Method and system for evaluating thermal resistance performance of asphalt mixture

Abstract

The invention discloses an asphalt mixture thermal resistance performance evaluation method and system. The method comprises the following steps: preparing a specified size test piece and arranging internal and surface temperature probes; performing heat insulation on the side surface and the lower surface of the test piece, placing the test piece in a closed test box, simulating a radiation environment through a controllable heat source, and performing a steady-state or dynamic cycle test; synchronously acquiring temperature, radiation and environment data; thermal resistance performance indexes such as thermal response delay time, a thermal diffusion coefficient, apparent thermal resistance, a normalized thermal resistance index and a thermal uniformity index are calculated; and training a machine learning model by using the data sets of the test pieces with different material parameters, and establishing a mapping relation between the material parameters and the thermal resistance performance to realize performance prediction. The system comprises a test piece and sensing module, an environment simulation and heat source control module, a data acquisition module and a thermal resistance performance calculation and prediction module. According to the method, the thermal resistance performance of the asphalt mixture is accurately and comprehensively evaluated, and a basis is provided for design and optimization of pavement materials.

Description

A method and system for evaluating the thermal resistance performance of asphalt mixtures Technical Field

[0001] This application relates to the field of civil engineering material testing technology, and in particular to a method and system for evaluating the thermal resistance performance of asphalt mixtures. Background Technology

[0002] With the increasing severity of the urban heat island effect, porous asphalt mixtures with good thermal resistance properties have shown great potential in reducing pavement temperature and mitigating the heat island effect. Currently, the main method for evaluating the thermal resistance properties of asphalt mixtures is to measure their thermal conductivity, with commonly used methods including the heat flow meter method and the laser flash method.

[0003] Chinese Patent, Publication No. CN107870179A, Publication Date: April 13, 2018, discloses a method for measuring the contact thermal resistance of asphalt concrete. The method employs a heat flow sensing test to measure the contact thermal resistance between asphalt concrete and a heat exchange pipe. This is achieved by embedding a temperature sensor, a heat flow sensor, and a seamless steel pipe arranged in a circuitous pattern within the specimen. The temperature sensor is embedded in the middle of the specimen at distances of 20mm, 40mm, 60mm, and 80mm from the surface, respectively. The heat flow sensor is fixed to the outer wall of the middle section of the seamless steel pipe, 80mm from the specimen surface. After specimen preparation, the specimen is wrapped with thermal insulation material to ensure heat transfer is primarily along the direction of gravity. Infrared lamps are used as the heat source, and the flow of the heat exchange medium within the pipe can be controlled at different velocities and temperatures. After measuring the temperature difference and heat flow at the contact interface, the contact thermal resistance can be calculated according to the definition of contact thermal resistance.

[0004] The shortcomings of the above technical solutions are as follows: 1. They focus on the "contact thermal resistance" between asphalt concrete and the internally buried pipes. Measurement relies on a limited number of temperature sensors (4) at fixed depths and a heat flow sensor, estimating interface parameters through data fitting and extrapolation. This method cannot obtain the continuous temperature gradient inside the specimen, the overall temperature distribution on the upper surface, or the temperature of the lower surface, making it difficult to calculate more comprehensive thermal resistance performance indicators such as thermal diffusivity, thermal response delay, and thermal uniformity. 2. Using infrared lamps as heat sources has limited control precision in terms of power, height, and incident angle, and relies on manual operation. The test environment is open or simply enclosed, and the insulation of the sides and lower surfaces of the specimen may be insufficient, making it susceptible to environmental fluctuations. The entire system lacks real-time monitoring and closed-loop feedback control of ambient temperature and humidity, and actual irradiance, making it difficult to accurately simulate and maintain steady-state or complex dynamic cyclic thermal boundary conditions. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the purpose of this application is to provide a method and system for evaluating the thermal resistance performance of asphalt mixtures, which can achieve high-precision and automated evaluation and prediction of the comprehensive thermal resistance performance of asphalt mixtures under multi-dimensional and controllable thermal environments.

