Asphalt pavement structure performance analysis method and system

Abstract

The invention discloses an asphalt pavement structure performance analysis method and system, relates to the technical field of road engineering, and aims to solve the problem that an asphalt pavement structure performance analysis model is insufficient in self-adaptability, and the asphalt pavement structure performance analysis method comprises the following steps: obtaining test measurement data of a drop weight deflectometer; combining the adjacent structural layers according to the similarity of the structural layer materials; different numerical simulation models are established according to different configuration parameter combinations, and structural performance evaluation indexes are calculated respectively; different samples are established according to different numerical simulation models, weights of different samples are calculated and normalized, and a sample training structure performance index calculation model is established and input; and calculating a structural performance evaluation index of the target road section according to test measurement data of the drop weight deflectometer of the target road section by applying the structural performance index calculation model. According to the asphalt pavement structure performance analysis method and system provided by the invention, the attention of the trained model to a specific sample can be improved, and the adaptive ability and prediction precision of the trained model are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of road engineering, and in particular to an asphalt pavement structure performance analysis method and system. Background Art

[0002] The mechanical response of asphalt pavement structure under load is an important indicator for evaluating the residual value of pavement structure. The rapid calculation of asphalt pavement structure performance indicators is of great significance for improving the pavement maintenance level. The drop weight deflectometer test is a commonly used method for evaluating pavement structure performance. Based on the pavement deflection measured by the drop weight deflectometer, the pavement structure layer parameters are inverted, and the mechanical response of the pavement structure layer is calculated by the forward modeling method. In order to improve the efficiency of pavement structure performance evaluation, it is a common technical method to construct a neural network model with deflection input and structure layer strain as output. However, when constructing a pavement structure performance evaluation model based on a neural network, the weights of different samples are relatively rarely considered, resulting in problems with the applicability of the trained model.

[0003] Due to the complex working conditions of the pavement drop-weight deflectometer test, how to select appropriate calculation indicators and establish a reasonable sample weight calculation method requires engineering and technical personnel to determine it based on summarizing engineering laws. This is a key technical issue in constructing an adaptive analysis model for pavement structure performance.

[0004] For example, there is a Chinese patent with the announcement number CN107742018B, which involves an incremental method for analyzing asphalt pavement structure based on the stress and strain dependence model of pavement materials. According to the asphalt pavement structure form and material type to be analyzed, the thickness of each layer, Poisson's ratio and the stress and strain dependence model of the pavement material modulus are determined. The modulus of each layer of the asphalt pavement structure is characterized by the stress and strain dependence model of the pavement material modulus. The calculation and analysis system is established in combination with the calculated load and interlayer bonding conditions. The incremental method is used as the basis for analysis. The modulus matrix of each point inside the pavement structure layer is obtained through numerical calculation of finite element software. On this basis, the stress, strain and displacement of any point inside the asphalt pavement structure are calculated according to the elastic layered system theory. This method takes into account the impact of material nonlinearity on pavement mechanics calculations, making the force analysis of asphalt pavement structure more reasonable and complete; however, the patent does not take into account the similarity between different samples in the sample database and the target working conditions, and lacks adaptability. Summary of the invention

[0005] In order to solve the problem of insufficient adaptability of asphalt pavement structure performance analysis models, the present invention proposes an asphalt pavement structure performance analysis method and system. During the neural network model training process, the similarity between sample conditions and target road section working conditions is considered, and the weight of similar samples in model training is increased, thereby improving the applicability of the model to current road conditions and the model prediction accuracy.

[0006] In order to achieve the above object, the present invention adopts the following technical scheme: a method for analyzing the structural performance of an asphalt pavement, comprising the following steps: D1, obtain the measurement data of the drop weight deflectometer test and merge adjacent structural layers according to the similarity of the structural layer materials; D2, establish different numerical simulation models according to different configuration parameter combinations and calculate the structural performance evaluation indicators respectively; D3, establish different samples according to different numerical simulation models, calculate the weights of different samples and normalize them, establish and input the sample training structure performance index calculation model; D4, applying the structural performance index calculation model, calculates the structural performance evaluation index of the target section based on the drop weight deflectometer test measurement data of the target section.

