Asphalt pavement moisture prediction method based on moisture prediction model, equipment and medium

Through the method based on the moisture prediction model, using core extraction and environmental box simulation data, the diffusion model is constructed for iterative optimization, which solves the problems of high cost of burying sensors and inaccurate measurements, and realizes high reliability, non-destructive measurement of asphalt pavement moisture and pavement surface Damage evaluation.

CN118940367AInactive Publication Date: 2024-11-12INNER MONGOLIA UNIV OF TECH

Patent Information

Application Number
CN202410965886.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-18
Publication Date
2024-11-12
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

In the prior art, embedded sensors are used for asphalt pavement moisture prediction, which are costly, susceptible to environmental impact and inaccurate measurements, and lack effective moisture analysis methods.

Method used

Using a method based on the moisture prediction model, the measured data is obtained through core extraction and environmental box simulation, a diffusion model is constructed, and the characteristic diffusion coefficient is used for iterative optimization to achieve non-destructive measurement and prediction of asphalt pavement moisture.

Benefits of technology

It realizes high-reliability non-destructive measurement of asphalt pavement moisture, reduces cost investment, improves measurement accuracy and trust, and can effectively predict the moisture content in areas below the pavement surface layer, which promotes the evaluation of road damage and the determination of maintenance time.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118940367A_ABST
    Figure CN118940367A_ABST
Patent Text Reader

Abstract

The invention relates to an asphalt pavement moisture prediction method based on a moisture prediction model, equipment and a medium. The method comprises the following steps: firstly, coring an asphalt pavement, cutting to obtain a disc-shaped test piece, putting the test piece into an environment box to simulate factors such as temperature and humidity in an actual environment, and measuring the moisture content of each test piece along with time change; a moisture prediction model with moisture changing along with diffusion time and diffusion distance is constructed, the model is preliminarily solved through a spatial discretization method, the solved moisture content is a three-dimensional curved surface containing undetermined parameters, and a gradient descent algorithm is used for fitting a test result and a preliminary solving result to obtain a feature diffusion coefficient. And finally, completing prediction verification through a verification group test so as to judge the water damage degree of the asphalt pavement and provide a reference for determining the optimal maintenance time. According to the method, the moisture content in the asphalt pavement is predicted through the diffusion model, so that the water damage degree of the asphalt pavement is judged, and a certain reference value can be provided for implementation of subsequent maintenance engineering.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of civil engineering, and in particular to a method, equipment and medium for predicting moisture in an asphalt pavement based on a moisture prediction model. Background Art

[0002] Asphalt pavement is widely used at home and abroad, and plays an important role in highways, urban roads, airport runways and other fields. Asphalt pavement is affected by rainfall and groundwater during use, and its internal soluble substances are reduced, resulting in water damage to the asphalt pavement. Some studies have shown that water will accelerate the aging process of asphalt binders and have irreversible effects on the molecular structure, colloidal structure and carbonyl index of asphalt binders. With precipitation in the form of rain, snow and other forms, moisture will gradually penetrate from the asphalt surface layer downward, causing loosening and damage to the pavement structure. Especially in the case of large temperature changes, after moisture enters the pavement, it will expand and contract when the temperature changes, aggravating the cracking and damage of the pavement; in cold areas, after moisture penetrates into the pavement, it will freeze and expand, causing the cracks and pores of the asphalt pavement to expand, and even causing the pavement to fluctuate and break, seriously affecting the flatness and service life of the pavement. At the same time, the infiltration of moisture will soften the asphalt binder of the asphalt pavement, reduce the shear strength and stability of the pavement, and easily cause rutting and deformation. Therefore, how to analyze the moisture content of asphalt pavement is a problem.

[0003] Chinese patent application publication number CN103808914A discloses an airport asphalt concrete pavement performance acquisition system and control evaluation method. The system includes buried sensors, data acquisition terminals and analysis control terminals, wherein each buried sensor is connected to a data acquisition terminal, and the data acquisition terminal is wirelessly connected to the analysis control terminal installed in the control center; the buried sensor includes a columnar housing, a single-chip controller, a temperature sensor, a humidity sensor, an acceleration sensor, a vibration sensor, a plurality of limiters, a lithium battery, a lower fixed base and an upper detachable cover; the limiter includes a fixed sleeve, a retractable limit rod, a high-strength spring, at least one fixed limit pin, at least one retractable limit pin and a limit sleeve.

[0004] The above application uses a universal humidity sensor buried at different depths below the road surface to test its moisture content, but its cost is high, and its performance is affected by the environment, such as temperature changes, pollutants, chemicals, etc., which may cause a decrease in accuracy; at the same time, if it is improperly installed or in an inappropriate environment, it may affect its measurement results; some humidity sensors may be fragile and easily damaged or physically destroyed, affecting the final measurement results.