[0006] To achieve the above objectives, this application adopts the following technical solution: This application provides a method for evaluating the thermal resistance performance of asphalt mixtures, the method comprising: S1, preparing asphalt mixture specimens of specified dimensions, pre-embedding or drilling temperature measuring channels along the thickness direction of the asphalt mixture specimens, and arranging at least three equally spaced internal temperature probes in the temperature measuring channels to monitor the internal temperature gradient of the asphalt mixture specimens; simultaneously, uniformly arranging at least two surface temperature probes on the upper and lower surfaces of the asphalt mixture specimens respectively; S2, placing the asphalt mixture specimen with the temperature probes arranged on a specimen platform; and tightly wrapping the asphalt mixture specimen with thermal insulation material. All sides and lower surfaces of the asphalt mixture specimen are exposed, with only the upper surface exposed to the controlled heat source; S3, the asphalt mixture specimen is placed in a sealed test chamber, the inner wall of which has a high reflectivity layer and integrates an ambient temperature and humidity sensor, a radiation sensor, and an infrared thermal imager; according to the target simulated working conditions, the power, height, and incident angle of the controlled heat source are set; real-time feedback is obtained through the radiation sensor integrated in the test chamber, and the power, height, and incident angle of the controlled heat source are adjusted according to the feedback radiation parameters to achieve a preset constant or periodically changing irradiance; S4, a steady-state test or a dynamic cyclic test is performed, the steady-state test... To apply constant radiation until the temperature field of the asphalt mixture specimen stabilizes, a dynamic cyclic test is conducted by periodically turning the controllable heat source on and off to simulate alternating heating and cooling processes; S5. During the test, test data are collected and recorded synchronously and continuously: the upper and lower surface temperatures of the asphalt mixture specimen obtained by the surface temperature probe; the internal temperature distribution along the thickness direction of the asphalt mixture specimen obtained by the internal temperature probe; the full-field temperature distribution image of the upper surface of the asphalt mixture specimen obtained by the infrared thermal imager; and the ambient temperature and humidity data inside the test chamber obtained by the ambient temperature and humidity sensor; S6. Based on step S... 5. Based on the collected test data, calculate the thermal resistance performance index of the asphalt mixture. The thermal resistance performance index of the asphalt mixture includes at least one of the following: thermal response delay time, thermal diffusivity, apparent thermal resistance, normalized thermal resistance index, and thermal uniformity index; S7. Test asphalt mixture specimens with different material compositions, gradations, and porosities according to steps S1 to S6 to obtain a thermal resistance performance index dataset; use this thermal resistance performance index dataset to train a machine learning prediction model to establish a nonlinear mapping relationship between asphalt mixture material parameters and thermal resistance performance index; predict the thermal resistance performance index of the asphalt mixture using the machine learning prediction model.

[0007] As a preferred technical solution, in step S6, the method for calculating the thermal response delay time includes: recording the moment when the controllable heat source is turned on or when the asphalt mixture specimen begins to be heated; monitoring the moment when the temperature at the center point of the upper surface of the asphalt mixture specimen begins to rise significantly; recording the moment when the temperature at the center point of the lower surface of the asphalt mixture specimen first rises to a preset temperature threshold from the moment heating begins; and subtracting the moment when the upper surface temperature begins to change from the moment the lower surface temperature reaches the threshold, the resulting time difference is the thermal response delay time. As a preferred technical solution, in step S6, the method for calculating the thermal diffusivity includes: based on the internal temperature distribution data along the thickness direction of the asphalt mixture specimen obtained by the internal temperature probe, fitting the measured temperature change curves at different internal depths with the solution of the one-dimensional unsteady-state thermal conductivity differential equation to derive the optimal thermal diffusivity.

[0008] As a preferred technical solution, in step S6, the method for calculating the apparent thermal resistance includes: calculating the effective heat flux density applied to the upper surface of the asphalt mixture specimen under stable heating conditions; measuring the average temperature difference between the upper and lower surfaces of the asphalt mixture specimen at this time; dividing the average temperature difference between the upper and lower surfaces of the asphalt mixture specimen by the calculated effective heat flux density, and the resulting quotient is the apparent thermal resistance; in step S6, the method for calculating the normalized thermal resistance index includes: dividing the apparent thermal resistance of the asphalt mixture specimen by the thickness of the asphalt mixture specimen to obtain the normalized thermal resistance index; or, multiplying the apparent thermal resistance by a standard temperature difference constant and then dividing by the thickness of the asphalt mixture specimen to obtain the normalized thermal resistance index.

[0009] As a preferred technical solution, in step S6, the method for calculating the thermal uniformity index includes: based on the full-field temperature distribution image of the upper surface of the asphalt mixture specimen acquired by the infrared thermal imager, calculating the average temperature of all areas on the entire upper surface of the asphalt mixture specimen; calculating the degree of deviation of the temperature of all points on the entire upper surface of the asphalt mixture specimen from the average temperature, i.e., the standard deviation of temperature; dividing the standard deviation of temperature by the average temperature to obtain the coefficient of variation; and dividing the standard deviation by the coefficient of variation to obtain the thermal uniformity index.

[0010] As a preferred technical solution, in step S7, the step of testing asphalt mixture specimens with different material compositions, gradations, and porosities according to steps S1 to S6 to obtain a thermal resistance performance index dataset includes: establishing a data file for each asphalt mixture specimen, the data file including the material parameters of the asphalt mixture specimen and the calculated thermal resistance performance index of the asphalt mixture specimen; summarizing all asphalt mixture specimens to form a structured table; each row of the structured table represents an asphalt mixture specimen, and each column represents the material parameters or thermal resistance performance index, thereby constituting a thermal resistance performance index dataset for machine learning training; as a preferred technical solution, in step S7, the step of using the thermal resistance performance index dataset to train a machine learning prediction model and establish a nonlinear mapping relationship between asphalt mixture material parameters and thermal resistance performance index. The process includes: selecting all columns describing material parameters from the thermal resistance performance index dataset as input to the machine learning prediction model, and selecting the thermal resistance performance index to be predicted as the output of the machine learning prediction model; cleaning the thermal resistance performance index dataset before training and normalizing all data; randomly dividing the thermal resistance performance index dataset into training and testing sets; inputting the training set into the initial machine learning prediction model, and using random forest, gradient boosting decision tree, or neural network to find the mapping pattern from material parameters to thermal resistance performance index to obtain the machine learning prediction model; using cross-validation to validate the machine learning prediction model and calculating the evaluation index of the machine learning prediction model; if the difference between the evaluation index and the preset evaluation index is greater than the deviation threshold, the machine learning prediction model is optimized to obtain the optimized machine learning prediction model.