[0007] In this technical solution, firstly, a drop-hammer deflectometer test is carried out on asphalt pavement to obtain data. The deflection data obtained in the drop-hammer deflectometer test are usually 7 to 9 data, and the distance between the measuring point of each data and the load center is distributed between 0 and 2m. The pavement structure layer geometric dimension information and structural layer material information are obtained by consulting the pavement design data or by means of core sampling, ground penetrating radar detection, etc., which includes: the pavement structure geometric dimension information mainly includes the surface layer thickness, base layer thickness and other information of the asphalt pavement. Then, the distribution range of the structural layer strength information is estimated, and the adjacent structural layers are merged according to the similarity of the structural layer materials.

[0008] Then, the value range of the configuration parameters of the numerical simulation model of the drop weight deflectometer test is determined, and the numerical simulation model of the drop weight deflectometer test on asphalt pavement is established. The structural performance evaluation index of the pavement structure layer under different combinations of numerical simulation model parameters is calculated; then, a database sample is formed, the weight is calculated and normalized, and then the sample is input into the structural performance index calculation model, and the model is trained until convergence.

[0009] Finally, the drop weight deflectometer test results of the target section are input into the trained structural performance index calculation model, and the pavement structure performance evaluation index in the corresponding test is calculated to realize the evaluation of the pavement structure performance.

[0010] This method takes into account the similarities between different samples in the sample database and the target working conditions, thereby improving the adaptability of the trained pavement structure performance analysis model and improving the model prediction accuracy.

[0011] The present invention is further configured as follows: Step D1 comprises: D11, obtain the test measurement data of the drop weight deflectometer of asphalt pavement, including deflection data and structural layer material information; D12, estimate the distribution range of the strength information of the structural layer according to the material information of the structural layer; D13, the number of equivalent structural layers after merging adjacent structural layers is three.

[0012] In the present technical solution, the data measured in the drop-weight deflectometer test of the asphalt pavement include deflection data, pavement structure composition information, structural geometric composition information and structural layer material information.

[0013] The distribution range of the structural layer strength information is estimated based on the structural layer material information measured by the drop weight deflectometer test. The configuration parameter value range of the numerical simulation model of the drop weight deflectometer test includes the number of pavement structure layers, the distribution range of the structural layer thickness, and the distribution range of the structural layer modulus parameters.

[0014] Adjacent structural layers with similar material composition are merged into the same structural layer. The merged pavement structural layer can usually be divided into three layers, which correspond to the common surface layer, base layer and roadbed structural layer combination in pavement structure design. They include: the surface layer in the asphalt pavement structure is usually asphalt mixture material, and the base layer is cement stabilized gravel base layer or asphalt stabilized gravel base layer. According to the composition and design parameters of the structural layer materials, the strength parameters such as the modulus of the structural layer can be estimated to a certain extent, and then the value distribution range of the strength parameters such as the modulus of the structural layer can be determined; adjacent structural layers can be merged according to the similarity of the structural layer materials, such as adjacent asphalt mixture structural layers can be merged into asphalt surface layers, and adjacent cement stabilized gravel structural layers can be merged into water-stabilized base layers.

[0015] The present invention is further configured as follows: Step D2 comprises: D21, determine the value range of configuration parameters of the numerical simulation model of the drop weight deflectometer test, the configuration parameters include structural layer thickness and structural layer modulus parameters; D22, select different configuration parameter combinations within the range of configuration parameter values ​​to establish numerical simulation models respectively, and calculate the structural performance evaluation indexes of the pavement structure layer of different numerical simulation models.

[0016] In the present technical solution, the value range of the configuration parameters of the numerical simulation model is determined based on the target section pavement structure size distribution, the load distribution of the drop weight deflectometer test, the structural layer modulus parameter distribution, and the number of structural layers, to ensure that the value range of the configuration parameters of the numerical simulation model covers the parameter distribution range corresponding to the target section, and the value range of the configuration parameters of the numerical simulation model exceeds the range of the parameter values ​​of the section to be analyzed by 20% before and after.

[0017] The value range of the numerical simulation model configuration can also be determined according to common pavement structure forms and industry technical specifications. The thickness of the middle surface layer is 7 to 20 cm, the thickness of the base layer is 30 to 45 cm, the dynamic modulus of the surface layer is 2000 to 22000 MPa, the dynamic modulus of the base layer is 3000 to 23000 MPa, and the subgrade modulus is 100 to 400 MPa.