[0005] In summary, there is currently a lack of an asphalt pavement moisture analysis method to solve or partially solve the aforementioned problems. Summary of the invention

[0006] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide an asphalt pavement moisture prediction method, equipment, and medium based on a moisture prediction model to solve or partially solve the problem of high degree of damage to the pavement existing in the moisture prediction scheme based on buried sensors.

[0007] The purpose of the present invention can be achieved by the following technical solutions:

[0008] One aspect of the present invention provides a method for predicting moisture in an asphalt pavement based on a moisture prediction model, comprising the following steps:

[0009] Step S1, obtaining measured penetration time-moisture content data of an asphalt pavement, wherein the measured penetration time-moisture content data is obtained by coring the asphalt pavement and performing environmental chamber simulation;

[0010] Step S2, based on the pre-constructed moisture prediction model, under the premise of preset initial conditions and boundary conditions, taking the characteristic diffusion coefficient as a parameter, solving the diffusion model to obtain the penetration time-depth-moisture content parameter data, obtaining the predicted penetration time-moisture content parameter data through dimensionality reduction processing, taking maximizing the similarity between the measured penetration time-moisture content data and the predicted penetration time-moisture content parameter data as the optimization goal, taking the characteristic diffusion coefficient as the optimization object, and performing iterative optimization until the stop condition is met;

[0011] Step S3, based on the iteratively optimized characteristic diffusion coefficient, the moisture prediction model is used to predict the moisture of the asphalt pavement.

[0012] As a preferred technical solution, the process of obtaining measured penetration time-moisture content data by coring and environmental chamber simulation of asphalt pavement includes the following steps:

[0013] Step S101, coring the asphalt pavement to obtain a disc-shaped sample with a diameter-to-thickness ratio greater than a preset value;

[0014] Step S102, placing the disc-shaped sample in an environmental box matching the temperature and humidity of the asphalt pavement, and obtaining actual penetration time-moisture content data through measurement.

[0015] As a preferred technical solution, the moisture prediction model is:

[0016]

[0017] Among them, x is the depth from the surface layer of the asphalt pavement, t is the time of water infiltration, C is the water content at a certain point in the asphalt pavement, and D is the characteristic diffusion coefficient.

[0018] As a preferred technical solution, the initial conditions and boundary conditions are as follows:

[0019] Initial condition: C(x,0) = 0

[0020] Boundary condition: C(a,t) = C2; C(b,t) = 0

[0021] Among them, t≥0 is the time of water infiltration, a < x < b, x is the depth from the surface layer of the asphalt pavement, a and b are the value boundaries, C is the water content at a certain point in the asphalt pavement, and C2 is the water content in contact with the outside at the surface layer of the asphalt pavement.

[0022] As a preferred technical solution, the process of solving the diffusion model to obtain the parametric data of penetration time-depth-water content includes the following steps:

[0023] Step S201, based on the water prediction model, the initial conditions and boundary conditions, calculate the parametric data of penetration time-depth-water content including the characteristic diffusion coefficient by constructing a partial differential equation.

[0024] As a preferred technical solution, the process of obtaining the parametric data of predicted penetration time-water content through dimensionality reduction processing includes the following steps:

[0025] Step S202, for the parametric data of penetration time-depth-water content, integrate in the depth dimension to obtain the parametric data of predicted penetration time-water content with the characteristic diffusion coefficient as a parameter.

[0026] As a preferred technical solution, the iterative optimization process includes the following steps:

[0027] Step S203, initialize the characteristic diffusion coefficient;

[0028] Step S204, calculate the goodness of fit between the measured penetration time-water content data and the predicted parametric data of penetration time-water content, and solve and correct the characteristic diffusion coefficient;

[0029] Step S205, repeat Step S204 until the stop condition is met.

[0030] As a preferred technical solution, the stop condition is:

[0031] The maximum number of iterations is reached or the goodness of fit between the measured penetration time-moisture content data and the predicted penetration time-moisture content parameter data is less than a preset value.

[0032] Another aspect of the present invention provides an electronic device, comprising: one or more processors and a memory, wherein the memory stores one or more programs, and the one or more programs include instructions for executing the aforementioned asphalt pavement moisture prediction method based on the moisture prediction model.

[0033] Another aspect of the present invention provides a computer-readable storage medium comprising one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the aforementioned asphalt pavement moisture prediction method based on the moisture prediction model.