[0011] As a preferred technical solution, in step S7, if the difference between the evaluation index and the preset evaluation index is greater than the deviation threshold, the machine learning prediction model is optimized to obtain the optimized machine learning prediction model. This includes: diagnosing the root causes of poor performance of the machine learning prediction model, including: insufficient data volume, poor data quality, insufficient data representativeness, weak feature correlation, model being too simple, model being too complex, and improper model parameter settings; and taking one or more corresponding optimization measures based on the root causes: preparing and testing more asphalt mixture specimens with different formulations; reviewing experimental data, identifying and correcting or eliminating obvious abnormal test records; attempting to create new, physically meaningful combination features based on existing material parameters; and using statistical methods or the feature importance assessment function built into the machine learning prediction model to identify and eliminate features that contribute little to the prediction target. Redundant features are used to simplify the input to the machine learning prediction model. For models that are too simple, a more complex model is chosen, or the complexity of the existing model is increased. For models that are too complex, the complexity is reduced, or regularization constraints are added. A set of hyperparameters that achieves optimal performance on the validation set is obtained using grid search or random search methods. For neural networks, the learning rate and training epochs are adjusted. After applying optimization measures, the model is retrained using an updated dataset and a new model configuration. If the difference between the new evaluation metric and the preset evaluation metric is still greater than the deviation threshold, the "diagnosis-optimization-retraining" cycle is repeated until the model performance meets the requirements. The optimized machine learning prediction model is then obtained.

[0012] This application provides an asphalt mixture thermal resistance performance evaluation system, which includes: a specimen and sensing module for preparing asphalt mixture specimens of a specified size, pre-embedding or drilling temperature measurement channels in the thickness direction of the specimens, arranging at least three equally spaced internal temperature probes, and uniformly arranging at least two surface temperature probes on the upper and lower surfaces of the specimens to monitor their internal temperature gradient and surface temperature; an environmental simulation and heat source control module, including a sealed test chamber with a high reflectivity layer on the inner wall, a platform for placing the specimens inside the test chamber, and wrapping the sides and lower surface of the specimens with thermal insulation material, exposing only the upper surface; the module also includes a controllable heat source, an environmental temperature and humidity sensor integrated inside the chamber, a radiation sensor, and an infrared thermal imager, used for... The target operating condition is set and the power, height, and incident angle of the heat source are adjusted to maintain a preset constant or periodically varying irradiance. A data acquisition module is used to synchronously and continuously acquire and record test data from the surface temperature probe, internal temperature probe, infrared thermal imager, and environmental temperature and humidity sensor during steady-state or dynamic cyclic testing. A thermal resistance performance calculation and prediction module is used to calculate at least one thermal resistance performance index among thermal response delay time, thermal diffusivity, apparent thermal resistance, normalized thermal resistance index, and thermal uniformity index based on the acquired test data. A machine learning prediction model is trained using a dataset of thermal resistance performance indexes obtained from tests of specimens with different material parameters to predict the thermal resistance performance index of asphalt mixtures.

[0013] Compared with existing technologies, the advantages of this application are as follows: By using equally spaced internal temperature probes arranged along the thickness direction, surface temperature probes on the upper and lower surfaces, and an infrared thermal imager (S1, S5), this solution can simultaneously acquire complete temperature field data from the surface to the interior and from point to surface. This makes the calculation no longer limited to a single contact thermal resistance, but can derive multiple indicators such as thermal response delay time, thermal diffusivity, apparent thermal resistance, normalized thermal resistance index, and thermal uniformity index (S6), thereby reflecting the thermophysical properties of asphalt mixtures more comprehensively and deeply.

[0014] This application employs a sealed test chamber with high internal reflectivity and a controllable heat source with integrated radiation sensors and real-time feedback control (S3). This scheme can accurately set and maintain constant or periodically varying irradiance (S3, S4). Simultaneously, the non-test surfaces of the specimen are tightly wrapped with thermal insulation material (S2), minimizing environmental interference and ensuring that experimental data are obtained under highly controlled and repeatable thermal conditions, laying a solid foundation for accurately evaluating material performance.

[0015] This application constructs a dataset of thermal resistance performance indicators by systematically testing specimens with different material parameters (composition, gradation, porosity) and training a machine learning prediction model (S7). This approach achieves a leap from "measurement" to "prediction." Once the model is trained, its thermal resistance performance can be quickly predicted based on the basic material parameters of asphalt mixtures, without the need for complex physical experiments each time. This greatly improves evaluation efficiency and provides a powerful data-driven tool for the design and performance optimization of new materials. Attached Figure Description

[0016] Figure 1 is a schematic diagram of the overall structure of the asphalt mixture thermal resistance performance evaluation device; Figure 2 is a schematic diagram of the structure of the side and bottom surfaces of the rutted slab specimen after being wrapped with thermal insulation material; Figure 3 is a schematic diagram of the structure of the side and bottom surfaces of the Marshall specimen after being wrapped with thermal insulation material; Figure 4 is a schematic diagram of the structure of the specimen placement platform; Figure 5 is a schematic diagram of the temperature change curves of the upper and lower surfaces of the specimen during the test; Wherein: 1, test chamber; 2, iodine tungsten lamp; 3, lifting and adjusting mechanism; 4, specimen platform; 5, asphalt mixture specimen; 6a, upper surface temperature probe; 6b, lower surface temperature probe; 7, chamber door; 8, data acquisition device; 9, ceramic fiber cotton; 10, observation window. Detailed Implementation

[0017] To enable those skilled in the art to better understand the present application, the technical solutions in specific embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0018] This application provides a method for evaluating the thermal resistance performance of asphalt mixtures. The method includes: S1, preparing asphalt mixture specimens of specified dimensions, pre-embedding or drilling temperature measuring channels along the thickness direction of the asphalt mixture specimens, and arranging at least three equally spaced internal temperature probes in the temperature measuring channels to monitor the internal temperature gradient T(z, t) of the asphalt mixture specimens, where z is the depth coordinate.