[0018] The method of selecting the combination of configuration parameters of the numerical simulation model can uniformly select the parameters within the parameter distribution range, cyclically combine the values ​​of different parameters into a parameter combination of a certain working condition, and establish a corresponding numerical simulation model to calculate the pavement structure performance evaluation index under the corresponding working condition. The pavement structure performance index extraction position is usually the mechanical response of the structural layer corresponding to the load action center in the drop weight deflectometer test.

[0019] The present invention is further configured as follows: Step D3 comprises: D31, integrating the structural performance evaluation index and the corresponding configuration parameters calculated by each numerical simulation model into a sample, and storing the samples in the database in the same format and order; D32, calculate the weight of each sample in the database according to the similarity between the target road section and the samples in the database, and normalize the sample weights; D33, establish a structural performance index calculation model, input samples with different weights into the structural performance index calculation model, and train the structural performance index calculation model until convergence.

[0020] In this technical solution, the weights of different samples in the database are calculated based on the similarity between different samples in the database and the drop weight deflectometer test conditions of the road section to be analyzed, and the indicators used to analyze the similarity of the samples mainly include the structural form of the road.

[0021] The weight is calculated as: In the formula, α m is the weight of the mth sample, is the correlation weight between the kth structural geometric form index and the structural performance evaluation index of the mth sample, is the distance-based similarity weight of the m-th sample.

[0022] Weight calculation method The calculation method is: In the formula, Correlation between the kth structural geometric form index of the mth sample and the ath pavement deflection calculated by the drop weight deflectometer test simulation model, max a {} means to select the largest index among all the correlation indexes as

[0023] Weight calculation method The calculation method is: In the formula, is the distance-based similarity weight of the mth sample, d ij It is a sample similarity measure based on Mahalanobis distance, and its calculation method is shown in the following formula: In the formula, and They are respectively the kth geometric dimension of the road section to be analyzed and the kth geometric dimension of the mth sample record in the database.

[0024] On the basis of calculating the weights of different samples, the sample weights in the database are homogenized: In the formula, is the weight of the i-th sample, is the weight of the homogenized ith sample, and A is the total number of samples in the database.

[0025] The present invention is further configured such that the structural performance index calculation model is a neural network model.

[0026] In this technical solution, the structural performance index calculation model adopts a neural network model, and the neural network model is generally multi-layered.

[0027] The present invention is further configured as follows: in step D32, the structural performance evaluation indexes in the samples analyzed for similarity include the geometric dimensions of the pavement structure, and the weights are calculated using a method of weight averaging based on the number of indexes.

[0028] In this technical solution, since the pavement structure geometric dimensions have multiple indicators, the present invention adopts a method of weight averaging based on the number of indicators to avoid the influence of the number of indicators on the weight value and the possible imbalance of weights of different types of indicators.

[0029] The present invention is further configured as follows: Step D33 includes the following steps: D331, randomly divide the samples in the database into training set, validation set and test set; D332, calculate the mean square error between the predicted value and the measured value of the calculation model of the structural performance index; D333, select the model with the smallest mean square error between the predicted value and the measured value on the validation set as the final model, and verify the model prediction accuracy on the test set.

[0030] In this technical solution, the calculation model of the structural performance index constructed by training first randomly divides the samples in the database into training set, validation set and test set, with the training set accounting for 70% to 90%, the validation set accounting for 20% to 5%, and the test set accounting for 10% to 5%; during the training process, the mean square error between the model prediction value and the measured value is used as the model evaluation standard, and the weight of the sample is considered in the calculation. The calculation formula is: Where Y i is the model prediction value, is the measured value, n is the number of samples input for a single model prediction, and w i is the sample weight.

[0031] Then, the mean square error between the predicted value and the measured value of the structural performance index calculation model is calculated as the model evaluation standard. The model with the smallest mean square error between the predicted value and the measured value is selected as the final model on the validation set, and the model prediction accuracy is verified on the test set.

[0032] The present invention is further configured as follows: in step D4, the model input includes the deflection information of the road surface, the thickness of the road surface layer, the thickness of the base layer, and the load applied by the drop weight deflectometer, and the model output is the structural performance evaluation index of the road surface structure layer, including the bending-tensile strain at the bottom of the road surface layer, the bending-tensile strain at the bottom of the base layer, and the compressive strain on the top surface of the roadbed.

