is [2 gen 3].

[0037] S6: Use the trained neural network to perform toughness calculations on the structure and materials of the newly built asphalt pavement. If the toughness index does not meet the set threshold, redesign it. In step S6, the toughness standard is set according to the designed asphalt pavement grade. For example, if it is a highway, the toughness index threshold should not exceed 0.7. If the road grade is reduced, the threshold can be appropriately increased. For example, for Class III and Class IV highways, the index threshold can be set to 0.8 and 0.9, respectively, indicating that the design of low-grade highways is not as strict as that of high-grade highways.

[0038] Example 1 A method for evaluating the low-temperature cracking toughness of asphalt pavement surface layers includes the following steps: S1: Obtain meteorological data for the area where the new asphalt pavement is located, including the lowest monthly average temperature in winter, the lowest annual temperature, the maximum daily temperature difference in a year, the longest duration of low temperature, wind speed, daily average solar radiation, ultraviolet intensity, and other environmental data.

[0039] This example uses the construction of a new expressway in Damxung County, Lhasa City, Tibet Autonomous Region as a case study. Considering construction factors, the minimum temperature is set at -37.5℃, the daily temperature range at 26.3℃, and the daily total solar radiation at 15 MJ / m². 2 The wind speed is 2 m / s; S2: Based on the proposed asphalt pavement structure material selection, determine the internal parameters such as surface layer modulus, surface layer thermal expansion coefficient, and surface layer thickness. Currently, the pavement structure used in some areas of Tibet is a 5cm asphalt mixture layer + 20cm semi-rigid base course. This structure is used as the initial design scheme.

[0040] S3: The ratio of the temperature stress at the top of the asphalt pavement surface layer to the splitting strength of the asphalt surface layer is taken as the temperature fatigue damage factor. The temperature fatigue damage factor and the temperature fatigue damage caused by the preset extreme low temperature conditions are used as toughness evaluation indicators.

[0041] S4: Based on the meteorological data and internal parameters obtained above, an orthogonal design table with seven levels and eight factors is set up for finite element simulation modeling to construct an orthogonal database, as shown in Table 1.

[0042] Table 1 S5: A neural network is used to perform regression fitting on external environmental data, pavement structure internal parameters, and toughness evaluation indicators in the database. After performance comparison, the GA-BP model is selected for regression prediction, with a fitting R0. 2 It can reach 0.985, which meets the forecast requirements.

[0043] R 2 The coefficient of determination (R²) is a core indicator for measuring the goodness of fit of a regression model. It represents the percentage of variation in the target variable that the model can explain. Its value typically ranges from 0 to 1; the closer it is to 1, the stronger the model's explanatory power. For example, R²... 2 =0.85 means that the model successfully captured 85% of the variation patterns in the data.

[0044] S6: The toughness index threshold was set to 0.5. The toughness of the newly built asphalt pavement structure and materials was verified using the trained neural network. It was found that the toughness index value for the predicted extreme case was 0.97, which did not meet the threshold. Therefore, this pavement structure scheme does not meet the requirements of the toughness design method and needs to be redesigned.

[0045] Redesign involves changing various parameters of the road design, such as materials and structural layer thickness. After the changes, new internal parameters (surface layer modulus, surface layer thickness, coefficient of thermal expansion, thermal conductivity, etc.) are applied. The above steps are repeated using a trained neural network model. If the toughness index value is less than the threshold, the toughness design requirements are met; otherwise, the pavement structure design is modified until the toughness design requirements are met.

[0046] The above description is merely a specific embodiment of the present invention, and the scope of protection of the present invention is not limited thereto. Any transformations or substitutions that can be conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for evaluating the low-temperature cracking toughness of asphalt pavement surface layers in Tibet, characterized in that, Includes the following steps: S1: Obtain meteorological data for the area where the new asphalt pavement is located, including the lowest monthly average temperature in winter, the lowest annual temperature, the maximum daily temperature difference in a year, the longest duration of low temperature, wind speed, daily average solar radiation, and ultraviolet intensity environmental data. S2: Determine the surface layer modulus, surface layer thermal expansion coefficient, and surface layer thickness internal parameters based on the proposed asphalt pavement structure material selection; S3: The ratio of the temperature stress at the top of the asphalt pavement surface layer to the splitting strength of the asphalt surface layer is taken as the temperature fatigue damage factor. The temperature fatigue damage factor and the temperature fatigue damage caused by the preset extreme low temperature conditions are used as toughness evaluation indicators. S4: Based on the meteorological data and internal parameters obtained above, set up a seven-level, eight-factor orthogonal design table for finite element simulation modeling to construct an orthogonal database; S5: Regression fitting is performed on the external environment data, pavement structure internal parameters and toughness evaluation index in the database using a neural network; S6: Use the trained neural network to perform toughness calculations on the structure and materials of the newly built asphalt pavement. If the toughness evaluation index does not meet the set threshold, then redesign.