[0019] At the same time, at least two surface temperature probes are evenly arranged on the upper and lower surfaces of the asphalt mixture specimen.

[0020] S2. Place the asphalt mixture specimen with the temperature probes set on the specimen platform. Tightly wrap all sides and the bottom surface of the asphalt mixture specimen with thermal insulation material, exposing only the top surface of the asphalt mixture specimen to a controlled heat source.

[0021] S3. Place the asphalt mixture specimen in a sealed test chamber. The inner wall of the chamber has a high-reflectivity layer and integrates an environmental temperature and humidity sensor, a radiation sensor, and an infrared thermal imager. An infrared thermal imager is added inside the sealed chamber and aimed at the upper surface of the specimen. This is used for non-contact, full-field monitoring of the uniformity of surface temperature distribution, hot spots, and the visualization of heat flow transfer. Environmental temperature and humidity sensors are placed inside the chamber and near the specimen to monitor the stability of the test microenvironment and provide a basis for data correction. A high-precision total solar radiation sensor is integrated on the same plane as the specimen placement. Before each test, the light source power and height are automatically fine-tuned through real-time feedback until a preset, precise irradiance value (e.g., 800 W / m²) is achieved.

[0022] Based on the target simulated operating conditions, the power, height, and incident angle of the controllable heat source are set. Real-time feedback is obtained through radiation sensors integrated within the test chamber. The power, height, and incident angle of the controllable heat source are adjusted according to the feedback radiation parameters to achieve a preset constant or periodically varying irradiance.

[0023] S4. Perform steady-state testing or dynamic cyclic testing. Steady-state testing involves applying constant radiation until the temperature field of the asphalt mixture specimen tends to stabilize. Dynamic cyclic testing involves controlling a controllable heat source to periodically open and close, simulating the alternating heating and cooling process.

[0024] S5. During the test, the following test data are collected and recorded synchronously and continuously: The surface temperature of the asphalt mixture specimen (upper and lower surfaces) is obtained from the surface temperature probe; the internal temperature distribution along the thickness direction of the asphalt mixture specimen is obtained from the internal temperature probe; the full-field temperature distribution image of the upper surface of the asphalt mixture specimen is obtained from the infrared thermal imager; and the ambient temperature and humidity data inside the test chamber are obtained from the ambient temperature and humidity sensor.

[0025] S6. Based on the test data collected in step S5, calculate the thermal resistance performance index of the asphalt mixture. The thermal resistance performance index of the asphalt mixture includes at least one of the following: thermal response delay time, thermal diffusivity, apparent thermal resistance, normalized thermal resistance index, and thermal uniformity index.

[0026] The five core indicators provided characterize the material from five key dimensions: transient response speed (thermal response delay time), intrinsic thermal properties (thermal diffusivity), overall thermal insulation capacity (apparent thermal resistance), normalized comparison benchmark (normalized thermal resistance index), and thermal field distribution characteristics (thermal uniformity index). This multi-indicator synergistic evaluation system overcomes the limitations of single indicators, revealing the essential thermal performance of materials more comprehensively and profoundly, and providing a full range of decision-making basis for material optimization design.

[0027] By calculating the thermal diffusivity, we can directly correlate it with the influence mechanism of material composition (such as asphalt, aggregate, and porosity) on the heat transfer rate, providing clear physical parameter guidance for controlling thermal properties by adjusting the mix proportion. The thermal uniformity index helps to discover local thermal bridging effects caused by material inhomogeneity, segregation, or defects, which is of guiding significance for production processes and quality control.

[0028] The normalized thermal resistance index, through integration and ratio calculation, can effectively offset systematic errors caused by environmental fluctuations, slight instability of the light source, or minor differences in initial conditions, making comparisons between test results from different batches and between different materials more reliable and convincing. Apparent thermal resistance is calculated based on calibrated net radiant power, resulting in more accurate results.

[0029] Thermal response delay time directly reflects the lag in the temperature rise of the pavement bottom under solar radiation. This is crucial for assessing the fatigue characteristics of pavement structures under daily cyclic temperatures and the thermal impact on the underlying subgrade, and has clear engineering application value.

[0030] This set of standardized, quantified, multi-dimensional indicators constitutes a "feature vector" describing the thermal resistance performance of materials. It serves as a high-quality, high-information-density data foundation for subsequent use of machine learning models to establish material formulation-performance prediction relationships and for intelligent material design and optimization. It drives the transformation of materials research from "experience-based trial and error" to a modern R&D model that is "data-driven and performance-oriented."

[0031] S7. Asphalt mixture specimens with different material compositions, gradations, and porosities are tested according to steps S1 to S6 to obtain a dataset of thermal resistance performance indicators. This dataset is used to train a machine learning prediction model to establish a nonlinear mapping relationship between asphalt mixture material parameters and thermal resistance performance indicators. The machine learning prediction model is then used to predict the thermal resistance performance indicators of the asphalt mixture.