[0033] In this technical solution, by inputting the deflection information of the road surface, the thickness of the road surface layer, the thickness of the base layer, and the load information of the drop weight deflectometer into the model, the structural performance evaluation index of the road surface structure layer including the bending-tensile strain at the bottom of the road surface layer, the bending-tensile strain at the bottom of the base layer, and the compressive strain on the top surface of the roadbed can be obtained.

[0034] The present invention also adopts the following technical solution: an asphalt pavement structure performance adaptive analysis system, using the above-mentioned asphalt pavement structure performance analysis method, characterized in that it includes: An information acquisition module, which acquires the measurement data of the drop weight deflectometer test; The simulation calculation module establishes a numerical simulation model and outputs performance evaluation indicators and configuration parameter composition samples; A simulation sample database module stores samples of different numerical simulation models; Model training module, which trains the structural performance index calculation model based on samples of different numerical simulation models; The index calculation module uses the structural performance index calculation model to calculate the structural performance evaluation index.

[0035] In the present technical scheme, the asphalt pavement structural performance adaptive analysis system uses the above-mentioned asphalt pavement structural performance analysis method. The information acquisition module obtains the deflection data and the information of the on-site structural layer in the drop weight deflectometer test. The simulation calculation module integrates the pavement deflection and structural performance evaluation indicators obtained under different working conditions and the corresponding calculations into a sample record, and stores them in the simulation sample database module in the same format and arrangement order. The model training module then trains the structural performance index calculation model based on the samples of different numerical simulation models. Finally, the pavement structural performance evaluation index calculation module applies the trained converged model to the analysis of the drop weight deflectometer test data to calculate the structural performance evaluation index.

[0036] The present invention is further configured as follows: the model training module includes: The database sample weight calculation module analyzes the similarity between the target road section conditions and the samples in the database and calculates the weights of different samples. The pavement structure performance index calculation model training module inputs samples with different weights into the structure performance index calculation model and trains the structure performance index calculation model until convergence.

[0037] In this technical solution, the database sample weight calculation module analyzes the similarity between the target road section conditions and the samples in the database to calculate the weights of various samples in the database. Then, the pavement structure performance index calculation model training module inputs samples with different weights into the model to be trained until the model converges.

[0038] The beneficial effects of the present invention are: 1) It can realize the rapid calculation of the mechanical response of the asphalt pavement structure in the drop weight deflectometer test and support the evaluation of the pavement structure performance; 2) Taking into account the structural similarity between the database samples and the road sections to be analyzed, a sample weight calculation method suitable for pavement structure performance analysis working conditions is designed, which improves the trained model's attention to specific samples and improves the trainable model's adaptability and prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of a method for analyzing the structural performance of an asphalt pavement according to the present invention.

[0040] Figure 2 It is a schematic diagram of the structure of the neural network model constructed in Example 1 of the present invention.

[0041] Figure 3 This is a comparison chart of the calculation accuracy of the calculation model obtained by considering the sample weight and not considering the sample weight in Example 1 of the present invention.

[0042] Figure 4It is a schematic flow chart of the method for adaptive rapid analysis of asphalt pavement structural performance in Example 2 of the present invention.

[0043] Figure 5 It is a schematic diagram of the on-site drop-weight deflectometer test in Example 2 of the present invention.

[0044] Figure 6 It is a comparison chart of the field measured values ​​and the calculated values ​​in Example 2 of the present invention.

[0045] Figure 7 It is a schematic diagram of the adaptive rapid analysis device for asphalt pavement structure performance in Example 3 of the present invention.

[0046] Figure 8 It is a schematic diagram of the structure of the electronic terminal in Example 5 of the present invention.

[0047] Figure numerals: 1: surface layer; 2: base layer; 3: roadbed; 4: strain sensor; 5: load action position. DETAILED DESCRIPTION

[0048] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0049] The present invention provides an adaptive rapid analysis method, device, storage medium and terminal for asphalt pavement structure performance, which can take into account the similarity between different samples in a sample database and target working conditions, thereby improving the applicability of the trained pavement structure performance analysis model and enhancing the model prediction accuracy.