[0034] Compared with the prior art, the present invention has at least one of the following beneficial effects:

[0035] (1) Realize non-destructive measurement of asphalt pavement moisture: In order to solve the problem of high degree of damage to the pavement caused by moisture prediction schemes based on buried sensors, the present invention does not need to bury sensors in the asphalt pavement. It only needs to perform coring and environmental chamber simulation on the pavement to be tested in advance to obtain the measured penetration time-moisture content data. Then, with the goal of minimizing the difference between the measured and predicted penetration time-moisture content data, the characteristic diffusion coefficient is iteratively optimized based on the constructed moisture prediction model, thereby realizing non-destructive measurement of asphalt pavement moisture. It does not require the asphalt pavement to be processed, which not only avoids damage to the asphalt pavement during measurement, but also saves labor costs and capital costs.

[0036] (2) Reliability: The present invention iteratively optimizes the characteristic diffusion coefficient through an iterative algorithm and predicts the moisture content of the surface layer of the asphalt pavement based on the moisture prediction model. It has high trust and reliability and can well predict the moisture content of each point in the area below the surface layer of the asphalt pavement.

[0037] (3) Broad application prospects: It has promoted the use of moisture content at various points below the surface layer of asphalt pavement to evaluate the damage of asphalt pavement in actual projects. It can be used to determine the optimal maintenance time of asphalt pavement. The moisture content can be used to judge the damage of the pavement, which is convenient for regular inspection and maintenance of the pavement, timely repair of cracks and damage on the pavement, and prevention of moisture from entering the interior of the pavement, thereby maintaining the integrity and stability of the pavement.

[0038] (4) Reducing computational overhead: Before the iterative optimization process, the penetration time-depth-moisture content parameter data is subjected to dimensionality reduction processing to obtain the predicted penetration time-moisture content parameter data, and then the characteristic diffusion coefficient is updated to convert the three-dimensional problem into a two-dimensional problem, thereby reducing the computational overhead. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a flow chart of a method for predicting moisture content of an asphalt pavement based on a moisture prediction model in an embodiment;

[0040] Figure 2 The absolute humidity and relative humidity correspondence table in the embodiment;

[0041] Figure 3 This is the moisture diffusion diagram of the 30cm area on the surface layer of the asphalt pavement in the embodiment;

[0042] Figure 4 A three-dimensional surface diagram obtained by solving the moisture prediction model in the embodiment;

[0043] Figure 5 It is a schematic diagram of integrating the three-dimensional curved surface in the thickness direction of the asphalt pavement in the embodiment;

[0044] Figure 6 This is a summary diagram of the quality changes of different samples in the embodiment;

[0045] Figure 7 It is a fitting diagram of the mass change of the sample in the environmental chamber in the embodiment;

[0046] Figure 8 The graph of water vapor concentration of different samples in the embodiment changing with time;

[0047] Fig. 9 It is a goodness of fit diagram of the test value and the predicted value at 85% water vapor concentration -25 degrees Celsius in the embodiment;

[0048] Fig.10 Plotting graphs for actual values ​​and fitted values ​​in the examples;

[0049] Fig.11 It is a diffusion coefficient diagram under different environmental conditions in the embodiment;

[0050] Fig.12 It is a goodness of fit diagram of the test value and the predicted value at 50% water vapor concentration -15 degrees Celsius in the embodiment;

[0051] Fig.13 It is a goodness of fit diagram of the experimental value and the predicted value at 100% water vapor concentration of -60 degrees Celsius in the embodiment. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0053] Example 1

[0054] In view of the problems existing in the aforementioned prior art, this embodiment provides an asphalt pavement moisture prediction method based on a moisture prediction model. The moisture of each point in the area below the surface layer of the asphalt pavement is set through boundary conditions and initial conditions. It is assumed that the moisture in the asphalt pavement is only obtained by penetration from the road surface, and the penetration direction is only one-dimensionally diffused in the direction perpendicular to the road surface, and the horizontal diffusion is not considered. The function equation uses a one-dimensional parabolic equation. The ultimate goal is to be able to predict the moisture content of each point in the area 30cm below the surface layer of the asphalt pavement. Different from the traditional general humidity sensor that measures the moisture content below the surface layer of the asphalt pavement, this method can measure the moisture content of each point in the area below the surface layer of the asphalt pavement while saving the cost of measurement. After obtaining the prediction results, the moisture content of each point in the asphalt pavement can be used to infer the degree of damage to the asphalt mixture. It can be used as a reference for judging the optimal maintenance time in the asphalt pavement maintenance project, so as to extend the service life of the asphalt pavement.

[0055] See also Figure 1 ,The asphalt pavement moisture prediction method based on the moisture prediction model includes the following steps:

[0056] Step S1, obtaining dynamic humidity data / curve (i.e. measured penetration time-water content data). This step specifically includes the following sub-steps:

[0057] Step S101, coring and processing the core sample from the asphalt pavement to be tested.