[0032] Specifically, the test chamber 1 is constructed of polyurethane foam board with an inner wall lined with aluminum foil. The front of the test chamber 1 has an openable and closable door 7 with a transparent observation window 10. An iodine-tungsten lamp 2 serves as a heat source inside the test chamber 1, its vertical distance from the asphalt mixture specimen 5 is adjusted by a lifting mechanism 3, and the input power is adjusted by a power regulator. The bottom of the test chamber 1 has a specimen platform 4 made of ceramic material. The asphalt mixture specimen 5 is placed on the platform, and K-type thermocouples are arranged on the upper and lower surfaces of the specimen 5 as upper surface temperature probes 6a and 6b, respectively. A data acquisition device 8 is connected to the upper and lower surface temperature probes 6a and 6b to record temperature data. The sides and bottom of the asphalt mixture specimen 5 are tightly wrapped with ceramic fiber cotton 9.

[0033] When implementing this method, porous asphalt mixture specimens are first formed according to standard methods. For rutted slab specimens (300mm×300mm×50mm, as shown in Figure 2), three temperature probes are evenly arranged on the diagonals of the upper and lower surfaces of the specimen; for Marshall specimens (Φ101.6mm×63.5mm, as shown in Figure 3), two temperature probes are arranged on the upper and lower surfaces of the specimen. After wrapping the specimens with heat insulation cotton, they are placed on specimen platform 4 (as shown in Figure 4).

[0034] Based on the simulated solar radiation intensity (e.g., 1000 W / m²), the iodine-tungsten lamp power was set to 1200 W and the height to 60 cm using a reference table calibrated in the pre-experiment. The heat source and data acquisition system were activated, and temperature change data were recorded over 60 minutes.

[0035] During data processing, the temperature-time curve is first plotted (as shown in Figure 5) to calculate the thermal response delay time.

[0036] This application utilizes equally spaced internal temperature probes arranged along the thickness direction, surface temperature probes on the upper and lower surfaces, and an infrared thermal imager (S1, S5). This scheme can simultaneously acquire complete temperature field data from the surface to the interior and from point to surface. This allows the calculation to go beyond a single contact thermal resistance, and instead derive multiple indicators such as thermal response delay time, thermal diffusivity, apparent thermal resistance, normalized thermal resistance index, and thermal uniformity index (S6), thereby reflecting the thermophysical properties of asphalt mixtures more comprehensively and deeply.

[0037] This application employs a sealed test chamber with high internal reflectivity and a controllable heat source with integrated radiation sensors and real-time feedback control (S3). This scheme can accurately set and maintain constant or periodically varying irradiance (S3, S4). Simultaneously, the non-test surfaces of the specimen are tightly wrapped with thermal insulation material (S2), minimizing environmental interference and ensuring that experimental data are obtained under highly controlled and repeatable thermal conditions, laying a solid foundation for accurately evaluating material performance.

[0038] This application constructs a dataset of thermal resistance performance indicators by systematically testing specimens with different material parameters (composition, gradation, porosity) and training a machine learning prediction model (S7). This approach achieves a leap from "measurement" to "prediction." Once the model is trained, its thermal resistance performance can be quickly predicted based on the basic material parameters of asphalt mixtures, without the need for complex physical experiments each time. This greatly improves evaluation efficiency and provides a powerful data-driven tool for the design and performance optimization of new materials.

[0039] Furthermore, in step S6, the method for calculating the thermal response delay time includes: recording the moment when the controllable heat source is turned on or when the asphalt mixture specimen begins to be heated; monitoring the moment when the temperature at the center point of the upper surface of the asphalt mixture specimen begins to rise significantly; and recording the moment when the temperature at the center point of the lower surface of the asphalt mixture specimen first rises to a preset temperature threshold from the moment heating begins. The time difference obtained by subtracting the moment when the upper surface temperature begins to change from the moment the lower surface temperature reaches the threshold is the thermal response delay time. Furthermore, in step S6, the method for calculating the thermal diffusivity includes: based on the internal temperature distribution data along the thickness direction of the asphalt mixture specimen obtained by the internal temperature probe, fitting the measured temperature change curves at different internal depths with the solution of the one-dimensional unsteady-state thermal conductivity differential equation to derive the optimal thermal diffusivity.

[0040] Furthermore, in step S6, the method for calculating the apparent thermal resistance includes: calculating the effective heat flux density applied to the upper surface of the asphalt mixture specimen under a stable heating state; measuring the average temperature difference between the upper and lower surfaces of the asphalt mixture specimen at this time; and dividing the average temperature difference between the upper and lower surfaces of the asphalt mixture specimen by the calculated effective heat flux density, the quotient of which is the apparent thermal resistance.

[0041] In step S6, the normalized thermal resistance index is calculated by dividing the apparent thermal resistance of the asphalt mixture specimen by the thickness of the asphalt mixture specimen to obtain the normalized thermal resistance index. Alternatively, the apparent thermal resistance is multiplied by a standard temperature difference constant and then divided by the thickness of the asphalt mixture specimen to obtain the normalized thermal resistance index.