[0050] Example 1 This embodiment provides a method for self-adaptive rapid analysis of asphalt pavement structure layer performance, comprising the following steps.

[0051] Step S11, analyzing the road section characteristics, determining the distribution range of the geometric dimensions of the pavement structure of the road section to be analyzed, and further determining the distribution of simulation model parameters based on finite elements according to the distribution of the geometric dimensions of the pavement structure of the road section to be analyzed.

[0052] In this embodiment, the distribution range of the pavement structure geometry to be analyzed and the distribution range of the determined model parameters are shown in Table 1; Table 1 Distribution range of geometric dimensions of the road section to be analyzed and distribution range of the determined model parameters Structure layer type Pavement structure size range of the road section to be analyzed Determined distribution range of simulation model parameters Surface 16cm~24cm 12cm~28cm Grassroots 36cm~42cm 32cm~46cm .

[0053] Step S12, establishing a finite element model based on the determined model geometric parameters and the estimated pavement structure layer modulus to simulate the mechanical response of the pavement structure in the drop weight deflectometer test; calculating the mechanical response of the pavement structure under the FWD load within a given parameter distribution range, and organizing the parameters given by the model calculation and the calculated structural mechanical response as a sample record and storing it in a sample database.

[0054] In this embodiment, the selected pavement structure performance evaluation index is the bending and tensile strain at the bottom of the base layer under load.

[0055] In this embodiment, the calculation parameters of some sample records are shown in Table 2, and some calculated sample records are shown in Table 3; Table 2 Calculation parameters of some numerical simulation models Sample Record Number Surface thickness Base thickness Surface modulus Base modulus A1 12 32 3000 10000 A2 12 32 3000 11000 A3 12 34 3000 10000 A4 14 32 3000 10000 A5 14 34 4000 10000 ; Table 3 Pavement deflection and structural layer strain calculated by some numerical models

[0056] Step S13, calculate the similarity between the model parameters of different sample records in the sample database and the structural parameters of the road section to be analyzed according to the structural characteristics of the road section to be analyzed, evaluate the correlation between different pavement structural parameters and pavement deflection under load, determine the weights of different pavement structural parameter indicators according to the correlation, and then calculate the weight coefficients of different samples in the database.

[0057] First, based on the correlation between the pavement structure size parameters and the pavement structure response parameters in the drop weight deflectometer test, the pavement structure size parameters selected in this embodiment are the pavement surface layer thickness and the base layer thickness, and the pavement structure response parameters include the pavement deflection calculated in the drop weight deflectometer test.

[0058] In this embodiment, the correlation between the pavement structure size parameters and the pavement deflection under load is analyzed as shown in Table 4. Since the calculated correlation parameters are all negative values, the opposite number is taken in the weight calculation process to ensure that the calculated parameter weight and sample weight are positive values; Table 4 Calculation parameters of some numerical simulation models

[0059] Since there are many forms of pavement structure geometric dimensions in the road section to be analyzed, and there are certain differences between the sample records in the database and the different structural geometric dimensions, it is necessary to first determine a pavement structure form for calculating the sample weight and training the corresponding model; and for other pavement structure forms in the road section to be analyzed, repeat the following operations to obtain the corresponding analysis model.

[0060] In this embodiment, the geometric parameters of the road surface model structure are first selected as shown in Table 5; Table 5 Example of the structural dimensions of the road section to be analyzed selected in the embodiment The structure number of the road section to be analyzed Surface thickness Base thickness T1 16 36 .

[0061] Then the similarity of the structural size parameters of different samples is calculated, and the calculation method is shown in the following formula: In the formula, is the distance-based similarity weight of the kth geometric dimension index of the mth sample, It is a sample similarity measure based on Mahalanobis distance, and its calculation method is shown in the following formula: In the formula, and are the kth geometric dimension of the road section to be analyzed and the kth geometric dimension of the mth sample record in the database. The calculated distance-based similarity weight of the sample is brought into the following formula to further calculate the sample weight: In the formula, α m is the weight of the mth sample, is the correlation weight of the kth structural geometric form index of the mth sample, where The calculation method is: In the formula, Correlation between the kth structural geometric form index of the mth sample and the pavement deflection calculated by the ath drop weight deflectometer test simulation model, max a {} means to select the largest index among all the correlation indexes as

[0062] Step S14, based on the calculation of each sample weight, further perform homogenization on the samples in the database to provide a basis for subsequent training based on the database samples. The calculation formula for homogenization of the sample weights in the database is: In the formula, is the weight of the i-th sample, is the weight of the homogenized ith sample, and A is the total number of samples in the database.