[0058] The core sample is obtained by coring the actual on-site asphalt pavement, and the obtained core sample is cut to obtain a disc-shaped sample. Preferably, considering that the moisture at each point below the surface layer of the asphalt pavement mainly penetrates vertically downward from the road surface, the aspect ratio of the sample is greater than 10 to better simulate the actual environment where moisture penetrates only from one direction. It should be noted that the minimum value of the aspect ratio can be adjusted according to actual needs.

[0059] Step S102: environmental chamber simulation.

[0060] The samples that meet the requirements are placed in an environmental chamber to simulate the temperature and humidity in the actual environment. According to the temperature and relative humidity in the environment, the corresponding value of the absolute humidity in the environmental chamber can be obtained by looking up the table. The corresponding table of absolute humidity and relative humidity is shown in the figure below. Figure 2 shown.

[0061] Through actual measurement, the mass change of the specimen in the environmental chamber is obtained as the specimen changes over time. Since the mass change of the specimen in the environmental chamber is only determined by the diffusion amount of water, the mass change rate is converted into the moisture content of the specimen. The obtained data is sorted and the dynamic humidity data / curve is drawn.

[0062] Step S2, constructing a moisture model with characteristic diffusion coefficient as a model parameter, with the goal of minimizing the difference between the predicted two-dimensional images of moisture content C changing with time t (i.e., the parameter-containing data of the predicted penetration time-moisture content), and iteratively optimizing the characteristic diffusion coefficient. This step specifically includes the following sub-steps:

[0063] Step S201, based on Fick's second law, establish a moisture prediction model with time t, depth x as independent variables and moisture content C as dependent variable. The main significance of the moisture prediction model is that moisture will continue to diffuse into the asphalt pavement under the action of temperature and concentration gradient. Considering that the pavement thickness is much smaller than the length and width, it is assumed that the moisture in the 30cm area below the asphalt pavement surface layer only diffuses vertically from the asphalt surface layer from top to bottom.

[0064] The mathematical form of the moisture prediction model is: Where x represents the depth from the surface layer of the asphalt pavement, t represents the time of water penetration, C represents the water content at a certain point in the asphalt pavement, and D is the characteristic diffusion coefficient. This moisture prediction model is a function of the water content C of each point in the 30cm area of ​​the asphalt pavement surface layer changing with the depth x of the asphalt pavement surface and time t. D is the characteristic diffusion coefficient, which is a fixed value that changes with different materials. Its value can represent the diffusion speed (the larger the D value, the faster the diffusion).

[0065] The core of establishing a moisture prediction model is to solve the characteristic diffusion coefficient, so some initialization settings are required, including proposing the mathematical form of the moisture prediction model, setting the initial values ​​and boundary conditions, etc. The solution of the moisture prediction model requires some initialization settings to define its initial values ​​and boundary conditions. After studying the actual asphalt pavement, it was found that the moisture in the 30cm area of ​​the asphalt pavement surface layer is mainly obtained by vertical one-dimensional penetration from the asphalt pavement surface layer. Therefore, the mathematical form of the moisture prediction model is modeled as an equation for solving a one-dimensional parabola to simulate the actual situation.

[0066] When solving the moisture prediction model, you need to first select a suitable solution grid and define the initial conditions and boundary conditions. Define the initial conditions: Write a function that returns the initial conditions. The initial conditions are applied to the first time value t = 0 to define the moisture content C2 (x, 0). Its meaning is that the moisture concentration in the external environment in contact with the asphalt pavement is a constant value, and its value can be obtained by measuring the actual environment. Assume that the asphalt pavement does not contain moisture before the moisture begins to diffuse, that is, C1 (x, 0) = 0. Define the boundary conditions: When defining the distance x from the surface layer of the asphalt pavement, the unit is cm, and its value range starts from 0cm and increases by 3cm in steps until it does not exceed 30cm. Specifically, it generates a vector containing equally spaced values, starting from 0 and increasing by one value every 3 units until it reaches or slightly exceeds 30cm. When defining the time t, the unit is h, and its value range starts from 0h and increases by 2h in steps until it does not exceed 250h. Specifically, it generates a vector containing equally spaced values ​​starting at 0h and increasing by 2 units until it reaches or slightly exceeds 250h.