[0042] Furthermore, in step S6, the calculation method for the thermal uniformity index includes: based on the full-field temperature distribution image of the upper surface of the asphalt mixture specimen acquired by an infrared thermal imager, calculating the average temperature of all areas on the entire upper surface of the asphalt mixture specimen. Calculating the deviation of the temperature at all points on the entire upper surface of the asphalt mixture specimen from the average temperature, i.e., the standard deviation of the temperature. Dividing the standard deviation of the temperature by the average temperature yields the coefficient of variation. Dividing this by the coefficient of variation yields the thermal uniformity index.

[0043] Furthermore, in step S7, asphalt mixture specimens with different material compositions, gradations, and porosities are tested according to steps S1 to S6 to obtain a thermal resistance performance index dataset. This includes: establishing a data file for each asphalt mixture specimen, which includes the material parameters of the asphalt mixture specimen and the calculated thermal resistance performance index of that specimen; and summarizing all asphalt mixture specimens to form a structured table. Each row of the structured table represents an asphalt mixture specimen, and each column represents a material parameter or thermal resistance performance index, thus constituting a thermal resistance performance index dataset for machine learning training.

[0044] Further, in step S7, training a machine learning prediction model using the thermal resistance performance index dataset to establish a nonlinear mapping relationship between asphalt mixture material parameters and thermal resistance performance indices includes: selecting all columns describing material parameters in the thermal resistance performance index dataset as input to the machine learning prediction model, and selecting the thermal resistance performance index to be predicted as output of the machine learning prediction model. Before training, the thermal resistance performance index dataset is cleaned and all data is normalized. The thermal resistance performance index dataset is randomly divided into training and testing sets. The training set is input into the initial machine learning prediction model, and random forest, gradient boosting decision tree, or neural network are used to find the mapping pattern from material parameters to thermal resistance performance indices to obtain the machine learning prediction model. Cross-validation is used to validate the machine learning prediction model, and the evaluation index of the machine learning prediction model is calculated. If the difference between the evaluation index and the preset evaluation index is greater than the deviation threshold, the machine learning prediction model is optimized to obtain the optimized machine learning prediction model.

[0045] Furthermore, in step S7, if the difference between the evaluation index and the preset evaluation index is greater than the deviation threshold, the machine learning prediction model is optimized to obtain the optimized machine learning prediction model, which includes diagnosing the root cause of the poor performance of the machine learning prediction model. The root cause includes: insufficient data volume, poor data quality, insufficient data representativeness, weak feature correlation, model that is too simple, model that is too complex, and improper model parameter settings.

[0046] Based on the root cause, take one or more corresponding optimization measures: Prepare and test more asphalt mixture specimens with different formulations. Review experimental data, identify and correct or remove obvious abnormal test records. Attempt to create new, physically meaningful combinations of features based on existing material parameters. Use statistical methods or the feature importance assessment function built into the machine learning prediction model to identify and remove redundant features that contribute little to the prediction target, simplifying the input of the machine learning prediction model. If the model is too simple, choose a more complex machine learning prediction model, or increase the complexity of the existing machine learning prediction model. If the model is too complex, reduce the model complexity, or add regularization constraints to the machine learning prediction model. Obtain a set of hyperparameter combinations that enable the machine learning prediction model to achieve optimal performance on the validation set through grid search or random search methods. For neural networks, adjust the learning rate of the machine learning prediction model and increase the number of training epochs.

[0047] After applying optimization measures, the machine learning prediction model is retrained using the updated dataset and new model configuration. If the difference between the new evaluation metric and the preset evaluation metric is still greater than the deviation threshold, the "diagnosis-optimization-retraining" cycle is repeated until the performance of the machine learning prediction model meets the requirements. The optimized machine learning prediction model is then obtained.

[0048] This application provides an asphalt mixture thermal resistance performance evaluation system, which includes: a specimen and a sensing module for preparing asphalt mixture specimens of a specified size, pre-embedding or drilling temperature measurement channels in the thickness direction of the specimen, arranging at least three equally spaced internal temperature probes, and uniformly arranging at least two surface temperature probes on the upper and lower surfaces of the specimen to monitor its internal temperature gradient and surface temperature.

[0049] The environmental simulation and heat source control module includes a sealed test chamber with a high-reflectivity inner wall. Inside the chamber is a platform for placing test specimens, and the sides and bottom surfaces of the specimens are wrapped with thermal insulation material, exposing only the top surface. The module also includes a controllable heat source, an integrated environmental temperature and humidity sensor, a radiation sensor, and an infrared thermal imager. These are used to set and adjust the power, height, and incident angle of the heat source according to the target operating conditions to maintain a preset constant or periodically varying irradiance.

[0050] The data acquisition module is used to synchronously and continuously acquire and record test data from surface temperature probes, internal temperature probes, infrared thermal imagers, and environmental temperature and humidity sensors during steady-state testing or dynamic cyclic testing.

[0051] The thermal resistance performance calculation and prediction module is used to calculate at least one of the following thermal resistance performance indicators based on the collected test data: thermal response delay time, thermal diffusivity, apparent thermal resistance, normalized thermal resistance index, and thermal uniformity index. It also uses a dataset of thermal resistance performance indicators obtained from tests of specimens with different material parameters to train a machine learning prediction model, so as to predict the thermal resistance performance indicators of asphalt mixtures through the machine learning prediction model.