2. The method for evaluating the low-temperature cracking toughness of asphalt pavement surface layers according to claim 1, characterized in that, In step S1, the data is the meteorological data of the region for the next 10 years as predicted during the design. If there is no predicted data, historical data from the past 10 years will be used. The data accuracy is at the county / district level, and the data interval is in hours.

3. The method for evaluating the low-temperature cracking toughness of asphalt pavement surface layers according to claim 1, characterized in that, In step S2, the pavement structure is selected as follows: ① For Class III highways and below: 18~20cm semi-rigid base course + 4~5cm asphalt mixture layer; ② For Class II highways: 20~30cm semi-rigid base course + 6cm asphalt mixture lower layer + 4cm asphalt mixture upper layer; ③ For Class I highways and above: 32~40cm semi-rigid base course + 7~12cm asphalt mixture or asphalt macadam lower layer + 6cm asphalt mixture intermediate layer + 4cm asphalt mixture upper layer.

4. The method for evaluating the low-temperature cracking toughness of asphalt pavement surface layers according to claim 1, characterized in that, In step S4, the process of designing the seven-level, eight-factor orthogonal design table includes: The eight factors are minimum temperature, daily temperature range, total daily solar radiation, wind speed, road thickness, surface layer modulus, surface layer thermal expansion coefficient, and ultraviolet intensity. The lowest temperature is obtained by taking the lowest monthly average temperature and the annual maximum daily temperature difference in the county where the road is located during winter. The average temperature is used as the fourth level, the average temperature plus half of the annual maximum daily temperature difference is the first level, the average temperature minus half of the annual maximum daily temperature difference is the seventh level, and the intermediate levels are divided evenly. The daily temperature range is set as the seventh level with the maximum annual daily temperature range, the first level with the minimum annual daily temperature range, and the intermediate levels are evenly divided. Wind speed levels are divided into 0, 1, 2, 3, 4, 5, and 6, corresponding to levels one through seven. The total daily solar radiation is 2.5, 5, 7.5, 10, 12.5, 15, and 17.5, corresponding to levels one through seven. The surface layer thickness is determined automatically by the preset road structure or by using recommended values: 3, 4, 5, 6, 7, 3+5, 4+6, corresponding to the first to the seventh level. The surface layer modulus is determined automatically by the preset pavement structure or by using recommended values: 3000, 3500, 4500, 5500, 6500, 7500, and 8500, corresponding to the first to seventh levels. The surface layer has thermal expansion coefficients of 1.25, 1.5, 1.75, 2, 2.25, 2.5, and 2.75, corresponding to levels one through seven. Ultraviolet intensity was analyzed based on the ultraviolet spectrum.

5. The method for evaluating the low-temperature cracking toughness of asphalt pavement surface layers according to claim 1, characterized in that, In step S4, the finite element simulation process includes: Step S4-1: Construct a three-dimensional geometric model using the COMSOL Multiphysics finite element simulation model; Step S4-2: Construct a thermo-mechanical coupling model and calculate the maximum temperature fatigue damage factor for each set of working conditions; Step S4-3: Simulate the temperature fatigue damage to the road surface caused by different low-temperature durations by setting the calculation time; Step S4-4: Export the temperature fatigue damage database corresponding to all operating conditions.

6. The method for evaluating the low-temperature cracking toughness of asphalt pavement surface layer according to claim 1, characterized in that, In step S5, the neural network model used is the GA-BP model, with the following parameter settings: number of generations: 5; population size: 10; real number encoding: [1e...]. -6 [1], the selection function parameter is 0.09, the crossover function parameter is 2, and the mutation function parameter is [2 gen 3].

7. The method for evaluating the low-temperature cracking toughness of asphalt pavement surface layers according to claim 1, characterized in that, In step S5, the training set and test set settings are designed with the data parameters being orthogonal, and the selection of the training set and test set should be random.

8. The method for evaluating the low-temperature cracking toughness of asphalt pavement surface layers according to claim 1, characterized in that, In step S6, the toughness standard is set according to the designed asphalt pavement grade. If it is a highway, the toughness index threshold should not exceed 0.

7. If the road grade is reduced, the threshold should be appropriately increased.

Images