[0067] The boundary range is defined in the area of ​​30cm of the asphalt pavement surface layer, and the solution interval is a≤x≤b, where a=0cm, b=30cm, which means that the point to be solved is located within the range of 30cm below the asphalt pavement surface layer. The boundary of time t is set to 0-250h. Its connotation is that after the survey of the actual on-site environment, the diffusion of moisture in the surface layer of the asphalt pavement is more obvious within 250h. Therefore, when the change of moisture content in the surface layer of the asphalt pavement predicted by the moisture prediction model changes with time, the size of the time grid is 0-250h. For the convenience of calculation, it is assumed that the moisture in the asphalt pavement is only obtained by penetration from the road surface, and the size of the moisture content is only related to x (depth from the road surface), t (penetration time), and C (here C represents the moisture concentration at the surface of the asphalt pavement, that is, x=0, t=0).

[0068] Before solving, we need to define C(a, t), that is, to make an initial selection of C(a, t) for subsequent calculations. For the left boundary at time t=0, for x=0, set its moisture content equal to the moisture content of the external environment that the road surface contacts; that is, C2(0, t) is a constant value that can be obtained through measurement. For the range where x is greater than 0, set its moisture content to 0, that is, set it at time t=0, and there is no moisture within 30cm below the surface layer of the asphalt pavement. The mathematical form of the initial conditions and boundary conditions is:

[0069] Initial conditions: t=0, x=0, C=C2; x>0, C=C1=0, that is, C(x,0)=0

[0070] Boundary conditions: t ≥ 0, x = 0, C = C2,; x = +∞, C = C1, i.e., C(a, t) = C2; C(b, t) = 0.

[0071] The moisture diffusion diagram of the 30 cm area of the asphalt pavement surface layer is as Figure 3 shown.

[0072] Calculate the obtained moisture prediction model and draw a three-dimensional image of the moisture content C of each point in the 30 cm area of the asphalt pavement surface layer changing with time t and depth x (i.e., the infiltration time-depth-moisture content parametric data). The specific calculation process is as follows.

[0073] For the solution of the moisture prediction model, the diffusion law involved is shown in Equation (1). The spatial order is two-order and has typical parabolic characteristics. Therefore, the method for solving one-dimensional parabolic-elliptic partial differential equations can be used to solve it.

[0074]

[0075] In the problem of moisture diffusion into the 30 cm area of the asphalt pavement surface layer, t ≥ 0, a < x < b, and its initial conditions and boundary conditions are as follows:

[0076] Initial conditions: t = 0, x = 0, C = C2; x > 0, C = C1 = 0, i.e., C(x, 0) = 0 (2)

[0077] Boundary conditions: t ≥ 0, x = 0, C = C2,; x = +∞, C = C1, i.e., C(a, t) = (3)

[0078] C2; C(b, t) = 0.

[0079] The physical meanings of Equation (2) and Equation (3) are as follows: At the beginning of diffusion, the water vapor concentration in the asphalt pavement surface layer is zero. At the contact surface between the asphalt pavement and the external atmospheric environment, the moisture concentration is C2. At the other end of the asphalt pavement far from the contact surface (where the pavement thickness x = b), it does not contact moisture and the moisture concentration is zero. Thus, a one-dimensional diffusion system of moisture vertically downward into the 30 cm area of the asphalt pavement surface layer is formed. Convert Equation (1), Equation (2), and Equation (3) into the general form

[0080]

[0081] Initial conditions: C(x, 0) = C0(x) (5) Boundary conditions:

[0082] In Equation (4) is the flux, is the source term. Comparing Equation Set (1), (2), and (3) with Equation Set (4), (5), and (6), we can obtain:

[0083]

[0084] C0(x)=0 (8)

[0085] p a (x, t, C) = C(a, t) - C0, q a (x, t) = 0; p b (x, t, C) = C (b, t), q (x, t) = 0; (9)

[0086] The above problem can be solved by using the partial differential equation solving toolbox integrated in the relevant software. Let the solution be C(x, t). Since the diffusion coefficient is unknown, the actual solution is the equation C(x, t, D) containing the parameter D (i.e., the penetration time-depth-water content including parameter data). According to the previous research on the water content in asphalt pavement, the diffusion coefficient is a constant related to the material, and can be preliminarily selected based on the research results of previous researchers on asphalt pavement. The characteristic diffusion coefficient solved by moisture prediction can also be regarded as a correction to the diffusion coefficient.

[0087] Step S202, the three-dimensional image obtained in step S201 is integrated along the depth direction of the road surface to reduce the dimension of the predicted result, so as to obtain a two-dimensional image of the moisture content C changing with time t (i.e., the parameter data of the predicted penetration time-moisture content).