[0052] It should be noted that the terms "first," "second," and similar terms used in this application specification and claims do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Similarly, "an" or "a" and similar terms do not indicate a quantity limitation, but rather indicate the presence of at least one. "A plurality" or "several" indicates at least two. Unless otherwise stated, terms such as "front," "back," "left," "right," "lower," and / or "upper" are for illustrative purposes only and are not limited to a location or spatial orientation. Terms such as "comprising" or "including" indicate that the elements or objects preceding "comprising" encompass the elements or objects listed following "comprising" or "including" and their equivalents, and do not exclude other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect.

[0053] The singular forms “a,” “the,” and “the” used in this application specification and appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0054] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A method for evaluating the thermal resistance performance of asphalt mixtures, characterized in that, The method includes: S1, preparing asphalt mixture specimens of specified dimensions, pre-embedding or drilling temperature measuring channels along the thickness direction of the asphalt mixture specimens, and arranging at least three equally spaced internal temperature probes in the temperature measuring channels to monitor the internal temperature gradient of the asphalt mixture specimens; simultaneously, uniformly arranging at least two surface temperature probes on the upper and lower surfaces of the asphalt mixture specimens respectively; S2, placing the asphalt mixture specimen with the temperature probes arranged on a specimen platform; tightly wrapping all sides and the lower surface of the asphalt mixture specimen with heat insulation material, exposing only the upper surface of the asphalt mixture specimen to a controllable heat source. S3. Place the asphalt mixture specimen in a sealed test chamber. The inner wall of the test chamber has a high reflectivity layer and integrates an ambient temperature and humidity sensor, a radiation sensor, and an infrared thermal imager. Based on the target simulated working conditions, set the power, height, and incident angle of the controllable heat source. Real-time feedback is obtained through the integrated radiation sensor within the test chamber. Adjust the power, height, and incident angle of the controllable heat source according to the feedback radiation parameters to achieve a preset constant or periodically varying irradiance. S4. Perform a steady-state test or a dynamic cyclic test. The steady-state test involves applying constant radiation until the temperature field of the asphalt mixture specimen tends to... For stability, the dynamic cycle test involves periodically turning a controllable heat source on and off to simulate alternating heating and cooling processes; S5, during the test, test data is collected and recorded synchronously and continuously: the upper and lower surface temperatures of the asphalt mixture specimen obtained by the surface temperature probe; the internal temperature distribution along the thickness direction of the asphalt mixture specimen obtained by the internal temperature probe; the full-field temperature distribution image of the upper surface of the asphalt mixture specimen obtained by the infrared thermal imager; and the ambient temperature and humidity data inside the test chamber obtained by the ambient temperature and humidity sensor; S6, based on the test data collected in step S5, calculation... The thermal resistance performance index of asphalt mixture is calculated. The thermal resistance performance index of asphalt mixture includes at least one of the following: thermal response delay time, thermal diffusivity, apparent thermal resistance, normalized thermal resistance index, and thermal uniformity index; S7, asphalt mixture specimens with different material compositions, gradations, and porosities are tested according to steps S1 to S6 to obtain a thermal resistance performance index dataset; a machine learning prediction model is trained using this thermal resistance performance index dataset to establish a nonlinear mapping relationship between asphalt mixture material parameters and thermal resistance performance index; the thermal resistance performance index of asphalt mixture is predicted using the machine learning prediction model.

2. The method for evaluating the thermal resistance performance of asphalt mixtures according to claim 1, characterized in that, In step S6, the method for calculating the thermal response delay time includes: recording the moment when the controllable heat source is turned on or when the asphalt mixture specimen begins to be heated; monitoring the moment when the temperature at the center point of the upper surface of the asphalt mixture specimen begins to rise significantly; recording the moment when the temperature at the center point of the lower surface of the asphalt mixture specimen first rises to a preset temperature threshold from the moment the heating begins; and subtracting the moment when the upper surface temperature begins to change from the moment when the lower surface temperature reaches the threshold, the resulting time difference is the thermal response delay time.

3. The method for evaluating the thermal resistance performance of asphalt mixtures according to claim 1, characterized in that, In step S6, the method for calculating the thermal diffusivity includes: based on the internal temperature distribution data along the thickness direction of the asphalt mixture specimen obtained by the internal temperature probe, fitting the measured temperature change curves at different internal depths with the solution of the one-dimensional unsteady-state thermal conductivity differential equation to derive the optimal thermal diffusivity.

4. The method for evaluating the thermal resistance performance of asphalt mixtures according to claim 1, characterized in that, In step S6, the method for calculating the apparent thermal resistance includes: calculating the effective heat flux density applied to the upper surface of the asphalt mixture specimen under stable heating conditions; measuring the average temperature difference between the upper and lower surfaces of the asphalt mixture specimen at this time; dividing the average temperature difference between the upper and lower surfaces of the asphalt mixture specimen by the calculated effective heat flux density, and the resulting quotient is the apparent thermal resistance; in step S6, the method for calculating the normalized thermal resistance index includes: dividing the apparent thermal resistance of the asphalt mixture specimen by the thickness of the asphalt mixture specimen to obtain the normalized thermal resistance index; or, multiplying the apparent thermal resistance by a standard temperature difference constant and then dividing by the thickness of the asphalt mixture specimen to obtain the normalized thermal resistance index.