[0063] Step S15, constructing a neural network model, the model input is the pavement deflection calculated by the drop weight deflectometer test simulation model, the model output is the pavement structure layer strain calculated by the drop weight deflectometer test simulation model, the neural network model structure is a feedforward neural network structure, and the structure type and model hyperparameters can be determined according to actual needs.

[0064] In this embodiment, the constructed neural network model structure includes two hidden layers, wherein the number of units in the first hidden layer is 70, and the number of units in the second hidden layer is 35. The neural network model structure constructed in this embodiment is as follows: Figure 2 shown.

[0065] Step S16, randomly divide the weighted database samples into a training set and a validation set, where the training set accounts for 85% and the validation set accounts for 15%. Input the data samples in the training set into the constructed neural network model, compare the results output by the model with the results in the sample database, and calculate the prediction mean square error of the model according to the following formula: Where Y i is the model prediction value, is the true value, n is the number of samples input for a single model prediction, and w i is the sample weight.

[0066] According to the calculated error, the back propagation method is used to correct the neural network model parameters. After all samples in the database are input into the model, the samples on the validation set are input into the neural network model, and the mean square error between the calculation results of the neural network model on the validation set and the true value is calculated as the loss error of this model training, that is, the evaluation index of the model prediction accuracy.

[0067] Input the samples in the database into the neural network model again, repeat the above operation, and get the loss error of the neural network model. After multiple trainings, the parameters of the neural network model are continuously optimized, the prediction accuracy of the model on the validation set will continue to increase, and the loss error of the model will continue to decrease. The model with the smallest loss error on the validation set will be used as the final trained model.

[0068] Step S17, applying the trained model to the road section to be analyzed, calculating the mechanical response of the pavement structure in the drop weight deflectometer test, and then evaluating the performance status of the pavement structure.

[0069] Step S18, in this embodiment, since the road section to be analyzed also uses a numerical simulation model to calculate the structural mechanical response of the pavement structure under the load of the drop weight deflectometer, this embodiment compares the calculation results of the numerical simulation model with the prediction results of the neural network model. At the same time, the embodiment also inputs the database samples without considering the sample weight, that is, the weight of all samples is 1, into the constructed neural network model, and trains a control group model to illustrate that the trained model has better model prediction accuracy.

[0070] In this embodiment, the deviation distribution range between the model prediction value obtained by considering the sample weight training and the true value is significantly smaller than the model prediction accuracy without considering the sample weight, and the deviation distribution is also more concentrated within a smaller threshold, such as Figure 3 As shown, the upper figure is the calculation accuracy of the model considering the sample weight, and the lower figure is the calculation accuracy of the model not considering the sample weight. The results show that the prediction accuracy of the model trained by considering the sample weight is significantly higher than that of the model trained by not considering the sample weight, which fully demonstrates the effectiveness of the method of the present invention and its superiority over the traditional method.

[0071] Example 2 This embodiment provides a method for adaptively analyzing the performance of asphalt pavement structure layers. Figure 4 , including the following steps: Step S21, a strain sensor is arranged at the bottom of the base of the asphalt pavement test section, and an on-site drop-weight deflectometer test is carried out to obtain deflection data, and the strain of the bottom of the base under the load is also monitored. The schematic diagram of the test section structure for carrying out on-site drop-weight deflectometer test and structural strain monitoring in this embodiment is shown in FIG. Figure 5 shown.

[0072] Step S22, when using this method to calculate the strain at the bottom of the base layer in the drop weight deflectometer experiment as a structural performance evaluation index, the calculation model construction method is similar to that of Example 1.

[0073] Step S23, comparing the structural performance evaluation index calculated by the calculation method of the present method, namely the bending-tensile strain at the bottom of the base layer, with the bending-tensile strain at the bottom of the base layer measured on site to evaluate the effectiveness of the present method. Figure 6 By comparing the calculated results of this method with the field measured results, it can be found that there is a good correlation between the two, and the numerical points are evenly distributed on both sides of the straight line y=x, indicating that the calculated results have good accuracy and verifying the effectiveness of this method.