[0088] Since in step S1, the water absorption curve of the material obtained in the laboratory is usually a mass-time curve, the mass-time curve can be converted into a concentration-time curve by subtracting the mass from the volume, but the curve describes the volume concentration change of the sample. In step S2, the C(x, t, D) solved by the moisture prediction model is the surface concentration change, so the function I is defined as the integral of C(x, t, D), and its purpose is to calculate the numerical integral of the moisture prediction model on the x grid, see formula (10)

[0089]

[0090] See also Figure 5 In step S2, the numerical integral of the moisture prediction model on the x grid is calculated. Specifically, it does the following: First, x is a vector describing the spatial grid. C is a function of thickness x and time t, which contains the system solution obtained by solving the moisture prediction model. The solution results in a three-dimensional surface such as Figure 4 The second is to specify the integration for the second dimension (i.e. along the x-axis). The mathematical expression of the integration is The schematic diagram of integrating a three-dimensional surface is as follows Figure 5As shown. The main purpose is that the result obtained in step S1 is a two-dimensional curve of the change of moisture content over time. When solving the moisture prediction model, the nonlinear fitting toolbox integrated in the software needs to be used to pass the result of the partial differential equation solving toolbox to the nonlinear fitting toolbox for fitting with the result obtained in step S1 to achieve the solution of the characteristic diffusion coefficient D.

[0091] Step S203, initializing parameters: selecting appropriate initial parameter values ​​of the characteristic diffusion coefficient, which are selected within a certain range based on measured data.

[0092] Step S204, calculating the gradient: calculating the goodness of fit between the measured penetration time-water content data and the predicted penetration time-water content data including parameters, solving and correcting the characteristic diffusion coefficient.

[0093] Step S205, repeat iteration: Repeat step S204 until the stop condition is met, such as reaching the maximum number of iterations or the objective function changes very little. By fitting the image curve predicted by the moisture prediction model, a moisture prediction model that best matches the moisture content of the asphalt pavement surface layer area in the actual project is obtained.

[0094] In the present embodiment, the gradient descent method is adopted to solve the characteristic diffusion coefficient in the moisture prediction model by iterative optimization. The gradient descent method adjusts the numerical value of the parameter by continuous iteration to minimize or maximize the value of the objective function (loss function). The gradient descent method gradually adjusts the numerical value of the parameter along the gradient direction of the objective function, i.e. the direction in which the function value grows fastest, until a local or global optimal solution is reached. In each iteration, the direction and step size of parameter update are determined by calculating the partial derivative (gradient) of the objective function to the parameter. The gradient descent method fits the function I containing the parameter D to the concentration-time curve obtained by the experiment, adopts the nonlinear fitting toolbox integrated in the software, and the result of the partial differential equation solving toolbox solving is transferred to the nonlinear fitting toolbox, so as to realize the solution to the characteristic diffusion coefficient D.

[0095] Preferably, under the premise of no conflict, other iterative optimization methods can be used to solve the characteristic diffusion coefficient.

[0096] Preferably, the obtained two-dimensional image is compared with the curve obtained in step S102 and the verification group on the same coordinate axis, and the goodness of fit between the results obtained by the moisture prediction model and the experimental test results of the verification group is used to ensure that the moisture prediction model finally obtained can accurately predict the moisture content of the area below the surface layer of the asphalt pavement.

[0097] Step S3, predicting the moisture content based on the iterated moisture prediction model of step S2, and predicting the pavement damage based on the prediction result, so as to determine the optimal maintenance time.

[0098] The effectiveness of this method will be illustrated by practical examples in the following text.

[0099] In order to predict the moisture content of each point in the 30cm area below the surface layer of the asphalt pavement under different environmental conditions, first take the sample under the environmental conditions of 85% water vapor concentration -25 degrees Celsius as an example, and investigate the on-site environment of the asphalt pavement to obtain the humidity and temperature conditions. Then set the temperature of the environmental box to 25 degrees Celsius and the water vapor concentration to 85% (simulating the daily climate in southern my country). Pre-process the samples obtained by coring the asphalt pavement on site to obtain samples with an aspect ratio greater than 10 (to ensure that the moisture in the sample is mainly infiltrated from a single direction), so as to simulate the moisture content of each point in the area 30cm below the surface layer of the actual asphalt pavement, which mainly comes from the vertical downward diffusion of moisture in contact with the surface layer of the asphalt pavement. Measure the initial mass of the sample to be placed in the environmental box and record the results. Put the sample into the environmental box with the previously set temperature and humidity, and take it out every 2 hours to measure its mass, for a total of 250 hours and 125 sets of data. According to the changes of the sample over time and the mass changes of the specimen, the mass change graphs of the samples under different environmental climates are summarized and drawn as shown in the figure. Figure 6 The simulated southern climate data is plotted after nonlinear fitting as shown in Figure 7 As shown in the figure, since the mass of the sample in the environmental chamber is only affected by the water immersion, the change in mass can be used to convert the water content in the sample. The data of the change in the mass of the sample in the environmental chamber over time is plotted in a two-dimensional image, and the data of the change in the water vapor concentration of the samples under different environments over time are summarized in Figure 8 Medium (85% Rh25 degrees represents a water vapor concentration of 85% and a temperature of 25 degrees Celsius, simulating the climate in southern my country; 50% Rh15 degrees represents a water vapor concentration of 50% and a temperature of 15 degrees Celsius, simulating the climate in northern my country; 100% Rh60 degrees represents a water vapor concentration of 100% and a temperature of 60 degrees Celsius, simulating extreme climate).