5. The method for evaluating the thermal resistance performance of asphalt mixtures according to claim 1, characterized in that, In step S6, the method for calculating the thermal uniformity index includes: based on the full-field temperature distribution image of the upper surface of the asphalt mixture specimen acquired by the infrared thermal imager, calculating the average temperature of all areas on the entire upper surface of the asphalt mixture specimen; calculating the degree of deviation of the temperature of all points on the entire upper surface of the asphalt mixture specimen from the average temperature, i.e., the standard deviation of temperature; dividing the standard deviation of temperature by the average temperature to obtain the coefficient of variation; and dividing the standard deviation by the coefficient of variation to obtain the thermal uniformity index.

6. The method for evaluating the thermal resistance performance of asphalt mixtures according to claim 1, characterized in that, In step S7, the process of testing asphalt mixture specimens with different material compositions, gradations, and porosities according to steps S1 to S6 to obtain a thermal resistance performance index dataset includes: establishing a data file for each asphalt mixture specimen, the data file including the material parameters of the asphalt mixture specimen and the calculated thermal resistance performance index of the asphalt mixture specimen; summarizing all asphalt mixture specimens to form a structured table; each row of the structured table represents an asphalt mixture specimen, and each column represents the material parameters or thermal resistance performance index, thereby constituting a thermal resistance performance index dataset for machine learning training.

7. The method for evaluating the thermal resistance performance of asphalt mixtures according to claim 6, characterized in that, In step S7, the step of training a machine learning prediction model using the thermal resistance performance index dataset to establish a nonlinear mapping relationship between asphalt mixture material parameters and thermal resistance performance index includes: selecting all columns describing material parameters in the thermal resistance performance index dataset as input to the machine learning prediction model, and selecting the thermal resistance performance index to be predicted as output of the machine learning prediction model; cleaning the thermal resistance performance index dataset before training and normalizing all data; randomly dividing the thermal resistance performance index dataset into training and testing sets; inputting the training set into the initial machine learning prediction model, and using random forest, gradient boosting decision tree, or neural network to find the mapping law from material parameters to thermal resistance performance index to obtain the machine learning prediction model; using cross-validation to validate the machine learning prediction model and calculating the evaluation index of the machine learning prediction model; if the difference between the evaluation index and the preset evaluation index is greater than the deviation threshold, then optimizing the machine learning prediction model to obtain the optimized machine learning prediction model.

8. The method for evaluating the thermal resistance performance of asphalt mixtures according to claim 7, characterized in that, In step S7, if the difference between the evaluation index and the preset evaluation index is greater than the deviation threshold, the machine learning prediction model is optimized to obtain the optimized machine learning prediction model. This includes: diagnosing the root cause of the poor performance of the machine learning prediction model, which may include: insufficient data volume, poor data quality, insufficient data representativeness, weak feature correlation, overly simple model, overly complex model, or improper model parameter settings; and taking one or more corresponding optimization measures based on the root cause: preparing and testing more asphalt mixture specimens with different formulations; reviewing experimental data, identifying and correcting or removing obvious abnormal test records; attempting to create new, physically meaningful combination features based on existing material parameters; and using statistical methods or the feature importance assessment function built into the machine learning prediction model to identify and remove redundant features that contribute little to the prediction target. The process involves optimizing the input to the machine learning prediction model. For models that are too simple, a more complex model is chosen, or the complexity of the existing model is increased. For models that are too complex, the complexity is reduced, or regularization constraints are added. A set of hyperparameters that optimizes the model's performance on the validation set is obtained using grid search or random search methods. For neural networks, the learning rate and training epochs are adjusted. After applying optimization measures, the model is retrained using the updated dataset and new model configuration. If the difference between the new evaluation metric and the preset evaluation metric is still greater than the deviation threshold, the "diagnosis-optimization-retraining" cycle is repeated until the model's performance meets the requirements. The optimized machine learning prediction model is then obtained.

9. A system for evaluating the thermal resistance performance of asphalt mixtures, characterized in that, The system includes: a specimen and sensing module for preparing asphalt mixture specimens of specified dimensions, pre-embedding or drilling temperature measurement channels in the thickness direction of the specimens, arranging at least three equally spaced internal temperature probes, and uniformly arranging at least two surface temperature probes on the upper and lower surfaces of the specimens to monitor their internal temperature gradient and surface temperature; an environmental simulation and heat source control module, including a sealed test chamber with a high-reflectivity inner wall, a platform for placing the specimens inside the test chamber, and using thermal insulation material to wrap the sides and lower surface of the specimens, exposing only their upper surface; the module also includes a controllable heat source, an environmental temperature and humidity sensor integrated into the chamber, a radiation sensor, and an infrared thermal imager, used to set and adjust the heat source according to the target working conditions. The power, height, and incident angle of the source are set to maintain a preset constant or periodically varying irradiance; a data acquisition module is used to synchronously and continuously acquire and record test data from the surface temperature probe, internal temperature probe, infrared thermal imager, and environmental temperature and humidity sensor during steady-state testing or dynamic cyclic testing; a thermal resistance performance calculation and prediction module is used to calculate at least one thermal resistance performance index among thermal response delay time, thermal diffusivity, apparent thermal resistance, normalized thermal resistance index, and thermal uniformity index based on the acquired test data, and to train a machine learning prediction model using a dataset of thermal resistance performance indexes obtained from tests of specimens with different material parameters, so as to predict the thermal resistance performance index of asphalt mixtures through the machine learning prediction model.

Images