[0074] Example 3 The embodiment provides an asphalt pavement structure performance adaptive rapid analysis device, referring to Figure 7 ,include: The pavement deflection information acquisition module S31 can obtain the pavement deflection information in the drop-hammer deflectometer test; the asphalt pavement drop-hammer deflectometer test numerical simulation module S32 can simulate and calculate the mechanical response of the asphalt pavement structure in the drop-hammer deflectometer test, and output the asphalt pavement deflection and the mechanical response of the pavement structure layer under the load; the asphalt pavement structure mechanical response analysis database module S33 can store the calculation results of the drop-hammer test numerical simulation model and the sorted sample records; the database sample similarity analysis and weight calculation module S34 can calculate the similarity between the sample records in the database and the actual pavement structure of the section to be analyzed, and further determine the weight of each sample record; the asphalt pavement structure performance adaptive prediction model training module S35 can construct an asphalt pavement structure performance adaptive prediction model based on the neural network model, and train and optimize the neural network model parameters based on the weighted database samples; the asphalt pavement structure performance adaptive calculation and analysis module S36 applies the optimal model obtained by training to the measured data of the section to be analyzed, predicts the mechanical response of the section to be analyzed under the drop-hammer deflectometer load, and evaluates the pavement structure performance of the section to be analyzed.

[0075] It should be noted that the module provided in this embodiment is similar to the method provided above in implementation manner, so it will not be described in detail.

[0076] It should also be noted that it should be understood that the division of the various modules of the above device is only a division of logical functions, and in actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated.

[0077] These modules may all be implemented in the form of software calling through processing elements, or all be implemented in the form of hardware, or some modules may be implemented in the form of software calling through processing elements and some modules may be implemented in the form of hardware.

[0078] For example, the road deflection information acquisition module can be a separate processing element, or it can be integrated into a chip of the above device, or it can be stored in the memory of the above device in the form of program code, and a processing element of the above device can call and execute the functions of the road deflection acquisition module. The implementation of other modules is similar.

[0079] Furthermore, all or part of these modules can be integrated together or implemented independently.

[0080] The processing element here may be an integrated circuit having a signal processing capability.

[0081] In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in a processor element or an instruction in software form.

[0082] For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASICs), or one or more digital signal processors (DSPs), or one or more field programmable gate arrays (FPGAs). For another example, when a module is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processors that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0083] Example 4 This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, a method for self-adaptive rapid analysis of asphalt pavement structure performance is implemented.

[0084] Those skilled in the art will appreciate that all or part of the steps of implementing the above method embodiments may be accomplished by hardware associated with a computer program.

[0085] The aforementioned computer program may be stored in a computer-readable storage medium.

[0086] When the program is executed, the steps of the above-mentioned method embodiments are performed.

[0087] The aforementioned storage media include but are not limited to ROM, RAM, magnetic disks and optical disks.

[0088] Example 5 This embodiment provides an electronic terminal, with reference to FIG. 8 , which includes a processor 41 , a memory 42 and a communicator 43 .

[0089] The memory 42 is connected to the processor 41 and the communicator 43 through the system bus and completes the communication between them. The memory 42 is used to store computer programs, the communicator 43 is used to communicate with other devices, and the processor 41 is used to run the computer program so that the electronic terminal executes each step of the above-mentioned asphalt pavement structure performance adaptive rapid analysis method.

[0090] The system bus mentioned above may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus.

[0091] The system bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0092] The communication interface enables communication between the database access device and other devices (such as clients, read-write libraries, and read-only libraries).

[0093] The memory may include a random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage.

[0094] The processor can be a general-purpose processor, for example, a central processing unit (CPU), a network processor (NP), or a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components.

[0095] In summary, the adaptive rapid analysis method, device, storage medium and terminal equipment of asphalt pavement structural performance provided by the present invention can realize the rapid calculation of the mechanical response of the structure of the asphalt pavement in the drop weight deflectometer test, and support the evaluation of the pavement structural performance. This method takes into account the structural similarity between the database samples and the road sections to be analyzed, and innovatively designs a sample weight calculation method suitable for pavement structural performance analysis working conditions, which improves the trained model's attention to specific samples, improves the adaptive ability and prediction accuracy of the trained model, and solves the problem of insufficient adaptability of existing analysis methods. It is of great significance for improving the accuracy of pavement structural performance evaluation.