[0100] Regarding the setting of initial conditions and boundary conditions of the moisture prediction model, when simulating the southern climate (85% Rh25℃), C2=19.56g / m 3 , simulated northern climate (50% Rh 15℃), C2 = 6.41 g / m 3 When simulating extreme climate (100% Rh 60℃), C2=129.54g / m 3 For the left boundary at time t=0, for x=0, the moisture content is equal to the moisture content of the external environment that the road surface contacts, that is, C(0, t)=C2. For the range where x is greater than 0, the moisture content is 0, that is, at time t=0, there is no moisture within 30cm below the surface layer of the asphalt pavement, that is, C(x, 0)=0.

[0101] In the iterative optimization, the gradient descent method is used to compare the two-dimensional curve obtained by the moisture prediction model (the two-dimensional curve is obtained by reducing the dimension of the three-dimensional surface obtained by the moisture prediction model) with the curve data obtained by the validation group on the same coordinate axis, such as Fig. 9 The goodness of fit R between the moisture prediction model prediction value and the validation group test value is obtained. 2 =0.98, to ensure that the final output function equation is the optimal solution. (Goodness of fit is an indicator used to measure the degree to which a statistical model fits the data. Goodness of fit is used to evaluate the degree to which a model fits the observed data, that is, the model's ability to explain the data. The higher the goodness of fit, the better the model fits the data, and the model's predictive ability may be stronger.)

[0102] The purpose of performing dimensionality reduction processing on the predicted results is to obtain a two-dimensional image in order to verify the deviation between the actual measured data and the predicted results. The goodness of fit result obtained in this example is 0.98, which shows the accuracy of the asphalt pavement moisture prediction method based on the moisture prediction model provided in this embodiment.

[0103] When the results obtained by the diffusion equation are consistent with those obtained by the validation group, its accuracy can be verified and used to predict the moisture content of the asphalt pavement surface layer.

[0104] According to the moisture content of each point in the asphalt pavement, the damage of the asphalt pavement can be analyzed. The result can be used as a reference to determine the best maintenance time of the asphalt pavement.

[0105] This method has the following beneficial effects:

[0106] (1) For the first time, the moisture prediction model was used to predict the moisture content of the surface layer of asphalt pavement. Compared with the traditional asphalt pavement method of using universal humidity sensors to measure the moisture content of each point in the area below the surface layer of the asphalt pavement, this method can greatly reduce the cost of measuring the moisture content of each point in the area below the surface layer of the asphalt pavement. The gradient descent method is used to fit the actual data, which has a high degree of trust and reliability. In addition, the gradient descent method provides intuitive parameter estimation results, making it easier to interpret and understand the fitting model. By analyzing the size and sign of the parameters, the relationship between the variables and their degree of influence can be inferred. The diffusion equation used in the moisture prediction model is a mathematical model that describes the diffusion or propagation of matter or energy in space. It is usually used to describe the diffusion process of liquids, gases, heat, etc. in space. The diffusion equation can be used to well predict the moisture content of each point in the area below the surface layer of the asphalt pavement.

[0107] (2) It realizes non-destructive measurement and avoids the tedious process of burying humidity sensors in asphalt pavement. Humidity sensors buried in asphalt pavement are easily affected by the environment and may be damaged under vehicle load. This greatly promotes the use of moisture content at various points below the surface layer of asphalt pavement to evaluate the damage of asphalt pavement in actual engineering, and has certain reference value for determining the optimal maintenance time of asphalt pavement.

[0108] Example 2

[0109] This embodiment provides an electronic device, including: one or more processors and a memory, wherein the memory stores one or more programs, and the one or more programs include instructions for executing the asphalt pavement moisture prediction method based on the moisture prediction model as described in Example 1.

[0110] Example 3

[0111] This embodiment provides a computer-readable storage medium, including one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the asphalt pavement moisture prediction method based on the moisture prediction model as described in Example 1.