[0096] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A method for analyzing the structural performance of an asphalt pavement, comprising the following steps: D1, obtain the measurement data of the drop weight deflectometer test and merge adjacent structural layers according to the similarity of the structural layer materials; D2, establish different numerical simulation models according to different configuration parameter combinations and calculate the structural performance evaluation indicators respectively; D3, establish different samples according to different numerical simulation models, calculate the weights of different samples and normalize them, establish and input the sample training structure performance index calculation model; D4, applying the structural performance index calculation model, calculates the structural performance evaluation index of the target section based on the drop weight deflectometer test measurement data of the target section.

2. The asphalt pavement structure performance analysis method according to claim 1, characterized in that: The step D1 comprises: D11, obtain the test measurement data of the drop weight deflectometer of asphalt pavement, including deflection data and structural layer material information; D12, estimate the distribution range of the strength information of the structural layer according to the material information of the structural layer; D13, the number of equivalent structural layers after merging adjacent structural layers is three.

3. The asphalt pavement structure performance analysis method according to claim 1, characterized in that: The step D2 comprises: D21, determining the value range of configuration parameters of the falling weight deflectometer test numerical simulation model, wherein the configuration parameters include structural layer thickness and structural layer modulus parameters; D22, select different configuration parameter combinations within the range of configuration parameter values ​​to establish numerical simulation models respectively, and calculate the structural performance evaluation indexes of the pavement structure layer of different numerical simulation models.

4. The asphalt pavement structure performance analysis method according to claim 1, characterized in that: The step D3 comprises: D31, integrating the structural performance evaluation index and the corresponding configuration parameters calculated by each numerical simulation model into a sample, and storing the samples in the database in the same format and order; D32, calculate the weight of each sample in the database according to the similarity between the target road section and the samples in the database, and normalize the sample weights; D33, establish a structural performance index calculation model, input samples with different weights into the structural performance index calculation model, and train the structural performance index calculation model until convergence.

5. The asphalt pavement structure performance analysis method according to claim 1, characterized in that: The structural performance index calculation model is a neural network model.

6. The asphalt pavement structure performance analysis method according to claim 4, characterized in that: In the step D32, the structural performance evaluation indexes in the samples analyzed for similarity include the geometric dimensions of the pavement structure, and the weights are calculated using a method of weight averaging based on the number of indexes.

7. The asphalt pavement structure performance analysis method according to claim 4, characterized in that: The step D33 comprises the following steps: D331, randomly divide the samples in the database into training set, validation set and test set; D332, calculate the mean square error between the predicted value and the measured value of the calculation model of the structural performance index; D333, select the model with the smallest mean square error between the predicted value and the measured value on the validation set as the final model, and verify the model prediction accuracy on the test set.

8. The asphalt pavement structure performance analysis method according to claim 1, characterized in that: In step D4, the model input includes the deflection information of the road surface, the thickness of the road surface layer, the thickness of the base layer, and the load of the drop weight deflectometer. The model output is the structural performance evaluation index of the road surface structure layer, including the bending-tensile strain at the bottom of the road surface layer, the bending-tensile strain at the bottom of the base layer, and the compressive strain on the top surface of the roadbed.

9. An adaptive analysis system for asphalt pavement structure performance, using an asphalt pavement structure performance analysis method according to any one of claims 1 to 8, characterized in that: include: An information acquisition module, which acquires the measurement data of the drop weight deflectometer test; The simulation calculation module establishes a numerical simulation model and outputs performance evaluation indicators and configuration parameter composition samples; A simulation sample database module stores samples of different numerical simulation models; Model training module, which trains the structural performance index calculation model based on samples of different numerical simulation models; The index calculation module uses the structural performance index calculation model to calculate the structural performance evaluation index.

10. The asphalt pavement structure performance adaptive analysis system according to claim 9, characterized in that: The model training module includes: The database sample weight calculation module analyzes the similarity between the target road section conditions and the samples in the database and calculates the weights of different samples. The pavement structure performance index calculation model training module inputs samples with different weights into the structure performance index calculation model and trains the structure performance index calculation model until convergence.

Images