[0112] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for predicting moisture in asphalt pavement based on a moisture prediction model, characterized in that: It includes the following steps: Step S1: Obtain the measured penetration time - moisture content data of the asphalt pavement, where the measured penetration time - moisture content data is obtained by coring the asphalt pavement and simulating it in an environmental chamber; Step S2: Based on the pre - constructed moisture prediction model, under the premise of preset initial conditions and boundary conditions, using the characteristic diffusion coefficient as a parameter, solve the diffusion model to obtain the penetration time - depth - moisture content data with parameters. Through dimensionality reduction processing, obtain the predicted penetration time - moisture content data with parameters. Taking maximizing the similarity between the measured penetration time - moisture content data and the predicted penetration time - moisture content data with parameters as the optimization objective, and taking the characteristic diffusion coefficient as the optimization object, perform iterative optimization until the stop condition is met; Step S3: Based on the iteratively optimized characteristic diffusion coefficient, use the moisture prediction model to realize the moisture prediction of the asphalt pavement.

2. The asphalt pavement moisture prediction method based on the moisture prediction model according to claim 1 is characterized in that: The process of obtaining the measured penetration time - moisture content data by coring the asphalt pavement and simulating it in an environmental chamber includes the following steps: Step S101: Core the asphalt pavement to obtain a disc - shaped specimen with a diameter - to - thickness ratio greater than a preset value; Step S102: Place the disc - shaped specimen in an environmental chamber that matches the temperature and humidity of the asphalt pavement, and obtain the measured penetration time - moisture content data through measurement.

3. The asphalt pavement moisture prediction method based on the moisture prediction model according to claim 1 is characterized in that: The moisture prediction model is: where x is the depth from the surface layer of the asphalt pavement, t is the time of moisture intrusion, C is the moisture content at a certain point in the asphalt pavement, and D is the characteristic diffusion coefficient.

4. The asphalt pavement moisture prediction method based on the moisture prediction model according to claim 1 is characterized in that: The initial conditions and boundary conditions are: Initial condition: C(x,0) = 0 Boundary condition: C(a,t) = C2; C(b,t) = 0 where t≥0 is the time of moisture intrusion, a < x < b, x is the depth from the surface layer of the asphalt pavement, a and b are the value boundaries, C is the moisture content at a certain point in the asphalt pavement, and C2 is the moisture content of the surface layer of the asphalt pavement in contact with the outside.

5. The asphalt pavement moisture prediction method based on the moisture prediction model according to claim 1 is characterized in that: The process of solving the diffusion model to obtain the penetration time - depth - moisture content data with parameters includes the following steps: Step S201: Based on the moisture prediction model, the initial conditions and boundary conditions, solve by constructing a partial differential equation to obtain the penetration time - depth - moisture content data with parameters including the characteristic diffusion coefficient.

6. The asphalt pavement moisture prediction method based on the moisture prediction model according to claim 1 is characterized in that: The process of obtaining the predicted penetration time - moisture content data with parameters through dimensionality reduction processing includes the following steps: Step S202: For the penetration time - depth - moisture content data with parameters, integrate in the depth dimension to obtain the predicted penetration time - moisture content data with parameters using the characteristic diffusion coefficient as a parameter.

7. The asphalt pavement moisture prediction method based on the moisture prediction model according to claim 1 is characterized in that: The iterative optimization process includes the following steps: Step S203: Initialize the characteristic diffusion coefficient; Step S204: Calculate the goodness of fit between the measured penetration time - moisture content data and the predicted penetration time - moisture content data with parameters, solve and correct the characteristic diffusion coefficient; Step S205: Repeat Step S204 until the stop condition is met.

8. The asphalt pavement moisture prediction method based on the moisture prediction model according to claim 1 is characterized in that: The stop condition is: The maximum number of iterations is reached or the goodness of fit between the measured penetration time-moisture content data and the predicted penetration time-moisture content parameter data is less than a preset value.

9. An electronic device, characterized in that: include: One or more processors and a memory, wherein the memory stores one or more programs, and the one or more programs include instructions for executing the asphalt pavement moisture prediction method based on the moisture prediction model as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: It comprises one or more programs executed by one or more processors of an electronic device, wherein the one or more programs comprise instructions for executing the asphalt pavement moisture prediction method based on the moisture prediction model as described in any one of claims 1-8.

Citation Information

Patent Citations

  • Airport asphalt concrete pavement performance acquisition system and control evaluation method

    CN103808914A

  • Asphalt mixture water-gas motion parameter calculation method, storage medium and system

    CN112699529A

  • Evaluation method based on internal moisture concentration of asphalt mixture

    CN114675012A

  • Unsaturated hydraulic concrete moisture diffusion coefficient inversion method based on actually measured wet expansion strain

    CN115901548A

Cited By

  • Device for detecting performance of hot mix plant recycled asphalt

    CN119198445A